{ "schemaVersion": "deck-intelligence.v1", "id": "019dd923-5e88-73ef-bd5c-f812573a947a", "slug": "eea7524c557036f4", "title": "2020 Air Street Capital The State of AI Report 2020", "subtitle": "Air Street Capital", "author": "Air Street Capital", "pageCount": 177, "kind": "consulting-deck", "orientation": "landscape", "aspectRatio": 1.777, "document": { "id": "019dd923-5e88-73ef-bd5c-f812573a947a", "slug": "eea7524c557036f4", "title": "2020 Air Street Capital The State of AI Report 2020", "rawTitle": "2020 Air Street Capital The State of AI Report 2020", "authorId": "AirStreetCapital", "authorName": "Air Street Capital", "authorDisplay": "Air Street Capital", "kind": { "slug": "consulting-deck", "label": "Consulting deck" }, "sourceType": { "slug": "vc_research", "label": "VC research" }, "fileName": "2020-Air-Street-Capital-The-State-of-AI-Report-2020.pdf", "sourceUrl": "https://stateof.ai", "sourceDate": "2020-10-01 00:00:00", "presentationDate": null, "ingestedAt": "2026-04-25 11:55:48+00", "status": "ready", "pageCount": 177, "orientation": "landscape", "aspectRatio": 1.777, "targetCompany": null, "targetTicker": null, "metadata": { "local_root": "/tmp/legacy-images/eea7524c557036f4", "origin_corpus": "consulting" } }, "coverage": { "pageCount": 177, "pagesIndexed": 177, "pagesWithImages": 177, "pagesWithMetadata": 177, "imageCoveragePct": 100, "metadataCoveragePct": 100, "componentCount": 1040, "chartCount": 104, "metricSlides": 0, "toolMatches": 279, "documentToolMatches": 0, "frameworkMatches": 24, "arcMatches": 2, "beatMatches": 7, "loopMatches": 22, "patternMatches": 0, "storymakersMatch": true, "scoreAvailable": true, "reviewAvailable": true, "activistThesisAvailable": false, "pitchdeckMetadataAvailable": false, "pitchdeckProfileAvailable": false }, "blocks": [], "matches": { "arcs": [ { "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "evidence": "Facts (Research/Talent/Industry pp.10-129) -> Implications (Politics pp.130-170) -> Action (Predictions p.172).", "isPrimary": true, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:28.977035+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "evidence": "Sections layer rising stakes from research curiosity to geopolitical conflict; predictions resolve.", "isPrimary": false, "confidence": 40, "extraction": { "at": "2026-04-29 13:07:28.977035+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "beats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 174, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 171, "name": "The Action (Now What)", "beatId": "019dd95a-0682-776c-8e34-f7c6d2e63d67", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Action (Now What)", "slug": "triple-take-the-action-now-what", "status": "active", "canonId": "019dd9b8-0a5e-735a-bf53-3a1ba629b7aa", "version": 1, "description": null }, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "isPrimaryArc": true, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 }, { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null, "categoryName": "Analysis & Information", "categorySlug": "analysis-information" }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 14, "from": 11, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null, "categoryName": "Comparison", "categorySlug": "comparison" }, "matchId": "019dd95a-07fe-70ce-8d3c-eb99cad9ccc2", "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null, "categoryName": "Comparison", "categorySlug": "comparison" }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null, "categoryName": "Comparison", "categorySlug": "comparison" }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 92, "from": 90, "loop": { "name": "36_maturity_curve", "slug": "36-maturity-curve", "status": "active", "bestFor": "Digital transformation, capability building, benchmarking", "canonId": "019dd956-748d-7329-ad3a-5c550b66e275", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current Maturity Level -> The Gap to Next Level -> Required Capabilities -> The Roadmap", "description": "Show where you are on a progression and what it takes to reach the next level", "familyLabel": null, "categoryName": "Consulting Framework", "categorySlug": "consulting-framework" }, "matchId": "019dd95a-07fe-70ce-8d3d-134b6ed954d8", "evidence": "FDA TPLC framework -> SPIRIT-AI -> first CMS reimbursement progression.", "position": 11, "objective": "Track AI medical regulation maturing toward reimbursement", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null, "categoryName": "Analysis", "categorySlug": "analysis" }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 110, "from": 107, "loop": { "name": "16_david_goliath", "slug": "16-david-goliath", "status": "active", "bestFor": "Startups, disruptive innovation, competitive displacement", "canonId": "019dd956-6c5d-704a-9c33-1aa083ca432e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Giant's Weakness -> Our Slingshot (Unique Edge) -> The Topple", "description": "Frame your initiative as the nimble underdog taking on a slow, bloated incumbent", "familyLabel": null, "categoryName": "Competitive", "categorySlug": "competitive" }, "matchId": "019dd95a-07fe-70ce-8d3d-19b5d7018f7a", "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null, "categoryName": "Causation", "categorySlug": "causation" }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 143, "from": 141, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-2af0a5465c72", "evidence": "Three slides clustering legislative responses to algorithmic risks.", "position": 17, "objective": "Show regulators waking up to deepfakes, ADM and model bias", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 152, "from": 150, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-3126d540f400", "evidence": "DoD principles, NeurIPS ethics, Google fairness/privacy efforts.", "position": 19, "objective": "Survey emerging responsible AI codes and corporate efforts", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } }, { "to": 170, "from": 168, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null, "categoryName": "Logical Reasoning", "categorySlug": "logical-reasoning" }, "matchId": "019dd95a-07fe-70ce-8d3d-3d1ca750c538", "evidence": "Tax-code automation incentive, jobs at risk, climate change call to action.", "position": 22, "objective": "Surface labor and climate implications of AI", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "tools": [ { "tool": { "name": "Chunking", "slug": "chunking", "status": "active", "bestFor": null, "canonId": "019dd956-abab-72ce-b64c-2924cc83cf39", "version": 1, "bodyDocId": null, "description": "Information grouped into chunks is remembered better - structure content in digestible units", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-012633b6d15c", "evidence": "Bullets chunked under Research/Talent/Industry/Politics headers.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Contrast Pairs", "slug": "contrast-pairs", "status": "active", "bestFor": null, "canonId": "019dd956-9b77-70d3-88b1-f34e26f60860", "version": 1, "bodyDocId": null, "description": "Creating tension through opposites: Before↔After, Problem↔Solution, Pain↔Relief", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "loop", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-1e8182ce6905", "evidence": "PyTorch vs TensorFlow pair set up across loop.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Inductive Reasoning", "slug": "inductive-reasoning", "status": "active", "bestFor": null, "canonId": "019dd956-99c6-705e-94cc-76a504c3da37", "version": 1, "bodyDocId": null, "description": "Grouping similar evidence to form a conclusion (Evidence → Pattern → Conclusion)", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-31d3149e0ebd", "evidence": "Loop builds case from multiple model-size data points.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Parallel Structure", "slug": "parallel-structure", "status": "active", "bestFor": null, "canonId": "019dd956-9f34-762a-93b7-1d8e84bbd153", "version": 1, "bodyDocId": null, "description": "Using similar grammatical structure for related ideas", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-86ce81800807", "evidence": "Parallel slide titles 25-26 use same conversational format.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Contrast Pairs", "slug": "contrast-pairs", "status": "active", "bestFor": null, "canonId": "019dd956-9b77-70d3-88b1-f34e26f60860", "version": 1, "bodyDocId": null, "description": "Creating tension through opposites: Before↔After, Problem↔Solution, Pain↔Relief", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "loop", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-0b5794aad0e6", "evidence": "Academia vs Big Tech contrast set up across loop.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chunking", "slug": "chunking", "status": "active", "bestFor": null, "canonId": "019dd956-abab-72ce-b64c-2924cc83cf39", "version": 1, "bodyDocId": null, "description": "Information grouped into chunks is remembered better - structure content in digestible units", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-88131b332432", "evidence": "Eight numbered predictions chunked clearly.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Slide recipe: Cover", "slug": "cover-slide", "status": "active", "bestFor": null, "canonId": "019df269-2a08-767d-9128-e7431e6fe8a3", "version": 1, "bodyDocId": "019df269-29e5-7738-b5fa-6892e5dc8e65", "description": "The first ten seconds. Build it deliberately.", "familyLabel": "slide-recipe", "categoryName": "Slide", "categorySlug": "slide" }, "agent": "designer", "layer": "slide", "agents": [ "designer" ], "matchId": "a221ed8c-243a-41de-95a5-26dd8ebbae65", "evidence": "This is a cover slide with a title, subtitle, and authors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:43.944352+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 1, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "b5664eb2-ca52-4a2d-a3db-3d9c476c32a7", "evidence": "The slide mentions the report's goal of triggering an informed conversation.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:44.066895+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 3, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Executive summary", "slug": "executive-summary", "status": "active", "bestFor": null, "canonId": "019dd9e1-4b63-72d5-9734-da757d91ca61", "version": 1, "bodyDocId": null, "description": "A condensed view of the entire argument on a single slide / block.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "5bbd142b-c824-40fa-81df-5fa2853f9563", "evidence": "This slide provides an overview of the report's contents and goals.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:44.035883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 3, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecd-fda0baa96ded", "evidence": "Title 'Executive Summary' is topic, but bullets state insights per section.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Core Message Extraction", "slug": "core-message-extraction", "status": "active", "bestFor": "Any slide that carries an argument or finding. The diagnostic: if the slide has a chart on it, it needs a core message. Run extraction whenever the current title is a topic (Q3 performance, Customer satisfaction) rather than a claim, or when a reader could reasonably draw more than one conclusion from the visible evidence.", "canonId": "019dd956-bc71-707d-8867-1b2e810175cf", "version": 1, "bodyDocId": "f85613cb-4b77-462f-bfe6-3b862516e1a6", "description": "The discipline of distilling raw analysis, data, or content into a single declarative sentence — the slide's core message — that the audience would walk away with if they remembered nothing else. Process, not artefact: extraction questions run on the raw content until one sentence falls out.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-0696e92e3c67", "evidence": "Distills 170+ pages into one screen of takeaways.", "pageRefs": null, "priority": null, "whenToUse": "Any slide that carries an argument or finding. The diagnostic: if the slide has a chart on it, it needs a core message. Run extraction whenever the current title is a topic (Q3 performance, Customer satisfaction) rather than a claim, or when a reader could reasonably draw more than one conclusion from the visible evidence.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": "Solves the two failure modes of slide writing simultaneously: the topic title (which outsources interpretation to the audience and produces three different conclusions in three different heads) and the multi-thesis stapler (which exceeds working memory and gets remembered as fragments). Aligns with how readers actually consume decks — Heath & Heath's Find the Core, the journalist's nut graf, and Minto's governing thought all converge on the same cognitive economics.", "antipattern": "Topic titles disguised as core messages (Customer satisfaction, no verb, no stake); multi-thesis sentences stapled with and; hedged sentences (may show signs of possible softening); chart-caption titles (Revenue fell 4%) that restate what the chart already shows instead of naming the so-what; meta-commentary (this slide explores...) that describes the slide rather than concluding it.", "cardinality": null, "narrativePurpose": "Forces the slide author to do the interpretive work before the slide ships, so the audience does not have to do it during the meeting. Wins the WYSIATI battle on purpose — the loudest fragment on the page becomes the author's chosen sentence rather than whichever chart label happens to be largest." }, { "tool": { "name": "Executive summary", "slug": "executive-summary", "status": "active", "bestFor": null, "canonId": "019dd9e1-4b63-72d5-9734-da757d91ca61", "version": 1, "bodyDocId": null, "description": "A condensed view of the entire argument on a single slide / block.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "95723de4-9d02-4d23-a1d8-35f83c598bc8", "evidence": "This slide summarizes key findings across multiple areas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:44.216829+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "0fd3b959-a09f-4118-86a6-c652d73c16b5", "evidence": "This slide introduces a new section with a headline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:44.277952+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 8, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "7821bb09-c318-42ac-9e9b-f025b7b837b3", "evidence": "This slide introduces a new section with a headline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:44.371248+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 10, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-08dcac027626", "evidence": "Title states insight: 'Only 15% of papers publish their code'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-0eb68e9ef56b", "evidence": "Specific 15% statistic anchors abstract claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-12d7f6d3f385", "evidence": "Callout: 'centralization of AI talent and compute as a huge problem.'", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-1531bc16f3c3", "evidence": "Title: PyTorch outpacing TensorFlow as leading indicator.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "6f65baca-aaa2-444d-a594-ce9c3e37a0a4", "evidence": "The slide title and content seem to be directed towards an audience interested in industry trends and technology.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:51.993862+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-19f71d5defe1", "evidence": "Multi-line time series tracks framework share by conference.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f4a2666a-a4b3-4edd-9093-42223cc4866d", "evidence": "Although not explicitly a 2x2 matrix, the legend and chart could suggest a categorization that might be presented in a 2x2 format, but it's not clear.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:52.05529+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-213958fa3249", "evidence": "Title: '47% of these implementations are based on PyTorch vs 18%'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-24f05ecb4e1d", "evidence": "47% vs 18% specific deltas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-2998205cd191", "evidence": "Memorable framing: 'Welcome to the Billion Parameter club'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "145187e6-5805-4f45-bfd6-2fe7d34a9e06", "evidence": "The use of a bar chart to display parameter counts suggests consideration of effective chart selection for communication.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.114739+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-2ef38db2891f", "evidence": "'Billion Parameter club' membership metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-37fcd4f23f9f", "evidence": "Title states the empirical scaling law conclusion.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Causal Chain", "slug": "causal-chain", "status": "active", "bestFor": null, "canonId": "019dd956-9e98-7269-907b-8cf727e9499d", "version": 1, "bodyDocId": null, "description": "Showing how one event leads to another in sequence", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "f27d948d-0b39-42d9-bbeb-7d16a1de85a8", "evidence": "The description of how bigger models, datasets, and compute budgets drive performance implies a causal relationship.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.17496+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-3a10cdcb86ec", "evidence": "Power-law scaling chart appropriate for log relationship.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-3fb0e01e620a", "evidence": "'$1 per 1,000 parameters' makes abstract cost concrete.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-416998989edf", "evidence": "Per-parameter pricing analogy to commodity goods.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b4ef34b1-8767-4fa2-b07d-a118d5c11c81", "evidence": "The use of specific cost estimates ($1 per 1,000 parameters) and ranges ($2.5k - $50k) shows quantification of costs.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:52.235963+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-45a4c69cf11a", "evidence": "'22 TPU v3 core years' translates abstract compute to time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "3da3a978-1eba-4545-92b8-56c5ea6edf8c", "evidence": "The inclusion of a detailed description and a chart to support the data suggests efficient use of data-ink.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:52.295754+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-4be16ed9d761", "evidence": "Title: outrageous costs for incremental gains.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 95, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Anchoring Bias", "slug": "anchoring-bias", "status": "active", "bestFor": null, "canonId": "019dd956-9254-73de-8f1b-c7250f8f4c9c", "version": 1, "bodyDocId": null, "description": "First piece of information becomes reference point for all subsequent judgments", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "1b0cb19a-4381-4cf4-8556-b2ca121c846c", "evidence": "The use of an extremely large cost estimate (over one hundred billion billion dollars) may serve as an anchor for the reader.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:52.355725+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Annotation", "slug": "annotation", "status": "active", "bestFor": null, "canonId": "019dd956-b5b9-7668-bf0e-3af99823ec5c", "version": 1, "bodyDocId": null, "description": "Adding explanatory labels to highlight key data points", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-50a3656d4004", "evidence": "Red rectangle annotation calls attention to key benchmark.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-555308ea27ed", "evidence": "'$100B billion' framing plays on shock.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Von Restorff Effect", "slug": "von-restorff-effect", "status": "active", "bestFor": null, "canonId": "019dd956-beab-73d8-9f62-30d0332e926e", "version": 1, "bodyDocId": null, "description": "Items that stand out are noticed and remembered - make key points visually distinct", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-4ee6c0474b28", "evidence": "Red box highlights ImageNet rows in big table.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-59502023c802", "evidence": "Title states 'larger model needs less data'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Logical chain", "slug": "logical-chain", "status": "active", "bestFor": null, "canonId": "019dd9e1-5104-702f-8953-85757af6553d", "version": 1, "bodyDocId": null, "description": "If-then, premise-conclusion, deductive chain.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d0dfbabe-6da4-44fa-a598-8fb905c3cb2b", "evidence": "The description of how larger models require fewer samples to reach the same performance implies a logical sequence of ideas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.415745+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-5c9ea3db7d85", "evidence": "Callout draws implication: advantage to large companies.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Contrast Principle", "slug": "contrast-principle", "status": "active", "bestFor": null, "canonId": "019dd956-bf0f-73ae-b9df-e60508cdaa65", "version": 1, "bodyDocId": null, "description": "Higher visual weight draws attention - use contrast strategically for emphasis", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "456e0e7b-80ce-4403-ba48-5160fa496b4f", "evidence": "The comparison of model performance across different languages and data availability highlights contrasts.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.477032+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Story Moments", "slug": "story-moments", "status": "active", "bestFor": null, "canonId": "019dd956-ae4e-768d-a18b-108e6cda511e", "version": 1, "bodyDocId": null, "description": "Design key emotional beats: The Shock, The Vision, The Proof, The Choice, The Call", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-63064d4e6955", "evidence": "Concrete Hausa/Uzbek case proves abstract scaling claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-65e78059e26a", "evidence": "Title: deep learning getting more efficient over time.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 22, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-6ad24a7d8699", "evidence": "Specific '2x every 16 months' rule.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 22, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-6ee61f0a123e", "evidence": "Title states caveat: small models can trump large.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 23, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Unexpected Pattern", "slug": "unexpected-pattern", "status": "active", "bestFor": null, "canonId": "019dd956-bc0f-71af-8d32-3c1a2da38c40", "version": 1, "bodyDocId": null, "description": "Breaking a pattern to create surprise and attention", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-734a00d60872", "evidence": "Counterpoint to scaling-law narrative breaks the pattern.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 23, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-76dda5b811d2", "evidence": "Title: new transformer LMs unlocking NLP use cases.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 24, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "ad5485af-1339-4da8-a6be-2de9f64ce16c", "evidence": "The components mention a specific model and its capabilities, indicating that the audience is likely familiar with machine learning and programming languages.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.049989+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-7a0d6756c1d7", "evidence": "Specific 90% C++->Java translation accuracy.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "8e934809-a3af-4bb1-897b-7dff806ef4b7", "evidence": "The slide presents a narrative about a machine translation model, using a clear structure to convey information.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.086372+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-7df0984aa874", "evidence": "Conversational title 'Computer, please...' frames as dialog.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-816aaf21e7ad", "evidence": "Conversational 'Computer, can you...' framing parallels prior slide.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "44ea2cc3-2e25-4b93-8c80-97e80ef76d77", "evidence": "The flowchart showing Broken Program -> Evaluator -> Feedback -> DrRepair -> Repair clearly illustrates a process flow.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:57.15479+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-8bf28f06f37e", "evidence": "Title: 'NLP benchmarks take a beating'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "1c969bcf-011b-4ea6-985a-d35516b2042d", "evidence": "The slide features bar charts and a horizontal bar chart, indicating a consideration of chart types for effective communication.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.220375+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-8c56eadec94a", "evidence": "'Take a beating' is dramatic verb choice.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-917e74bb800a", "evidence": "Callout: GPT-3 has knowledge gaps.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Analytical method", "slug": "analytical-method", "status": "active", "bestFor": null, "canonId": "b42c4911-7946-49b5-951c-5f3d23f4e2bd", "version": 1, "bodyDocId": null, "description": "Diagnostic / structuring methods (5-whys, fishbone, issue tree, hypothesis-driven).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "650aa7ce-bbea-46a1-b52c-8879c423e811", "evidence": "The slide discusses the transformer's ability to generalize and its applications.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.322596+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 29, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-970d6bfc407a", "evidence": "GPT-2 fed images example illustrates generality.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 29, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-98fd088e9fa3", "evidence": "Title: 'Biology is experiencing its AI moment'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-a2e7ae945501", "evidence": "21,000 papers in 2020 anchors the claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "8a798384-cbd8-4bcd-aa7c-f9c501e70af2", "evidence": "The slide features a bar chart with a clear and simple design, effectively using data-ink.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.3898+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-9e7bdc8fb983", "evidence": "'AI moment' analogy frames whole biology section.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-a4f4504bdb06", "evidence": "Cell painting visuals carry the case.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "f33822dd-0c36-4592-aada-6bce06aa2381", "evidence": "The slide features a hexagonal grid of cell images, using visual anchors to organize the information.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.458594+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Causal Chain", "slug": "causal-chain", "status": "active", "bestFor": null, "canonId": "019dd956-9e98-7269-907b-8cf727e9499d", "version": 1, "bodyDocId": null, "description": "Showing how one event leads to another in sequence", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "be9a5a4c-8d62-4d63-b527-407d697a7223", "evidence": "The slide discusses causal relationships between biological perturbations and cellular morphology.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.534196+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Story Moments", "slug": "story-moments", "status": "active", "bestFor": null, "canonId": "019dd956-ae4e-768d-a18b-108e6cda511e", "version": 1, "bodyDocId": null, "description": "Design key emotional beats: The Shock, The Vision, The Proof, The Choice, The Call", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-ab6a73678ab4", "evidence": "COVID drug screen as flagship use case.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-adca83b5eb99", "evidence": "Callout: patients can be treated before vision loss.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 33, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-b3e24aa6318c", "evidence": "Title states clinical performance gains.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 34, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-b78109c4e216", "evidence": "9.40% specific delta cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 34, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Authority Bias", "slug": "authority-bias", "status": "active", "bestFor": null, "canonId": "019dd956-c38d-70b9-a284-b2b479763356", "version": 1, "bodyDocId": null, "description": "Expert opinions are weighted more heavily - cite credible sources", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-bbdb4001ed41", "evidence": "Photo + Judea Pearl/Bengio name-drop authority.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Logical chain", "slug": "logical-chain", "status": "active", "bestFor": null, "canonId": "019dd9e1-5104-702f-8953-85757af6553d", "version": 1, "bodyDocId": null, "description": "If-then, premise-conclusion, deductive chain.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "660603a4-7717-4f08-b92d-c59f981b9d3c", "evidence": "The slide discusses a logical chain of reasoning for applying causal inference in ML.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.67594+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-bd879ae4172a", "evidence": "'Top 25% of doctors' concrete benchmark.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 36, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "AIDA Model", "slug": "aida-model", "status": "active", "bestFor": null, "canonId": "019dd956-86fc-773c-813f-d416714b8413", "version": 1, "bodyDocId": null, "description": "Attention, Interest, Desire, Action - classic persuasion sequence", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "5c1d1f4b-cba9-4af0-af7e-19f128f71f9b", "evidence": "The slide presents a problem (Model explainability is an important area of AI safety), presents a solution (A new approach aims to incorporate causal structure between input features into model explanations), and provides evidence (Asymmetric Shapley Values (ASV) are proposed to incorporate this causal information).", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:46.80324+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 37, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-c29ee74c7b4a", "evidence": "98% synthesizable molecules statistic.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 38, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-c78c6743458f", "evidence": "Title: 'ML will generate a synthesis plan faster than you can'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 39, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-cb914c1b66e6", "evidence": "Title states GNN-led antibiotic discovery.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-cc8e1982feeb", "evidence": "10x MSE improvement quantified.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 42, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d39aec7824c5", "evidence": "Specific 72%/33%/16% hit rates by target.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 43, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d80a56e9c735", "evidence": "4 million contributors data point.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 45, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Singularity Effect", "slug": "singularity-effect", "status": "active", "bestFor": null, "canonId": "019dd956-dcd3-7287-94d8-5a49496c2ddb", "version": 1, "bodyDocId": null, "description": "People care more about individuals than statistics - use specific examples", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d47c7908ae38", "evidence": "Loss of smell as single memorable predictor of COVID.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 45, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-df61de4be204", "evidence": "Pro bono COVID antiviral framing builds goodwill.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 46, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-e2850c31c7af", "evidence": "Wimbledon nostalgia angle anchors abstract synthesis.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 47, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-e758c599d693", "evidence": "Title: attention turning to vision tasks.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 48, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "626d0b8b-66c9-4107-996e-424671aec862", "evidence": "Diagram showing input image, visible ground/depth prediction vs hidden ground/geometry prediction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:56.641776+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 49, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "c6cb1007-d2be-4e0a-a8d5-e489f3e362c8", "evidence": "Process flow showing Data generation and Stereo training stages", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:56.719918+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 50, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-e811011e6691", "evidence": "Title: 'won't drain your battery'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 56, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-ed0800a4d1d7", "evidence": "4-9x parameter reduction quantified.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 56, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-f34c33cd1bbe", "evidence": "Title: federated learning research booming.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 58, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-f657ce4b6664", "evidence": "5x growth and 2020 H1 surpasses all 2019.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 58, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "1ecdc9fb-dc1e-4b11-afb3-46083244e06b", "evidence": "GP models benefit from several key features, like depth and convolutions, inspired from neural networks (NNs).", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.060047+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-f890e57562dd", "evidence": "15 mins -> 40 sec specific speedup.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "31b8520a-0f4f-498a-8c41-441ec53a5022", "evidence": "GP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.095024+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-fd927c8c14dd", "evidence": "'Quantum supremacy' dramatic framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 62, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "a3dcebdc-fd75-40e0-8074-f82c9d975c0e", "evidence": "Introduction | Research | Talent | Industry | Politics | Predictions", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:01.220759+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 63, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-03878557c33d", "evidence": "Title: 'The Great Brain Drain'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "e95d919e-3284-4c7b-a412-98d40bc9151d", "evidence": "chart/bar-horizontal: Universities losing AI professors chart/bar-horizontal: Companies hiring AI professors", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.28902+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-05ea1b5a2df4", "evidence": "Brain drain metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Problem-Agitate-Solve (PAS)", "slug": "problem-agitate-solve-pas", "status": "active", "bestFor": null, "canonId": "019dd956-869f-7564-8d42-de54ed29ab90", "version": 1, "bodyDocId": null, "description": "Identify problem, agitate the pain, present the solution", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "caccc75e-dac5-4949-8ad9-efb1cc13ff0e", "evidence": "You Cannot Serve Two Masters: The Harms of Dual Affiliation — Ben Recht, David A. Forsyth, and Alexei Efros", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.358521+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 65, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-0fdef4a916c0", "evidence": "Callout questions adequacy of dual appointments.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 65, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-1348a2c02cb3", "evidence": "Title: 'departures correlate with reduced graduate entrepreneurship'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Annotation", "slug": "annotation", "status": "active", "bestFor": null, "canonId": "019dd956-b5b9-7668-bf0e-3af99823ec5c", "version": 1, "bodyDocId": null, "description": "Adding explanatory labels to highlight key data points", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-17183707af9a", "evidence": "Red boxes mark a/b/c rows in regression table.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Von Restorff Effect", "slug": "von-restorff-effect", "status": "active", "bestFor": null, "canonId": "019dd956-beab-73d8-9f62-30d0332e926e", "version": 1, "bodyDocId": null, "description": "Items that stand out are noticed and remembered - make key points visually distinct", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-1aa87e4f91eb", "evidence": "Red highlight makes key coefficients pop.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-1e3e9679d757", "evidence": "EUR100M / 50 professors specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 67, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-20f4d319886f", "evidence": "$100M Roux Institute donation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 68, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "e13e0144-f513-4e80-a733-b89aabe6b1a1", "evidence": "chart/bar-vertical: Bar chart showing percentage of authors with undergraduate degrees from China (2009-2019)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.584286+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 70, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-248abc4687b6", "evidence": "29% of NeurIPS authors educated in China.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 70, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-285f4304f838", "evidence": "Title: 54% of Chinese grads move to USA.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 71, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-2d51924b8432", "evidence": "Title: US strong talent retainer post-PhD.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 72, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-33638cdacd15", "evidence": "90% retention statistic.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 72, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "498b0bdf-d1bf-4641-929f-d71e24dc6aa8", "evidence": "legend: Private sector (large co), Private sector (small co), Academic, Government/Nonprofit", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:56.02336+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 73, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-3466d1fe84d5", "evidence": "Title: foreign vs American grad placement contrast.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 73, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-3a98774747a5", "evidence": "55%/40% UK/China sector splits.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 74, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-3e796159f533", "evidence": "China 27%, Europe 11%, India 11% feeders.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 75, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Waterfall chart", "slug": "waterfall-chart", "status": "active", "bestFor": null, "canonId": "e2799da3-725f-44a9-a093-570f1f388e64", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ab826dad-25f7-4377-8b17-74a8c447e5b2", "evidence": "chart/waterfall: Waterfall chart showing percentages: 31% (US), 27% (China), 11% (India), 11% (Europe), 4% (Iran), 3% (Canada), 13% (Others)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:56.150097+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 75, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-41cceb38a085", "evidence": "Callout pulls implication of H-1B suspension.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 76, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-47ed1576e33d", "evidence": "Title: American institutions dominate NeurIPS.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 77, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-4915c6ade0f5", "evidence": "Ranked bar chart of top contributors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 77, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-4d8bcd8ceb73", "evidence": "Title: ICML 2020 cements US lead.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 78, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-53cfb90f551d", "evidence": "Stanford 10x enrollment growth.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 79, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-54b06cfed15f", "evidence": "Title: 'Demand outstrips supply'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 80, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-5b7c2247f8f2", "evidence": "3x more postings than views.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 80, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-5ef0efa0c171", "evidence": "Callout ties COVID to LinkedIn posting drop.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 81, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9ae644a9-434c-4f01-9619-449cc6a0c769", "evidence": "title/headline: Section 3: Industry", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:56.467722+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 82, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-6334a7dee2c0", "evidence": "Title states first phase 1 trial milestone.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 83, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6723d58f387f", "evidence": "EUR250M deal, 29 months timeline specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 84, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "AIDA Model", "slug": "aida-model", "status": "active", "bestFor": null, "canonId": "019dd956-86fc-773c-813f-d416714b8413", "version": 1, "bodyDocId": null, "description": "Attention, Interest, Desire, Action - classic persuasion sequence", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "471ac015-ba9b-4bcf-bad8-14d4d587cd34", "evidence": "The components mention a 'platform strategies' and 'asset focused' company building approach which hints at the AIDA model's Attention, Interest, Desire, and Action stages.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.736566+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6bd69a90d122", "evidence": "$239M mega round figure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "SCQA Framework", "slug": "scqa-framework", "status": "active", "bestFor": null, "canonId": "019dd956-7bcb-763f-bdf6-1dfedf3bd36c", "version": 1, "bodyDocId": "019df22a-a0b0-719e-937e-2ac31e53d32c", "description": "Situation, Complication, Question, Answer - structured problem framing", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "9f3caf15-92f6-496f-be0e-595957f9a75e", "evidence": "The slide mentions a specific prediction and outcome, which aligns with the SCQA framework's Situation, Complication, Question, and Answer structure.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.796804+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 86, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6e41aba64d40", "evidence": "100,000-fold candidate increase.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 87, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "9c526ea9-d892-4aef-93cc-f72f82b994f3", "evidence": "The slide presents a case study with data and results, which follows a narrative arc to convey the story.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.857223+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 87, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "6f1d8824-6edb-4a78-8dff-23891f83da8e", "evidence": "The slide shows a process flow from input to output, which implies a before-after transformation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.918496+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 88, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f3249e98-e341-4f4a-847c-46303d3c6b43", "evidence": "Although not explicitly a 2x2 matrix, the diagram/hub-spoke shows multiple factors influencing metabolic response, which could be interpreted as a matrix-like structure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:58.978775+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-72b5231657f4", "evidence": "1,100 participants, 77% accuracy stats.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Hypothesis-Driven Structure", "slug": "hypothesis-driven-structure", "status": "active", "bestFor": null, "canonId": "019dd956-833f-7419-9932-be80362019a7", "version": 1, "bodyDocId": null, "description": "Start with a hypothesis and organize evidence to prove/disprove it", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "a4933b61-8c3a-423e-bd49-28a1d39e980c", "evidence": "The slide presents a regulatory framework diagram and a spectrum of ML/AI-based algorithms, which implies a structured and hypothesis-driven approach.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:59.039429+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-76b85eae62df", "evidence": "Callout explains TPLC regulatory shift.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-7a8ce7a20b26", "evidence": "Specific 1% of 20,000 studies meet quality.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d01a7386-7b4e-462b-9f46-08251682d2dc", "evidence": "The slide presents information in a list/bullet format, which is a common presentation technique.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.101414+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f4778304-04e2-4c51-b2ab-cd6628e2bb6f", "evidence": "The slide mentions a critical step towards implementation in clinical medicine, which implies a call to action.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.16585+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 92, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-7f44b8ec21a7", "evidence": "Callout explains why CMS approval is critical.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 92, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-81ca965fe8c4", "evidence": "US map visualises legislative spread.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 93, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-84325ef6aaf9", "evidence": "Title: 'driverless cars are still not so driverless'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 94, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-883973d61621", "evidence": "Only 3 of 66 companies allowed.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 94, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "5a8f1c27-ddfe-4050-bd1f-c095deacb1e3", "evidence": "The slide presents a bar chart and a metric, which shows a comparison and could be related to a chart selection guide.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.288769+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-8ece501aad52", "evidence": "0.000737% of human miles vivid contrast.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-9228dd9ff452", "evidence": "'Microscopic' framing emphasises gap.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-9586b599740e", "evidence": "Title: 'Sketchy metrics: Tracking AV progress'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Anchoring Bias", "slug": "anchoring-bias", "status": "active", "bestFor": null, "canonId": "019dd956-9254-73de-8f1b-c7250f8f4c9c", "version": 1, "bodyDocId": null, "description": "First piece of information becomes reference point for all subsequent judgments", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "c9838607-9b31-47c4-baaa-44a4452879c4", "evidence": "The slide mentions a potentially misleading metric (miles/disengagement), which could be related to anchoring bias.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.350687+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Lie Factor", "slug": "lie-factor", "status": "active", "bestFor": null, "canonId": "019dd956-b501-7740-8c7c-af7ecca16e9c", "version": 1, "bodyDocId": null, "description": "Ensure visual representation accurately reflects data magnitude", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-9aed41eaa60f", "evidence": "Calls out misleading 8,679% YoY metric.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "986de603-b020-4d6c-a4f6-82e96616b8e4", "evidence": "The components mention 'Amazon', 'Zoox', and 'autonomous ride-hailing', indicating an audience calibration.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.685233+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "4988c880-37d2-4f90-b2e1-b19cb9026b15", "evidence": "The quote from Amazon 'We’re acquiring Zoox to help bring their vision of autonomous ride-hailing to reality' can be seen as a call to action.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.719072+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-9cbcc3ff8a10", "evidence": "$1.3B Zoox/Amazon deal cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "6baf4b87-908f-44d8-92b7-e45e731830a8", "evidence": "The bubble chart showing funding rounds for various companies instantiates a chart selection guide.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.787974+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 98, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-a088c3168f86", "evidence": "$7B raised since July 2019.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 98, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "f82a8526-1630-4905-a961-617e227e2096", "evidence": "The components mention 'DiDi's self-driving unit raised >$500M' and 'launched its ride hailing service to consumers', indicating a before-after bridge.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.855943+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 99, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Capital Allocation Framework", "slug": "capital-allocation-framework", "status": "active", "bestFor": null, "canonId": "d2901a21-50b9-4151-ab1b-55b432772ea1", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "30ae4ce1-a669-4630-a467-b4e5dbc411f2", "evidence": "The components mention 'Waymo, Aurora, and GM Cruise have acquired LiDAR companies or built sensors in-house', indicating a capital allocation framework.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.924003+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-a57f61ad6c73", "evidence": "Callout explains LiDAR vertical integration moat.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-a84c63083138", "evidence": "$2B / $3.4B SPAC valuations.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 101, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Market Segmentation Pyramid", "slug": "market-segmentation-pyramid", "status": "active", "bestFor": null, "canonId": "1c9f3543-c96f-4a5b-b4a4-9b2a412c0aaf", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9291303f-a055-4751-b177-b5f7eb30f997", "evidence": "The components mention 'Velodyne will list shy of $2B valuation' and 'Luminar at $3.4B', indicating a market segmentation pyramid.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.990912+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 101, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-ad52e19434ae", "evidence": "Title: 'Supervised learning and the cost of edge cases'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-b0a5af0a4d55", "evidence": "Anti-Moore's-Law analogy in callout.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Problem Statement Canvas", "slug": "problem-statement-canvas", "status": "active", "bestFor": null, "canonId": "019dd956-839d-744b-ad11-9bf3e3642a18", "version": 1, "bodyDocId": null, "description": "Structured definition of the problem: What? Who? Where? When? Scale?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "0505b6ee-d387-4123-807d-36cc3c04ff5c", "evidence": "The components mention 'the hardest part of building AI is how it deals with situations that happen uncommonly', indicating a problem statement canvas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:50.059038+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "474a31e4-dda4-4137-a3ca-a37503faa2b1", "evidence": "The components mention '11 major datasets and 2 updates since 2019', indicating a data story arc.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:50.126475+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 103, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Monroe's Motivated Sequence", "slug": "monroes-motivated-sequence", "status": "active", "bestFor": null, "canonId": "019dd956-8758-700f-83cf-d260a54ad24f", "version": 1, "bodyDocId": null, "description": "Five-step persuasion: Attention, Need, Satisfaction, Visualization, Action", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "7d9cef01-2718-4193-97e3-2347bed618f7", "evidence": "The components mention 'Tesla currently recognises significant deferred revenue for its Autopilot feature', indicating a motivated sequence.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:50.264013+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 106, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-b7c36953ed3a", "evidence": "Title: AI compute providers on gen 2+ products.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 107, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-bbfb2099f055", "evidence": "59.4B transistors / 823 sqmm specifics.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 107, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-bcde5de70619", "evidence": "Title: M2000 faster training drops cost.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Comparison frame", "slug": "comparison-frame", "status": "active", "bestFor": null, "canonId": "019dd9e1-484b-7399-a4fc-d14644b99f93", "version": 1, "bodyDocId": null, "description": "Two or more states/subjects placed side by side to expose a gap.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "4faec66b-c36d-460b-8053-f1af26f86395", "evidence": "The components mention 'The reported 16x faster training time for the image classification model EfficientNet-B4 on the M2000 vs. NVIDIA DGX-A100', indicating a comparison frame.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:50.364827+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Contrast Principle", "slug": "contrast-principle", "status": "active", "bestFor": null, "canonId": "019dd956-bf0f-73ae-b9df-e60508cdaa65", "version": 1, "bodyDocId": null, "description": "Higher visual weight draws attention - use contrast strategically for emphasis", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-c7f61c64a03e", "evidence": "$259,600 vs $3,000,000 stark visual cost contrast.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-ca57d778cd60", "evidence": "Visual equation '1 chip = 16 boxes' makes ratio concrete.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Hierarchy", "slug": "visual-hierarchy", "status": "active", "bestFor": null, "canonId": "019dd956-b2ca-706b-937f-e52441086fd7", "version": 1, "bodyDocId": null, "description": "Order of perception through Size > Color > Position > Weight > Space", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-c2f6502c9012", "evidence": "Single chip vs grid of 16 boxes plus '=' sign and price tags.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-cc76bc902134", "evidence": "Specific 3.7x speedup vs prior gen.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 109, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Waterfall chart", "slug": "waterfall-chart", "status": "active", "bestFor": null, "canonId": "e2799da3-725f-44a9-a093-570f1f388e64", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "3caccf96-90f7-4063-9b87-facb10c50665", "evidence": "The slide mentions a comparison of TPU v3 and TPU v4 training performance across models, which is often visualized using a waterfall chart.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.20688+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 109, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "04bd1378-2252-4a0c-806d-0314bbc129aa", "evidence": "The slide compares performance across 8 models, which could be represented in a 2x4 or 2x2 matrix, but the exact structure is not clear.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.276346+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 110, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-d3d773aaa610", "evidence": "2.5x speedup vs V100.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 110, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-d61b76661837", "evidence": "Title: rise of MLOps signals industry shift.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chartjunk Elimination", "slug": "chartjunk-elimination", "status": "active", "bestFor": null, "canonId": "019dd956-b49e-7453-ab01-b60f338eeda1", "version": 1, "bodyDocId": null, "description": "Remove decorative elements that don't convey data: 3D effects, shadows, unnecessary gridlines", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "6a394b14-28e3-4ce4-8a01-1d583498ee8a", "evidence": "The slide uses a combination of bar and line charts to display GitHub project growth analysis and Google Search Trends for 'MLOps', which is an example of effective chart selection and junk elimination.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.34563+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-d951efff9228", "evidence": "MLOps framed as 'DevOps for ML'.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "338fb3f8-e47e-45dd-8902-c96422fb4988", "evidence": "components: list/numbered: Sample (Hong Kong Monetary Authority*) 1. Board and Senior Management accountable for the outcome of AI applications", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:20:48.917267+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 112, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-de2334eb1b7e", "evidence": "Title: AI drives revenue and reduces costs by function.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 113, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-e05d30596c26", "evidence": "Function-by-function comparison view.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 113, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-e46d25a90315", "evidence": "11 use cases for high performers vs 3 for average.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 114, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Heatmap matrix", "slug": "heatmap-matrix", "status": "active", "bestFor": null, "canonId": "019de52d-2363-7248-9aa8-cd0bcf077891", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "704e76a1-e1f2-441c-9793-c45895d4143c", "evidence": "The slide uses a heatmap to display organizations' AI capabilities across different industries, which is an example of a heatmap matrix.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.445951+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 114, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "fe71d316-b762-41cf-83be-2f3acaa0a435", "evidence": "The slide uses a bar chart to display UiPath ARR growth, which is a common chart type for showing growth over time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.514726+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 115, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-e881347ba74e", "evidence": "7,000 enterprise customers UiPath.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 115, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-eff20cde4b5b", "evidence": ">90% success on 8-turn calls.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 116, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-f29bd8756312", "evidence": "1 week intervention timeline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 117, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-f4a7fce29219", "evidence": "40% improvement metric cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 118, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-fa27e1fa8328", "evidence": "30x increase in pandemic phishing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 120, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "062fe0a1-276f-40ef-96cb-3013abc220e0", "evidence": "The slide seems to be targeted at an audience interested in AI applications, specifically in identity document verification.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.090808+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 121, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "7771a736-a551-4853-adf1-cfeda0554aa8", "evidence": "The use of AI feature map visualization and annotated French ID card as visual aids.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:57:04.131173+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 121, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "f9970606-2f7f-4346-962a-a144c6c330c5", "evidence": "The use of a bar chart to display adverse media coverage per geography.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.255115+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-fcd6ecdf2fc0", "evidence": "85% risk data coverage.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "1db4bb10-633a-40eb-8d79-a00eff501aaa", "evidence": "The slide presents a narrative about AI-powered content analysis for risk classification.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.20968+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "d26c760c-1f05-45e2-b92a-e496ac0064d9", "evidence": "The slide presents a comparison of lead times before and after using machine translation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.33772+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chartjunk Elimination", "slug": "chartjunk-elimination", "status": "active", "bestFor": null, "canonId": "019dd956-b49e-7453-ab01-b60f338eeda1", "version": 1, "bodyDocId": null, "description": "Remove decorative elements that don't convey data: 3D effects, shadows, unnecessary gridlines", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "cf75213f-f6f9-4156-a5b8-e537e9a6b3b4", "evidence": "The use of simple and clear charts to display data.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.377361+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-00823accbdc8", "evidence": "20 weeks -> 2 weeks lead time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Aha! Moment", "slug": "aha-moment", "status": "active", "bestFor": null, "canonId": "019dd956-a5b5-7763-820e-b299c5f27a32", "version": 1, "bodyDocId": null, "description": "The moment when the audience realizes the value - design for insight revelation", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "31d47980-5771-4838-9199-3b1303f29a1a", "evidence": "The slide presents a surprising fact about BERT language model's adoption.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.460148+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 124, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9726c492-b718-4408-b4fb-6cc7777ce8cf", "evidence": "The use of a screenshot to highlight the differences in search results.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.500606+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 124, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9fb87768-d034-43b1-b01d-49688f1c3e74", "evidence": "The slide implies that the solution can be implemented to achieve significant labour reduction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.580508+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 125, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-04c4b0e27f30", "evidence": "70% direct labour reduction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 125, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Big Idea Formula", "slug": "big-idea-formula", "status": "active", "bestFor": "Any narrative-driven persuasive deck whose job is to move the audience: keynotes, fundraising pitches, capital-allocation board asks, all-hands behaviour-change talks, conference talks meant to shift mental models. Use it before structuring; pair with Pyramid when the deck must also withstand line-by-line scrutiny.", "canonId": "019dd956-7dd1-74d7-9866-ffdef19ecfbf", "version": 1, "bodyDocId": "c30d7484-69e0-4e21-82cf-20fa0966869e", "description": "Nancy Duarte's articulation of how to frame a presentation's central message: the speaker's unique point of view plus what's at stake for the audience, expressed as one complete declarative sentence. The narrative-driven counterpart to Minto's analytical governing thought.", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "ddbbc789-031a-41c8-ad22-31a3fdc36e14", "evidence": "The slide presents a key idea about automating CNC machine programming.", "pageRefs": null, "priority": "Core", "whenToUse": "Any narrative-driven persuasive deck whose job is to move the audience: keynotes, fundraising pitches, capital-allocation board asks, all-hands behaviour-change talks, conference talks meant to shift mental models. Use it before structuring; pair with Pyramid when the deck must also withstand line-by-line scrutiny.", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.703251+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": "Solves the topic-not-thesis trap and the stake-less recommendation trap simultaneously. A single declarative sentence with a verb forces the speaker's stance into the room; routing the consequence through the word 'you' translates internal urgency into the audience's currency. The result is short enough to be repeated after the meeting — and that repetition is the persuasion mechanism.", "antipattern": "A topic dressed as a thesis ('Sustainability Roadmap 2026'), a stake attached to the speaker rather than the audience, two ideas spliced with 'and', abstract verbs (leverage / unlock / empower) that survive deletion, or writing the Big Idea after the deck is already built so the slides don't actually defend it.", "cardinality": null, "narrativePurpose": "Forces the deck to commit to a single declarative thesis before slide one — a sentence that names the speaker's stance and routes a consequence through the audience's currency. Acts as the filter every slide, anecdote, and chart must ladder up to." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-0a570d7926a8", "evidence": ">95% productivity target.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "24fe0dac-49b5-4d0a-8b5f-6b3e9c8e4b5a", "evidence": "The slide presents a diagram of the 3D File -> CNC Machine -> Finished Parts process.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.66245+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Audience Segmentation", "slug": "audience-segmentation", "status": "active", "bestFor": null, "canonId": "019dd956-8079-7046-8e6b-ecdf1bf15590", "version": 1, "bodyDocId": null, "description": "Grouping audience by attitude toward your message: Champions, Neutrals, Skeptics", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "aa836a9c-6ec8-4604-aedc-93b00927b87e", "evidence": "The slide seems to be targeted at an audience interested in NLP applications.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.821192+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-0df4d7861454", "evidence": "5M pip installs / 2,500 models / 164 languages.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "42815d8d-5692-463b-901b-f0a365bb99a1", "evidence": "The slide presents a bar chart with a clear and simple design.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.782008+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-12a1937c50b8", "evidence": "'Cambrian explosion' biological metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Iconography", "slug": "iconography", "status": "active", "bestFor": null, "canonId": "019dd9e1-4d71-77c5-a8d6-20a534104995", "version": 1, "bodyDocId": null, "description": "Icon grids and pictograms standing in for text.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "75285175-9646-4a46-bbee-280a4dd3d618", "evidence": "The slide presents a grid of industry logos.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.900446+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 128, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null }, { "tool": { "name": "Social Proof", "slug": "social-proof", "status": "active", "bestFor": null, "canonId": "019dd956-990d-7620-82b3-6dd286f34568", "version": 1, "bodyDocId": null, "description": "People adapt behavior based on what others do - show adoption and testimonials", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "f5499e89-9015-458a-9cc5-bc67867a57d9", "evidence": "The slide mentions the number of downloads and contributors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.939883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 128, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-161823e24473", "evidence": "Title: AI funding remains strong despite COVID.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "37d80e1d-c29c-454e-871d-68fcc102d513", "evidence": "The use of a bar chart and line chart to display funding data.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:05.021319+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-1899be92cca9", "evidence": "$25B+, 350+ deals.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Storytelling Effect", "slug": "storytelling-effect", "status": "active", "bestFor": null, "canonId": "019dd956-981b-74f9-9096-5c8df03dab5a", "version": 1, "bodyDocId": null, "description": "Stories are remembered up to 22x better than facts alone", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "2dbb4282-6a8a-41ad-9d36-58f6d758a94f", "evidence": "The slide presents a narrative about funding trends.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:05.059254+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d3aff7c4-f141-42e8-8e5a-d4061bf5b812", "evidence": "The slide is clearly a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:05.138416+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 130, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "5034c473-5339-445c-9f90-7d1f1cc157f3", "evidence": "The slide mentions specific researchers and their work.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:05.217939+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 131, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-21dfe70c07ee", "evidence": "50% of world allows it.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "263a3bd8-78d4-4e3e-b996-e9c7777fd3c8", "evidence": "The slide presents a narrative about facial recognition adoption.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:05.299403+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ed0-1dd22d777e68", "evidence": "World map visualises facial-recognition prevalence.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Anadiplosis", "slug": "anadiplosis", "status": "active", "bestFor": null, "canonId": "019dd956-a0a5-75f9-a98c-7b7d8ee377a1", "version": 1, "bodyDocId": null, "description": "Ending one clause with a word that begins the next", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "b9437467-b4b6-4ad1-8ef9-e7cbe9805ac1", "evidence": "These examples are likely just the tip of the iceberg.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:51.949034+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "2c8cdab9-f3f9-494f-bb2e-1fb378d3fc40", "evidence": "Facial recognition: From potential risks to wrongful arrests", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:51.982068+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-2aaa2c756640", "evidence": "Wrongful arrests framed emotionally.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-271adfb71b54", "evidence": "'Tip of the iceberg' metaphor in callout.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-2f33b142cff0", "evidence": "$650M settlement specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 134, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Aha! Moment", "slug": "aha-moment", "status": "active", "bestFor": null, "canonId": "019dd956-a5b5-7763-820e-b299c5f27a32", "version": 1, "bodyDocId": null, "description": "The moment when the audience realizes the value - design for insight revelation", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "cf66170a-a225-45ac-af75-64274fa569cf", "evidence": "A New York Times investigation revealed that Clearview scraped billions of images and then licensed their “search engine for faces” to over 600 law enforcement agencies.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.085881+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-32b60627f4a2", "evidence": "Billions of images, 600 agencies.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-34160a750f56", "evidence": "'Search engine for faces' framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Evidence Matrix", "slug": "evidence-matrix", "status": "active", "bestFor": null, "canonId": "019dd956-7ecc-74c2-b46c-0b9c1f4c73d7", "version": 1, "bodyDocId": null, "description": "Planning visual assets: Claim → Evidence Needed → Data Source → Visual Type", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "1b2e565d-7407-4a10-ad6f-9b40421a249b", "evidence": "The ImageNet team recruited 12 graduate students representing 4 countries of origin, male and female genders, and a handful of racial groups from diverse backgrounds to systematically identify offensive cate", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:52.183939+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 137, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-38cdd6b8e936", "evidence": "Callout: ImageNet creators reduce bias.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 137, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-3d30f02a937e", "evidence": "'Move fast and break things' analogy critiqued.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 138, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-435098f2846e", "evidence": "Title: research agendas overlap with US Army.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 144, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-475f94d77ecc", "evidence": "$800M military AI investment.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 145, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "13133e77-01cb-49b7-a09b-702eabc08224", "evidence": "The top AI systems from Heron Systems and Lockheed Martin both made use of Deep Reinforcement Learning - the same approach applied by DeepMind in their work on Go and StarCraft II and OpenAI in their work on", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.427096+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-4a23826e583b", "evidence": "'AlphaDogfight' brand-extends AlphaGo metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-4e960ede1d8b", "evidence": "Callout: dual-use AI migrates from games to military.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f1fceed3-5cf8-4985-8a95-bb76f4cf62da", "evidence": "The US continues to emphasize importance of AI leadership to its military", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.495095+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 149, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-5212ac8e0be6", "evidence": "2024 specific milestone.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 149, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "9dc775cd-9276-40c5-9bb2-be2c9cdcf30a", "evidence": "The US Department of Defense, The US Intelligence Community, China, and the OECD all develop or adopt their own AI Policy documents.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.562856+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 150, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "95cce96d-88c5-4041-b48f-a14704a8c9e0", "evidence": "Proliferating educational content and tools through the TensorFlow community.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.664152+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 152, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-5608dcb22f05", "evidence": "Callout flags Huawei chip supply running out.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 154, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "2cf0d011-cc42-4cb3-b8bd-abccb560720a", "evidence": "TSMC grows in importance to the US semiconductor strategy.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.798614+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 155, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "c4f1453e-adcb-403a-997d-34baae6b469e", "evidence": "Performance comparison table for different process nodes.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.86527+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 156, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-59cc2faad8c3", "evidence": "TSMC R&D = SMIC revenue comparison.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 156, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "381985e4-6b71-4765-8a7c-2cd694a508ac", "evidence": "China is the world's largest importer of semiconductors, totalling $200B/year.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.141172+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "cd44405a-6c79-48e3-8d70-44a2151ef64d", "evidence": "chart/line: Cumulative total return of SMIC vs TSM from Oct 2019 to Jul 2020.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.170822+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-5ee1dbf0bcb9", "evidence": "$29B Chinese fund / Cambricon $370M IPO.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "f8c9decc-a7ab-49ed-8371-8a11379bc982", "evidence": "TSMC employees are offered as much as 2.5x their annual salary and bonuses to leave.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.232369+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6355d4859f76", "evidence": "100 engineers / 2.5x salary.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9c531313-1ad6-4576-918c-ba5cbd582087", "evidence": "list/bullet: Government-backed Quanxin Integrated Circuit Manufacturing (QXIC, founded 2019) and Wuhan Hongxin Semiconductor Manufacturing Co. (HSMC, founded 2017) are are led by ex-TSMC executives and have each hired 50", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.264253+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "de38d332-3b30-4ad6-a43e-5f0fa93b57c4", "evidence": "chart/line: Chip production swings to Asia", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:55.356199+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-67a02a7c0f89", "evidence": "Callout contrasts US design vs manufacturing share.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f0ab2e40-e6f5-42ea-85bc-986b3d1352ca", "evidence": "The CHIPS for America Act would earmark $22B to subsidise US manufacturing of chips.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.325461+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "11cc0c2b-33a2-4ae7-91e3-117792d1ec40", "evidence": "list/bullet: December 2016: US and Germany block $723M bid by China’s Fujian Grand Chip Investment Fund (China) for Aixtron.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.417345+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 160, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Audience frame", "slug": "audience-frame", "status": "active", "bestFor": null, "canonId": "2a1eefbd-5be8-4600-898b-49dbfcf73ec8", "version": 1, "bodyDocId": null, "description": "Tools for modelling and segmenting the audience (persona, JTBD, segmentation).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "6557fe4a-294f-4cf9-a235-6f211d8346bc", "evidence": "The State of AI Report and AI Nationalism essay predicted that political leaders would start to question whether acquisitions of key AI startups should be blocked.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.480002+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 161, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null }, { "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ab7b6de1-4799-4c2a-8d71-091e43be013d", "evidence": "The slide quotes Hermann Hauser, a founder and investor, lending credibility to the argument.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-17 03:53:51.398039+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6984a298eb94", "evidence": "Founder quote about Arm being 'torn'.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Problem-Agitate-Solve (PAS)", "slug": "problem-agitate-solve-pas", "status": "active", "bestFor": null, "canonId": "019dd956-869f-7564-8d42-de54ed29ab90", "version": 1, "bodyDocId": null, "description": "Identify problem, agitate the pain, present the solution", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "d6e1de4f-753a-4153-8f77-792b098deb60", "evidence": "The slide presents a problem (the likely sale of Arm to NVIDIA), agitation (job losses, destruction of Arm's neutrality), and an implied need for a solution.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-17 03:53:51.359706+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "682f4a92-a264-43bb-af95-412d91438cb8", "evidence": "chart/bar-vertical: Federal budget for non-defense AI R&D", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.542119+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 163, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6f82c9802d77", "evidence": "$1,503M AI budget figure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 163, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-75a5a2e2bd73", "evidence": "$100B bipartisan act.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-713a78db3bc1", "evidence": "Quote: 'superpower of the future' framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Monroe's Motivated Sequence", "slug": "monroes-motivated-sequence", "status": "active", "bestFor": null, "canonId": "019dd956-8758-700f-83cf-d260a54ad24f", "version": 1, "bodyDocId": null, "description": "Five-step persuasion: Attention, Need, Satisfaction, Visualization, Action", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "2c8ad468-6cdb-4ac6-ba08-961dfc30eb96", "evidence": "The bi-partisan act explicitly frames a race to lead in AI: “The country that wins the race in key technologies—such as artificial intelligence, quantum computing, advanced communications, and advanced manufacturing—will be the superpower of the future.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.602981+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Journey frame", "slug": "journey-frame", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ab2-7354-8c61-721dee94c0d1", "version": 1, "bodyDocId": null, "description": "Sequential stages over time (timeline, before-during-after).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "04dd000e-aaa3-4519-beb4-01f8c36ead5b", "evidence": "AI pilot zones are encouraged to “carry out AI based policy experiments” and “carry out AI based social experiments”.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.663168+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 165, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "8fc7f342-184e-42ca-9a68-c13128bbcea0", "evidence": "Pressure on the UK to choose between the US and China.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.727183+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 166, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Map / spatial", "slug": "map-spatial", "status": "active", "bestFor": null, "canonId": "019dd9e1-4f52-724d-8b4e-b5ed8925a0a0", "version": 1, "bodyDocId": null, "description": "Geographic or spatial visualisation.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "1b3d27cb-6e92-404a-9984-9c12b70df1ca", "evidence": "image/map: World map showing AI strategy status and focus", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.787202+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 167, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ed0-7bb067e530ec", "evidence": "World map shows national AI strategy adoption.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 167, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "989c4b53-2b33-4076-b2fe-f6369389467c", "evidence": "chart/line: Effective tax rates for U.S. companies, by type of expenditure", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.84925+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 168, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-7c6dd2d5227f", "evidence": "Callout: tax code drives socially undesirable automation.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 168, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "ed95d14d-c371-45d9-8577-924a40e68b48", "evidence": "chart/bar-stacked: Change in workforce due to AI adoption, % of respondents", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.911199+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 169, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "fce72d03-ed96-4fab-8b98-ece35d56c923", "evidence": "Bengio, Hassabis, Ng and other AI research leaders unite at NeurIPS 2019 in a call to action for climate change", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:06.57547+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Authority Bias", "slug": "authority-bias", "status": "active", "bestFor": null, "canonId": "019dd956-c38d-70b9-a284-b2b479763356", "version": 1, "bodyDocId": null, "description": "Expert opinions are weighted more heavily - cite credible sources", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-80c4a54edd76", "evidence": "Bengio, Hassabis, Ng named as endorsers.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "fd2b35c1-d26a-441d-9958-6b3d2316c791", "evidence": "list/bullet: Automatic monitoring with remote sensing (e.g. deforestation, climate disasters).", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.535538+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ff11b82e-9d3a-4dfd-bfa3-bc2a3ecb076d", "evidence": "Section 5: Predictions", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.655403+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 171, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-84d1b923990a", "evidence": "Eight crisp predictive statements.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Closing Techniques", "slug": "closing-techniques", "status": "active", "bestFor": null, "canonId": "019dd956-af13-774b-a4fd-7db0a3a88d51", "version": 1, "bodyDocId": null, "description": "Strong endings: Callback, challenge, vision, clear CTA, memorable line", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-8d1e7f203895", "evidence": "Predictions act as forward-looking close.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b259a1e8-8559-4221-bb46-635ef97d16bd", "evidence": "Section 6: Conclusion", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.773252+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 173, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Storytelling: the closing ask", "slug": "closing-ask", "status": "active", "bestFor": null, "canonId": "019df269-24e0-70ab-8dd9-983832c1c665", "version": 1, "bodyDocId": "019df269-24bc-7554-b88d-245127606b8d", "description": "The single most important slide in the deck — after the cover. The final", "familyLabel": "narrative-block", "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "agent": "storyteller", "layer": "slide", "agents": [ "storyteller" ], "matchId": "34e3f689-2033-4073-8f24-64eda4f5ce16", "evidence": "callout: We would appreciate any and all feedback on how we could improve this Report further, as well as contribution suggestions for next year's edition.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.851971+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 174, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "d2682600-c126-429f-a3c8-ab94c52a493b", "evidence": "The slide is explicitly titled Executive Summary and organizes key findings into thematic categories.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "frameworkName": "executive-summary" }, { "matchId": "ac9c4ac3-b62d-4393-b946-39bc8d290a4f", "evidence": "The slide maps predictions to grades and evidence in a structured table format.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:29.357+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 9, "frameworkName": "evidence-matrix" }, { "matchId": "019dd95a-1ca5-70bb-bba0-7887504640e9", "evidence": "Title cites 'empirical scaling laws of neural language models'.", "framework": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "frameworkName": "Neural Scaling Laws" }, { "matchId": "76731a61-e1bb-4626-81f7-a4a472910e0a", "evidence": "Empirical scaling laws of neural language models", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "frameworkName": "Scaling Laws" }, { "matchId": "03c95046-1288-44a7-809f-957bac66e973", "evidence": "Charts demonstrate the relationship between model parameters, compute, and performance.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "frameworkName": "Scaling Laws" }, { "matchId": "39c933a2-744a-43a6-be17-8b12357cc35d", "evidence": "The slide explicitly discusses causal inference as a methodology for ML.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "frameworkName": "causal-inference" }, { "matchId": "019dd95a-1ca5-70bb-bba0-7f6ba50cda91", "evidence": "Slide titled 'Causal Inference', Judea Pearl pictured.", "framework": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "frameworkName": "Causal Inference (Pearl Ladder)" }, { "matchId": "54022e5b-b551-4f18-98bd-af02d2e8bda7", "evidence": "Diagram illustrating the interaction between model owners, clients, and startups using encryption.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 60, "frameworkName": "federated-learning-workflow" }, { "matchId": "9c5c3748-c463-407e-a865-3bdec5d7acc8", "evidence": "The slide links professor departures to student outcomes over a specific time horizon.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "frameworkName": "causal-chain" }, { "matchId": "f59dc316-82f9-41a4-82bb-9a3e68d25976", "evidence": "Flow chart showing migration between career stages", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:39.312+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 71, "frameworkName": "Sankey Diagram" }, { "matchId": "10999e37-7235-4cf6-b9e7-5b9fa572e5a5", "evidence": "The slide explicitly describes spinning off drug assets into independent entities.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "frameworkName": "breakup-spinoff" }, { "matchId": "5de1efea-c4ed-43e6-9b40-002dd4be742d", "evidence": "Diagram showing central human figure with various data inputs radiating outwards.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "frameworkName": "hub-spoke" }, { "matchId": "019dd95a-1ca5-70bb-bba0-836d67706a3c", "evidence": "Callout names FDA TPLC regulatory approach.", "framework": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "frameworkName": "FDA Total Product Lifecycle (TPLC)" }, { "matchId": "d864565c-121c-4ddb-89bd-03fc3c360b14", "evidence": "Diagram showing the total product lifecycle approach for AI/ML software", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "frameworkName": "FDA TPLC Regulatory Framework" }, { "matchId": "6b1d9cd2-6564-4c4f-837f-2527c06d4759", "evidence": "Mentioned as an international framework for clinical trial reporting", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "frameworkName": "CONSORT-AI" }, { "matchId": "da2dbdbb-fbab-4100-b43c-b85d2e9dcdb0", "evidence": "Mentioned as an international framework for clinical trial protocols", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "frameworkName": "SPIRIT-AI" }, { "matchId": "83cc3738-c80b-41c3-886a-85dca485e7ac", "evidence": "Mentions owning key technology components of their value chain.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "frameworkName": "value-chain" }, { "matchId": "019dd95a-1ca5-70bb-bba0-87de388e10d2", "evidence": "Callout: 'own a key technology component of their value chain'.", "framework": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "frameworkName": "Value Chain" }, { "matchId": "092fcd99-dd06-4461-91a1-86777b38ea5a", "evidence": "The text explicitly mentions 'supervised ML seems to follow an S-Curve' and the chart depicts this trajectory.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "frameworkName": "S-Curve" }, { "matchId": "2106a44d-0ccc-429c-a92f-b5da285918bd", "evidence": "Chronological list of dataset releases", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:43.092+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 103, "frameworkName": "timeline" }, { "matchId": "d32ea224-1a08-428e-a04b-cbff2d67405c", "evidence": "The slide depicts the autonomous driving stack as a sequential process flow.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 104, "frameworkName": "process" }, { "matchId": "4068c7a2-1cce-465e-a435-28ecdbdb0749", "evidence": "Mentions Environmental, Social and Governance pillars", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 119, "frameworkName": "ESG Framework" }, { "matchId": "8a21a9bb-9a1f-4dd5-b772-3526cb3c8cb7", "evidence": "The text explicitly mentions 'These examples are likely just the tip of the iceberg.'", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "frameworkName": "iceberg-model" }, { "matchId": "8c8d565d-0abd-4ff4-b52e-eb1d5cf2462c", "evidence": "The slide explicitly references the People + AI Guidebook and lists its core modules.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 152, "frameworkName": "People + AI Guidebook" } ], "patterns": [] }, "storymakers": { "model": "claude-cli-sonnet-4.6", "backend": null, "rawJson": { "arc": { "id": 11, "slug": "triple-take", "beats": [ { "name": "The Facts (What)", "end_page": 129, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "start_page": 10 }, { "name": "The Implications (So What)", "end_page": 170, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "start_page": 130 }, { "name": "The Action (Now What)", "end_page": 174, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "start_page": 171 } ], "evidence": "Facts (Research/Talent/Industry pp.10-129) -> Implications (Politics pp.130-170) -> Action (Predictions p.172).", "confidence": 60 }, "loops": [ { "id": 4, "slug": "04-tale-two-worlds", "end_page": 14, "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "start_page": 11 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 23, "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "start_page": 15 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 29, "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "start_page": 24 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 46, "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "start_page": 30 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 57, "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "start_page": 47 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 62, "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "start_page": 58 }, { "id": 4, "slug": "04-tale-two-worlds", "end_page": 69, "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "start_page": 64 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 76, "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "start_page": 70 }, { "id": 39, "slug": "39-benchmark-gap", "end_page": 81, "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "start_page": 77 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 89, "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "start_page": 83 }, { "id": 36, "slug": "36-maturity-curve", "end_page": 92, "evidence": "FDA TPLC framework -> SPIRIT-AI -> first CMS reimbursement progression.", "position": 11, "objective": "Track AI medical regulation maturing toward reimbursement", "confidence": 60, "start_page": 90 }, { "id": 10, "slug": "10-iceberg", "end_page": 106, "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "start_page": 93 }, { "id": 16, "slug": "16-david-goliath", "end_page": 110, "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "start_page": 107 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 115, "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "start_page": 111 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 128, "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "start_page": 116 }, { "id": 26, "slug": "26-ripple-effect", "end_page": 140, "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "start_page": 131 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 143, "evidence": "Three slides clustering legislative responses to algorithmic risks.", "position": 17, "objective": "Show regulators waking up to deepfakes, ADM and model bias", "confidence": 60, "start_page": 141 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 149, "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "start_page": 144 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 152, "evidence": "DoD principles, NeurIPS ethics, Google fairness/privacy efforts.", "position": 19, "objective": "Survey emerging responsible AI codes and corporate efforts", "confidence": 60, "start_page": 150 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 160, "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "start_page": 153 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 167, "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "start_page": 161 }, { "id": 2, "slug": "02-pattern-hunter", "end_page": 170, "evidence": "Tax-code automation incentive, jobs at risk, climate change call to action.", "position": 22, "objective": "Surface labor and climate implications of AI", "confidence": 60, "start_page": 168 } ], "slide_tools": [ { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title 'Executive Summary' is topic, but bullets state insights per section.", "confidence": 70 }, { "id": 116, "slug": "chunking", "layer": "Loop", "evidence": "Bullets chunked under Research/Talent/Industry/Politics headers.", "confidence": 80 }, { "id": 153, "slug": "core-message-extraction", "layer": "Slide", "evidence": "Distills 170+ pages into one screen of takeaways.", "confidence": 70 } ], "page_number": 7 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states insight: 'Only 15% of papers publish their code'.", "confidence": 90 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific 15% statistic anchors abstract claim.", "confidence": 85 }, { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: 'centralization of AI talent and compute as a huge problem.'", "confidence": 80 } ], "page_number": 11 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: PyTorch outpacing TensorFlow as leading indicator.", "confidence": 90 }, { "id": 129, "slug": "chart-selection-guide", "layer": "Slide", "evidence": "Multi-line time series tracks framework share by conference.", "confidence": 85 }, { "id": 78, "slug": "contrast-pairs", "layer": "Loop", "evidence": "PyTorch vs TensorFlow pair set up across loop.", "confidence": 80 } ], "page_number": 13 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: '47% of these implementations are based on PyTorch vs 18%'.", "confidence": 90 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "47% vs 18% specific deltas.", "confidence": 85 } ], "page_number": 14 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Memorable framing: 'Welcome to the Billion Parameter club'.", "confidence": 85 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'Billion Parameter club' membership metaphor.", "confidence": 75 }, { "id": 74, "slug": "inductive-reasoning", "layer": "Loop", "evidence": "Loop builds case from multiple model-size data points.", "confidence": 75 } ], "page_number": 15 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states the empirical scaling law conclusion.", "confidence": 90 }, { "id": 129, "slug": "chart-selection-guide", "layer": "Slide", "evidence": "Power-law scaling chart appropriate for log relationship.", "confidence": 70 } ], "page_number": 16 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "'$1 per 1,000 parameters' makes abstract cost concrete.", "confidence": 85 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Per-parameter pricing analogy to commodity goods.", "confidence": 60 } ], "page_number": 17 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "'22 TPU v3 core years' translates abstract compute to time.", "confidence": 85 } ], "page_number": 18 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: outrageous costs for incremental gains.", "confidence": 95 }, { "id": 158, "slug": "von-restorff-effect", "layer": "Slide", "evidence": "Red box highlights ImageNet rows in big table.", "confidence": 80 }, { "id": 138, "slug": "annotation", "layer": "Slide", "evidence": "Red rectangle annotation calls attention to key benchmark.", "confidence": 80 }, { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "'$100B billion' framing plays on shock.", "confidence": 70 } ], "page_number": 19 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states 'larger model needs less data'.", "confidence": 85 }, { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout draws implication: advantage to large companies.", "confidence": 80 } ], "page_number": 20 }, { "tools": [ { "id": 122, "slug": "story-moments", "layer": "Slide", "evidence": "Concrete Hausa/Uzbek case proves abstract scaling claim.", "confidence": 60 } ], "page_number": 21 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: deep learning getting more efficient over time.", "confidence": 85 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific '2x every 16 months' rule.", "confidence": 80 } ], "page_number": 22 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states caveat: small models can trump large.", "confidence": 80 }, { "id": 152, "slug": "unexpected-pattern", "layer": "Slide", "evidence": "Counterpoint to scaling-law narrative breaks the pattern.", "confidence": 75 } ], "page_number": 23 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: new transformer LMs unlocking NLP use cases.", "confidence": 80 } ], "page_number": 24 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific 90% C++->Java translation accuracy.", "confidence": 80 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Conversational title 'Computer, please...' frames as dialog.", "confidence": 65 } ], "page_number": 25 }, { "tools": [ { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Conversational 'Computer, can you...' framing parallels prior slide.", "confidence": 65 }, { "id": 87, "slug": "parallel-structure", "layer": "Loop", "evidence": "Parallel slide titles 25-26 use same conversational format.", "confidence": 70 } ], "page_number": 26 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'NLP benchmarks take a beating'.", "confidence": 80 }, { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "'Take a beating' is dramatic verb choice.", "confidence": 60 } ], "page_number": 27 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: GPT-3 has knowledge gaps.", "confidence": 70 } ], "page_number": 28 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "GPT-2 fed images example illustrates generality.", "confidence": 70 } ], "page_number": 29 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'Biology is experiencing its AI moment'.", "confidence": 90 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'AI moment' analogy frames whole biology section.", "confidence": 80 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "21,000 papers in 2020 anchors the claim.", "confidence": 80 } ], "page_number": 30 }, { "tools": [ { "id": 165, "slug": "picture-superiority-effect", "layer": "Slide", "evidence": "Cell painting visuals carry the case.", "confidence": 70 } ], "page_number": 31 }, { "tools": [ { "id": 122, "slug": "story-moments", "layer": "Slide", "evidence": "COVID drug screen as flagship use case.", "confidence": 60 } ], "page_number": 32 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: patients can be treated before vision loss.", "confidence": 75 } ], "page_number": 33 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states clinical performance gains.", "confidence": 85 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "9.40% specific delta cited.", "confidence": 75 } ], "page_number": 34 }, { "tools": [ { "id": 169, "slug": "authority-bias", "layer": "Slide", "evidence": "Photo + Judea Pearl/Bengio name-drop authority.", "confidence": 75 } ], "page_number": 35 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "'Top 25% of doctors' concrete benchmark.", "confidence": 75 } ], "page_number": 36 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "98% synthesizable molecules statistic.", "confidence": 70 } ], "page_number": 38 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'ML will generate a synthesis plan faster than you can'.", "confidence": 80 } ], "page_number": 39 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states GNN-led antibiotic discovery.", "confidence": 80 } ], "page_number": 41 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "10x MSE improvement quantified.", "confidence": 75 } ], "page_number": 42 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific 72%/33%/16% hit rates by target.", "confidence": 75 } ], "page_number": 43 }, { "tools": [ { "id": 224, "slug": "singularity-effect", "layer": "Slide", "evidence": "Loss of smell as single memorable predictor of COVID.", "confidence": 65 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "4 million contributors data point.", "confidence": 80 } ], "page_number": 45 }, { "tools": [ { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "Pro bono COVID antiviral framing builds goodwill.", "confidence": 65 } ], "page_number": 46 }, { "tools": [ { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "Wimbledon nostalgia angle anchors abstract synthesis.", "confidence": 60 } ], "page_number": 47 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: attention turning to vision tasks.", "confidence": 80 } ], "page_number": 48 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'won't drain your battery'.", "confidence": 85 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "4-9x parameter reduction quantified.", "confidence": 80 } ], "page_number": 56 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: federated learning research booming.", "confidence": 80 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "5x growth and 2020 H1 surpasses all 2019.", "confidence": 80 } ], "page_number": 58 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "15 mins -> 40 sec specific speedup.", "confidence": 75 } ], "page_number": 61 }, { "tools": [ { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "'Quantum supremacy' dramatic framing.", "confidence": 60 } ], "page_number": 62 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'The Great Brain Drain'.", "confidence": 90 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Brain drain metaphor.", "confidence": 80 }, { "id": 78, "slug": "contrast-pairs", "layer": "Loop", "evidence": "Academia vs Big Tech contrast set up across loop.", "confidence": 75 } ], "page_number": 64 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout questions adequacy of dual appointments.", "confidence": 75 } ], "page_number": 65 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'departures correlate with reduced graduate entrepreneurship'.", "confidence": 90 }, { "id": 138, "slug": "annotation", "layer": "Slide", "evidence": "Red boxes mark a/b/c rows in regression table.", "confidence": 85 }, { "id": 158, "slug": "von-restorff-effect", "layer": "Slide", "evidence": "Red highlight makes key coefficients pop.", "confidence": 75 } ], "page_number": 66 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "EUR100M / 50 professors specific.", "confidence": 75 } ], "page_number": 67 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$100M Roux Institute donation.", "confidence": 75 } ], "page_number": 68 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "29% of NeurIPS authors educated in China.", "confidence": 80 } ], "page_number": 70 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 54% of Chinese grads move to USA.", "confidence": 85 } ], "page_number": 71 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: US strong talent retainer post-PhD.", "confidence": 85 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "90% retention statistic.", "confidence": 80 } ], "page_number": 72 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: foreign vs American grad placement contrast.", "confidence": 80 } ], "page_number": 73 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "55%/40% UK/China sector splits.", "confidence": 70 } ], "page_number": 74 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "China 27%, Europe 11%, India 11% feeders.", "confidence": 80 } ], "page_number": 75 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout pulls implication of H-1B suspension.", "confidence": 75 } ], "page_number": 76 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: American institutions dominate NeurIPS.", "confidence": 80 }, { "id": 129, "slug": "chart-selection-guide", "layer": "Slide", "evidence": "Ranked bar chart of top contributors.", "confidence": 70 } ], "page_number": 77 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: ICML 2020 cements US lead.", "confidence": 80 } ], "page_number": 78 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Stanford 10x enrollment growth.", "confidence": 80 } ], "page_number": 79 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'Demand outstrips supply'.", "confidence": 90 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "3x more postings than views.", "confidence": 80 } ], "page_number": 80 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout ties COVID to LinkedIn posting drop.", "confidence": 70 } ], "page_number": 81 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title states first phase 1 trial milestone.", "confidence": 85 } ], "page_number": 83 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "EUR250M deal, 29 months timeline specific.", "confidence": 75 } ], "page_number": 84 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$239M mega round figure.", "confidence": 75 } ], "page_number": 85 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "100,000-fold candidate increase.", "confidence": 75 } ], "page_number": 87 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "1,100 participants, 77% accuracy stats.", "confidence": 80 } ], "page_number": 89 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout explains TPLC regulatory shift.", "confidence": 75 } ], "page_number": 90 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific 1% of 20,000 studies meet quality.", "confidence": 80 } ], "page_number": 91 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout explains why CMS approval is critical.", "confidence": 80 } ], "page_number": 92 }, { "tools": [ { "id": 165, "slug": "picture-superiority-effect", "layer": "Slide", "evidence": "US map visualises legislative spread.", "confidence": 70 } ], "page_number": 93 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'driverless cars are still not so driverless'.", "confidence": 90 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Only 3 of 66 companies allowed.", "confidence": 80 } ], "page_number": 94 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "0.000737% of human miles vivid contrast.", "confidence": 85 }, { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "'Microscopic' framing emphasises gap.", "confidence": 65 } ], "page_number": 95 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'Sketchy metrics: Tracking AV progress'.", "confidence": 90 }, { "id": 137, "slug": "lie-factor", "layer": "Slide", "evidence": "Calls out misleading 8,679% YoY metric.", "confidence": 60 } ], "page_number": 96 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$1.3B Zoox/Amazon deal cited.", "confidence": 75 } ], "page_number": 97 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$7B raised since July 2019.", "confidence": 80 } ], "page_number": 98 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout explains LiDAR vertical integration moat.", "confidence": 70 } ], "page_number": 100 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$2B / $3.4B SPAC valuations.", "confidence": 75 } ], "page_number": 101 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: 'Supervised learning and the cost of edge cases'.", "confidence": 85 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Anti-Moore's-Law analogy in callout.", "confidence": 70 } ], "page_number": 102 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: AI compute providers on gen 2+ products.", "confidence": 80 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "59.4B transistors / 823 sqmm specifics.", "confidence": 80 } ], "page_number": 107 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: M2000 faster training drops cost.", "confidence": 90 }, { "id": 132, "slug": "visual-hierarchy", "layer": "Slide", "evidence": "Single chip vs grid of 16 boxes plus '=' sign and price tags.", "confidence": 80 }, { "id": 159, "slug": "contrast-principle", "layer": "Slide", "evidence": "$259,600 vs $3,000,000 stark visual cost contrast.", "confidence": 85 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "Visual equation '1 chip = 16 boxes' makes ratio concrete.", "confidence": 70 } ], "page_number": 108 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Specific 3.7x speedup vs prior gen.", "confidence": 75 } ], "page_number": 109 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "2.5x speedup vs V100.", "confidence": 75 } ], "page_number": 110 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: rise of MLOps signals industry shift.", "confidence": 90 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "MLOps framed as 'DevOps for ML'.", "confidence": 70 } ], "page_number": 111 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: AI drives revenue and reduces costs by function.", "confidence": 85 }, { "id": 129, "slug": "chart-selection-guide", "layer": "Slide", "evidence": "Function-by-function comparison view.", "confidence": 65 } ], "page_number": 113 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "11 use cases for high performers vs 3 for average.", "confidence": 75 } ], "page_number": 114 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "7,000 enterprise customers UiPath.", "confidence": 75 } ], "page_number": 115 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": ">90% success on 8-turn calls.", "confidence": 80 } ], "page_number": 116 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "1 week intervention timeline.", "confidence": 75 } ], "page_number": 117 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "40% improvement metric cited.", "confidence": 75 } ], "page_number": 118 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "30x increase in pandemic phishing.", "confidence": 80 } ], "page_number": 120 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "85% risk data coverage.", "confidence": 75 } ], "page_number": 122 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "20 weeks -> 2 weeks lead time.", "confidence": 80 } ], "page_number": 123 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "70% direct labour reduction.", "confidence": 75 } ], "page_number": 125 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": ">95% productivity target.", "confidence": 70 } ], "page_number": 126 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "5M pip installs / 2,500 models / 164 languages.", "confidence": 80 }, { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'Cambrian explosion' biological metaphor.", "confidence": 70 } ], "page_number": 127 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: AI funding remains strong despite COVID.", "confidence": 80 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$25B+, 350+ deals.", "confidence": 75 } ], "page_number": 129 }, { "tools": [ { "id": 165, "slug": "picture-superiority-effect", "layer": "Slide", "evidence": "World map visualises facial-recognition prevalence.", "confidence": 75 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "50% of world allows it.", "confidence": 75 } ], "page_number": 132 }, { "tools": [ { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'Tip of the iceberg' metaphor in callout.", "confidence": 70 }, { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "Wrongful arrests framed emotionally.", "confidence": 75 } ], "page_number": 133 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$650M settlement specific.", "confidence": 80 } ], "page_number": 134 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "Billions of images, 600 agencies.", "confidence": 80 }, { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "'Search engine for faces' framing.", "confidence": 70 } ], "page_number": 135 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: ImageNet creators reduce bias.", "confidence": 70 } ], "page_number": 137 }, { "tools": [ { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'Move fast and break things' analogy critiqued.", "confidence": 75 } ], "page_number": 138 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Title: research agendas overlap with US Army.", "confidence": 80 } ], "page_number": 144 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$800M military AI investment.", "confidence": 75 } ], "page_number": 145 }, { "tools": [ { "id": 121, "slug": "metaphor-analogy", "layer": "Slide", "evidence": "'AlphaDogfight' brand-extends AlphaGo metaphor.", "confidence": 75 }, { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: dual-use AI migrates from games to military.", "confidence": 80 } ], "page_number": 148 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "2024 specific milestone.", "confidence": 70 } ], "page_number": 149 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout flags Huawei chip supply running out.", "confidence": 70 } ], "page_number": 154 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "TSMC R&D = SMIC revenue comparison.", "confidence": 75 } ], "page_number": 156 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$29B Chinese fund / Cambricon $370M IPO.", "confidence": 75 } ], "page_number": 157 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "100 engineers / 2.5x salary.", "confidence": 75 } ], "page_number": 158 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Callout contrasts US design vs manufacturing share.", "confidence": 85 } ], "page_number": 159 }, { "tools": [ { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "Founder quote about Arm being 'torn'.", "confidence": 70 } ], "page_number": 162 }, { "tools": [ { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$1,503M AI budget figure.", "confidence": 75 } ], "page_number": 163 }, { "tools": [ { "id": 156, "slug": "emotional-appeal", "layer": "Slide", "evidence": "Quote: 'superpower of the future' framing.", "confidence": 75 }, { "id": 154, "slug": "concrete-language", "layer": "Slide", "evidence": "$100B bipartisan act.", "confidence": 75 } ], "page_number": 164 }, { "tools": [ { "id": 165, "slug": "picture-superiority-effect", "layer": "Slide", "evidence": "World map shows national AI strategy adoption.", "confidence": 70 } ], "page_number": 167 }, { "tools": [ { "id": 119, "slug": "so-what-test", "layer": "Slide", "evidence": "Callout: tax code drives socially undesirable automation.", "confidence": 75 } ], "page_number": 168 }, { "tools": [ { "id": 169, "slug": "authority-bias", "layer": "Slide", "evidence": "Bengio, Hassabis, Ng named as endorsers.", "confidence": 80 } ], "page_number": 170 }, { "tools": [ { "id": 118, "slug": "action-titles", "layer": "Slide", "evidence": "Eight crisp predictive statements.", "confidence": 80 }, { "id": 116, "slug": "chunking", "layer": "Loop", "evidence": "Eight numbered predictions chunked clearly.", "confidence": 80 }, { "id": 124, "slug": "closing-techniques", "layer": "Slide", "evidence": "Predictions act as forward-looking close.", "confidence": 75 } ], "page_number": 172 } ], "slides_seen": 177, "deck_summary": "A 177-page research compilation that loosely follows a Triple Take macro arc (Facts in Research/Talent/Industry -> Implications in Politics -> Action in Predictions) but the dominant pattern is repeated Pattern Hunter loops where each slide stands alone with an action-title insight + concrete stat + so-what callout. Storymakers slide-level discipline is rigorous; Block-level frameworks (SCQA, MECE, hero arc) are absent.", "secondary_arcs": [ { "id": 13, "slug": "mountain", "beats": [ { "name": "Setup & Stakes", "end_page": 9, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "start_page": 1 }, { "name": "Rising Action", "end_page": 129, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "start_page": 10 }, { "name": "Climax", "end_page": 170, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "start_page": 130 }, { "name": "Resolution", "end_page": 177, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "start_page": 171 } ], "evidence": "Sections layer rising stakes from research curiosity to geopolitical conflict; predictions resolve.", "confidence": 40 } ], "images_inspected": 6, "slide_frameworks": [ { "frameworks": [ { "name": "Neural Scaling Laws", "slug": "neural-scaling-laws", "evidence": "Title cites 'empirical scaling laws of neural language models'.", "confidence": 80 } ], "page_number": 16 }, { "frameworks": [ { "name": "Causal Inference (Pearl Ladder)", "slug": "causal-inference", "evidence": "Slide titled 'Causal Inference', Judea Pearl pictured.", "confidence": 70 } ], "page_number": 35 }, { "frameworks": [ { "name": "FDA Total Product Lifecycle (TPLC)", "slug": "fda-tplc", "evidence": "Callout names FDA TPLC regulatory approach.", "confidence": 80 } ], "page_number": 90 }, { "frameworks": [ { "name": "Value Chain", "slug": "value-chain", "evidence": "Callout: 'own a key technology component of their value chain'.", "confidence": 60 } ], "page_number": 100 } ] }, "matchedAt": "2026-04-26 02:19:17+00", "slidesSeen": 177, "deckSummary": "A 177-page research compilation that loosely follows a Triple Take macro arc (Facts in Research/Talent/Industry -> Implications in Politics -> Action in Predictions) but the dominant pattern is repeated Pattern Hunter loops where each slide stands alone with an action-title insight + concrete stat + so-what callout. Storymakers slide-level discipline is rigorous; Block-level frameworks (SCQA, MECE, hero arc) are absent.", "imagesInspected": 6, "extractionSeconds": 490.62405 }, "score": { "backend": "codex", "scoredAt": "2026-05-02 20:36:47.389+00", "subScores": { "scqa_arc": 43, "action_titles": 82, "mece_structure": 70, "closing_strength": 49, "evidence_quality": 86, "clarity_of_thesis": 48, "production_quality": 70, "visual_storytelling": 78 }, "totalScore": 66, "coveragePct": 100, "explanations": { "scqa_arc": "The deck opens with scope and definitions then moves into a long topical evidence tour, so SCQA tension appears locally but not as a deck-level Situation-Complication-Question-Answer arc.", "action_titles": "Action-title coverage is high at 157 of 177 slides, with slide 11 as a clear example: \"AI research is less open than you think: Only 15% of papers publish their code.\"", "mece_structure": "The major sections on slides 10, 63, 82, 130, and 171 are mostly distinct, but case studies and policy/industry implications frequently overlap.", "closing_strength": "Slides 171-174 end with predictions and a feedback ask, but there is no clear recommendation, decision, or next-step agenda.", "evidence_quality": "Most claims are supported by charts, tables, metrics, quotes, or cases, including slides 19, 64, 113, 129, and 156.", "clarity_of_thesis": "Slides 3 and 7 define the scope and themes, but the first five slides do not state a single declarative thesis for the report.", "production_quality": "The deck is professionally structured with high action-title density, but very long titles and 56 overcrowded slides weaken whitespace and consistency.", "visual_storytelling": "Slides 13, 19, 64, 113, and 167 use charts, tables, and maps appropriately, though many dense or overcrowded pages reduce focus." }, "slidesAnalyzed": 177 }, "review": { "backend": null, "verdict": "This is a strong long-form AI market intelligence report, but only a partial Storymakers exemplar because it has disciplined evidence headlines without a sharp opening thesis or decisive closing action.", "reviewedAt": "2026-05-01 21:54:27.325+00", "slidesSeen": 177, "suggestions": [ "Move a one-slide 2020 thesis into slides 2-3 that states the answer before the inventory, for example: AI progress is accelerating, but advantage is concentrating around compute, talent, deployment capability, and national strategy.", "Rewrite section dividers as claim-led MECE pillars instead of nouns, such as 'Research: scale is concentrating advantage' or 'Politics: AI is becoming a regulated strategic asset'.", "Add a synthesis slide before slide 172 that ties the four analytical pillars into 3-5 implications, then make the predictions feel like consequences rather than a separate list.", "Shorten the longest titles and push secondary evidence into callouts so each headline passes the one-breath action-title test." ], "closingScore": 61, "openingScore": 44, "topStrengths": [ "Strong annual-report architecture: agenda on slide 3, prediction scorecard on slides 8-9, then clear Research/Talent/Industry/Politics/Predictions chapters.", "Many action titles carry the insight and metric upfront, especially slides 11, 19, 66, 94, 111, and 129.", "The deck builds credibility through breadth of evidence, moving from technical research trends to talent flows, industry adoption, and geopolitics." ], "topWeaknesses": [ "The opening buries the answer behind cover, author bios, contributors, and definitions; the thesis is not explicit in the first 5 slides.", "The middle behaves more like a curated evidence catalog than a laddered SCQA argument, with limited connective tissue between subtopics.", "The ending provides predictions but not a synthesized 'so what now' for investors, policymakers, operators, or researchers." ], "narrativeScore": 68, "pillarCritique": "The section dividers are mostly MECE domain pillars: Research, Talent, Industry, Politics, and Predictions (slides 10, 63, 82, 130, 171). They still read as report taxonomy rather than claim-led pillars, so the chapters risk becoming well-organized topic dumps.", "closingCritique": "The close has a forward-looking move via slide 172's '8 predictions for the next 12 months' and a feedback ask on slide 174. It lacks a memorable final synthesis or clear call to action for readers before the disclosure/about/cover end matter on slides 175-177.", "openingCritique": "The first 5 slides establish identity, credibility, scope, and vocabulary, but they do not state the thesis or stakes. The closest hook is slide 3's agenda/compilation framing, while the executive answer does not arrive until slide 7.", "extractionSeconds": 99.484, "narrativeCritique": "The deck has a recognizable setup -> tension -> analysis -> resolution arc: front matter and definitions (slides 1-7), accountability and market tensions (slides 8-9), pillar analysis across Research/Talent/Industry/Politics (slides 10-170), then predictions and thanks (slides 171-174). It is missing a crisp synthesis act that turns the 160-slide evidence base into one explicit answer or recommendation.", "titleQualityScore": 79, "titleQualityCritique": "Most content slides use declarative, evidence-bearing titles, such as slide 11 'AI research is less open than you think: Only 15% of papers publish their code' and slide 13 on PyTorch outpacing TensorFlow. Quality is diluted by topic labels and overlong titles, including slide 5 'Definitions' and slide 172 '8 predictions for the next 12 months'." }, "activistThesis": null, "pitchdeck": { "metadata": null, "profile": null }, "slides": [ { "page": 1, "type": "setup", "function": "front_matter", "imagePath": "https://imgproxy.kitesheet.com/pijUJkdwEx27C51DHaAgo8YM_ejYqysWxUjWNG5i0JQ/rs:fit:1200:1200:0/q:82/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9lZWE3NTI0YzU1NzAzNmY0L3AwMDEuanBn", "rawType": "cover", "block": null, "metadata": { "slideType": "cover", "slideTypeCanon": { "name": "Cover", "slug": "cover", "status": "active", "canonId": "019de52d-05f2-729d-9f93-e0504f0a2410", "version": 1, "description": null }, "function": "front_matter", "functionCanon": { "name": "Front matter", "slug": "front_matter", "status": "active", "canonId": "019de52d-133d-73f0-9e26-201eda96b098", "version": 1, "description": null }, "density": "balanced", "densityScore": 26, "componentCount": 6, "textChars": 78, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:29.988+00", "seconds": 4.646, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/1", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-1", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.1, "x": 0.02, "y": 0.02 }, "kind": "paragraph", "text": "stateof.ai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0d3a7636-c1fd-4489-b0c2-198b3e525093", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.85, "y": 0.9 }, "kind": "paragraph", "text": "Ian Hogarth", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "70dac249-dfd7-4e7b-bf99-cffdedc4e204", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.05, "y": 0.9 }, "kind": "paragraph", "text": "Nathan Benaich", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "77b8cfc7-b4e0-49db-a1d2-11b1c31f63af", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.88, "y": 0.02 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cbd76f41-60a2-4e11-9666-46a07830d561", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "title", "text": "State of AI Report", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7828c5eb-7a51-4520-bba3-7b98b7387421", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.3, "x": 0.35, "y": 0.5 }, "kind": "title", "text": "October 1, 2020", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Subtitle", "slug": "subtitle", "status": "active", "canonId": "019de52c-fb09-739d-900c-815b0d9ac526", "version": 1, "description": null }, "framework": null }, "subkind": "subtitle", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1b093ba1-061c-4ff9-893c-43bf9d83061f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Slide recipe: Cover", "slug": "cover-slide", "status": "active", "bestFor": null, "canonId": "019df269-2a08-767d-9128-e7431e6fe8a3", "version": 1, "bodyDocId": "019df269-29e5-7738-b5fa-6892e5dc8e65", "description": "The first ten seconds. Build it deliberately.", "familyLabel": "slide-recipe", "categoryName": "Slide", "categorySlug": "slide" }, "agent": "designer", "layer": "slide", "agents": [ "designer" ], "matchId": "a221ed8c-243a-41de-95a5-26dd8ebbae65", "evidence": "This is a cover slide with a title, subtitle, and authors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:43.944352+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 1, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "thumbSrc": "https://imgproxy.kitesheet.com/IwpoDUsmDIthrweWM7AHa7sM3s5FjQs6H5rbgsHBP_Y/rs:fit:480:480:0/q:78/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9lZWE3NTI0YzU1NzAzNmY0L3AwMDEuanBn", "imagePathAlt": "https://imgproxy.kitesheet.com/ewOSg73pgmHUjrlBwD8sNU-n3n5cZNCKV1lJNUQPrY0/rs:fit:1200:1200:0/q:82/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9lZWE3NTI0YzU1NzAzNmY0L3AwMDEucG5n", "thumbSrcAlt": "https://imgproxy.kitesheet.com/tBzp1KwGI1857Z-IRIE4PJJu9nWHBW5SyDW1K3mOf9M/rs:fit:480:480:0/q:78/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9lZWE3NTI0YzU1NzAzNmY0L3AwMDEucG5n" }, { "page": 2, "type": "setup", "function": "other", "rawType": "team_bio", "block": null, "metadata": { "slideType": "team_bio", "slideTypeCanon": { "name": "Team bio", "slug": "team_bio", "status": "active", "canonId": "019de52d-0a7d-7394-ae5b-037478e71f1e", "version": 1, "description": null }, "function": "other", "functionCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-15e1-75bc-99dc-07732380e2ce", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 5, "textChars": 543, "nDataPoints": 2, "notes": null, "elementsJson": [ "headshot", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:28.394+00", "seconds": 2.925, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/2", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-2", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.35, "w": 0.16, "x": 0.635, "y": 0.215 }, "kind": "image", "text": "Ian Hogarth photo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9badc361-38e6-4119-bd8b-8a458103758c", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.16, "x": 0.208, "y": 0.215 }, "kind": "image", "text": "Nathan Benaich photo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fba164cc-626d-4e3c-8601-ce87b1b23fb3", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.35, "x": 0.52, "y": 0.63 }, "kind": "paragraph", "text": "Ian Hogarth: Ian is an angel investor in 60+ startups. He is a Visiting Professor at UCL... Ian was co-founder and CEO of Songkick... He studied engineering at Cambridge... He is the Chair of Phasecraft, a quantum software company.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4337a0c6-1fe2-47fd-b3e1-3db4e60a24a6", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.35, "x": 0.09, "y": 0.63 }, "kind": "paragraph", "text": "Nathan Benaich: Nathan is the General Partner of Air Street Capital, a venture capital firm investing in AI-first technology and life science companies. He founded RAAIS and London.AI... He studied biology at Williams College and earned a PhD from Cambridge.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a16900bc-9adf-4bd0-91a7-e48511a0ad9b", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.2, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "About the authors", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3b217433-5e45-4537-9566-52c79cde9ecf", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 3, "type": "setup", "title": "The slide serves as a table of contents and mission statement for the report.", "function": "front_matter", "rawType": "executive_summary", "block": null, "metadata": { "slideType": "executive_summary", "slideTypeCanon": { "name": "Executive summary", "slug": "executive_summary", "status": "active", "canonId": "019de52d-06be-73ff-81b3-f99fe7f0025e", "version": 1, "description": null }, "function": "front_matter", "functionCanon": { "name": "Front matter", "slug": "front_matter", "status": "active", "canonId": "019de52d-133d-73f0-9e26-201eda96b098", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 1191, "nDataPoints": 0, "notes": "The slide serves as a table of contents and mission statement for the report.", "elementsJson": [ "paragraph", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:34.444+00", "seconds": 8.963, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/3", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-3", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Consider this Report as a compilation of the most interesting things we’ve seen with a goal of triggering an informed conversation about the state of AI and its implication for the future.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa5f-ebd3a0dc4b77", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.8, "x": 0.05, "y": 0.58 }, "kind": "list", "text": "Research: Technology breakthroughs and their capabilities.\nTalent: Supply, demand and concentration of talent working in the field.\nIndustry: Areas of commercial application for AI and its business impact.\nPolitics: Regulation of AI, its economic implications and the emerging geopolitics of AI.\nPredictions: What we believe will happen in the next 12 months and a 2019 performance review to keep us honest.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.444+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0becf666-e2b2-433b-9bd4-677ca6acc4d0", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.15 }, "kind": "paragraph", "text": "Artificial intelligence (AI) is a multidisciplinary field of science and engineering whose goal is to create intelligent machines. We believe that AI will be a force multiplier on technological progress in our increasingly digital, data-driven world.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.444+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2f44ebed-4c91-4335-8065-3cfbbc91d06d", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.35 }, "kind": "paragraph", "text": "The State of AI Report is now in its third year. New to the 2020 edition are several invited content contributions from a range of well-known and up-and-coming companies and research groups.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.444+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "95c2882e-56c8-452f-ab8a-3b76bda9fd93", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.88 }, "kind": "paragraph", "text": "Collaboratively produced by Ian Hogarth (@soundboy) and Nathan Benaich (@nathanbenaich).", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.444+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e761d34b-d4d0-4b2c-89c8-55894ca71c1a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.05, "y": 0.03 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.444+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f311adc7-58e8-486a-b884-08dafd508da7", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "b5664eb2-ca52-4a2d-a3db-3d9c476c32a7", "evidence": "The slide mentions the report's goal of triggering an informed conversation.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:44.066895+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 3, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Executive summary", "slug": "executive-summary", "status": "active", "bestFor": null, "canonId": "019dd9e1-4b63-72d5-9734-da757d91ca61", "version": 1, "bodyDocId": null, "description": "A condensed view of the entire argument on a single slide / block.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "5bbd142b-c824-40fa-81df-5fa2853f9563", "evidence": "This slide provides an overview of the report's contents and goals.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:44.035883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 3, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 4, "type": "appendix", "title": "This is an acknowledgments slide.", "function": "other", "rawType": "appendix", "block": null, "metadata": { "slideType": "appendix", "slideTypeCanon": { "name": "Appendix", "slug": "appendix", "status": "active", "canonId": "019de52d-0762-71f8-b564-89ab863301cc", "version": 1, "description": null }, "function": "other", "functionCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-15e1-75bc-99dc-07732380e2ce", "version": 1, "description": null }, "density": "sparse", "densityScore": 17, "componentCount": 4, "textChars": 209, "nDataPoints": 0, "notes": "This is an acknowledgments slide.", "elementsJson": [ "logo_grid", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:29.29+00", "seconds": 3.799, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/4", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-4", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.3, "w": 0.9, "x": 0.05, "y": 0.2 }, "kind": "image", "text": "Company logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.29+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ea0c44c7-6511-4857-b0fa-013d2fd100b9", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.6 }, "kind": "paragraph", "text": "Jack Clark, Jeff Ding, Chip Huyen, Rebecca Kagan, Andrej Karpathy, Moritz Müller-Freitag, Torsten Reil, Charlotte Stix, and Nu (Claire) Wang.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.29+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d572072d-1df4-4824-9039-bdf51a3a52fd", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.35, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Thank you to our contributors", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.29+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9b6bdafb-66bb-4641-92a1-39d83cf6a631", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.35, "x": 0.05, "y": 0.53 }, "kind": "title", "text": "Thank you to our reviewers", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.29+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f4410e0a-08d0-4ab1-a507-ee1f1e1ce6ce", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 5, "type": "appendix", "title": "Part of the State of AI 2020 report.", "function": "establish_context", "rawType": "other", "block": null, "metadata": { "slideType": "other", "slideTypeCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-0af5-71ed-83de-d51ae245fc43", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 2, "textChars": 1329, "nDataPoints": 0, "notes": "Part of the State of AI 2020 report.", "elementsJson": [ "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:28.94+00", "seconds": 3.348, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/5", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-5", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Artificial intelligence (AI): A broad discipline with the goal of creating intelligent machines, as opposed to the natural intelligence that is demonstrated by humans and animals. It has become a somewhat catch all term that nonetheless captures the long term ambition of the field to build machines that emulate and then exceed the full range of human cognition.\n\nMachine learning (ML): A subset of AI that often uses statistical techniques to give machines the ability to \"learn\" from data without being explicitly given the instructions for how to do so. This process is known as \"training\" a \"model\" using a learning \"algorithm\" that progressively improves model performance on a specific task.\n\nReinforcement learning (RL): An area of ML concerned with developing software agents that learn goal-oriented behavior by trial and error in an environment that provides rewards or penalties in response to the agent’s actions (called a \"policy\") towards achieving that goal.\n\nDeep learning (DL): An area of ML that attempts to mimic the activity in layers of neurons in the brain to learn how to recognise complex patterns in data. The \"deep\" in deep learning refers to the large number of layers of neurons in contemporary ML models that help to learn rich representations of data to achieve better performance gains.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.94+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f355f6fd-9ac1-4cb2-916d-352d4f6f5458", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.15, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "Definitions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.94+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "86486b6b-4745-435e-a44b-90274119b02b", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 6, "type": "appendix", "title": "Part of the 'State of AI 2020' report.", "function": "establish_context", "rawType": "appendix", "block": null, "metadata": { "slideType": "appendix", "slideTypeCanon": { "name": "Appendix", "slug": "appendix", "status": "active", "canonId": "019de52d-0762-71f8-b564-89ab863301cc", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 35, "componentCount": 2, "textChars": 1089, "nDataPoints": 0, "notes": "Part of the 'State of AI 2020' report.", "elementsJson": [ "headline_text", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:30.554+00", "seconds": 4.957, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/6", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-6", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.9, "x": 0.05, "y": 0.2 }, "kind": "list", "text": "Algorithm: An unambiguous specification of how to solve a particular problem.\nModel: Once a ML algorithm has been trained on data, the output of the process is known as the model. This can then be used to make predictions.\nSupervised learning: A model attempts to learn to transform one kind of data into another kind of data using labelled examples. This is the most common kind of ML algorithm today.\nUnsupervised learning: A model attempts to learn a dataset's structure, often seeking to identify latent groupings in the data without any explicit labels. The output of unsupervised learning often makes for good inputs to a supervised learning algorithm at a later point.\nTransfer learning: An approach to modelling that uses knowledge gained in one problem to bootstrap a different or related problem, thereby reducing the need for significant additional training data and compute.\nNatural language processing (NLP): Enabling machines to analyse, understand and manipulate language.\nComputer vision: Enabling machines to analyse, understand and manipulate images and video.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.554+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "19662506-3b68-4879-9090-8f90adb155a1", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.15, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Definitions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.554+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aa6ebe6d-22d9-4b56-9353-7a5c33a3a1f1", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 7, "type": "setup", "title": "The slide uses a structured list format to categorize insights into four distinct pillars.", "function": "summarize", "rawType": "executive_summary", "block": null, "metadata": { "slideType": "executive_summary", "slideTypeCanon": { "name": "Executive summary", "slug": "executive_summary", "status": "active", "canonId": "019de52d-06be-73ff-81b3-f99fe7f0025e", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 5, "textChars": 1793, "nDataPoints": 1, "notes": "The slide uses a structured list format to categorize insights into four distinct pillars.", "elementsJson": [ "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:30.484+00", "seconds": 4.752, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/7", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-7", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.2, "w": 0.9, "x": 0.05, "y": 0.6 }, "kind": "list", "text": "Industry: The first trial of an AI-discovered drug begins in Japan and the first US medical reimbursement for AI-based imaging procedure is granted. Self-driving car mileage remains microscopic and open sourcing of data grows to crowdsource new solutions. Google, Graphcore, and NVIDIA continue to make major advances in their AI hardware platforms. NLP applications in industry continue to expand their footprint and are implemented in Google Search and Microsoft Bing.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "076feb2b-5d9c-482c-898c-c52437c814c8", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.9, "x": 0.05, "y": 0.4 }, "kind": "list", "text": "Talent: American institutions and corporations further their dominance of major academic conference papers acceptances. Multiple new institutions of higher education dedicated to AI are formed. Corporate-driven academic brain drain is significant and appears to negatively impact entrepreneurship. US AI ecosystem is fuelled by foreign talent and the contribution of researchers educated in China to world-class papers is clear.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "35f9931e-0e74-4a9b-bec8-77ba3feb395a", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.9, "x": 0.05, "y": 0.2 }, "kind": "list", "text": "Research: A new generation of transformer language models are unlocking new NLP use-cases. Huge models, large companies and massive training costs dominate the hottest area of AI today: Natural Language Processing. Biology is experiencing its “AI moment”: From medical imaging, genetics, proteomics, chemistry to drug discovery. AI is mostly closed source: Only 15% of papers publish their code, which harms accountability and reproducibility in AI.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "68c04e98-b803-40fa-8e55-c6da4f3958e6", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.8 }, "kind": "list", "text": "Politics: After two wrongful arrests involving facial recognition, ethical risks that researchers have been warning about come into sharp focus. Semiconductor companies continue to grow in geopolitical significance, particularly Taiwan’s TSMC. The US Military is absorbing AI progress from academia and industry labs. Nations pass laws to let them scrutinize foreign takeovers of AI companies and the UK’s Arm will be a key test.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c2b3b41e-63c7-4b2c-812a-73f9115ed5fe", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.25, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Executive Summary", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "92ce276f-f706-43d7-966c-215bb05a4289", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chunking", "slug": "chunking", "status": "active", "bestFor": null, "canonId": "019dd956-abab-72ce-b64c-2924cc83cf39", "version": 1, "bodyDocId": null, "description": "Information grouped into chunks is remembered better - structure content in digestible units", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-012633b6d15c", "evidence": "Bullets chunked under Research/Talent/Industry/Politics headers.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecd-fda0baa96ded", "evidence": "Title 'Executive Summary' is topic, but bullets state insights per section.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Core Message Extraction", "slug": "core-message-extraction", "status": "active", "bestFor": "Any slide that carries an argument or finding. The diagnostic: if the slide has a chart on it, it needs a core message. Run extraction whenever the current title is a topic (Q3 performance, Customer satisfaction) rather than a claim, or when a reader could reasonably draw more than one conclusion from the visible evidence.", "canonId": "019dd956-bc71-707d-8867-1b2e810175cf", "version": 1, "bodyDocId": "f85613cb-4b77-462f-bfe6-3b862516e1a6", "description": "The discipline of distilling raw analysis, data, or content into a single declarative sentence — the slide's core message — that the audience would walk away with if they remembered nothing else. Process, not artefact: extraction questions run on the raw content until one sentence falls out.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-0696e92e3c67", "evidence": "Distills 170+ pages into one screen of takeaways.", "pageRefs": null, "priority": null, "whenToUse": "Any slide that carries an argument or finding. The diagnostic: if the slide has a chart on it, it needs a core message. Run extraction whenever the current title is a topic (Q3 performance, Customer satisfaction) rather than a claim, or when a reader could reasonably draw more than one conclusion from the visible evidence.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": "Solves the two failure modes of slide writing simultaneously: the topic title (which outsources interpretation to the audience and produces three different conclusions in three different heads) and the multi-thesis stapler (which exceeds working memory and gets remembered as fragments). Aligns with how readers actually consume decks — Heath & Heath's Find the Core, the journalist's nut graf, and Minto's governing thought all converge on the same cognitive economics.", "antipattern": "Topic titles disguised as core messages (Customer satisfaction, no verb, no stake); multi-thesis sentences stapled with and; hedged sentences (may show signs of possible softening); chart-caption titles (Revenue fell 4%) that restate what the chart already shows instead of naming the so-what; meta-commentary (this slide explores...) that describes the slide rather than concluding it.", "cardinality": null, "narrativePurpose": "Forces the slide author to do the interpretive work before the slide ships, so the audience does not have to do it during the meeting. Wins the WYSIATI battle on purpose — the loudest fragment on the page becomes the author's chosen sentence rather than whichever chart label happens to be largest." }, { "tool": { "name": "Executive summary", "slug": "executive-summary", "status": "active", "bestFor": null, "canonId": "019dd9e1-4b63-72d5-9734-da757d91ca61", "version": 1, "bodyDocId": null, "description": "A condensed view of the entire argument on a single slide / block.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "95723de4-9d02-4d23-a1d8-35f83c598bc8", "evidence": "This slide summarizes key findings across multiple areas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:44.216829+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [ { "matchId": "d2682600-c126-429f-a3c8-ab94c52a493b", "evidence": "The slide is explicitly titled Executive Summary and organizes key findings into thematic categories.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:30.484+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 7, "frameworkName": "executive-summary" } ], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 8, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 2, "componentCount": 1, "textChars": 46, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:29.738+00", "seconds": 3.949, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/8", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-8", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.1, "w": 0.7, "x": 0.15, "y": 0.45 }, "kind": "title", "text": "Scorecard: Reviewing our predictions from 2019", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.738+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7e4b63cc-6911-4cd7-80dd-e4f37bca6869", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "0fd3b959-a09f-4118-86a6-c652d73c16b5", "evidence": "This slide introduces a new section with a headline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:44.277952+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 8, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 9, "type": "setup", "title": "The slide uses a traffic-light color coding system (Green/Yellow/Red) to grade the accuracy of past predictions.", "function": "summarize", "rawType": "comparison_table", "block": null, "metadata": { "slideType": "comparison_table", "slideTypeCanon": { "name": "Comparison table", "slug": "comparison_table", "status": "active", "canonId": "019de52d-0a06-73f8-8813-fe022fab1de2", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "sparse", "densityScore": 13, "componentCount": 2, "textChars": 106, "nDataPoints": 15, "notes": "The slide uses a traffic-light color coding system (Green/Yellow/Red) to grade the accuracy of past predictions.", "elementsJson": [ "comparison_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:29.357+00", "seconds": 3.508, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/9", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-9", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.85, "w": 0.96, "x": 0.02, "y": 0.1 }, "kind": "table", "text": "Our 2019 Prediction | Grade | Evidence", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.357+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6e2e664b-6ca9-42d1-937a-20525290b7ab", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.02, "y": 0.02 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.357+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6e1d510c-1732-4c8a-993e-ec71c003a39e", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "ac9c4ac3-b62d-4393-b946-39bc8d290a4f", "evidence": "The slide maps predictions to grades and evidence in a structured table format.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:29.357+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 9, "frameworkName": "evidence-matrix" } ], "arcBeats": [ { "to": 9, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 1, "name": "Setup & Stakes", "beatId": "019dd95a-0682-776c-8e34-fa40e6e023a1", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Setup & Stakes", "slug": "mountain-setup-stakes", "status": "active", "canonId": "019dd9b8-18b5-77f3-8e4d-5dbbdb6aea1d", "version": 1, "description": null }, "evidence": "Cover, exec summary and 2019 scorecard set the stage.", "position": 1, "isPrimaryArc": false, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 10, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 1, "componentCount": 1, "textChars": 19, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:29.996+00", "seconds": 3.967, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/10", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-10", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.28, "x": 0.36, "y": 0.45 }, "kind": "title", "text": "Section 1: Research", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:29.996+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c93305f6-4270-4184-9b24-8b23b7ba3d6a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "7821bb09-c318-42ac-9e9b-f025b7b837b3", "evidence": "This slide introduces a new section with a headline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:44.371248+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 10, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 11, "type": "analysis", "title": "The chart shows a peak in late 2019 followed by a decline to 15% by June 2020.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 786, "nDataPoints": 1, "notes": "The chart shows a peak in late 2019 followed by a decline to 15% by June 2020.", "elementsJson": [ "headline_text", "action_title", "paragraph", "area_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.593+00", "seconds": 6.464, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/11", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-11", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.08, "w": 0.15, "x": 0.75, "y": 0.47 }, "kind": "callout", "text": "Jun 2020 Percentage : 15%", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.593+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "773396b5-9ddf-4eb7-a9cc-b6497aea368f", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "This points to centralization of AI talent and compute as a huge problem.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-f1fa99569974", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.85, "x": 0.05, "y": 0.45 }, "kind": "chart", "text": "Code Availability over time", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Area chart", "slug": "area", "status": "active", "canonId": "019de52c-ff89-71e0-911e-5cab78b90f53", "version": 1, "description": null }, "framework": null }, "subkind": "area", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.593+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "38c31253-5792-40b9-baff-fec05f89011b", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Code Availability: 15%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-f77c21afa1f5", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Research paper code implementations are important for accountability, reproducibility and driving progress in AI. The field has made little improvement on this metric since mid-2016. Traditionally, academic groups are more likely to publish their code than industry groups. Notable organisation that don't publish all of their code are OpenAI and DeepMind. For the biggest tech companies, their code is usually intertwined with proprietary scaling infrastructure that cannot be released. This points to centralization of AI talent and compute as a huge problem.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.593+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "86b20b6a-e144-4c0d-ba85-ff57a94fc027", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "AI research is less open than you think: Only 15% of papers publish their code", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.593+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ba918ac-b71b-453d-bc23-68a39f3763b0", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-08dcac027626", "evidence": "Title states insight: 'Only 15% of papers publish their code'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-0eb68e9ef56b", "evidence": "Specific 15% statistic anchors abstract claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-12d7f6d3f385", "evidence": "Callout: 'centralization of AI talent and compute as a huge problem.'", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 14, "from": 11, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-eb99cad9ccc2", "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 12, "type": "appendix", "title": "The slide acts as a showcase/screenshot of the Papers With Code website interface.", "function": "establish_context", "rawType": "other", "block": null, "metadata": { "slideType": "other", "slideTypeCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-0af5-71ed-83de-d51ae245fc43", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 39, "componentCount": 5, "textChars": 418, "nDataPoints": 6, "notes": "The slide acts as a showcase/screenshot of the Papers With Code website interface.", "elementsJson": [ "headline_text", "screenshot", "icon_grid" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:31.578+00", "seconds": 5.371, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/12", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-12", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Hosting 3,000 State-of-the-Art leaderboards, 750+ ML components, and 25,000+ research along with code.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-0245a592cb2e", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.9, "x": 0.05, "y": 0.3 }, "kind": "image", "text": "Screenshot of the Papers With Code website showing 'Browse State-of-the-Art' and 'Methods' sections.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.578+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7b53ce31-e986-4dae-84c6-71e8b8d0ca21", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "State-of-the-Art leaderboards: 3,000", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-05d14e543512", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.8, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Hosting 3,000 State-of-the-Art leaderboards, 750+ ML components, and 25,000+ research along with code.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.578+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9b09b27c-3ce5-4272-9018-3ee87ffaf68b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Papers With Code tracks openly-published code and benchmarks model performance", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.578+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "17b9a666-e1f4-4bbb-a6e8-b80b37a3e65d", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 14, "from": 11, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-eb99cad9ccc2", "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 13, "type": "analysis", "title": "The slide uses a line chart to show the percentage of PyTorch mentions across various AI conferences over time.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 778, "nDataPoints": 6, "notes": "The slide uses a line chart to show the percentage of PyTorch mentions across various AI conferences over time.", "elementsJson": [ "headline_text", "action_title", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:31.824+00", "seconds": 5.59, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/13", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-13", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Of 20-35% of conference papers that mention the framework they use, 75% cite the use of PyTorch but not TensorFlow.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:13+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa5f-d089d16139f2", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.28, "y": 0.45 }, "kind": "chart", "text": "% PyTorch Papers of Total TensorFlow/PyTorch Papers", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.824+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e3cd01b8-0c02-48e4-ac99-3b88b62bbfaf", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "PyTorch adoption rate: 75%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:13+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa5f-d5ae27ebe942", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.96, "x": 0.02, "y": 0.25 }, "kind": "paragraph", "text": "Of 20-35% of conference papers that mention the framework they use, 75% cite the use of PyTorch but not TensorFlow. Of 161 authors who published more TensorFlow papers than PyTorch papers in 2018, 55% of them have switched to PyTorch. The opposite happened in 15% of cases. Meanwhile, we observe that TensorFlow, Caffe and Caffe2 are still the workhorse for production AI.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.824+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0c93fae2-dfa4-43c9-85b4-f81b4e1f40b7", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facebook’s PyTorch is fast outpacing Google’s TensorFlow in research papers, which tends to be a leading indicator of production use down the line", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.824+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cf63ba02-8f11-4931-af5b-395b35616eeb", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.65, "x": 0.02, "y": 0.03 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.824+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "85a3f312-f070-4efc-897b-953ca8467265", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Contrast Pairs", "slug": "contrast-pairs", "status": "active", "bestFor": null, "canonId": "019dd956-9b77-70d3-88b1-f34e26f60860", "version": 1, "bodyDocId": null, "description": "Creating tension through opposites: Before↔After, Problem↔Solution, Pain↔Relief", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "loop", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-1e8182ce6905", "evidence": "PyTorch vs TensorFlow pair set up across loop.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-1531bc16f3c3", "evidence": "Title: PyTorch outpacing TensorFlow as leading indicator.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "6f65baca-aaa2-444d-a594-ce9c3e37a0a4", "evidence": "The slide title and content seem to be directed towards an audience interested in industry trends and technology.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:51.993862+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-19f71d5defe1", "evidence": "Multi-line time series tracks framework share by conference.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 14, "from": 11, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-eb99cad9ccc2", "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 14, "type": "analysis", "title": "Data source: Papers With Code. The chart shows a clear shift from other frameworks to PyTorch between 2017 and 2020.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 64, "componentCount": 7, "textChars": 683, "nDataPoints": 6, "notes": "Data source: Papers With Code. The chart shows a clear shift from other frameworks to PyTorch between 2017 and 2020.", "elementsJson": [ "headline_text", "paragraph", "area_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:28.928+00", "seconds": 2.667, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/14", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-14", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.15, "w": 0.25, "x": 0.53, "y": 0.66 }, "kind": "callout", "text": "Jun 2020 data breakdown: PyTorch 47%, TensorFlow 18%, Others 35%", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.928+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8e83c145-e13c-488b-817f-518b6ab77f93", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "47% of these implementations are based on PyTorch vs. 18% for TensorFlow.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:17+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-812633e97884", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.85, "x": 0.05, "y": 0.38 }, "kind": "chart", "text": "Share of implementations by framework over time (2017-2020)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked area chart", "slug": "area-stacked", "status": "active", "canonId": "019de52c-ffb1-75fa-a614-76782f418cf1", "version": 1, "description": null }, "framework": null }, "subkind": "area-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.928+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0874c5e3-d8a7-4516-8e69-dbd008a2785e", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.15, "x": 0.18, "y": 0.41 }, "kind": "legend", "text": "Framework legend including PyTorch, TensorFlow, JAX, MXNet, Caffe2, and others", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.928+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cb25a03a-67b7-4cb4-80e0-b1fc2f13e8b2", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Share of implementations: 47%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:17+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-869f3a206b43", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.96, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "47% of these implementations are based on PyTorch vs. 18% for TensorFlow. PyTorch offers greater flexibility and a dynamic computational graph that makes experimentation easier. JAX is a Google framework that is more math friendly and favored for work outside of convolutional models and transformers.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.928+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "865ac090-2ac9-430b-a75f-a9d0d5b924f0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "PyTorch is also more popular than TensorFlow in paper implementations on GitHub", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:28.928+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "920ff8ff-dd18-4405-9ad8-99875cb13ccf", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f4a2666a-a4b3-4edd-9093-42223cc4866d", "evidence": "Although not explicitly a 2x2 matrix, the legend and chart could suggest a categorization that might be presented in a 2x2 format, but it's not clear.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:52.05529+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-213958fa3249", "evidence": "Title: '47% of these implementations are based on PyTorch vs 18%'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-24f05ecb4e1d", "evidence": "47% vs 18% specific deltas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 14, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 14, "from": 11, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-eb99cad9ccc2", "evidence": "Slides 13-14 explicitly contrast PyTorch overtaking TensorFlow with charts.", "position": 1, "objective": "Compare PyTorch vs TensorFlow trajectories", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 15, "type": "analysis", "title": "The chart highlights the exponential jump in model size, particularly with GPT-3.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 5, "textChars": 343, "nDataPoints": 17, "notes": "The chart highlights the exponential jump in model size, particularly with GPT-3.", "elementsJson": [ "headline_text", "action_title", "bar_chart_vertical", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:30.282+00", "seconds": 3.976, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/15", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-15", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.8, "x": 0.02, "y": 0.22 }, "kind": "callout", "text": "Huge models, large companies and massive training costs dominate the hottest area of AI today, NLP.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.282+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3788e1dc-5d92-4a37-b823-37bcbbd672e6", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.7, "x": 0.15, "y": 0.28 }, "kind": "chart", "text": "Bar chart of language model parameter counts", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.282+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b39842b4-b186-4f27-8957-1f80b424a044", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Parameter count: 175B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:13+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa5f-b9ce5a1346d2", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.7, "x": 0.02, "y": 0.93 }, "kind": "source-note", "text": "Note: The number of parameters indicates how many different coefficients the algorithm optimizes during the training process.", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.282+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d87c773d-3ce1-41a3-9bda-7b44033d41e2", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Language models: Welcome to the Billion Parameter club", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.282+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7c6baf8-8190-4e6a-b4b0-60f03659c9d9", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Inductive Reasoning", "slug": "inductive-reasoning", "status": "active", "bestFor": null, "canonId": "019dd956-99c6-705e-94cc-76a504c3da37", "version": 1, "bodyDocId": null, "description": "Grouping similar evidence to form a conclusion (Evidence → Pattern → Conclusion)", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-31d3149e0ebd", "evidence": "Loop builds case from multiple model-size data points.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-2998205cd191", "evidence": "Memorable framing: 'Welcome to the Billion Parameter club'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "145187e6-5805-4f45-bfd6-2fe7d34a9e06", "evidence": "The use of a bar chart to display parameter counts suggests consideration of effective chart selection for communication.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.114739+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-2ef38db2891f", "evidence": "'Billion Parameter club' membership metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 15, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 16, "type": "analysis", "title": "The charts demonstrate power-law relationships between model performance (test loss) and three key scaling factors: compute, dataset size, and parameter count.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 7, "textChars": 556, "nDataPoints": 3, "notes": "The charts demonstrate power-law relationships between model performance (test loss) and three key scaling factors: compute, dataset size, and parameter count.", "elementsJson": [ "headline_text", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.083+00", "seconds": 6.464, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/16", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-16", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Empirical scaling laws of neural language models show smooth power-law relationships, which means that as model performance increases, the model size and amount of computation has to increase more rapidly.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-2eed3a28a5ea", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.32, "x": 0.35, "y": 0.35 }, "kind": "chart", "text": "Test Loss vs Dataset Size", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "388d5180-8846-4537-bc08-82971101c7d2", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.32, "x": 0.68, "y": 0.35 }, "kind": "chart", "text": "Test Loss vs Parameters", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "96bed9b2-138f-4248-8886-1c9b4308b298", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.32, "x": 0.02, "y": 0.35 }, "kind": "chart", "text": "Test Loss vs Compute", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9d9395b9-d136-4922-8ca2-85705c45f600", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Test Loss", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-305665383bee", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "Empirical scaling laws of neural language models show smooth power-law relationships, which means that as model performance increases, the model size and amount of computation has to increase more rapidly.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2dd67ee1-9b98-410c-b764-8d4aba45748b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Bigger models, datasets and compute budgets clearly drive performance", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f6779238-49ef-4e86-be48-9ee9f0beae4d", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-37fcd4f23f9f", "evidence": "Title states the empirical scaling law conclusion.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Causal Chain", "slug": "causal-chain", "status": "active", "bestFor": null, "canonId": "019dd956-9e98-7269-907b-8cf727e9499d", "version": 1, "bodyDocId": null, "description": "Showing how one event leads to another in sequence", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "f27d948d-0b39-42d9-bbeb-7d16a1de85a8", "evidence": "The description of how bigger models, datasets, and compute budgets drive performance implies a causal relationship.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.17496+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-3a10cdcb86ec", "evidence": "Power-law scaling chart appropriate for log relationship.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "019dd95a-1ca5-70bb-bba0-7887504640e9", "evidence": "Title cites 'empirical scaling laws of neural language models'.", "framework": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "frameworkName": "Neural Scaling Laws" }, { "matchId": "76731a61-e1bb-4626-81f7-a4a472910e0a", "evidence": "Empirical scaling laws of neural language models", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:33.083+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 16, "frameworkName": "Scaling Laws" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 17, "type": "analysis", "title": "The slide uses a mix of bulleted data points and narrative paragraphs to illustrate the cost barrier to entry for AI model training.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 82, "componentCount": 8, "textChars": 1504, "nDataPoints": 9, "notes": "The slide uses a mix of bulleted data points and narrative paragraphs to illustrate the cost barrier to entry for AI model training.", "elementsJson": [ "headline_text", "action_title", "paragraph", "callout_box", "numbered_list" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.605+00", "seconds": 5.749, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/17", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-17", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Based on variables released by Google et al., you're paying circa $1 per 1,000 parameters. This means OpenAI's 175B parameter GPT-3 could have cost tens of millions to train.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-1f1848da78fc", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.3, "x": 0.18, "y": 0.515 }, "kind": "list", "text": "$2.5k - $50k (110 million parameter model)\n$10k - $200k (340 million parameter model)\n$80k - $1.6m (1.5 billion parameter model)", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ad60e401-5a11-40a7-90b6-fd58cc0b8552", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Training cost: $10M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-203880288d74", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.8, "x": 0.15, "y": 0.35 }, "kind": "paragraph", "text": "Just how much does it cost to train a model? Two correct answers are \"depends\" and \"a lot\". More quantitatively, here are current ballpark list-price costs of training differently sized BERT [4] models on the Wikipedia and Book corpora (15 GB).", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2df057fb-135b-4ba7-b040-6ce2d31db595", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.8, "x": 0.15, "y": 0.78 }, "kind": "paragraph", "text": "Not many companies - certainly not many startups - can afford this cost. Some argue that this is not a severe issue; let the Googles of the world pre-train and publish the large language models, and let the rest of the world fine-tune them (a much cheaper endeavor) to specific tasks. Others (e.g., Etchemendy and Li [6]) are not as sanguine.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4cb9c463-7bc5-4ea2-b168-989e987053fe", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.7, "x": 0.15, "y": 0.66 }, "kind": "paragraph", "text": "For example, based on information released by Google, we estimate that, at list-price, training the 11B-parameter variant of T5 [5] cost well above $1.3 million for a single run. Assuming 2-3 runs of the large model and hundreds of the small ones, the (list-)price tag for the entire project may have been $10 million.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e1ce7b1b-efdd-4a22-8835-0de21ba32cc9", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "Based on variables released by Google et al., you're paying circa $1 per 1,000 parameters. This means OpenAI's 175B parameter GPT-3 could have cost tens of millions to train. Experts suggest the likely budget was $10M.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fef45105-818d-4820-a992-a8d7933b6dc0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Tuning billions of model parameters costs millions of dollars", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.605+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cff22a54-40ad-4403-9a7d-dcc6f442e6bd", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-3fb0e01e620a", "evidence": "'$1 per 1,000 parameters' makes abstract cost concrete.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-416998989edf", "evidence": "Per-parameter pricing analogy to commodity goods.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b4ef34b1-8767-4fa2-b07d-a118d5c11c81", "evidence": "The use of specific cost estimates ($1 per 1,000 parameters) and ranges ($2.5k - $50k) shows quantification of costs.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:52.235963+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 17, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 18, "type": "analysis", "title": "The chart displays three metrics across three model sizes (37.5B, 150B, 600B weights).", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 5, "textChars": 499, "nDataPoints": 9, "notes": "The chart displays three metrics across three model sizes (37.5B, 150B, 600B weights).", "elementsJson": [ "headline_text", "line_chart", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:30.145+00", "seconds": 2.153, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/18", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-18", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This transformer-based model with conditional computation for machine translation has 600B parameters.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-6eb12ba8354b", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.7, "x": 0.15, "y": 0.35 }, "kind": "chart", "text": "Multi-axis line chart showing Quality gain, Training wall time, and Computation cost vs model size.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.145+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b54954b8-074c-4d34-81bf-35966e7b014e", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "TPU v3 core years: 22", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-73db6c073c38", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.28 }, "kind": "paragraph", "text": "This transformer-based model with conditional computation for machine translation has 600B parameters.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.145+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d7db8a6d-8a32-42e3-9583-7754da7be429", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "To achieve the needed quality improvements in machine translation, Google's final model trained for the equivalent of 22 TPU v3 core years or ~5 days with 2,048 cores non-stop", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:30.145+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e8f9933a-81d4-4bcd-bdf1-e2cc5d4f80b4", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-45a4c69cf11a", "evidence": "'22 TPU v3 core years' translates abstract compute to time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "3da3a978-1eba-4545-92b8-56c5ea6edf8c", "evidence": "The inclusion of a detailed description and a chart to support the data suggests efficient use of data-ink.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:52.295754+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 19, "type": "analysis", "title": "The table highlights the diminishing returns of scaling ML models, specifically noting that reducing ImageNet error to 1% would cost over $100 quintillion.", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 51, "componentCount": 5, "textChars": 671, "nDataPoints": 90, "notes": "The table highlights the diminishing returns of scaling ML models, specifically noting that reducing ImageNet error to 1% would cost over $100 quintillion.", "elementsJson": [ "headline_text", "paragraph", "data_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:31.078+00", "seconds": 2.584, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/19", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-19", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Without major new research breakthroughs, dropping the ImageNet error rate from 11.5% to 1% would require over one hundred billion billion dollars!", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-473246ee6a5b", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Computational cost: 10^112", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:14+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-488ba6700d42", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.28 }, "kind": "paragraph", "text": "Without major new research breakthroughs, dropping the ImageNet error rate from 11.5% to 1% would require over one hundred billion billion dollars! Many practitioners feel that progress in mature areas of ML is stagnant.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.078+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "84af67a2-2bef-4fde-9186-9182bf7fbb0d", "parentComponentId": null }, { "bbox": { "h": 0.48, "w": 0.54, "x": 0.23, "y": 0.41 }, "kind": "table", "text": "Table comparing Polynomial vs Exponential scaling costs across ImageNet, MS COCO, SQuAD 1.1, CoLLN 2003, and WMT 2014 benchmarks.", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.078+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b1d4b7eb-808a-45ca-9074-e5004a77efe2", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "We're rapidly approaching outrageous computational, economic, and environmental costs to gain incrementally smaller improvements in model performance", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.078+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1851e607-9e8d-43d4-b5f1-1aaf69cd73a2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-4be16ed9d761", "evidence": "Title: outrageous costs for incremental gains.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 95, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Anchoring Bias", "slug": "anchoring-bias", "status": "active", "bestFor": null, "canonId": "019dd956-9254-73de-8f1b-c7250f8f4c9c", "version": 1, "bodyDocId": null, "description": "First piece of information becomes reference point for all subsequent judgments", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "1b0cb19a-4381-4cf4-8556-b2ca121c846c", "evidence": "The use of an extremely large cost estimate (over one hundred billion billion dollars) may serve as an anchor for the reader.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:52.355725+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Annotation", "slug": "annotation", "status": "active", "bestFor": null, "canonId": "019dd956-b5b9-7668-bf0e-3af99823ec5c", "version": 1, "bodyDocId": null, "description": "Adding explanatory labels to highlight key data points", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-50a3656d4004", "evidence": "Red rectangle annotation calls attention to key benchmark.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-555308ea27ed", "evidence": "'$100B billion' framing plays on shock.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Von Restorff Effect", "slug": "von-restorff-effect", "status": "active", "bestFor": null, "canonId": "019dd956-beab-73d8-9f62-30d0332e926e", "version": 1, "bodyDocId": null, "description": "Items that stand out are noticed and remembered - make key points visually distinct", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-4ee6c0474b28", "evidence": "Red box highlights ImageNet rows in big table.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 20, "type": "analysis", "title": "Charts illustrate scaling laws in machine learning, specifically the relationship between model size (parameters), data (tokens), and compute budget.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 6, "textChars": 637, "nDataPoints": 2, "notes": "Charts illustrate scaling laws in machine learning, specifically the relationship between model size (parameters), data (tokens), and compute budget.", "elementsJson": [ "headline_text", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.537+00", "seconds": 5.039, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/20", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-20", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This has implications for problems where training data samples are expensive to generate, which likely confers an advantage to large companies entering new domains with supervised learning-based models.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-d69c0e536516", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Larger models require fewer samples to reach the same performance", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55b8420e-2c8b-4e51-a7b0-c6f858fe29cf", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.5, "y": 0.35 }, "kind": "chart", "text": "The optimal model size grows smoothly with the loss target and compute budget", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "679373a4-c7be-467e-8d50-ac32ad808488", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Test Loss", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-da5622fb448e", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "This has implications for problems where training data samples are expensive to generate, which likely confers an advantage to large companies entering new domains with supervised learning-based models.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0b0a902c-36b8-4d6c-9cca-edb41ecb5d84", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "A larger model needs less data than a smaller peer to achieve the same performance", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "85efa8db-5b64-467f-a097-360ad686a5ef", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-59502023c802", "evidence": "Title states 'larger model needs less data'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Logical chain", "slug": "logical-chain", "status": "active", "bestFor": null, "canonId": "019dd9e1-5104-702f-8953-85757af6553d", "version": 1, "bodyDocId": null, "description": "If-then, premise-conclusion, deductive chain.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d0dfbabe-6da4-44fa-a598-8fb905c3cb2b", "evidence": "The description of how larger models require fewer samples to reach the same performance implies a logical sequence of ideas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.415745+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-5c9ea3db7d85", "evidence": "Callout draws implication: advantage to large companies.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [ { "matchId": "03c95046-1288-44a7-809f-957bac66e973", "evidence": "Charts demonstrate the relationship between model parameters, compute, and performance.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:33.537+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "frameworkName": "Scaling Laws" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 21, "type": "analysis", "title": "The chart shows a positive correlation between model size and translation performance across varying levels of training data availability.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 733, "nDataPoints": 7, "notes": "The chart shows a positive correlation between model size and translation performance across varying levels of training data availability.", "elementsJson": [ "headline_text", "action_title", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.558+00", "seconds": 7.003, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/21", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-21", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Google made use of their large language models to deliver higher quality translations for languages with limited amounts of training data, for example Hansa and Uzbek.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-b8f0f7045af6", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.6, "x": 0.2, "y": 0.35 }, "kind": "chart", "text": "Line chart showing BLEU score improvement across different model sizes as training data availability decreases.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.558+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "948354f1-9f91-4a04-9a58-12d76aeab189", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.2, "x": 0.7, "y": 0.5 }, "kind": "legend", "text": "MoE(2048,36L) - 600B, MoE(2048,12L) - 200B, MoE(512E,36L) - 150B, MoE(512E,12L) - 50B, MoE(128E,36L) - 37B, MoE(128E,12L) - 12.5B, T(96L) - 2.3B", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.558+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1bc3075e-fd10-482e-99a5-22a55163c250", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "BLEU score", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-bd73ba2ecb02", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Google made use of their large language models to deliver higher quality translations for languages with limited amounts of training data, for example Hansa and Uzbek. This highlights the benefits of transfer learning.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.558+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c84e7bbd-c352-46db-9e8c-0c65b2235416", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Low resource languages with limited training data are a beneficiary of large models", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.558+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d18f5e13-e594-44a2-8fad-48e87a9883e2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Contrast Principle", "slug": "contrast-principle", "status": "active", "bestFor": null, "canonId": "019dd956-bf0f-73ae-b9df-e60508cdaa65", "version": 1, "bodyDocId": null, "description": "Higher visual weight draws attention - use contrast strategically for emphasis", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "456e0e7b-80ce-4403-ba48-5160fa496b4f", "evidence": "The comparison of model performance across different languages and data availability highlights contrasts.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.477032+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Story Moments", "slug": "story-moments", "status": "active", "bestFor": null, "canonId": "019dd956-ae4e-768d-a18b-108e6cda511e", "version": 1, "bodyDocId": null, "description": "Design key emotional beats: The Shock, The Vision, The Proof, The Choice, The Call", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-63064d4e6955", "evidence": "Concrete Hausa/Uzbek case proves abstract scaling claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 22, "type": "analysis", "title": "The slide uses two charts: a log-scale scatter plot for compute eras and a linear scatter plot for training efficiency.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 54, "componentCount": 6, "textChars": 523, "nDataPoints": 30, "notes": "The slide uses two charts: a log-scale scatter plot for compute eras and a linear scatter plot for training efficiency.", "elementsJson": [ "headline_text", "paragraph", "line_chart", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.41+00", "seconds": 3.684, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/22", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-22", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Since 2012 the amount of compute needed to train a neural network to the same performance on ImageNet classification has been decreasing by a factor of 2 every 16 months.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-0ed6af5f278f", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.52, "y": 0.35 }, "kind": "chart", "text": "Training efficiency factor", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.41+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2cb6a45c-35fe-460d-88e5-b35384a4e97a", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Two distinct eras of compute in training AI systems", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.41+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "69cb3217-7fe9-4e78-a059-525b73f189d9", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Training efficiency factor: 2x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-1354c74adcea", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "Since 2012 the amount of compute needed to train a neural network to the same performance on ImageNet classification has been decreasing by a factor of 2 every 16 months.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.41+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "84a8d640-7fee-4061-a757-f1fa2805a021", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.75, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Even as deep learning consumes more data, it continues to get more efficient", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.41+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bd9ada0c-bd9c-4284-9a3d-e0e727433721", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-65e78059e26a", "evidence": "Title: deep learning getting more efficient over time.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 22, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-6ad24a7d8699", "evidence": "Specific '2x every 16 months' rule.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 22, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 23, "type": "evidence", "title": "The slide uses a combination of a data table, a line chart showing F1 score vs data points, and a bar chart showing intent accuracy across low/high data regimes", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 77, "componentCount": 8, "textChars": 845, "nDataPoints": 16, "notes": "The slide uses a combination of a data table, a line chart showing F1 score vs data points, and a bar chart showing intent accuracy across low/high data regimes.", "elementsJson": [ "headline_text", "paragraph", "data_table", "line_chart", "bar_chart_grouped", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.712+00", "seconds": 3.887, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/23", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-23", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Their model outperforms Google's BERT model in conversational applications, especially in low data regimes, suggesting BERT is far from a silver bullet for all NLP tasks.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-9268d32c98f3", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.62, "y": 0.42 }, "kind": "chart", "text": "Intent Accuracy for Low vs High amount of data across various models.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "edf87f02-3c31-45c3-a425-b89c720034b7", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.32, "y": 0.42 }, "kind": "chart", "text": "F1 Score vs # of data points for UVE (PolyAI) and Span-BERT.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a9907ec-9e6b-4768-b520-065b3b5b83fb", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.06, "x": 0.92, "y": 0.45 }, "kind": "image", "text": "Legend of model logos.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9f92af05-dce5-4899-949d-d3d9c7d0ab56", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "1-vs-100 Accuracy: 68.2%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:15+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa60-9594a8639a74", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "PolyAI, a London-based conversational AI company, open-sourced their ConvERT model (a pre-trained contextual re-ranker based on transformers). Their model outperforms Google's BERT model in conversational applications, especially in low data regimes, suggesting BERT is far from a silver bullet for all NLP tasks.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aaaa0fe0-eeb3-4d85-90b5-3a3698c5353d", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.02, "y": 0.42 }, "kind": "table", "text": "Model comparison table showing ELMo, BERT, USE, and ConvERT with 1-vs-100 accuracy and model size.", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "97cec22b-05ff-4e5f-b717-755f0878f91e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Yet, for some use cases like dialogue small, data-efficient models can trump large models", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.712+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "26a23464-3282-4805-9d4a-5f7948a86d3b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-6ee61f0a123e", "evidence": "Title states caveat: small models can trump large.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 23, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Unexpected Pattern", "slug": "unexpected-pattern", "status": "active", "bestFor": null, "canonId": "019dd956-bc0f-71af-8d32-3c1a2da38c40", "version": 1, "bodyDocId": null, "description": "Breaking a pattern to create surprise and attention", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-734a00d60872", "evidence": "Counterpoint to scaling-law narrative breaks the pattern.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 23, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 23, "from": 15, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-ee7c2907ccd3", "evidence": "Slides 15-19 stack scaling-law evidence; 20-23 add caveats.", "position": 2, "objective": "Establish pattern that bigger models + more compute drive performance at rising cost", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 24, "type": "analysis", "title": "Includes screenshots of Hugging Face model interface and a Twitter post by Sharif Shameem demonstrating GPT-3 code generation.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 50, "componentCount": 6, "textChars": 672, "nDataPoints": 0, "notes": "Includes screenshots of Hugging Face model interface and a Twitter post by Sharif Shameem demonstrating GPT-3 code generation.", "elementsJson": [ "headline_text", "bullet_list", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:31.294+00", "seconds": 2.19, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/24", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-24", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "GPT-3, T5, BART are driving a drastic improvement in the performance of transformer models for text-to-text tasks like translation, summarization, text generation, text to code.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:17+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-a56f6fc7ead8", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "image", "text": "Twitter post by Sharif Shameem showing GPT-3 code generation", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.294+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a66de495-72c3-4a32-8dab-321a9b98e1e3", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "image", "text": "Hugging Face summarization interface", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.294+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e35793f8-1031-4748-bd06-28e5273c5471", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "GPT-3, T5, BART are driving a drastic improvement in the performance of transformer models for text-to-text tasks like translation, summarization, text generation, text to code.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.294+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0b6cabff-d771-4f6c-b701-ed114d393365", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "Here's a sentence describing what Google's home page should look and here's GPT-3 generating the code for it nearly perfectly. - Sharif Shameem", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:17+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-aa77105f429f", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "A new generation of transformer language models are unlocking new NLP use-cases", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:31.294+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "baf9f57c-5718-4f1a-b1a3-6cfedde8d545", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-76dda5b811d2", "evidence": "Title: new transformer LMs unlocking NLP use cases.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 24, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 25, "type": "evidence", "title": "The slide highlights the capability of TransCoder to translate code without expert knowledge, achieving high success rates in unit tests.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 859, "nDataPoints": 2, "notes": "The slide highlights the capability of TransCoder to translate code without expert knowledge, achieving high success rates in unit tests.", "elementsJson": [ "headline_text", "paragraph", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.006+00", "seconds": 3.83, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/25", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-25", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "An unsupervised machine translation model trained on GitHub projects with 1,000 parallel functions can translate 90% of these functions from C++ to Java and 57% of Python functions into C++ and successfully pass unit tests.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:16+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-484bdd4420a4", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.25, "y": 0.4 }, "kind": "image", "text": "Code comparison between Python input and Unsupervised C++ translation", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.006+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ca45e11b-28c0-4d30-b4fd-21f4a14f7021", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "translation success rate: 90%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:16+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-4c9f11890df5", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "An unsupervised machine translation model trained on GitHub projects with 1,000 parallel functions can translate 90% of these functions from C++ to Java and 57% of Python functions into C++ and successfully pass unit tests. No expert knowledge required, but no guarantees that the model didn't memorize the functions either.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.006+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f9247f0e-a0ab-4451-8b90-34d50e07fa47", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.5, "x": 0.25, "y": 0.85 }, "kind": "source-note", "text": "Figure 2: Example of unsupervised Python to C++ translation. TransCoder successfully translates the Python input function SumOfKsubArray into C++...", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.006+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "950d5188-ca05-418b-aa2c-563c8c749110", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "Computer, please convert my code into another programming language", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.006+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1f5058c1-6ff5-4dec-be91-b92caa5f1a60", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "ad5485af-1339-4da8-a6be-2de9f64ce16c", "evidence": "The components mention a specific model and its capabilities, indicating that the audience is likely familiar with machine learning and programming languages.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.049989+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-7a0d6756c1d7", "evidence": "Specific 90% C++->Java translation accuracy.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "8e934809-a3af-4bb1-897b-7dff806ef4b7", "evidence": "The slide presents a narrative about a machine translation model, using a clear structure to convey information.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.086372+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-7df0984aa874", "evidence": "Conversational title 'Computer, please...' frames as dialog.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 26, "type": "evidence", "title": "Includes a visual representation of the DrRepair process flow and a code snippet example.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 54, "componentCount": 6, "textChars": 845, "nDataPoints": 0, "notes": "Includes a visual representation of the DrRepair process flow and a code snippet example.", "elementsJson": [ "headline_text", "bullet_list", "flowchart", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.008+00", "seconds": 2.803, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/26", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-26", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Given a broken program and diagnostic feedback (compiler error message), DrRepair localizes an erroneous line and generates a repaired line.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:16+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-668066012b36", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.55, "x": 0.42, "y": 0.28 }, "kind": "diagram", "text": "Flowchart showing Broken Program -> Evaluator -> Feedback -> DrRepair -> Repair", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.008+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "09a8555e-e673-4a51-b030-cd5a515af2af", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.35, "x": 0.03, "y": 0.32 }, "kind": "list", "text": "The model jointly reasons over the broken source code and the diagnostic feedback using graph neural networks.\nThey use self-supervised learning to obviate the need for labelling by taking code from programming competitions and corrupting it into a broken program.\nA SOTA is set on DeepFix, which is a program repair benchmark for correct intro programming assignments in C.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.008+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0319349f-e941-4dd4-a85b-ad20df4cfe7c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.45, "y": 0.86 }, "kind": "other", "text": "Example taken from SpoC dataset (909A-45398788.cpp)", "attrs": null, "canon": { "kind": { "name": "Other", "slug": "other", "status": "active", "canonId": "019de52c-fa90-743c-8bf6-20d1fbe82b6d", "version": 1, "description": "Fallback when no canonical kind applies." }, "tool": null, "subkind": { "name": "Unclassified", "slug": "unclassified", "status": "active", "canonId": "019de52d-05ca-77db-81fb-db7d4234bbc0", "version": 1, "description": null }, "framework": null }, "subkind": "unclassified", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.008+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9278c4f1-4647-4e36-9dac-bddbacb435b4", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "Given a broken program and diagnostic feedback (compiler error message), DrRepair localizes an erroneous line and generates a repaired line.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.008+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "13f3a0fa-fca6-40f6-b490-00a37d2eceb4", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Computer, can you automatically repair my buggy programs too?", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.008+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "95f4d376-891f-42cb-836b-b3f2617864a3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Parallel Structure", "slug": "parallel-structure", "status": "active", "bestFor": null, "canonId": "019dd956-9f34-762a-93b7-1d8e84bbd153", "version": 1, "bodyDocId": null, "description": "Using similar grammatical structure for related ideas", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-86ce81800807", "evidence": "Parallel slide titles 25-26 use same conversational format.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-816aaf21e7ad", "evidence": "Conversational 'Computer, can you...' framing parallels prior slide.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "44ea2cc3-2e25-4b93-8c80-97e80ef76d77", "evidence": "The flowchart showing Broken Program -> Evaluator -> Feedback -> DrRepair -> Repair clearly illustrates a process flow.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:57.15479+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 26, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 27, "type": "analysis", "title": "The slide uses two different chart types (bar-vertical and bar-horizontal) to illustrate the same trend of AI models surpassing human performance.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 690, "nDataPoints": 14, "notes": "The slide uses two different chart types (bar-vertical and bar-horizontal) to illustrate the same trend of AI models surpassing human performance.", "elementsJson": [ "headline_text", "bar_chart_vertical", "bar_chart_horizontal", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.055+00", "seconds": 2.524, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/27", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-27", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The human benchmark on GLUE is reliably beat today (right) and the SuperGLUE human benchmark is almost surpassed too!", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-03bb7a051703", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.5, "y": 0.45 }, "kind": "chart", "text": "SuperGLUE benchmark performance by model", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.055+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "34f525f6-4b1d-4a97-b1c3-7afa1e0146e8", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.05, "y": 0.45 }, "kind": "chart", "text": "GLUE benchmark progress over time", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.055+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fc061b2c-58f3-426d-af7b-a03347a57059", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Benchmark Score: 90.6", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-0674d29cc8e7", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "It was only 12 months ago that the human GLUE benchmark was beat by 1 point. Now SuperGLUE is in sight. GLUE and it's more challenging sibling SuperGLUE are benchmarks that evaluate NLP systems at a range of tasks spanning logic, common sense understanding, and lexical semantics. The human benchmark on GLUE is reliably beat today (right) and the SuperGLUE human benchmark is almost surpassed too!", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.055+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7d424da6-828f-4081-9082-40f1ba0c1c63", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "NLP benchmarks take a beating: Over a dozen teams outrank the human GLUE baseline", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.055+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d271a801-b93a-47dc-90bd-a86546f31418", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-8bf28f06f37e", "evidence": "Title: 'NLP benchmarks take a beating'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "1c969bcf-011b-4ea6-985a-d35516b2042d", "evidence": "The slide features bar charts and a horizontal bar chart, indicating a consideration of chart types for effective communication.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.220375+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-8c56eadec94a", "evidence": "'Take a beating' is dramatic verb choice.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 28, "type": "analysis", "title": "The slide uses a grouped bar chart to compare performance across three categories (Commonsense, Linguistics, Knowledge) for four model sizes (Small, Medium, Lar", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 73, "componentCount": 7, "textChars": 1102, "nDataPoints": 12, "notes": "The slide uses a grouped bar chart to compare performance across three categories (Commonsense, Linguistics, Knowledge) for four model sizes (Small, Medium, Large, X-Large).", "elementsJson": [ "headline_text", "paragraph", "bullet_list", "bar_chart_grouped", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.739+00", "seconds": 9.072, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/28", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-28", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "GPT-3's performance is lopsided with large knowledge gaps.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-c253b57d862c", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.38, "x": 0.58, "y": 0.32 }, "kind": "chart", "text": "GPT-3 Few Shot Test Performance", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.739+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "22bbd888-d51d-49cd-afec-52dbb7f586e9", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.52, "x": 0.02, "y": 0.32 }, "kind": "list", "text": "Large models like GPT-3 that are pre-trained on vast language corpora obviate the need for task-specific fine-tuning on a specific dataset. This enables few-shot learning on new tasks. A new benchmark measures knowledge acquired during pre-training by evaluating in few-shot settings (% avg. weighted accuracy below).\nWhile the GPT-3 X-Large improves over random chance by over 20 percentage points on average, the model's accuracy ranges from 69% for US Foreign Policy to 26% for College Chemistry. Moreover, GPT-3's average confidence is a poor estimator of its accuracy and can be off by up to 24%.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.739+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b3949038-da74-4f78-ba4c-411c79538189", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Weighted accuracy: 69%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-c6ad4c9ab7a0", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "A multi-task language understanding challenge tests for world knowledge and problem solving ability across 57 tasks including maths, US history, law and more. GPT-3's performance is lopsided with large knowledge gaps.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.739+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5c773ab1-e9ca-489f-9a06-f2236f657958", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.13, "y": 0.93 }, "kind": "source-note", "text": "Figure note: “Small” (2.7B parameters), “Medium” (6.7B), “Large” (13B) and “X-Large” (175B).", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.739+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fea84cf5-1bb8-4d84-b896-bd2e23c41e30", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.95, "x": 0.02, "y": 0.13 }, "kind": "title", "text": "What’s next after SuperGLUE? More challenging NLP benchmarks zero-in on knowledge", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.739+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "40591bf6-4be3-4d64-adf6-f920b3cbea2e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-917e74bb800a", "evidence": "Callout: GPT-3 has knowledge gaps.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 29, "type": "evidence", "title": "The slide uses a visual example of image completion to demonstrate the transformer's ability to process non-textual data.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 5, "textChars": 677, "nDataPoints": 0, "notes": "The slide uses a visual example of image completion to demonstrate the transformer's ability to process non-textual data.", "elementsJson": [ "headline_text", "paragraph", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.138+00", "seconds": 5.323, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/29", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-29", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "For example, GPT-2 was trained on text but can be fed images in the form of a sequence of pixels to learn how to autocomplete images in an unsupervised manner.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-b804f2ff5548", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.5, "x": 0.25, "y": 0.48 }, "kind": "image", "text": "Grid of image completion examples showing Input, Completions, and Original images.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f52622d6-9301-48a6-ab24-7424b5219e49", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.34 }, "kind": "list", "text": "For example, GPT-2 was trained on text but can be fed images in the form of a sequence of pixels to learn how to autocomplete images in an unsupervised manner.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "71a7a398-ac48-4cc9-b170-715e9b5165b2", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.96, "x": 0.02, "y": 0.145 }, "kind": "paragraph", "text": "The transformer’s ability to generalise is remarkable. It can be thought of as a new layer type that is more powerful than convolutions because it can process sets of inputs and fuse information more globally.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2676bb89-ac43-4439-9fd7-11d3ea7b1125", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.02, "y": 0.03 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "56a6a5b4-8816-479c-bcee-987349a170c3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Analytical method", "slug": "analytical-method", "status": "active", "bestFor": null, "canonId": "b42c4911-7946-49b5-951c-5f3d23f4e2bd", "version": 1, "bodyDocId": null, "description": "Diagnostic / structuring methods (5-whys, fishbone, issue tree, hypothesis-driven).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "650aa7ce-bbea-46a1-b52c-8879c423e811", "evidence": "The slide discusses the transformer's ability to generalize and its applications.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.322596+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 29, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-970d6bfc407a", "evidence": "GPT-2 fed images example illustrates generality.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 29, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 29, "from": 24, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f35d665a8850", "evidence": "Sequence of GPT-3/T5/code translation/DrRepair examples.", "position": 3, "objective": "Show transformers unlocking new NLP/code/vision use cases", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 30, "type": "analysis", "title": "The 2020 data point includes an annualized projection indicated by a dashed box.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 62, "componentCount": 7, "textChars": 541, "nDataPoints": 21, "notes": "The 2020 data point includes an annualized projection indicated by a dashed box.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_vertical" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:32.703+00", "seconds": 2.858, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/30", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-30", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Publications involving AI methods (e.g. deep learning, NLP, computer vision, RL) in biology are growing >50% year-on-year since 2017.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-41844e1b2d4c", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.9, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Annual count of AI-related biology papers from 2000 to 2020", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.703+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d9bb7721-5405-46c7-8bf2-30c1dd7d6fe9", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.95, "x": 0.02, "y": 0.92 }, "kind": "disclaimer", "text": "PubMed.gov | stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Disclaimer", "slug": "disclaimer", "status": "active", "canonId": "91173477-d91f-4024-bbaf-d4f93e3f82dc", "version": 1, "description": "Legal/cautionary text that occupies most of the body of its slide." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.703+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "78609d08-6aa0-417e-902c-db763df87d50", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.15, "x": 0.7, "y": 0.44 }, "kind": "legend", "text": "2020 annualized", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.703+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8bb274d1-840c-4648-af36-b50b5825298c", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of publications: 21,000", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-46ce9d09f013", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "Publications involving AI methods (e.g. deep learning, NLP, computer vision, RL) in biology are growing >50% year-on-year since 2017. Papers published since 2019 account for 25% of all output since 2000.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.703+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4790fcd6-29cd-4de2-a939-1a5a4d534533", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Biology is experiencing its “AI moment”: Over 21,000 papers in 2020 alone", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:32.703+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1af2d055-d3cb-486e-b439-b49d7f490526", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-98fd088e9fa3", "evidence": "Title: 'Biology is experiencing its AI moment'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-a2e7ae945501", "evidence": "21,000 papers in 2020 anchors the claim.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "8a798384-cbd8-4bcd-aa7c-f9c501e70af2", "evidence": "The slide features a bar chart with a clear and simple design, effectively using data-ink.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.3898+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-9e7bdc8fb983", "evidence": "'AI moment' analogy frames whole biology section.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 31, "type": "evidence", "title": "The slide uses a visual comparison to show how techniques from computer vision are being applied to biological research.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 67, "componentCount": 9, "textChars": 460, "nDataPoints": 1, "notes": "The slide uses a visual comparison to show how techniques from computer vision are being applied to biological research.", "elementsJson": [ "headline_text", "paragraph", "logo_grid", "infographic" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.851+00", "seconds": 6.928, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/31", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-31", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Large labelled datasets offer huge potential for generating new biological knowledge about health and disease.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-5e4b4629a753", "parentComponentId": null }, { "bbox": { "h": 0.47, "w": 0.48, "x": 0.43, "y": 0.43 }, "kind": "image", "text": "Hexagonal grid of cell images", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5c0872aa-13af-472e-9087-b6559f5062ba", "parentComponentId": null }, { "bbox": { "h": 0.47, "w": 0.28, "x": 0.07, "y": 0.43 }, "kind": "image", "text": "Grid of ImageNet photos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "94835bc5-c510-4d18-9385-44a1bb8868b9", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.59, "y": 0.31 }, "kind": "image", "text": "RECURSION", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7568aa2-5011-4324-a52e-aa7e1dfd8b11", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.15, "y": 0.31 }, "kind": "image", "text": "IMAGENET", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bf4c6dca-2c51-441b-9865-3cd0c70b00ff", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.2, "x": 0.15, "y": 0.38 }, "kind": "paragraph", "text": ">14M labeled images", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55ec3738-5b84-4ff9-b0d8-e38eb5b203eb", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.55, "x": 0.38, "y": 0.38 }, "kind": "paragraph", "text": "RxRx.ai image datasets of cells treated with various chemical agents", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8ed39e73-f206-4157-97ac-843cb6d4f395", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "title", "text": "Large labelled datasets offer huge potential for generating new biological knowledge about health and disease.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "268de28a-b35f-4f65-8367-5eae9c4dee99", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "From physical object recognition to “cell painting”: Decoding biology through images", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.851+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ffb1c576-e42f-4b0c-8d65-e92fc275acb4", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ece-a4f4504bdb06", "evidence": "Cell painting visuals carry the case.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "f33822dd-0c36-4592-aada-6bce06aa2381", "evidence": "The slide features a hexagonal grid of cell images, using visual anchors to organize the information.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:57.458594+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 32, "type": "evidence", "title": "Includes a chord diagram of biological interactions and two line/contour plots showing drug perturbation effects.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 880, "nDataPoints": 2, "notes": "Includes a chord diagram of biological interactions and two line/contour plots showing drug perturbation effects.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "line_chart", "photo", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.419+00", "seconds": 3.282, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/32", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-32", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.04, "w": 0.95, "x": 0.03, "y": 0.22 }, "kind": "callout", "text": "Embeddings from experimental data illuminate biological relationships and predict COVID-19 drug successes.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ef5b50d9-15f7-4dfe-95d4-8da97cc9d9f2", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.23, "x": 0.75, "y": 0.28 }, "kind": "chart", "text": "Chord diagram of biological relationships.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Dot plot", "slug": "dot", "status": "active", "canonId": "019de52d-0235-738b-bd2b-78f8e24d7cb0", "version": 1, "description": null }, "framework": null }, "subkind": "dot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cf1203f0-ed47-4e7e-b7c9-d20aa8921b53", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.28, "x": 0.46, "y": 0.65 }, "kind": "chart", "text": "SARS-CoV-2 HRCE and Storm-JAK perturbation plots.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6440ac2c-58e8-437e-87af-487459cb087e", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.28, "x": 0.46, "y": 0.28 }, "kind": "image", "text": "Diagram showing microscopy images processed through a neural network architecture.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "903668fb-7d40-4a03-99cb-b5e2a7a11736", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.03, "y": 0.28 }, "kind": "list", "text": "Deep learning models trained to identify biologically-perturbed cells imaged by fluorescent microscopy can identify 100s-1000s of relevant features of cellular morphology. Applying these features makes it possible to relate the biology induced by genetic changes, immune/cytokine perturbations, and drugs. These models were applied to experiments on COVID-19 infection and cytokine storm, identifying repurposable candidates and correctly predicting 4 randomized clinical trial results from in vitro data: rxrx.ai.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a4bbb7da-01db-44a7-a939-8bd5d52b83e0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.85, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Deep learning on cellular microscopy accelerates biological discovery with drug screens", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.419+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8c42722c-65a1-4bb6-a584-5a18e9ae6047", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Causal Chain", "slug": "causal-chain", "status": "active", "bestFor": null, "canonId": "019dd956-9e98-7269-907b-8cf727e9499d", "version": 1, "bodyDocId": null, "description": "Showing how one event leads to another in sequence", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "be9a5a4c-8d62-4d63-b527-407d697a7223", "evidence": "The slide discusses causal relationships between biological perturbations and cellular morphology.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.534196+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Story Moments", "slug": "story-moments", "status": "active", "bestFor": null, "canonId": "019dd956-ae4e-768d-a18b-108e6cda511e", "version": 1, "bodyDocId": null, "description": "Design key emotional beats: The Shock, The Vision, The Proof, The Choice, The Call", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-ab6a73678ab4", "evidence": "COVID drug screen as flagship use case.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 33, "type": "evidence", "title": "The slide uses a combination of medical imaging segmentation and a predictive line chart to demonstrate early intervention capabilities.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 71, "componentCount": 7, "textChars": 984, "nDataPoints": 12, "notes": "The slide uses a combination of medical imaging segmentation and a predictive line chart to demonstrate early intervention capabilities.", "elementsJson": [ "headline_text", "bullet_list", "screenshot", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.306+00", "seconds": 2.955, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/33", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-33", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This means that patients could receive the treatment they need before exAMD conversion to save their eyesight.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-22a51bb1c264", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.38, "x": 0.568, "y": 0.635 }, "kind": "chart", "text": "Line chart showing system prediction of conversion over months since first presentation.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4a5e1d2a-8eec-42a6-a1fb-dc519dea0694", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.38, "x": 0.568, "y": 0.335 }, "kind": "image", "text": "Segmentation maps showing eye scan features.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "132cba22-01bb-4a5a-bc4b-ed2a3b4e26b7", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.48, "x": 0.03, "y": 0.38 }, "kind": "list", "text": "Anatomical changes can be identified by comparing segmentation maps that label each pixel with their corresponding automatic features.\nSuch changes can be seen to occur in a normal eye before it converts to exAMD and pushes the patient into a high-risk subgroup.\nThis means that patients could receive the treatment they need before exAMD conversion to save their eyesight.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "682b1789-f33d-46c8-9cd2-1a0b01e20785", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "System prediction", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-26616e0f09dd", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.48, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "After diagnosis of 'wet' age-related macular degeneration (exAMD) in one eye, a computer vision system can predict whether a patient's second eye will convert from healthy to exAMD within six months. The system uses 3D eye scans and predicted semantic segmentation maps.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0a68dfd9-bf8f-4d47-bac4-1262dfcc37e4", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.94, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Ophthalmology advances as the sandbox for deep learning applied to medical imaging", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dc1df05a-494f-4b03-ad16-a2d219ccae2e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-adca83b5eb99", "evidence": "Callout: patients can be treated before vision loss.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 33, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 34, "type": "evidence", "title": "Includes ROC curves comparing AI performance against human readers (radiologists).", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 78, "componentCount": 9, "textChars": 719, "nDataPoints": 12, "notes": "Includes ROC curves comparing AI performance against human readers (radiologists).", "elementsJson": [ "headline_text", "action_title", "paragraph", "data_table", "process_diagram", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.116+00", "seconds": 2.731, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/34", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-34", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The system outperformed human radiologists and could generalise to US data when trained on UK data only.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-dc09a697ee06", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.25, "x": 0.73, "y": 0.43 }, "kind": "chart", "text": "ROC curve for US dataset", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "537aa3f3-639f-4dda-83b4-b4b754c0a879", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.25, "x": 0.47, "y": 0.43 }, "kind": "chart", "text": "ROC curve for UK dataset", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b21be200-12a5-4bb4-a422-c7a85c223bff", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.4, "x": 0.05, "y": 0.65 }, "kind": "diagram", "text": "Evaluation methodology", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3a52c69c-a8cd-467d-835f-8bf2929f9b30", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.2, "x": 0.26, "y": 0.43 }, "kind": "diagram", "text": "Ground-truth determination process", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5d2d73f1-04f8-4768-8ef9-0e1edcf0c11d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Sensitivity: 9.40%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-e299af34ebca", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "The AI system, an ensemble of three deep learning models operating on individual lesions, individual breasts and the full case, was trained to produce a cancer risk score between 0 and 1 for the entire mammography case. The system outperformed human radiologists and could generalise to US data when trained on UK data only.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e698f1d5-33d1-4269-8193-9daab0fa8583", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.2, "x": 0.05, "y": 0.43 }, "kind": "table", "text": "Test datasets summary table", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d3bf1fed-cba9-4c04-be85-bd78c45f69a0", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "AI-based screening mammography reduces false positives and false negatives in two large, clinically-representative datasets from the US and UK", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.116+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "63007f23-9dbf-47b1-be75-c51e1d9a9311", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-b3e24aa6318c", "evidence": "Title states clinical performance gains.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 34, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-b78109c4e216", "evidence": "9.40% specific delta cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 34, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 35, "type": "analysis", "title": "Features a photo of Judea Pearl, a pioneer in causal inference.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 1226, "nDataPoints": 0, "notes": "Features a photo of Judea Pearl, a pioneer in causal inference.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "headshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.16+00", "seconds": 5.578, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/35", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-35", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Causal inference explicitly addresses this issue. Many pioneers in the field including Judea Pearl (pictured) and Yoshua Bengio believe that this will be a powerful new way to enable ML systems to generalize better.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-e7f6428ffd69", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.25, "x": 0.66, "y": 0.4 }, "kind": "image", "text": "Judea Pearl", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6e598e6d-05d7-44ba-91f8-476bb2af9207", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.6, "x": 0.02, "y": 0.65 }, "kind": "list", "text": "Causal inference explicitly addresses this issue. Many pioneers in the field including Judea Pearl (pictured) and Yoshua Bengio believe that this will be a powerful new way to enable ML systems to generalize better, be more robust and contribute more to decision making.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "03e2ca92-6534-495e-9300-50a7b4c75d6f", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.6, "x": 0.02, "y": 0.38 }, "kind": "list", "text": "Many jobs require us to understand the impact of a policy change. For example whether a doctor should give a patient a particular course of treatment. This is not something that correlation-based ML systems are designed for. Once a policy change has been made, the relationship between the input and output variables will differ from the training data.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "30207f7b-86a2-4b7b-818f-da4b5613e28d", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.6, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "Most ML applications utilise statistical techniques to explore correlations between variables. This requires that experimental conditions remain the same and that the trained ML system is applied on the same kind of data as the training data. This ignores a major component of how humans learn - by reasoning about cause and effect.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a983bb58-13ab-47e4-8592-68bd1da890f4", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Causal Inference: Taking ML beyond correlation", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1d229020-573c-4195-a769-038a52af2c17", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Authority Bias", "slug": "authority-bias", "status": "active", "bestFor": null, "canonId": "019dd956-c38d-70b9-a284-b2b479763356", "version": 1, "bodyDocId": null, "description": "Expert opinions are weighted more heavily - cite credible sources", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-bbdb4001ed41", "evidence": "Photo + Judea Pearl/Bengio name-drop authority.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Logical chain", "slug": "logical-chain", "status": "active", "bestFor": null, "canonId": "019dd9e1-5104-702f-8953-85757af6553d", "version": 1, "bodyDocId": null, "description": "If-then, premise-conclusion, deductive chain.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "660603a4-7717-4f08-b92d-c59f981b9d3c", "evidence": "The slide discusses a logical chain of reasoning for applying causal inference in ML.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:57.67594+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [ { "matchId": "39c933a2-744a-43a6-be17-8b12357cc35d", "evidence": "The slide explicitly discusses causal inference as a methodology for ML.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:36.16+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "frameworkName": "causal-inference" }, { "matchId": "019dd95a-1ca5-70bb-bba0-7f6ba50cda91", "evidence": "Slide titled 'Causal Inference', Judea Pearl pictured.", "framework": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 35, "frameworkName": "Causal Inference (Pearl Ladder)" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 36, "type": "analysis", "title": "The charts compare algorithm accuracy vs doctor accuracy for associative (top) and counterfactual (bottom) models.", "function": "diagnose", "rawType": "diagnosis", "block": null, "metadata": { "slideType": "diagnosis", "slideTypeCanon": { "name": "Diagnosis", "slug": "diagnosis", "status": "active", "canonId": "019de52d-0cd7-7212-898f-0464041ff8d8", "version": 1, "description": null }, "function": "diagnose", "functionCanon": { "name": "Diagnose", "slug": "diagnose", "status": "active", "canonId": "019de52d-1657-7223-acc5-7437d878c850", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 89, "componentCount": 9, "textChars": 1251, "nDataPoints": 88, "notes": "The charts compare algorithm accuracy vs doctor accuracy for associative (top) and counterfactual (bottom) models.", "elementsJson": [ "headline_text", "bullet_list", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:34.495+00", "seconds": 3.767, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/36", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-36", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "When compared to the standard associative algorithm and 44 doctors using a test set of clinical vignettes, the counterfactual algorithm places in the top 25% of doctors, achieving expert clinical accuracy.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-c3ed86999c00", "parentComponentId": null }, { "bbox": { "h": 0.22, "w": 0.28, "x": 0.68, "y": 0.33 }, "kind": "chart", "text": "Doctor vs. posterior ranking", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c65f485d-d83c-4f60-8832-992a716b9227", "parentComponentId": null }, { "bbox": { "h": 0.22, "w": 0.28, "x": 0.68, "y": 0.63 }, "kind": "chart", "text": "Doctor vs. counterfactual ranking", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f38894e8-283b-4aaf-9aa9-a82c53caad73", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.6, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Existing AI approaches to diagnosis are purely associative, identifying diseases that are strongly correlated with a patient's symptoms. The inability to disentangle correlation from causation can result in suboptimal or dangerous diagnoses.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "263d85db-d52f-4039-bf8f-0750c928675f", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.6, "x": 0.05, "y": 0.68 }, "kind": "list", "text": "This is shown in the figures on the right where the bottom chart (counterfactual) has more blue points (algorithm>doctor) above the dashed red line (doctor=algorithm) than the top chart (associative).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2684a428-98a1-4be1-ac2f-a80eeacc7010", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.6, "x": 0.05, "y": 0.39 }, "kind": "list", "text": "To overcome this, diagnosis can be reformulated as a counterfactual inference task that uses counterfactual diagnostic algorithms.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3d340e65-c207-4f32-9a59-f6b01617598f", "parentComponentId": null }, { "bbox": { "h": 0.18, "w": 0.6, "x": 0.05, "y": 0.48 }, "kind": "list", "text": "When compared to the standard associative algorithm and 44 doctors using a test set of clinical vignettes, the counterfactual algorithm places in the top 25% of doctors, achieving expert clinical accuracy. In contrast, the standard associative algorithm achieves an accuracy placing in the top 48% of doctors.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e3408b3b-7618-4471-a973-253fb957dbd4", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Clinical accuracy: 25%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-c760eb5f9eeb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Causal reasoning is a vital missing ingredient for applying AI to medical diagnosis", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.495+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4c030f07-4ccb-4496-98a4-4000d1e56a16", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-bd879ae4172a", "evidence": "'Top 25% of doctors' concrete benchmark.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 36, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 37, "type": "analysis", "title": "The slide contrasts standard Shapley values (which assume feature independence) with ASV, demonstrating the difference on an income classifier model.", "function": "present_solution", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_solution", "functionCanon": { "name": "Present solution", "slug": "present_solution", "status": "active", "canonId": "019de52d-15b8-71da-bd23-2b767850cbce", "version": 1, "description": null }, "density": "dense", "densityScore": 74, "componentCount": 7, "textChars": 1052, "nDataPoints": 36, "notes": "The slide contrasts standard Shapley values (which assume feature independence) with ASV, demonstrating the difference on an income classifier model.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_grouped" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.844+00", "seconds": 3.091, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/37", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-37", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Asymmetric Shapley Values (ASV) are proposed to incorporate this causal information.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-e74a27cde8ba", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.4, "x": 0.55, "y": 0.4 }, "kind": "chart", "text": "Explaining income classifier on Adult Census data set", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.844+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ccf4e12e-c5fc-46ea-8189-fe73ac2863bb", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.03, "y": 0.4 }, "kind": "list", "text": "Explainability is critical to the iterative development of new AI systems. Exposing how models work and why they succeed or fail helps developers to improve their design.\nShapley values that respect the data manifold explain the black-box relationship between the data features and model predictions.\nAsymmetric Shapley Values can incorporate any known causal hierarchies among features (e.g. age and education), which helps expand our toolkit of viable approaches to AI explainability in real-world contexts.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.844+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f296d464-855f-4809-b7be-2fe195d3ec61", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Global Shapley value", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa61-e8bdebc0aa7a", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.03, "y": 0.28 }, "kind": "paragraph", "text": "A flaw with Shapley values, one current approach to explainability, is that they assume the model's input features are uncorrelated. Asymmetric Shapley Values (ASV) are proposed to incorporate this causal information.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.844+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "34d7c267-a38a-4ca0-b0ec-04f4411de01d", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.15, "x": 0.82, "y": 0.95 }, "kind": "source-note", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.844+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "29337772-77f8-4bfe-8834-bc8d5dfda241", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.95, "x": 0.03, "y": 0.13 }, "kind": "title", "text": "Model explainability is an important area of AI safety: A new approach aims to incorporate causal structure between input features into model explanations", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.844+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "54d3ebcf-6467-4e0d-ba3a-bd75069326b7", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "AIDA Model", "slug": "aida-model", "status": "active", "bestFor": null, "canonId": "019dd956-86fc-773c-813f-d416714b8413", "version": 1, "bodyDocId": null, "description": "Attention, Interest, Desire, Action - classic persuasion sequence", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "5c1d1f4b-cba9-4af0-af7e-19f128f71f9b", "evidence": "The slide presents a problem (Model explainability is an important area of AI safety), presents a solution (A new approach aims to incorporate causal structure between input features into model explanations), and provides evidence (Asymmetric Shapley Values (ASV) are proposed to incorporate this causal information).", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:46.80324+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 37, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 38, "type": "evidence", "title": "The slide uses a combination of box plots for distribution visualization and a summary table for performance metrics.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 83, "componentCount": 8, "textChars": 1244, "nDataPoints": 40, "notes": "The slide uses a combination of box plots for distribution visualization and a summary table for performance metrics.", "elementsJson": [ "bullet_list", "boxplot", "data_table", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.09+00", "seconds": 6.154, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/38", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-38", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Molecules generated using REACTOR are synthetically-accessible and drug-like by default, even without explicit consideration of these constraints as optimization objectives.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-79213129f27c", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.22, "x": 0.75, "y": 0.3 }, "kind": "chart", "text": "Drug-Likeness of DRD2-optimized Molecules", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Box plot", "slug": "box", "status": "active", "canonId": "019de52d-020e-73d9-b4cf-5793a03e9b74", "version": 1, "description": null }, "framework": null }, "subkind": "box", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6f850603-6629-4071-a511-cda91aab0d5b", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.22, "x": 0.5, "y": 0.3 }, "kind": "chart", "text": "Synthetic Accessibility Score of DRD2-optimized Molecules", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Box plot", "slug": "box", "status": "active", "canonId": "019de52d-020e-73d9-b4cf-5793a03e9b74", "version": 1, "description": null }, "framework": null }, "subkind": "box", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f9b4f3ed-b095-4d9f-8932-61b29b688b3e", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.45, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "RL agent designs molecules using step-wise transitions defined by chemical reaction templates. REACTOR frames molecular building blocks as initial states and chemical reactions as the actions that alter these states. Molecules generated using REACTOR are synthetically-accessible and drug-like by default, even without explicit consideration of these constraints as optimization objectives (top graphs). REACTOR generates a higher proportion of unique molecules that are also predicted to be active by the underlying reward model (bottom table).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "806edbbd-fec5-4b72-8e9e-d7ee67faf81e", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Activity: 98%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-7ddeebbf66a1", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.5, "y": 0.9 }, "kind": "source-note", "text": "Computation of Scaffold Similarity requires the presence of a ring system, thus the N/A.", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "db0de47e-3a67-485b-9737-9c59987f2e41", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.5, "y": 0.68 }, "kind": "table", "text": "Performance metrics table comparing REACTOR to Hill Climbing, ORGAN, JTVAE, GCPN, and MolDQN across Activity, Diversity, Scaff. Similarity, and Uniqueness.", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9246f1cd-cc0d-4bc5-97f4-40f3272a9030", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Reinforcement learning helps ensure that molecules you discover in silico can actually be synthesized in the lab. This helps chemists avoid dead ends during drug discovery.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.09+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55aa64f8-c25f-4196-8339-5e5040c44acc", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-c29ee74c7b4a", "evidence": "98% synthesizable molecules statistic.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 38, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 39, "type": "evidence", "title": "The slide uses a chemical reaction diagram to illustrate the synthesis process and a bar chart to show the progression of model accuracy.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 6, "textChars": 600, "nDataPoints": 4, "notes": "The slide uses a chemical reaction diagram to illustrate the synthesis process and a bar chart to show the progression of model accuracy.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_horizontal", "other" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.266+00", "seconds": 5.227, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/39", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-39", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Molecular transformer is 10% more accurate than the best human chemists.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-1068f0cad293", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.48, "x": 0.48, "y": 0.45 }, "kind": "chart", "text": "Test set accuracy for chemical synthesis plans (%)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ff5da799-f127-4d62-8cc2-e69185de5ad0", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.4, "x": 0.05, "y": 0.45 }, "kind": "image", "text": "Chemical reaction diagram", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "24733757-da67-4355-a81a-6cbc79da33ac", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Repurposing the transformer architecture by treating chemistry as a machine translation problem unlocks efficient chemical synthesis planning to accelerate drug discovery workflows.\nModel benchmarked on a freely available set of one million reactions reported in US patents.\nMolecular transformer is 10% more accurate than the best human chemists.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a5738521-2dde-4352-a548-92087db0607c", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Test set accuracy: 90.40%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-14fff226d7d8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Have your desired molecule? ML will generate a synthesis plan faster than you can", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "285766e9-c10c-4d30-baa2-e51134eee883", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-c78c6743458f", "evidence": "Title: 'ML will generate a synthesis plan faster than you can'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 39, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 40, "type": "analysis", "title": "The slide contrasts traditional molecular vectorization with GNN-based graph representation.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 66, "componentCount": 7, "textChars": 1104, "nDataPoints": 0, "notes": "The slide contrasts traditional molecular vectorization with GNN-based graph representation.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.184+00", "seconds": 4.105, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/40", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-40", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Graph neural networks (GNNs) are an emerging family of methods that are designed to process 3D data (i.e. non-Euclidean space).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-00ad8b2edfb5", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.38, "x": 0.585, "y": 0.365 }, "kind": "image", "text": "Comparison diagram showing traditional molecular vectorization vs GNN approach.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0769abdd-8aa5-49e3-86ef-8bb31400b1d5", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.55, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "Convolutional neural networks are designed to learn features from images that are represented as a regular grid of independent pixels in 2D space.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "34ede6b8-3473-42f7-ac16-200673837cda", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.55, "x": 0.05, "y": 0.21 }, "kind": "list", "text": "Most deep learning methods focus on learning from 2D input data (i.e. Euclidean space). Graph neural networks (GNNs) are an emerging family of methods that are designed to process 3D data (i.e. non-Euclidean space).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7095c017-0ef9-4b57-a118-31f65eebe489", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.55, "x": 0.05, "y": 0.62 }, "kind": "list", "text": "Researchers have adapted and continue to optimise various 2D models to operate in the 3D domain. In the following slides, we profile several studies that illustrate the expressive power of GNNs to problems in biology and chemistry.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7c56bae0-d4ad-4ad0-adef-c87c17c6ea43", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.55, "x": 0.05, "y": 0.45 }, "kind": "list", "text": "Now consider a chemical molecule, which is described as a graph of atoms that are connected to other atoms by bonds. Using a 2D neural network approach would not make use of the information that is explicitly encoded in molecular graph.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e06b07f7-fb3c-44e1-89cb-fd4fb5238415", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Graph neural networks: Solving problems by making use of 3D input data", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.184+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ba18a270-9ad0-4ff1-b531-e8869a642a88", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 41, "type": "evidence", "title": "The slide uses a funnel diagram to contrast the scale of AI-driven screening vs conventional screening.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 66, "componentCount": 7, "textChars": 703, "nDataPoints": 12, "notes": "The slide uses a funnel diagram to contrast the scale of AI-driven screening vs conventional screening.", "elementsJson": [ "headline_text", "paragraph", "process_diagram", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:33.555+00", "seconds": 2.425, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/41", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-41", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A graph neural network was trained on empirical data of molecules and their binary antibiotic toxicity.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-d2c700854498", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.25, "x": 0.68, "y": 0.65 }, "kind": "chart", "text": "Efficacy of halicin in mice over time (CFU/ml)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9e0da49f-fe84-47d5-aae8-0e85731bdfa8", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.55, "x": 0.12, "y": 0.38 }, "kind": "diagram", "text": "Workflow showing directed message passing neural network, training set, iterative model re-training, and chemical landscape funnel.", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d1f2e0f9-9c80-4b76-bfc6-cffe4ff0ac69", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.25, "x": 0.68, "y": 0.37 }, "kind": "image", "text": "Halicin chemical structure and mouse model illustration", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b0d40986-da35-422b-94fb-bb4b2b3e0e62", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "CFU/ml", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-d45ce7cab324", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.95, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "A graph neural network was trained on empirical data of molecules and their binary antibiotic toxicity. This model then virtually screened millions of potentially antibiotic compounds to find a structurally different antibiotic, halicin, with broad-spectrum activities in mice.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c646798a-d629-4d22-bcc9-22f8940f727c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Graph networks learn to guide antibiotic drug screening, leading to new drugs in vivo", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:33.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7b60adaa-5980-4ebb-b4ac-05baaa4dafeb", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-cb914c1b66e6", "evidence": "Title states GNN-led antibiotic discovery.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 42, "type": "evidence", "title": "The slide features a technical diagram of the PNA architecture and a box plot comparing its performance against MPNN, GAT, GCN, and GIN models.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 89, "componentCount": 11, "textChars": 1199, "nDataPoints": 6, "notes": "The slide features a technical diagram of the PNA architecture and a box plot comparing its performance against MPNN, GAT, GCN, and GIN models.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "bar_chart_grouped" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.523+00", "seconds": 5.181, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/42", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-42", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "PNA layer shows an 10x improvement of the mean-squared error (MSE) on multitask graph-based property prediction relative to other state-of-the-art graph networks (MPNN, GAT, CCN, GIN).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-560439244046", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.35, "x": 0.63, "y": 0.48 }, "kind": "chart", "text": "Log of mean-squared error on graph property prediction", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Box plot", "slug": "box", "status": "active", "canonId": "019de52d-020e-73d9-b4cf-5793a03e9b74", "version": 1, "description": null }, "framework": null }, "subkind": "box", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "97b0dbf9-a5d6-413d-8449-45c8b2fddf70", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.35, "x": 0.63, "y": 0.25 }, "kind": "image", "text": "Diagram of PNA architecture showing aggregators and scalers", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1cb50ef2-8916-46ba-9608-5762ee823da1", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.6, "x": 0.03, "y": 0.46 }, "kind": "list", "text": "Local graph properties cannot be understood with a single graph aggregator; multiple operations must be used jointly.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3e677629-717b-4f07-9b91-7d30d78898fd", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.6, "x": 0.03, "y": 0.56 }, "kind": "list", "text": "Using 4 aggregators (mean, min, max, std), along with 3 degree scalers, authors generalize previous work on GNNs and prove mathematically that PNA is the most expressive GNN.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55f77689-99c4-488c-869e-24c0fc08a70d", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.6, "x": 0.03, "y": 0.7 }, "kind": "list", "text": "PNA layer shows an 10x improvement of the mean-squared error (MSE) on multitask graph-based property prediction relative to other state-of-the-art graph networks (MPNN, GAT, CCN, GIN).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "61ced9a8-449d-469c-aee8-86c82977a632", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.6, "x": 0.03, "y": 0.32 }, "kind": "list", "text": "Chemical property prediction from molecular structures helps scale drug discovery in silico. GNNs are an emerging and highly expressive model for learning these molecular representations.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7cc7362f-8cd0-4b94-955c-ea2607361b99", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.6, "x": 0.03, "y": 0.22 }, "kind": "list", "text": "Principal Neighborhood Aggregation combines different aggregators and scalers to improve graph-based chemical property prediction.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a97fd609-268f-4c21-b734-847aadc1452c", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Mean-squared error (MSE): 10x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-5a0dcc0823e7", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.15, "x": 0.82, "y": 0.95 }, "kind": "source-note", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1e41e7b6-14d5-4510-8ea5-48a310ea5650", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Enhancing chemical property prediction using graph neural networks", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.523+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e6d84114-d86d-4c5b-b269-57b218b13539", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-cc8e1982feeb", "evidence": "10x MSE improvement quantified.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 42, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 43, "type": "evidence", "title": "The slide compares GNN-based hit rates (72%, 33%, 16%) against a ~1% industry standard, supported by a process diagram and stacked bar charts.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 79, "componentCount": 8, "textChars": 1001, "nDataPoints": 12, "notes": "The slide compares GNN-based hit rates (72%, 33%, 16%) against a ~1% industry standard, supported by a process diagram and stacked bar charts.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "bar_chart_100pct", "line_chart", "logo_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.182+00", "seconds": 3.601, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/43", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-43", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Graph neural networks trained on DEL data and applied to three different protein targets produce hit rates at 30 µM of 72% (sEH), 33% (ERα), and 16% (c-KIT).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-99e4563f2413", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.55, "x": 0.42, "y": 0.62 }, "kind": "chart", "text": "Hit rate comparison across sEH, ERα, and c-KIT targets", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart (100%)", "slug": "bar-stacked-100pct", "status": "active", "canonId": "019de52c-ff14-71b8-8d88-860f8d7da84b", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked-100pct", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8b36b804-e338-470b-93fb-44c098c81184", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.55, "x": 0.42, "y": 0.38 }, "kind": "diagram", "text": "Machine learning model training process flow", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b5d210cf-30d5-49e6-ab7b-11c40a989b0e", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.3, "x": 0.05, "y": 0.92 }, "kind": "image", "text": "Google Research, X-Chem, Zebi AI logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "31bd92fd-397c-4b2e-bdb1-4f5fafd936db", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.35, "x": 0.05, "y": 0.38 }, "kind": "list", "text": "Graph neural networks trained on DEL data and applied to three different protein targets produce hit rates at 30 µM of 72% (sEH), 33% (ERα), and 16% (c-KIT). This is in contrast to traditional high-throughput small molecule screening (without ML), which normally reports hit rates of ~1%.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c8ea0ef7-2573-48f4-83da-e262addd630d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Hit rate: 72%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-9f1dae2aed30", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "DELs are composed of millions to billions of small molecules with unique DNA tags attached, which can be seen as building blocks for larger molecules. By training a GNN on binding affinity between drugs and a target, researchers can find hits to three drug targets from ~88 M synthesizable or inexpensive purchasable compounds.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5f267a48-3824-420a-b1cd-de8139ce6aac", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "AI sifts through chemical space using DNA-encoded small molecule libraries (DEL)", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.182+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "39579dbf-4b3c-438d-981d-c684637f866e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d39aec7824c5", "evidence": "Specific 72%/33%/16% hit rates by target.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 43, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 44, "type": "evidence", "title": "The chart shows performance (mean squared error) of different models (UniRep fusion, baseline, Doc2Vec, RGN) across various protein tasks.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 615, "nDataPoints": 32, "notes": "The chart shows performance (mean squared error) of different models (UniRep fusion, baseline, Doc2Vec, RGN) across various protein tasks.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.42+00", "seconds": 3.635, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/44", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-44", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.45, "w": 0.5, "x": 0.25, "y": 0.45 }, "kind": "chart", "text": "Quantitative function mean squared error across protein tasks", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Dot plot", "slug": "dot", "status": "active", "canonId": "019de52d-0235-738b-bd2b-78f8e24d7cb0", "version": 1, "description": null }, "framework": null }, "subkind": "dot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.42+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3f5ff9ea-404a-4b22-b26f-f4ff0cfbb10b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.3, "y": 0.45 }, "kind": "image", "text": "Protein structure icons", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.42+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c820dc74-8abd-49e5-865c-e2e89e6083b8", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.1, "x": 0.25, "y": 0.48 }, "kind": "legend", "text": "UniRep fusion, Our best baseline, Doc2Vec, RGN", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.42+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c551c2fe-60b3-44cb-b343-9bdc9ee388aa", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "mean squared error", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-9c634bb82974", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "Proteins are biological molecules that can be described as crystal structures (167k available today) or their amino acid (AA) sequences (24 million available today). Similar to the process of learning word vectors, this work shows that AA sequence representations learned by an RNN can predict a variety of structural and functional properties for diverse proteins.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.42+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "451eca77-9adf-404b-b283-837e376c3b9a", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Language models show promise in learning to predict protein properties from amino acid sequences alone", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.42+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4054ec3d-205f-4a32-a377-68a9b4abf4c6", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 45, "type": "evidence", "title": "Includes logos for ZOE, King's College London, and MGH.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 688, "nDataPoints": 12, "notes": "Includes logos for ZOE, King's College London, and MGH.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_horizontal", "line_chart", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:34.253+00", "seconds": 2.43, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/45", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-45", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Loss of smell is the most predictive symptom of COVID-19", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-74ecc42c178e", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.05, "y": 0.45 }, "kind": "chart", "text": "Loss of smell is the most predictive symptom of COVID-19", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "779b2d17-50cf-4b2b-b4a3-589fdf55b7d0", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.52, "y": 0.45 }, "kind": "chart", "text": "ROC predictions for risk of a positive test in the UK test set (b) and US validation set (c)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ad2b80c-3772-42c4-a3fd-7d9b81bf64dc", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Odds ratio: 4 million", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa63-7a380a5b6e5f", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "The COVID Symptom Study app collects and analyzes the health of over 4 million global contributors to discover new symptoms, predict COVID hotspots and using AI, eventually predict COVID-19 without a physical test. ZOE is running the world's largest clinical study to validate the prediction model.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c56a7b93-6d5a-420a-8143-780fccec32ee", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "COVID-19: Analyzing symptoms from over 4 million contributors detects novel disease symptom ahead of public health community and could inform diagnosis without tests", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:34.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ceec2b71-eb4f-4b2c-a2c9-387bd8484da3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d80a56e9c735", "evidence": "4 million contributors data point.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 45, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Singularity Effect", "slug": "singularity-effect", "status": "active", "bestFor": null, "canonId": "019dd956-dcd3-7287-94d8-5a49496c2ddb", "version": 1, "bodyDocId": null, "description": "People care more about individuals than statistics - use specific examples", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-d47c7908ae38", "evidence": "Loss of smell as single memorable predictor of COVID.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 45, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 46, "type": "evidence", "title": "The slide highlights the 'COVID Moonshot' initiative, showcasing the speed of AI-driven molecule design compared to traditional methods.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 55, "componentCount": 5, "textChars": 1050, "nDataPoints": 6, "notes": "The slide highlights the 'COVID Moonshot' initiative, showcasing the speed of AI-driven molecule design compared to traditional methods.", "elementsJson": [ "paragraph", "bullet_list", "screenshot", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.238+00", "seconds": 3.105, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/46", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-46", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "An international team of scientists are working pro-bono, with no IP claims, to crowdsource a COVID antiviral.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa62-f169e5dab7de", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.68 }, "kind": "image", "text": "Design a Compound, We Will Make It. Methodology, Compound Tracker, Discuss. COVID Moonshot logo.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.238+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8acef24b-e086-4dbc-ad02-cdaf152b061f", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.9, "x": 0.05, "y": 0.26 }, "kind": "list", "text": "An international team of scientists are working pro-bono, with no IP claims, to crowdsource a COVID antiviral.\nPostEra’s synthesis technology allowed the consortium to design ‘recipes’ for 2,000 molecules in under 48 hours. Human chemists would have taken 3-4 weeks to achieve the same task.\nMoonshot has received over 10,000 submissions from 365 contributors around the world, testing almost 900 compounds and identifying 3 lead series.\nMoonshot has found several compounds with high potency and begun live viral assays and preparation for animal testing. The hope is to have a candidate shown to be efficacious in animals within 6 months.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.238+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "817f8e61-9c4b-469b-85b4-c26007156fcc", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.2, "x": 0.14, "y": 0.93 }, "kind": "paragraph", "text": "Learn more: postera.ai/covid", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.238+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3d325fb4-3fb7-4b38-84f8-61b04a142ba4", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Drug discovery goes open source to tackle COVID-19. This is a rare example of where AI is being actively used on a clearly-defined problem that’s part of the COVID-19 response.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.238+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "47d6bad8-2e88-418c-8249-908a97c705d5", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-df61de4be204", "evidence": "Pro bono COVID antiviral framing builds goodwill.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 46, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 46, "from": 30, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f421814661f3", "evidence": "p.30 'AI moment' framing then 16 case-study slides on bio/medical applications.", "position": 4, "objective": "Survey AI breakthroughs across biology, drug discovery and COVID", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 47, "type": "evidence", "title": "The slide explains a technical process involving pose estimation and image-to-image translation to create realistic sports video.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 5, "textChars": 696, "nDataPoints": 0, "notes": "The slide explains a technical process involving pose estimation and image-to-image translation to create realistic sports video.", "elementsJson": [ "headline_text", "paragraph", "process_diagram", "screenshot" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.525+00", "seconds": 5.234, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/47", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-47", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Combining a model of player and tennis ball trajectories, pose estimation, and unpaired image-to-image translation to create a realistic controllable tennis match video between any players you wish!", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-1ff99abda0cd", "parentComponentId": null }, { "bbox": { "h": 0.52, "w": 0.92, "x": 0.04, "y": 0.38 }, "kind": "image", "text": "Technical process flow diagrams showing pose estimation, image-to-image translation, and final sprite generation.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.525+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a5eae733-ef8f-42cc-88b4-b7e79f627c90", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.25 }, "kind": "list", "text": "Combining a model of player and tennis ball trajectories, pose estimation, and unpaired image-to-image translation to create a realistic controllable tennis match video between any players you wish!", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.525+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0f7ebd35-370b-4a39-9fa7-dd38dde05098", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.43, "x": 0.11, "y": 0.094 }, "kind": "paragraph", "text": "For more examples, head to cs.stanford.edu/~haotianz/research/vid2player/", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.525+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d020b9a3-380c-44f0-bf84-0b6a23e001f8", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Missed out on strawberries and cream this year? A controllable synthetic video version of Wimbledon tennis matches", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.525+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed3ad1a1-9f65-4236-aba1-6d104bdb6e86", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-e2850c31c7af", "evidence": "Wimbledon nostalgia angle anchors abstract synthesis.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 47, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 48, "type": "analysis", "title": "The slide highlights the shift from traditional CNN-based architectures like Faster R-CNN to transformer-based approaches in computer vision.", "function": "present_solution", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_solution", "functionCanon": { "name": "Present solution", "slug": "present_solution", "status": "active", "canonId": "019de52d-15b8-71da-bd23-2b767850cbce", "version": 1, "description": null }, "density": "balanced", "densityScore": 48, "componentCount": 5, "textChars": 996, "nDataPoints": 0, "notes": "The slide highlights the shift from traditional CNN-based architectures like Faster R-CNN to transformer-based approaches in computer vision.", "elementsJson": [ "bullet_list", "process_diagram", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.156+00", "seconds": 2.776, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/48", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-48", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "callout", "text": "A transformer-based object detection model matches the performance of the best object detection models while removing hand-coded prior knowledge and using half the compute budget.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.156+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2a7e432f-55d1-4046-952c-40b4f1348b19", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.4, "x": 0.55, "y": 0.52 }, "kind": "image", "text": "Visual examples of object detection and segmentation outputs.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.156+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a3a445a8-b04d-44af-be33-5f81c6f82bde", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "image", "text": "DETR architecture diagram showing CNN, transformer encoder-decoder, and bipartite matching loss.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.156+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "158d2943-cf62-45a9-afe6-603e3a032a3b", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.32 }, "kind": "list", "text": "Popular models like Faster R-CNN require various means of hand-encoding prior knowledge into the architecture in order to make predictions relative to some initial guesses. A new framework, DEtection TRansformer (DETR), uses 2D images features from a CNN, flattens them into a sequence, and uses transformers to model pairwise interactions between the features. DETR is trained end-to-end with a loss function that matches predicted and ground-truth objects. The model is simpler because it drops multiple hand-designed priors and its attention decoder helps with interpretability.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.156+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55cc2207-b900-4bd0-b16e-2737b91b3518", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Attention turns to computer vision tasks like object detection and segmentation", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.156+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9d762031-4875-44fa-bdb1-a89c96390d59", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-e758c599d693", "evidence": "Title: attention turning to vision tasks.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 48, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 49, "type": "analysis", "title": "The slide contrasts traditional visible-only depth/ground prediction with the new 'Footprints' method that infers hidden geometry.", "function": "present_solution", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_solution", "functionCanon": { "name": "Present solution", "slug": "present_solution", "status": "active", "canonId": "019de52d-15b8-71da-bd23-2b767850cbce", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 4, "textChars": 734, "nDataPoints": 0, "notes": "The slide contrasts traditional visible-only depth/ground prediction with the new 'Footprints' method that infers hidden geometry.", "elementsJson": [ "headline_text", "bullet_list", "photo", "infographic" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.149+00", "seconds": 2.455, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/49", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-49", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.04, "w": 0.85, "x": 0.02, "y": 0.22 }, "kind": "callout", "text": "Footprints: A method for estimating the visible and hidden traversable space from a single RGB image.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.149+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "205c68e5-1f9a-4fb1-80aa-49f0020d29f8", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.55, "x": 0.42, "y": 0.35 }, "kind": "image", "text": "Diagram showing input image, visible ground/depth prediction vs hidden ground/geometry prediction.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.149+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b8063076-e13e-4465-bc79-00e619ceac7f", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.45, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "Neural networks can predict geometry and semantic meaning of a scene from a single color image. However, most methods aim to predict the geometry of surfaces that are visible to the camera. This doesn't enable path planning for robots or augmented reality agents.\nFootprints allows an agent to know where it can walk or roll, beyond the immediately visible surfaces. This enables virtual characters to more realistically explore their environments in AR applications.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.149+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "75006290-d7c9-473e-9bb2-a41cd3e5d6e0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.75, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Computer vision predicts where an agent can walk beyond what is seen", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.149+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "48c84d18-4c38-4ccd-854d-7ee25aee7a67", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "626d0b8b-66c9-4107-996e-424671aec862", "evidence": "Diagram showing input image, visible ground/depth prediction vs hidden ground/geometry prediction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:56.641776+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 49, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 50, "type": "analysis", "title": "The slide illustrates a technical process flow for AI training data generation.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 5, "textChars": 780, "nDataPoints": 0, "notes": "The slide illustrates a technical process flow for AI training data generation.", "elementsJson": [ "bullet_list", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.264+00", "seconds": 2.39, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/50", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-50", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Here, single image depth prediction networks (monodepth) can be used to convert any single image into training data for stereo networks.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-8bc4ebd924c0", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.55, "x": 0.42, "y": 0.35 }, "kind": "diagram", "text": "Process flow showing Data generation and Stereo training stages", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9f225924-a457-409f-ad3f-2851f239e2b8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.4, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Training state-of-the-art stereo matching networks on a collection of single images.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0fe11845-b6e7-458f-937e-a849e31d5898", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.02, "y": 0.29 }, "kind": "paragraph", "text": "Stereo matching networks estimate depth from a calibrated stereo pair of images. Training data for such networks requires left and right image pair and ground truth depth. Such data is very difficult to collect, involving special hardware like LiDAR as well as careful calibration and synchronization of cameras. Here, single image depth prediction networks (monodepth) can be used to convert any single image into training data for stereo networks.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f47290fd-ecf9-4ccf-a050-6ba58bb95400", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.45, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Computer vision learns stereo from single images", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a5163e6d-339b-4625-8bec-173f1abd8eea", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "c6cb1007-d2be-4e0a-a8d5-e489f3e362c8", "evidence": "Process flow showing Data generation and Stereo training stages", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:56.719918+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 50, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 51, "type": "evidence", "title": "The slide describes a technical process involving CNNs and geometric cues for image processing.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 1144, "nDataPoints": 0, "notes": "The slide describes a technical process involving CNNs and geometric cues for image processing.", "elementsJson": [ "headline_text", "bullet_list", "photo", "flowchart" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.852+00", "seconds": 6.909, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/51", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-51", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "By making simultaneous use of geometric cues and a deep segmentation network, it is possible to find the direction of the vertical in a spherical image and rotate it, such that the image is level with the ground plane.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-daf790526159", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.5, "x": 0.465, "y": 0.6 }, "kind": "diagram", "text": "Flowchart showing Input Image -> Segmentation CNN -> Vertical Axis Segmentation -> Spherical Rotation -> Levelled Image, with a parallel path for Vanishing Points (VP) Renderer.", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3a2a4192-ac1f-45b5-a876-baf71bf6a68a", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.5, "x": 0.465, "y": 0.3 }, "kind": "image", "text": "Misaligned 360° image resulting in heavy distortions (left) and correctly levelled result (right)", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d40a3b17-6c1b-454d-89eb-e67115fd4ea1", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.3 }, "kind": "list", "text": "360° cameras are a powerful tool for rapidly documenting entire scenes, but do not consistently return level images. This negatively impacts the performance of computer vision models.\nBy making simultaneous use of geometric cues and a deep segmentation network, it is possible to find the direction of the vertical in a spherical image and rotate it, such that the image is level with the ground plane. This system significantly outperforms the previous state-of-the-art method.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "00582028-f445-4f8c-8428-e5fe85b10227", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "State-of-the-art geometry-guided deep learning method for levelling misaligned 360° images.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d96f0049-ee0a-4fd6-abf3-96590bb37509", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Enabling the use of consumer-grade 360° cameras in construction using deep learning", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f7212aa2-477d-4a09-842c-9aae81603152", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 52, "type": "evidence", "title": "The slide showcases performance metrics across 8 different tasks from the DeepMind Control Suite, comparing Dreamer against other agents.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 784, "nDataPoints": 8, "notes": "The slide showcases performance metrics across 8 different tasks from the DeepMind Control Suite, comparing Dreamer against other agents.", "elementsJson": [ "headline_text", "bullet_list", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.544+00", "seconds": 4.514, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/52", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-52", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Dreamer exceeds previous model-based and model-free agents in terms of data-efficiency, computation time, and final performance.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-e719e45ac10c", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.52, "x": 0.47, "y": 0.28 }, "kind": "chart", "text": "Performance comparison of Dreamer vs other agents across 8 tasks", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.544+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e6a83740-8c70-4d13-8d8a-67955d11e146", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.25, "x": 0.03, "y": 0.88 }, "kind": "image", "text": "University of Toronto, Google Brain, DeepMind", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.544+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "da3e0175-32f1-466c-872a-94ebfafa8efe", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.03, "y": 0.28 }, "kind": "list", "text": "Dreamer predicts both actions and state values by training purely in an imagined latent space using pixel inputs. The agent learns policies in an efficient manner by backpropagating the analytic value gradients through the latent dynamics. When compared against existing representation learning methods on the DeepMind Control Suite with image inputs, Dreamer exceeds previous model-based and model-free agents in terms of data-efficiency, computation time, and final performance.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.544+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0d66272d-ecf5-40ac-8516-e02ad857a8dd", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Episode Return", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-e8e20c8cf81c", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.45, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Learning dynamic behaviors through latent imagination", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.544+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1a4cff1d-2eec-4b27-80f1-4b886c9071d7", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 53, "type": "evidence", "title": "The slide features a visual demonstration of future state prediction samples (wait, nudge, turn right, go straight) using semantic segmentation and optical flow", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 60, "componentCount": 6, "textChars": 913, "nDataPoints": 2, "notes": "The slide features a visual demonstration of future state prediction samples (wait, nudge, turn right, go straight) using semantic segmentation and optical flow-like visualizations.", "elementsJson": [ "headline_text", "bullet_list", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.196+00", "seconds": 6.075, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/53", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-53", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A probabilistic driving that uses future state prediction performs better than one trained to directly optimize control without being supervised with future scene predictions.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-bc42147d36ef", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.5, "x": 0.25, "y": 0.5 }, "kind": "image", "text": "Visual representation of future sample predictions (wait, nudge, turn right, go straight) showing segmentation and flow maps.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.196+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fd00f77f-307c-43f0-a734-d531e368a40a", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.96, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Predicting how a given driving situation will unfold, ranging from what the driver will do and the behavior of dynamic agents in the scene, can help an autonomous agent to learn how to drive from videos.\nA probabilistic driving that uses future state prediction performs better than one trained to directly optimize control without being supervised with future scene predictions. This translated to a 33% steering and 46% speed improvement over the baseline.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.196+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "caa900c5-ef4d-4b3f-812a-e7b12f64391a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "steering improvement: 33%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-c371d764dbcb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.02 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.196+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d914947e-c216-4f82-ad4e-f50a56e4a70a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Learning to drive by predicting and reasoning about the future", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.196+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b460edfd-7aac-46aa-9d2c-87061865ee95", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 54, "type": "evidence", "title": "The slide explains the LoRRA architecture and provides two examples of its application.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 817, "nDataPoints": 3, "notes": "The slide explains the LoRRA architecture and provides two examples of its application.", "elementsJson": [ "headline_text", "bullet_list", "flowchart", "photo", "data_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.555+00", "seconds": 6.282, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/54", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-54", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "LoRRA is an approach that reads text in an image and jointly reasons about the image and text content to answer a question from a fixed or by selecting one of the OCR strings derived from the image.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-6a7ee73d0e4c", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.51, "x": 0.46, "y": 0.28 }, "kind": "diagram", "text": "LoRRA model architecture diagram", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ea929206-1990-4edd-baf5-d939cc1be622", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.51, "x": 0.46, "y": 0.65 }, "kind": "image", "text": "Examples of LoRRA in action", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b3e109fb-1e48-422f-8121-7d5c948db20e", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.42, "x": 0.03, "y": 0.28 }, "kind": "list", "text": "While progress has been made in visual question answering (VQA), today’s systems cannot read and reason about text in an image. LoRRA is an approach that reads text in an image and jointly reasons about the image and text content to answer a question from a fixed or by selecting one of the OCR strings derived from the image. The system is trained on a new dataset that includes 45,336 questions on 28,408 images.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "94f4da18-7ef2-4156-a217-556ee66608dc", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.85, "x": 0.03, "y": 0.22 }, "kind": "title", "text": "Look, Read, Reason & Answer (LoRRA), a novel model for answering questions based on text in images.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0fc4f2a3-8823-478d-a602-cea7af15d019", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.45, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Visual Question Answering about everyday images", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7bfd6b34-f1fb-4a93-8d15-b23364cb1afb", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 55, "type": "evidence", "title": "The slide showcases the capabilities of SinGAN across five tasks: Paint to image, Editing, Harmonization, Super-resolution, and Animation.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 40, "componentCount": 5, "textChars": 596, "nDataPoints": 0, "notes": "The slide showcases the capabilities of SinGAN across five tasks: Paint to image, Editing, Harmonization, Super-resolution, and Animation.", "elementsJson": [ "headline_text", "bullet_list", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.756+00", "seconds": 5.293, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/55", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-55", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "SinGAN is an unconditional generative scheme that generates diverse realistic samples beyond textures.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-54961f8bb70a", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.55, "x": 0.42, "y": 0.25 }, "kind": "image", "text": "Visual demonstration of SinGAN tasks: Paint to image, Editing, Harmonization, Super-resolution, and Animation.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.756+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "40fe3053-52c9-43fd-80fb-84b38f6b678c", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.03, "y": 0.3 }, "kind": "list", "text": "SinGAN is a powerful tool for a wide range of image manipulation tasks. The model learns image patch statistics across multiple scales using adversarial training. However, the model has limited semantic diversity.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.756+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "12378dff-5664-4b80-86e2-5291e493a5df", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "SinGAN is an unconditional generative scheme that generates diverse realistic samples beyond textures.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.756+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fc48f697-8873-484b-97f2-27622b4f118e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Learning a multi-purpose generative model from a single natural image", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.756+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "540073bf-c2b8-47ac-b3f0-824d0f2e1dae", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 56, "type": "analysis", "title": "Charts show performance trade-offs between accuracy (COCO AP) and efficiency (parameters, latency).", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 71, "componentCount": 8, "textChars": 604, "nDataPoints": 15, "notes": "Charts show performance trade-offs between accuracy (COCO AP) and efficiency (parameters, latency).", "elementsJson": [ "headline_text", "action_title", "line_chart", "data_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.476+00", "seconds": 2.941, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/56", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-56", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "EfficientDet-D7 achieves state-of-the-art on COCO object detection task with 4-9x fewer model parameters than the best-in-class and can run 2-4x faster on GPUs and 5-11x faster on CPUs than other detectors.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-28a9de477677", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.3, "x": 0.35, "y": 0.35 }, "kind": "chart", "text": "GPU Latency comparison chart", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1282bda9-0829-455e-ad6f-d45ecf287b21", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.3, "x": 0.68, "y": 0.35 }, "kind": "chart", "text": "CPU Latency comparison chart", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "37042e57-8a89-43f1-8e96-5022f324c9b4", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.3, "x": 0.02, "y": 0.35 }, "kind": "chart", "text": "Model Size comparison chart", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "774ff0aa-cb70-452c-ac25-2e5c93e4521e", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "COCO AP: 4-9x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-2cf3e03140af", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "EfficientDet-D7 achieves state-of-the-art on COCO object detection task with 4-9x fewer model parameters than the best-in-class and can run 2-4x faster on GPUs and 5-11x faster on CPUs than other detectors.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "41d56840-002d-48af-8a28-a84733b340bf", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.8, "x": 0.15, "y": 0.6 }, "kind": "table", "text": "Embedded data tables within charts", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ff051c64-ca8d-43dc-a55d-649be80de8a5", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "On-device computer vision models that won’t drain your battery", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.476+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bf641e64-8e5b-40d0-923b-d6705c4cae3a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-e811011e6691", "evidence": "Title: 'won't drain your battery'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 56, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-ed0800a4d1d7", "evidence": "4-9x parameter reduction quantified.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 56, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 57, "type": "analysis", "title": "The chart shows the progression of algorithm complexity and accuracy, starting from an empty program to a complex evolved algorithm.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 61, "componentCount": 6, "textChars": 999, "nDataPoints": 1, "notes": "The chart shows the progression of algorithm complexity and accuracy, starting from an empty program to a complex evolved algorithm.", "elementsJson": [ "headline_text", "paragraph", "bullet_list", "line_chart", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:35.873+00", "seconds": 2.269, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/57", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-57", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Preliminary work on a simplified image classification problem shows how to remove this human bias by using evolutionary methods to automatically find the code for complete ML algorithms.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-6692875e83fa", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.55, "x": 0.42, "y": 0.38 }, "kind": "chart", "text": "Chart showing accuracy improvement over time with algorithm evolution", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.873+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fbbd5bf1-94f3-44c1-930d-fbd612cf914a", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.05, "y": 0.35 }, "kind": "list", "text": "AutoML-Zero evaluates candidate algorithms from a sparse search space starting from an empty program.\nWhile computationally intensive, evolutionary search distributed over many machines discovers more complex and effective techniques (orange labels) to improve accuracy (y-axis) over time (x-axis) on a “toy” binary image classification task from CIFAR-10.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.873+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "33f70508-fbfb-491b-bfcd-a27438482675", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Best Accuracy Found", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-680e2047b230", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "To date, AutoML runs neural architecture search by combining complex, handwritten building blocks. Preliminary work on a simplified image classification problem shows how to remove this human bias by using evolutionary methods to automatically find the code for complete ML algorithms.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.873+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ef25a1f9-a427-4700-99e7-9d1cff715445", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Evolving entire algorithms from basic mathematical operations alone with AutoML-Zero", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:35.873+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "36529e04-7aa3-4f82-a605-b4eb822c2069", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 57, "from": 47, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-f820616b6fcc", "evidence": "Wimbledon synth video, object detection, depth, AutoML-Zero examples in sequence.", "position": 5, "objective": "Catalog computer vision and generative model advances", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 58, "type": "analysis", "title": "The chart uses a dual-axis representation (bar for total count, line for percentage share).", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 7, "textChars": 486, "nDataPoints": 10, "notes": "The chart uses a dual-axis representation (bar for total count, line for percentage share).", "elementsJson": [ "headline_text", "paragraph", "bar_chart_vertical", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:36.178+00", "seconds": 2.345, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/58", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-58", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Almost 5x growth in the number of papers that mention federated learning from 2018 to 2019. More papers have been published in the first half of 2020 than in all of 2019.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-45fe485d5216", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "chart", "text": "Total papers", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cd2ec6b6-29bb-4de4-ab20-494223844e0d", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "chart", "text": "% Google", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "833b0239-cf98-4085-8fd8-77dc782a16f0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.3, "x": 0.35, "y": 0.35 }, "kind": "legend", "text": "Total papers | % Google", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b39259f4-3026-43ad-abd7-f47302335746", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of research papers: 1050", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-493f2daee620", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Almost 5x growth in the number of papers that mention federated learning from 2018 to 2019. More papers have been published in the first half of 2020 than in all of 2019.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e5ad17e2-cdc5-4e1a-9ed4-067cf32a575b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Kicked off by Google in 2016, federated learning research is now booming", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:36.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2b9672e9-1df8-4299-94c2-57bd4ccb032f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ece-f34c33cd1bbe", "evidence": "Title: federated learning research booming.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 58, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-f657ce4b6664", "evidence": "5x growth and 2020 H1 surpasses all 2019.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 58, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 59, "type": "evidence", "title": "The slide displays screenshots of training loss/accuracy graphs for different platforms.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 49, "componentCount": 6, "textChars": 481, "nDataPoints": 4, "notes": "The slide displays screenshots of training loss/accuracy graphs for different platforms.", "elementsJson": [ "headline_text", "paragraph", "screenshot", "line_chart" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.277+00", "seconds": 4.417, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/59", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-59", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This enables the training of arbitrary neural models on private data living on a web browser or mobile device.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-3af7f1640964", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.15, "x": 0.12, "y": 0.4 }, "kind": "chart", "text": "Android training loss graph", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1d7d8480-c31a-4745-87b1-df8a636e9e44", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.15, "x": 0.27, "y": 0.4 }, "kind": "chart", "text": "iOS training loss graph", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ef6f69b-0a7b-4999-96b3-9f4420d79134", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.45, "y": 0.4 }, "kind": "chart", "text": "Javascript training loss and accuracy graphs", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a7435aba-8185-4405-a615-40e213d79b95", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "This enables the training of arbitrary neural models on private data living on a web browser or mobile device.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e0831154-b953-4271-8b0e-e32b013abc9e", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "OpenMined, the leading open-source community for privacy-preserving ML, demonstrates the first open-source federated learning platform for web, mobile, server, and IoT", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "195fe81d-099e-4d0b-b78e-5fa141c122c6", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 60, "type": "analysis", "title": "The slide uses a diagram to illustrate the technical workflow of federated learning and encrypted inference.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 1159, "nDataPoints": 0, "notes": "The slide uses a diagram to illustrate the technical workflow of federated learning and encrypted inference.", "elementsJson": [ "headline_text", "action_title", "bullet_list", "process_diagram", "logo_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.113+00", "seconds": 4.175, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/60", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-60", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "While the pooling of medical data should lead to improved medical knowledge and clinical care, it is also an area with strong safeguards around privacy.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa64-f6faf3b6ba8b", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.4, "x": 0.55, "y": 0.3 }, "kind": "diagram", "text": "Technical workflow diagram showing Model Owner, Client, Network, and Medical Startup interactions with encryption/decryption.", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "49471839-2a4e-4223-9b27-f9ec038cfd44", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.4, "x": 0.05, "y": 0.9 }, "kind": "image", "text": "OpenMined, TUM, Imperial College London, CrypTen logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "08b77c03-8253-4e14-be83-c43b9f6c433a", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.05, "y": 0.31 }, "kind": "list", "text": "The 5P Project (Kaissis, Ziller, Passerat-Palmbach, Braren, Rueckert et al., Technical University of Munich, Imperial College London and OpenMined) demonstrates federated learning and encrypted inference on paediatric chest X-rays in a clinical setting.\nLarge academic consortia (German Cancer Consortium Joint Imaging Platform) and mixed consortia including startups and established industry (London Medical Imaging and AI Centre for Value Based Healthcare)\nProspective testing and first production roll-outs are expected within the next year.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d2d945e3-4916-47e3-b62e-1214f637cf0b", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "While the pooling of medical data should lead to improved medical knowledge and clinical care, it is also an area with strong safeguards around privacy. New techniques enable privacy-preserving innovation.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b412ac77-d7fe-41e1-8c57-78b5c5b0044a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Prospective testing begins for privacy-preserving AI applied to medical imaging", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "047cd009-1534-4c78-9769-535ce769c1b6", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "54022e5b-b551-4f18-98bd-af02d2e8bda7", "evidence": "Diagram illustrating the interaction between model owners, clients, and startups using encryption.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:38.113+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 60, "frameworkName": "federated-learning-workflow" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 61, "type": "analysis", "title": "The slide compares CNNs and TICK-GP models on digit recognition and compares SVGP vs VISH on flight delay prediction.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 72, "componentCount": 7, "textChars": 1202, "nDataPoints": 4, "notes": "The slide compares CNNs and TICK-GP models on digit recognition and compares SVGP vs VISH on flight delay prediction.", "elementsJson": [ "headline_text", "bullet_list", "other", "bar_chart_grouped" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.784+00", "seconds": 6.661, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/61", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-61", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "GP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-89f44346e83c", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.3, "x": 0.65, "y": 0.65 }, "kind": "chart", "text": "Fig B: Comparison of SVGP and VISH training time and NLPD error.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.784+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "06f32cf4-5f12-4c06-af04-f0f046cd0471", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.3, "x": 0.65, "y": 0.33 }, "kind": "chart", "text": "Fig A: Comparison of CNN and TICK-GP uncertainty calibration on digit recognition.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.784+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3fd8a9ff-cd19-48c1-bb28-8053a4130765", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.55, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "GP models benefit from several key features, like depth and convolutions, inspired from neural networks (NNs).\nHowever, GPs have better calibrated uncertainty compared to NNs (Fig A), which here are shown to be confidently wrong more often (attributing no mass to the true label - shown in blue).\nGP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.\nThis GP method circumvents the need to invert large matrices, significantly speeding up training time and enabling a quicker response and adaptation to emerging events.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.784+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c7b61347-d0e6-4af6-b0ed-f6071fc0bd37", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Training time: 40 sec", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-8eb118a6ddb6", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "GPs are becoming more accurate and faster to train, whilst retaining their favourable properties, like calibrated uncertainty, making them more relevant for real-world applications today.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.784+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3889c934-ca8c-45d3-a1d2-a2ebdfacce03", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Gaussian Processes (GPs) Strike Back: Quantified uncertainty and faster training speed", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.784+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "93549c4d-020a-4d8c-96a1-5f8341017157", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "1ecdc9fb-dc1e-4b11-afb3-46083244e06b", "evidence": "GP models benefit from several key features, like depth and convolutions, inspired from neural networks (NNs).", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.060047+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-f890e57562dd", "evidence": "15 mins -> 40 sec specific speedup.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "31b8520a-0f4f-498a-8c41-441ec53a5022", "evidence": "GP training time is reduced from 15 mins to just 40 sec (Fig B) when predicting delay of commercial flights with a dataset of 6 million data points.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.095024+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 61, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 62, "type": "evidence", "title": "The slide evaluates a past prediction against actual outcomes.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 54, "componentCount": 5, "textChars": 1112, "nDataPoints": 2, "notes": "The slide evaluates a past prediction against actual outcomes.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.954+00", "seconds": 5.452, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/62", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-62", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.1, "w": 0.65, "x": 0.02, "y": 0.22 }, "kind": "callout", "text": "In our 2019 Predictions we predicted: “Google has a major breakthrough in quantum computing hardware, triggering the formation of at least 5 new startups trying to do quantum machine learning.”", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.954+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6de1dec8-a827-402b-b32f-e4abc614ed62", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "Google had a monumental breakthrough achieving 'quantum supremacy' where their Sycamore 54-qubit quantum processor performed a target computation in 200 seconds that they estimate would have taken the world's most powerful classical supercomputer 10,000 years.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-1151276ddd8a", "parentComponentId": null }, { "bbox": { "h": 0.297, "w": 0.227, "x": 0.723, "y": 0.355 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.954+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "59f7ed69-a37a-4b4d-9370-48dac3337595", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.65, "x": 0.02, "y": 0.35 }, "kind": "list", "text": "Google had a monumental breakthrough achieving ‘quantum supremacy’ where their Sycamore 54-qubit quantum processor performed a target computation in 200 seconds that they estimate would have taken the world’s most powerful classical supercomputer 10,000 years.\nThis result is the first experimental challenge against the extended Church-Turing thesis, which states that classical computers can efficiently implement any “reasonable” model of computation\nIt is extremely unclear whether quantum computers will be useful for ML related tasks anytime soon, but multiple startups are now exploring the possibility.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.954+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a2c8891-04dd-4cbb-926f-85e5f95c9763", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "2019 Prediction outcome: Google quantum supremacy", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.954+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b54020de-3f72-4f5b-8361-7d5049a04c11", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ece-fd927c8c14dd", "evidence": "'Quantum supremacy' dramatic framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 62, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 62, "from": 58, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3c-fd470c64e823", "evidence": "Three short bursts of new research direction case studies.", "position": 6, "objective": "Map emerging research themes: federated learning, GPs, quantum", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 63, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 13, "componentCount": 4, "textChars": 110, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.379+00", "seconds": 5.609, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/63", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-63", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.45, "x": 0.02, "y": 0.03 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.379+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10c03085-6019-4ad4-b2a4-b6fa377089c0", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.06, "x": 0.92, "y": 0.03 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.379+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "73b09a4c-1c78-4b3b-8acf-282655bb3133", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.09, "x": 0.89, "y": 0.93 }, "kind": "paragraph", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.379+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7621cfc1-cfa7-441a-9e52-f2f3e354e97d", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.24, "x": 0.38, "y": 0.45 }, "kind": "title", "text": "Section 2: Talent", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.379+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "53d66dc4-d905-43be-82e3-8ab3c72305af", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "a3dcebdc-fd75-40e0-8074-f82c9d975c0e", "evidence": "Introduction | Research | Talent | Industry | Politics | Predictions", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:01.220759+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 63, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 64, "type": "analysis", "title": "The slide highlights the trend of academic talent moving to industry, specifically in the AI sector.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 68, "componentCount": 7, "textChars": 732, "nDataPoints": 28, "notes": "The slide highlights the trend of academic talent moving to industry, specifically in the AI sector.", "elementsJson": [ "headline_text", "action_title", "paragraph", "bar_chart_horizontal", "bar_chart_horizontal", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.91+00", "seconds": 2.927, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/64", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-64", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Google, DeepMind, Amazon, Microsoft have hired 52 tenured and tenure-track professors from US Universities between 2004 and 2018.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-a932818e2563", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "chart", "text": "Companies hiring AI professors", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "97a6860b-8113-4a37-b6ba-d85b8e1ac5c0", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Universities losing AI professors", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9d7bb1be-7091-4138-a5b3-27bf82e0bc3a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of AI professors: 52", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:22+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-af9505b884de", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.2 }, "kind": "paragraph", "text": "Google, DeepMind, Amazon, Microsoft have hired 52 tenured and tenure-track professors from US Universities between 2004 and 2018. Carnegie Mellon, U. Washington and Berkeley have lost 38 professors during the same period. Note that no AI professor left in 2004, whereas 41 AI professors left in 2018 alone.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f01fffdb-8fda-4f84-a151-641a26b5433e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.9 }, "kind": "source-note", "text": "Figure note: Graphs include AI professors who completely left academia or reported a dual industry and academic affiliation", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6f4868ae-8451-4a78-873f-7f18e398d480", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "The Great Brain Drain: AI professors depart US universities for technology companies", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "808ab7b2-f134-44bb-855b-8b905c1da92b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Contrast Pairs", "slug": "contrast-pairs", "status": "active", "bestFor": null, "canonId": "019dd956-9b77-70d3-88b1-f34e26f60860", "version": 1, "bodyDocId": null, "description": "Creating tension through opposites: Before↔After, Problem↔Solution, Pain↔Relief", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "loop", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-0b5794aad0e6", "evidence": "Academia vs Big Tech contrast set up across loop.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-03878557c33d", "evidence": "Title: 'The Great Brain Drain'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "e95d919e-3284-4c7b-a412-98d40bc9151d", "evidence": "chart/bar-horizontal: Universities losing AI professors chart/bar-horizontal: Companies hiring AI professors", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.28902+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-05ea1b5a2df4", "evidence": "Brain drain metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 64, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 65, "type": "analysis", "title": "The slide contrasts a positive industry-academic partnership (DeepMind/Cambridge) with a critical perspective on dual affiliation (Berkeley faculty op-ed).", "function": "frame_problem", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "frame_problem", "functionCanon": { "name": "Frame problem", "slug": "frame_problem", "status": "active", "canonId": "019de52d-13db-753e-ad70-e9b2d0dab398", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 79, "componentCount": 9, "textChars": 1212, "nDataPoints": 0, "notes": "The slide contrasts a positive industry-academic partnership (DeepMind/Cambridge) with a critical perspective on dual affiliation (Berkeley faculty op-ed).", "elementsJson": [ "headline_text", "paragraph", "headshot", "logo_grid", "callout_box" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.669+00", "seconds": 3.267, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/65", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-65", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.1, "w": 0.5, "x": 0.45, "y": 0.4 }, "kind": "callout", "text": "You Cannot Serve Two Masters: The Harms of Dual Affiliation", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6a00627c-491e-4dda-8e59-103c81967922", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "This model assumes people can slice their time and attention like a computer, but people can't do this.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:27+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-56046661abcd", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.15, "x": 0.05, "y": 0.43 }, "kind": "image", "text": "DeepMind and Cambridge logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aabd1091-695a-488f-adc0-200a414359b4", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "New professorships may free the ladder for young academic talent to rise. Meanwhile, some companies including Facebook champion the dual academic/industry affiliation as the solution. Some academics don't buy it.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0039bbfa-3ae2-4447-9849-dd3c77792ee2", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.3, "x": 0.05, "y": 0.72 }, "kind": "paragraph", "text": "New DeepMind scholarships launch at Queen Mary to encourage more women in AI", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "33111b7f-f5e4-4cbb-99c8-26a23ba6a62a", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.5, "x": 0.45, "y": 0.6 }, "kind": "paragraph", "text": "Facebook would like to have computer science faculty in AI committed to work 80% of their time in industrial jobs and 20% of their time at their university. They call this scheme \"co-employment\" or \"dual affiliation.\" This model assumes people can slice their time and attention like a computer, but people can't do this.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9177dc58-114d-4c54-8316-da2e96e23d64", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.3, "x": 0.05, "y": 0.53 }, "kind": "paragraph", "text": "Following an international search, Professor Neil Lawrence has been appointed as the inaugural DeepMind Professor of Machine Learning at Cambridge, supported by a benefaction from the world-leading British AI company.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d9540e2e-d206-4bc3-91f5-0897ddc0bc6a", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "You Cannot Serve Two Masters: The Harms of Dual Affiliation — Ben Recht, David A. Forsyth, and Alexei Efros", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:27+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-5973ee8c4009", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Tech companies endow AI professorships in return for poaching, but is this really enough?", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.669+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3dd1a991-b70e-407e-a0e1-78ea724faf1b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Problem-Agitate-Solve (PAS)", "slug": "problem-agitate-solve-pas", "status": "active", "bestFor": null, "canonId": "019dd956-869f-7564-8d42-de54ed29ab90", "version": 1, "bodyDocId": null, "description": "Identify problem, agitate the pain, present the solution", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "caccc75e-dac5-4949-8ad9-efb1cc13ff0e", "evidence": "You Cannot Serve Two Masters: The Harms of Dual Affiliation — Ben Recht, David A. Forsyth, and Alexei Efros", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.358521+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 65, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-0fdef4a916c0", "evidence": "Callout questions adequacy of dual appointments.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 65, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 66, "type": "analysis", "title": "The slide uses a timeline diagram and a regression table to illustrate the causal link between professor departures and student entrepreneurship.", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 61, "componentCount": 6, "textChars": 925, "nDataPoints": 5, "notes": "The slide uses a timeline diagram and a regression table to illustrate the causal link between professor departures and student entrepreneurship.", "elementsJson": [ "headline_text", "bullet_list", "data_table", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.284+00", "seconds": 4.856, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/66", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-66", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "4-6 years after the departure of tenured AI professors, graduates are 4% less likely to start an AI company.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-2a069215bd88", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.45, "x": 0.05, "y": 0.65 }, "kind": "diagram", "text": "Timeline of departures relative to graduation year", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Timeline", "slug": "timeline", "status": "active", "canonId": "019de52d-03c5-73fc-84aa-4180cbf0f5cf", "version": 1, "description": null }, "framework": null }, "subkind": "timeline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c44c9830-3061-4347-af79-19b8af9e0213", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "4-6 years after the departure of tenured AI professors, graduates are 4% less likely to start an AI company (a).\nThis finding does not hold when professors leave 1-3 years before a student’s graduation, suggesting that the interaction between professor and student is important (b).\nThere is also no significant correlation between AI professor departure and the formation of non-AI companies by graduates at the same university (c).\nThere is ongoing debate within the AI community about how significant this is.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6b57605e-9308-475e-ab56-2802bd44f417", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI startup formation probability: 4%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-2dd560ed7296", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.5, "y": 0.6 }, "kind": "table", "text": "Regression results table showing coefficients for AI Entrepreneur and Non-AI (IT) outcomes.", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bd284c1b-01a8-499b-b56f-a35e0196a95a", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "The loss of AI professors seems to matter: Departures correlate with reduced graduate entrepreneurship across 69 US universities", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f64c1ceb-43bf-46d2-8edd-0f26e92eb7db", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-1348a2c02cb3", "evidence": "Title: 'departures correlate with reduced graduate entrepreneurship'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Annotation", "slug": "annotation", "status": "active", "bestFor": null, "canonId": "019dd956-b5b9-7668-bf0e-3af99823ec5c", "version": 1, "bodyDocId": null, "description": "Adding explanatory labels to highlight key data points", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-17183707af9a", "evidence": "Red boxes mark a/b/c rows in regression table.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Von Restorff Effect", "slug": "von-restorff-effect", "status": "active", "bestFor": null, "canonId": "019dd956-beab-73d8-9f62-30d0332e926e", "version": 1, "bodyDocId": null, "description": "Items that stand out are noticed and remembered - make key points visually distinct", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-1aa87e4f91eb", "evidence": "Red highlight makes key coefficients pop.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "9c5c3748-c463-407e-a865-3bdec5d7acc8", "evidence": "The slide links professor departures to student outcomes over a specific time horizon.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:40.284+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 66, "frameworkName": "causal-chain" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 67, "type": "evidence", "title": "The slide highlights a specific institutional investment in AI talent.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 51, "componentCount": 6, "textChars": 558, "nDataPoints": 3, "notes": "The slide highlights a specific institutional investment in AI talent.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.69+00", "seconds": 3.234, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/67", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-67", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The Eindhoven Artificial Intelligence Systems Institute in The Netherlands plans to recruit 50 professors (!)", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-f1cbf7059d92", "parentComponentId": null }, { "bbox": { "h": 0.285, "w": 0.5, "x": 0.452, "y": 0.315 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.69+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "88027635-187c-4564-9645-885405e0a0c8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "The Eindhoven Artificial Intelligence Systems Institute in The Netherlands plans to recruit 50 professors (!)", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.69+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "23bbae28-4b00-4e90-97d4-1f1b17a5bf7d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Investment: €100M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa65-f4cd316f9d86", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.45, "x": 0.04, "y": 0.3 }, "kind": "paragraph", "text": "The Eindhoven University of Technology (TUE) has committed to spending €100M over 5 years to create a new institute that will focus on the use of smart algorithms in machines, like robots and autonomous cars. TUE was ranked 120 in the QS World University Ranking 2021.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.69+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "49207c41-9b5f-4868-b458-bbeef8bb719b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Can €100M buy you 50 professors for a new AI Institute?", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.69+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e944da56-e81c-4b36-8a40-554c14955dea", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-1e3e9679d757", "evidence": "EUR100M / 50 professors specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 67, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 68, "type": "evidence", "title": "Slide from State of AI 2020 report.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 48, "componentCount": 5, "textChars": 756, "nDataPoints": 1, "notes": "Slide from State of AI 2020 report.", "elementsJson": [ "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.892+00", "seconds": 2.431, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/68", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-68", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "11 corporate partners joined, including The Jackson Laboratory (a major non-profit biomedical research center).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-0e189591f046", "parentComponentId": null }, { "bbox": { "h": 0.48, "w": 0.43, "x": 0.52, "y": 0.34 }, "kind": "image", "text": "Group photo of Roux Institute partners and leadership", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.892+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "da65664b-f5fa-4052-aea0-4ed2b8efe09b", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "11 corporate partners joined, including The Jackson Laboratory (a major non-profit biomedical research center). The initial program portfolio will cover two broad disciplines: applied analytics, computer science, data science, data visualization, and machine learning. AI-first biology, which includes cover bioinformatics, biotechnology, genomics, health data analytics, and precision medicine.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.892+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2336c5b9-728f-4bc6-97f7-f28474b4384d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Donation amount: $100M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-122cd2f41214", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "A $100M donation from Silver Lake founder to mint the Roux Institute at Northeastern University: New graduate degrees that focus on AI applied to the digital and life sciences", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.892+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a3210356-ac87-42dc-814e-ae5273bc3e34", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-20f4d319886f", "evidence": "$100M Roux Institute donation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 68, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 69, "type": "evidence", "title": "The slide validates a previous prediction by showcasing the successful launch of MBUZAI.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 61, "componentCount": 6, "textChars": 985, "nDataPoints": 3, "notes": "The slide validates a previous prediction by showcasing the successful launch of MBUZAI.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.988+00", "seconds": 3.491, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/69", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-69", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "In our 2019 Predictions we predicted: “Institutes of higher education establish purpose-built AI degrees to fill talent void.”", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-7f584f63d21f", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.43, "x": 0.52, "y": 0.33 }, "kind": "image", "text": "Campus photo of MBUZAI", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "812ce82d-e4df-4eb1-b6ab-b87d6cba12d2", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "MBUZAI received 2,223 applicants of 97 nationalities. Those admitted come from 31 countries, with a majority hailing from outside the MENA region.\nThe Interim President of MBUZAI is Professor Sir Michael Brady who was formerly Professor of Information Engineering at Oxford University.\nThe Board of Trustees includes Kai-Fu Lee of Sinovation Ventures and former President of Google China, and Daniela Rus of MIT’s CSAIL.\nSorbonne University Abu Dhabi launched their own bachelor’s degree in maths and data science for AI.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3cb5bdec-9e43-4f80-b0bb-dbe540f27361", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "applicants: 2,223", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-8016f6868c0e", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "In our 2019 Predictions we predicted: “Institutes of higher education establish purpose-built AI degrees to fill talent void.” Mohamed bin Zayed University of AI (MBUZAI) is a new research-based institute of higher education.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7fb9f004-d6f7-42a3-a524-e41702d03a3d", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "2019 Prediction outcome: Abu Dhabi opens the “World’s first AI University”", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.988+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "be3a8763-b17a-44af-b402-81c67cc4dc8f", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 69, "from": 64, "loop": { "name": "04_tale_two_worlds", "slug": "04-tale-two-worlds", "status": "active", "bestFor": "Competitive analysis, benchmarking, case for change", "canonId": "019dd956-674f-750b-b15d-f082e2717f4e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current State -> Desired State / Benchmark -> The Gap & Implication", "description": "Show the gap between two states to drive urgency or highlight opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-01ee6de4b667", "evidence": "Brain-drain charts, endowed professorships, new institutes responding.", "position": 7, "objective": "Contrast academia vs industry pull on AI talent", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 70, "type": "analysis", "title": "The slide tracks the origin of AI research talent, specifically those with undergraduate degrees from China.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 51, "componentCount": 6, "textChars": 502, "nDataPoints": 11, "notes": "The slide tracks the origin of AI research talent, specifically those with undergraduate degrees from China.", "elementsJson": [ "headline_text", "action_title", "bar_chart_vertical", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.513+00", "seconds": 3.997, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/70", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-70", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "29% of authors with papers accepted at NeurIPS 2019 earned their undergraduate degree in China.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-98130fb447eb", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.6, "x": 0.2, "y": 0.4 }, "kind": "chart", "text": "Bar chart showing percentage of authors with undergraduate degrees from China (2009-2019)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.513+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "08e26335-db84-46b4-a7e4-69ecbc740d96", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "29% of authors with papers accepted at NeurIPS 2019 earned their undergraduate degree in China.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.513+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "34c9bc93-147d-4e80-b761-2cf20850ccd5", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Percentage of NeurIPS authors with undergraduate degrees from China: 29%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-9c6d1a947a3b", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.5, "x": 0.02, "y": 0.02 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.513+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "524f5c32-cf40-4c9d-99d7-1370ba2363a7", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Chinese-educated researchers make increasingly significant contributions at NeurIPS", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.513+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4fbf6b3f-e8dd-48f4-ad96-b153688c77a9", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "e13e0144-f513-4e80-a733-b89aabe6b1a1", "evidence": "chart/bar-vertical: Bar chart showing percentage of authors with undergraduate degrees from China (2009-2019)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:01.584286+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 70, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-248abc4687b6", "evidence": "29% of NeurIPS authors educated in China.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 70, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 71, "type": "analysis", "title": "The chart visualizes talent migration patterns for AI researchers.", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 28, "componentCount": 4, "textChars": 265, "nDataPoints": 24, "notes": "The chart visualizes talent migration patterns for AI researchers.", "elementsJson": [ "headline_text", "callout_box", "sankey" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.312+00", "seconds": 3.601, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/71", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-71", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.6, "x": 0.05, "y": 0.28 }, "kind": "callout", "text": "The US attracts over half of foreign NeurIPS 2019 authors by the time they finish undergrad.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.312+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8f9db606-52df-443f-b1cb-4334e077643d", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "chart", "text": "Share in each country of each stage of career", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Sankey diagram", "slug": "sankey", "status": "active", "canonId": "019de52d-0145-711d-a871-198a7a06733b", "version": 1, "description": null }, "framework": null }, "subkind": "sankey", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.312+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "68a93a5e-7fef-425e-8a94-12b9617288ee", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Share of researchers: 54%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-688712190c64", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "But after leaving university in China, 54% of graduates who go on to publish at NeurIPS move to the USA", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.312+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a20622c6-3c62-4f69-be60-31c587b9a4d4", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-285f4304f838", "evidence": "Title: 54% of Chinese grads move to USA.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 71, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [ { "matchId": "f59dc316-82f9-41a4-82bb-9a3e68d25976", "evidence": "Flow chart showing migration between career stages", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:39.312+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 71, "frameworkName": "Sankey Diagram" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 72, "type": "resolution", "function": "summarize", "rawType": "key_takeaways", "block": null, "metadata": { "slideType": "key_takeaways", "slideTypeCanon": { "name": "Key takeaways", "slug": "key_takeaways", "status": "active", "canonId": "019de52d-06e7-7674-a891-9f7de3b11b6d", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 6, "textChars": 348, "nDataPoints": 4, "notes": null, "elementsJson": [ "headline_text", "action_title", "paragraph", "pie_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:37.923+00", "seconds": 2.15, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/72", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-72", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Almost 90% of Chinese and non-Chinese students who earn an American PhD are retained in the US for work.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-4849a72a8c77", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.3, "x": 0.2, "y": 0.35 }, "kind": "chart", "text": "Chinese PhD students retention", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Pie chart", "slug": "pie", "status": "active", "canonId": "019de52c-ffdb-71bb-a800-5c5ab9f1c373", "version": 1, "description": null }, "framework": null }, "subkind": "pie", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.923+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "33319b04-79df-4596-8648-45dd30f3779a", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.3, "x": 0.55, "y": 0.35 }, "kind": "chart", "text": "International PhD students retention", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Pie chart", "slug": "pie", "status": "active", "canonId": "019de52c-ffdb-71bb-a800-5c5ab9f1c373", "version": 1, "description": null }, "framework": null }, "subkind": "pie", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.923+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b45fbfbb-6f57-4a39-9794-df2c60ef47c5", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Retention rate: 90%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:23+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-4ca76cea4f18", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Almost 90% of Chinese and non-Chinese students who earn an American PhD are retained in the US for work.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.923+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ef2b5e78-0d89-4183-bb3a-9ee2810220a8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "The US is an incredibly strong talent retainer post-PhD", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:37.923+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d69f7df7-1004-49e0-b5c6-c39f9c08321a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-2d51924b8432", "evidence": "Title: US strong talent retainer post-PhD.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 72, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-33638cdacd15", "evidence": "90% retention statistic.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 72, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 73, "type": "resolution", "title": "The slide highlights the role of H1B visa sponsorship in career outcomes for AI PhD graduates.", "function": "quantify_impact", "rawType": "key_takeaways", "block": null, "metadata": { "slideType": "key_takeaways", "slideTypeCanon": { "name": "Key takeaways", "slug": "key_takeaways", "status": "active", "canonId": "019de52d-06e7-7674-a891-9f7de3b11b6d", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "balanced", "densityScore": 41, "componentCount": 5, "textChars": 460, "nDataPoints": 8, "notes": "The slide highlights the role of H1B visa sponsorship in career outcomes for AI PhD graduates.", "elementsJson": [ "headline_text", "callout_box", "bar_chart_stacked", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.418+00", "seconds": 7.6, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/73", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-73", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.28 }, "kind": "callout", "text": "Foreign nationals are 2x more likely to join large companies, in part due to their H1B sponsoring power.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.418+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "263e5051-efd0-408d-9ee7-f23bb72b8d9f", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.3, "y": 0.4 }, "kind": "chart", "text": "Stacked bar chart comparing career sectors for Foreign vs US Nationals", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.418+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6afb0097-a0ee-4dfc-8a63-aae09ff7edb5", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.4, "x": 0.3, "y": 0.38 }, "kind": "legend", "text": "Private sector (large co), Private sector (small co), Academic, Government/Nonprofit", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.418+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e6fdc912-4b33-4c64-a1de-dcd008991ee2", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Career placement percentage: 54%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-7ad72671c4b1", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Foreign national graduates of US AI PhD programs are most likely to end up in large companies whereas American nationals are more likely to end up in startups or academia", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.418+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "15d3d673-9e07-475e-b2ba-b46c0403cf57", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "498b0bdf-d1bf-4641-929f-d71e24dc6aa8", "evidence": "legend: Private sector (large co), Private sector (small co), Academic, Government/Nonprofit", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:56.02336+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 73, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-3466d1fe84d5", "evidence": "Title: foreign vs American grad placement contrast.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 73, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 74, "type": "resolution", "title": "Data source: CSET. Part of the state of AI 2020 report.", "function": "analyze_data", "rawType": "key_takeaways", "block": null, "metadata": { "slideType": "key_takeaways", "slideTypeCanon": { "name": "Key takeaways", "slug": "key_takeaways", "status": "active", "canonId": "019de52d-06e7-7674-a891-9f7de3b11b6d", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 40, "componentCount": 5, "textChars": 423, "nDataPoints": 10, "notes": "Data source: CSET. Part of the state of AI 2020 report.", "elementsJson": [ "headline_text", "bar_chart_vertical", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:38.247+00", "seconds": 2.103, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/74", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-74", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "55% of graduates moving to the UK take private sector jobs; 40% of those who move to China do the same.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-b748c04d3417", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.35, "x": 0.33, "y": 0.35 }, "kind": "chart", "text": "Destination countries of AI PhD students who leave the US post-graduation", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.247+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "97890c4f-1551-41aa-90a1-e4ac689405a3", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Percentage of AI PhD graduates: 18%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-b956f84d8944", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.9, "x": 0.05, "y": 0.28 }, "kind": "paragraph", "text": "55% of graduates moving to the UK take private sector jobs; 40% of those who move to China do the same.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.247+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "eecbc7c1-a4d2-4875-98e9-ae008caf2f09", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "The UK and China are the biggest beneficiaries of American-educated AI PhDs who leave the US after graduation", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:38.247+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "853c1888-ea6c-45fe-b14c-cdeb902fee96", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-3a98774747a5", "evidence": "55%/40% UK/China sector splits.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 74, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 75, "type": "analysis", "title": "The chart uses a waterfall-style visualization to show cumulative distribution of talent origins.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 49, "componentCount": 6, "textChars": 471, "nDataPoints": 7, "notes": "The chart uses a waterfall-style visualization to show cumulative distribution of talent origins.", "elementsJson": [ "waterfall_chart", "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.002+00", "seconds": 2.462, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/75", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-75", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "China (27%), Europe (11%), and India (11%) are the largest feeder nations for US institutions.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:26+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-1f894c456249", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.6, "x": 0.2, "y": 0.035 }, "kind": "chart", "text": "Waterfall chart showing percentages: 31% (US), 27% (China), 11% (India), 11% (Europe), 4% (Iran), 3% (Canada), 13% (Others)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Waterfall chart", "slug": "waterfall", "status": "active", "canonId": "019de52d-002e-733e-b5cf-dd3beb6692cc", "version": 1, "description": null }, "framework": null }, "subkind": "waterfall", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.002+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b9ac36f4-dc13-4208-8d00-0a4c288216c7", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI researcher origin: 27%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:26+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-23da9550ced4", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.5, "x": 0.25, "y": 0.085 }, "kind": "paragraph", "text": "Country from which an individual earned their undergrad", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.002+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b6609c8c-cac7-44ef-b180-b67868500236", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.8, "x": 0.05, "y": 0.22 }, "kind": "title", "text": "China (27%), Europe (11%), and India (11%) are the largest feeder nations for US institutions.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.002+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "936d3a8f-3a67-4216-98c3-dfdcca29122c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "The majority of top AI researchers working in the US were not trained in America", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.002+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fef40876-4e8b-4d54-9914-bc76bdd0bc7f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-3e796159f533", "evidence": "China 27%, Europe 11%, India 11% feeders.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 75, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Waterfall chart", "slug": "waterfall-chart", "status": "active", "bestFor": null, "canonId": "e2799da3-725f-44a9-a093-570f1f388e64", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ab826dad-25f7-4377-8b17-74a8c447e5b2", "evidence": "chart/waterfall: Waterfall chart showing percentages: 31% (US), 27% (China), 11% (India), 11% (Europe), 4% (Iran), 3% (Canada), 13% (Others)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:56.150097+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 75, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 76, "type": "analysis", "title": "Includes references to specific presidential proclamations and data on foreign student retention.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 75, "componentCount": 8, "textChars": 790, "nDataPoints": 12, "notes": "Includes references to specific presidential proclamations and data on foreign student retention.", "elementsJson": [ "headline_text", "paragraph", "screenshot", "bar_chart_vertical", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.107+00", "seconds": 7.537, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/76", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-76", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Note that 92% of top international US AI PhD graduates work in the US post-graduation and 80% intend to stay if they can.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-1eaf1ed8ddbf", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.25, "x": 0.32, "y": 0.45 }, "kind": "chart", "text": "Foreign students working in the US post-graduation", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "76f42fc0-cf96-4717-804a-e1613ea62f26", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.35, "x": 0.62, "y": 0.45 }, "kind": "chart", "text": "Post-graduation intent for foreign AI students at US graduate schools", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6b61d47b-f4b8-4118-8172-dd4aeef7d4d4", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.25, "x": 0.02, "y": 0.45 }, "kind": "image", "text": "Proclamation documents", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3a2a3e88-9ed4-4fd0-8df8-de83acfdcca5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "President Trump suspended the entry of aliens into the US during COVID-19 and then retreated. Note that 92% of top international US AI PhD graduates work in the US post-graduation and 80% intend to stay if they can.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "64a30cd8-dd9e-4871-998a-0e288ca7236f", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Retention rate of foreign AI PhD graduates: 92%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-23a29f993745", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.95, "x": 0.02, "y": 0.14 }, "kind": "paragraph", "text": "Given how dependent America's AI industry is on immigrants there has been a strong backlash to Trump's proclamation to suspend H1-B visas. Eight federal lawsuits and hundreds of universities object.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0b35e52a-374d-41d8-a8e8-2d01a81a1fac", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.02, "y": 0.03 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.107+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "43458778-f0d1-450f-a534-da362f7b2e7e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-41cceb38a085", "evidence": "Callout pulls implication of H-1B suspension.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 76, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 76, "from": 70, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-061965a0c0d4", "evidence": "Six slides on China feeders, retention, foreign nationals, H-1B backlash.", "position": 8, "objective": "Map global flows of AI talent into the US", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 77, "type": "setup", "title": "The chart uses fractional counts for paper authorship.", "function": "compare_peers", "rawType": "competitive_analysis", "block": null, "metadata": { "slideType": "competitive_analysis", "slideTypeCanon": { "name": "Competitive analysis", "slug": "competitive_analysis", "status": "active", "canonId": "019de52d-093e-776b-94c6-c1573f5f1a31", "version": 1, "description": null }, "function": "compare_peers", "functionCanon": { "name": "Compare peers", "slug": "compare_peers", "status": "active", "canonId": "019de52d-1453-72ba-a367-9ba3d34c1630", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 6, "textChars": 337, "nDataPoints": 24, "notes": "The chart uses fractional counts for paper authorship.", "elementsJson": [ "headline_text", "action_title", "bar_chart_horizontal" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.408+00", "seconds": 5.815, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/77", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-77", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Google, Stanford, CMU, MIT and Microsoft Research own the Top-5.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-cc20f086d7a3", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.6, "x": 0.2, "y": 0.28 }, "kind": "chart", "text": "Horizontal bar chart of NeurIPS 2019 paper counts by institution", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.408+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cf2982a9-5001-4b75-9e3c-2d28bea9b3d8", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.25, "x": 0.58, "y": 0.41 }, "kind": "legend", "text": "USA, UK, China, Europe, Canada", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.408+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4cc9bad1-2fac-46ce-b183-5178f74dda45", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Papers published at NeurIPS 2019: 90", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa66-d093fa39892b", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.5, "x": 0.05, "y": 0.22 }, "kind": "title", "text": "Google, Stanford, CMU, MIT and Microsoft Research own the Top-5.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.408+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ebdcdfd2-e0ec-4a92-b6cb-46780ff3f66a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "American institutions and corporations continue to dominate NeurIPS 2019 papers", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.408+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "313a892f-efcd-4d5f-b86b-1a2f5226ff59", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-47ed1576e33d", "evidence": "Title: American institutions dominate NeurIPS.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 77, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-4915c6ade0f5", "evidence": "Ranked bar chart of top contributors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 77, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 78, "type": "analysis", "title": "The chart displays the Publication Index position gains for top 20 organizations at ICML 2020 vs 2019.", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 52, "componentCount": 6, "textChars": 504, "nDataPoints": 18, "notes": "The chart displays the Publication Index position gains for top 20 organizations at ICML 2020 vs 2019.", "elementsJson": [ "headline_text", "bar_chart_vertical", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.029+00", "seconds": 2.205, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/78", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-78", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The top 20 most prolific organisations by ICML 2020 paper acceptances further cemented their position vs. ICML 2019.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-039adfeb10ae", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.9, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Bar chart of Publication Index position gains for 18 organizations", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.029+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7f997abc-8310-4e62-bf2a-154acf70e39d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Publication Index: 19.4", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-0791877e11aa", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "The top 20 most prolific organisations by ICML 2020 paper acceptances further cemented their position vs. ICML 2019. The chart below shows their Publication Index position gains vs. ICML 2019.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.029+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8a3da5f3-ccd1-463d-88fa-781f607d382f", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.2, "x": 0.05, "y": 0.93 }, "kind": "source-note", "text": "Credit: Gleb Chuvpilo", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.029+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bc8f1088-fac8-459c-b490-088cf895e978", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "The same is true at ICML 2020: American organisations cement their leadership position", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.029+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0ca7d31f-9e7f-4366-8758-c8d9a79bbbbf", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-4d8bcd8ceb73", "evidence": "Title: ICML 2020 cements US lead.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 78, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 79, "type": "analysis", "title": "The chart illustrates the surge in interest in AI/NLP education at Stanford over two decades.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 50, "componentCount": 6, "textChars": 401, "nDataPoints": 126, "notes": "The chart illustrates the surge in interest in AI/NLP education at Stanford over two decades.", "elementsJson": [ "headline_text", "bar_chart_stacked" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.386+00", "seconds": 5.445, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/79", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-79", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Stanford now teaches 10x the students per year as during 1999–2004, and twice as many as 2012–2014.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-405c0a7b47d1", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.6, "x": 0.2, "y": 0.3 }, "kind": "chart", "text": "Stanford NLP class enrollment", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dfb94930-a84e-4a60-b544-95d4924502c7", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.2, "x": 0.25, "y": 0.35 }, "kind": "legend", "text": "CS224D (DL NLP), CS224S (speech), CS276 (IR), CS224U (NLU), CS124 (ugrad HLT), CS224N (NLP)", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "957b1750-a4c9-4c53-9315-70195723cc7c", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Student enrollment: 10x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-46a0968233f4", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.9, "x": 0.04, "y": 0.22 }, "kind": "title", "text": "Stanford now teaches 10x the students per year as during 1999–2004, and twice as many as 2012–2014.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "711c04fc-3f99-4a0d-89b9-185a1ce4d8e3", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Leading Universities continue to expand AI course enrollment", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b99880d9-fe3f-4835-aac2-f3b5ffcf9dd8", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-53cfb90f551d", "evidence": "Stanford 10x enrollment growth.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 79, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 80, "type": "analysis", "title": "Data source: Indeed.com US data. Timeframe: 2015-2018.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 50, "componentCount": 6, "textChars": 543, "nDataPoints": 2, "notes": "Data source: Indeed.com US data. Timeframe: 2015-2018.", "elementsJson": [ "headline_text", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.146+00", "seconds": 4.883, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/80", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-80", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Analysis of Indeed.com US data shows almost 3x more job postings than job views for AI-related roles.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-5b072ca334a5", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.75, "x": 0.05, "y": 0.35 }, "kind": "chart", "text": "Line chart comparing postings per million vs searches per million from 2015 to 2018", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.146+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2a5e7a75-5e51-418e-9310-b3d3738715a9", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.15, "x": 0.82, "y": 0.38 }, "kind": "legend", "text": "postings per million (4-week moving average) 1,124.1; searches per million (4-week moving average) 398.3", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.146+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "573c18a6-161f-49e0-b953-b0c7533c1d31", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Job postings vs. searches: 3x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:24+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-5ea15703c6a5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Analysis of Indeed.com US data shows almost 3x more job postings than job views for AI-related roles. Job postings grew 12x faster than job viewings in the last from late 2016 to late 2018.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.146+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8ab69ef3-9bbf-4035-b9ef-71665d461f8a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.4, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Demand outstrips supply for AI talent", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.146+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8aa7b4c2-078c-4f11-ba12-481c30265167", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-54b06cfed15f", "evidence": "Title: 'Demand outstrips supply'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 80, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-5b7c2247f8f2", "evidence": "3x more postings than views.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 80, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 81, "type": "analysis", "title": "The chart uses Exponential Moving Averages (EMA) for the data series.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 7, "textChars": 506, "nDataPoints": 2, "notes": "The chart uses Exponential Moving Averages (EMA) for the data series.", "elementsJson": [ "headline_text", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.588+00", "seconds": 3.185, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/81", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-81", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Public job postings on LinkedIn that mention a deep learning framework were on a strong 2020 ramp up but took a hit due to COVID-19 since February 2020", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-91679d8fcaa8", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.75, "x": 0.15, "y": 0.35 }, "kind": "chart", "text": "Line chart showing job postings for TF/Keras and Pytorch over time", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.588+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6ce048ac-9ed3-4858-8788-62c392da8999", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.2, "x": 0.02, "y": 0.94 }, "kind": "disclaimer", "text": "Credit: François Chollet", "attrs": null, "canon": { "kind": { "name": "Disclaimer", "slug": "disclaimer", "status": "active", "canonId": "91173477-d91f-4024-bbaf-d4f93e3f82dc", "version": 1, "description": "Legal/cautionary text that occupies most of the body of its slide." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.588+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "358950e5-05e5-49db-810c-b7c8ed679e12", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.45, "x": 0.26, "y": 0.385 }, "kind": "legend", "text": "TF | Keras (EMA) - Pytorch (EMA)", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.588+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c2d417e0-f925-4d8a-853a-6579cd9f0b3a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Job postings", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-97962d739ea5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "Public job postings on LinkedIn that mention a deep learning framework were on a strong 2020 ramp up but took a hit due to COVID-19 since February 2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.588+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9d321caf-79e7-4d22-8985-d8eacd664f35", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.75, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "While hot, the AI talent market is not immune to the COVID-19 pandemic", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.588+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1234a81e-e7b9-489b-82b3-c56171836af7", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-5ef0efa0c171", "evidence": "Callout ties COVID to LinkedIn posting drop.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 81, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 81, "from": 77, "loop": { "name": "39_benchmark_gap", "slug": "39-benchmark-gap", "status": "active", "bestFor": "Performance improvement, competitive analysis, target setting", "canonId": "019dd956-75a4-7522-8d1a-af94da4703e9", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Our Performance -> Industry Average -> Best-in-Class -> The Gap = The Opportunity", "description": "Compare performance against best-in-class to quantify the opportunity", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-085e655f03be", "evidence": "Slide 80 explicitly: 3x more job postings than views; 79 expanding enrollment.", "position": 9, "objective": "Benchmark AI talent supply vs demand gap", "confidence": 65, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 82, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 13, "componentCount": 4, "textChars": 112, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:39.977+00", "seconds": 2.211, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/82", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-82", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.055, "x": 0.925, "y": 0.035 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.977+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1aed2d6e-1db9-4f2a-b0e1-2dc2182f0af6", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.45, "x": 0.02, "y": 0.035 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.977+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "936276f6-b333-4ba0-b319-a2dbd04d8100", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.061, "x": 0.919, "y": 0.95 }, "kind": "paragraph", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.977+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d9ad9bbf-44d0-4d6f-8502-8aeeb715521b", "parentComponentId": null }, { "bbox": { "h": 0.042, "w": 0.268, "x": 0.366, "y": 0.458 }, "kind": "title", "text": "Section 3: Industry", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:39.977+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ee25694e-afa7-4da0-b618-0a44ac8e8e53", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9ae644a9-434c-4f01-9619-449cc6a0c769", "evidence": "title/headline: Section 3: Industry", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:56:56.467722+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 82, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 83, "type": "evidence", "title": "The slide highlights the efficiency gains of AI in drug discovery, noting a 20% reduction in candidate testing compared to traditional methods.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 56, "componentCount": 5, "textChars": 1258, "nDataPoints": 3, "notes": "The slide highlights the efficiency gains of AI in drug discovery, noting a 20% reduction in candidate testing compared to traditional methods.", "elementsJson": [ "headline_text", "bullet_list", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.265+00", "seconds": 6.483, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/83", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-83", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Only 350 candidates were tested in real-world experiments to find DSP-1181. This is around 20% of the normal number of compound candidates that are tested in a typical campaign.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-b3b9e5d4f2ab", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.25, "x": 0.02, "y": 0.092 }, "kind": "image", "text": "Exscientia and Sumitomo Dainippon Pharma logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.265+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f84031d9-91a1-437d-803a-f2252ae33a73", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.96, "x": 0.02, "y": 0.29 }, "kind": "list", "text": "The drug, DSP-1181, acts as an agonist to the receptor for serotonin, a signaling molecule in the brain that mediates mood.\nWhile the mechanism of obsessive-compulsive disorder has not been definitively established, it is believed that increasing serotonin signaling using receptor agonists could improve OCD symptoms.\nThis study used AI techniques to generate tens of millions of potential molecules against the serotonin receptor and sift through the candidates to decide which ones to prioritise for synthesis and testing.\nOnly 350 candidates were tested in real-world experiments to find DSP-1181. This is around 20% of the normal number of compound candidates that are tested in a typical campaign.\nSeparately, Exscientia expanded its commercial partnerships by adding a €240M deal with Bayer on cardiovascular disease and oncology,", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.265+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "72514683-c619-471f-9f49-4f53876ee868", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.96, "x": 0.02, "y": 0.23 }, "kind": "paragraph", "text": "The result of a 12 month collaboration between British Exscientia and Japanese Sumitomo Dainippon Pharma.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.265+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bb9e449e-f037-44f3-85fc-3d6acb23f0b3", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "The first phase 1 clinical trial of an AI-designed drug begins in Japan to treat OCD patients", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.265+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "48159a3d-d37e-40bf-bc78-f2aaeb24dcce", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-6334a7dee2c0", "evidence": "Title states first phase 1 trial milestone.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 83, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 84, "type": "evidence", "title": "The slide uses a simple timeline structure to demonstrate the validation of AI in pharma.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 68, "componentCount": 7, "textChars": 907, "nDataPoints": 2, "notes": "The slide uses a simple timeline structure to demonstrate the validation of AI in pharma.", "elementsJson": [ "headline_text", "paragraph", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.731+00", "seconds": 2.634, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/84", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-84", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "29 months after signing a €250M deal to evaluate over one thousand combinations of immunological drug targets for potential synergistic effects with bispecific small molecules, Exscientia discovers a novel, first-in-class small molecule that Sanofi will now progress.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-c8aa60b12b69", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.05, "y": 0.92 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.731+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1c2031e0-6207-4794-80d9-80a7297dc285", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.25, "x": 0.18, "y": 0.4 }, "kind": "list", "text": "MAY 9, 2017: Exscientia announces EUR250 million collaboration for a multiple-product development and licence option agreement with Sanofi", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Timeline item", "slug": "timeline-item", "status": "active", "canonId": "019de52c-fd0e-74ce-9443-440fbd8c23b4", "version": 1, "description": null }, "framework": null }, "subkind": "timeline-item", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.731+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "831b446b-67cd-426e-aa88-5817161e76cc", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.25, "x": 0.55, "y": 0.4 }, "kind": "list", "text": "AUGUST 5, 2019: Sanofi Exercises option on Exscientia-designed bispecific small molecule for treatment of immunological conditions", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Timeline item", "slug": "timeline-item", "status": "active", "canonId": "019de52c-fd0e-74ce-9443-440fbd8c23b4", "version": 1, "description": null }, "framework": null }, "subkind": "timeline-item", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.731+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "987619ce-6338-4a11-bdc7-4f3ae0f2e502", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Deal value: €250M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa67-cdea7a231564", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "29 months after signing a €250M deal to evaluate over one thousand combinations of immunological drug targets for potential synergistic effects with bispecific small molecules, Exscientia discovers a novel, first-in-class small molecule that Sanofi will now progress.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.731+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c488077b-7db6-4cae-a0bc-084adbde72b5", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Emerging evidence that large pharma is validating AI-first therapeutic discovery outputs", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.731+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ab73f648-49c6-4fa8-b626-fac23892a31d", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6723d58f387f", "evidence": "EUR250M deal, 29 months timeline specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 84, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 85, "type": "analysis", "title": "The slide illustrates the 'platform strategy' model where parent AI companies spin off specific drug assets into independent entities.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 66, "componentCount": 7, "textChars": 739, "nDataPoints": 7, "notes": "The slide illustrates the 'platform strategy' model where parent AI companies spin off specific drug assets into independent entities.", "elementsJson": [ "headline_text", "paragraph", "logo_grid", "process_diagram" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.895+00", "seconds": 2.704, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/85", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-85", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Platform technologies give rise to promising drug assets that are spun off into independent entities following the successful “asset focused” company building approach of the life science sector and its investors.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:26+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-3771b9dc4702", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.8, "x": 0.1, "y": 0.55 }, "kind": "diagram", "text": "Flowchart showing Recursion to CereXis and Atomwise to X37 spin-offs", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "880ba881-84e4-45b9-bc22-78c3440dc763", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.35 }, "kind": "image", "text": "Recursion, Healx, Exscientia, Insitro, Atomwise logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "67746753-5939-4fb4-82c1-e4d34cefb51c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.45 }, "kind": "metric", "text": "$121M Series C, $56M Series C, $60M Series C, $143M Series B, $123M Series B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": null, "subkind": { "name": "Big number", "slug": "big-number", "status": "active", "canonId": "019de52c-fc70-70c8-b965-0584548a9204", "version": 1, "description": null }, "framework": null }, "subkind": "big-number", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "81dbe970-0b4b-4e4e-9929-9a81bf6436ce", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Funding amount: $239M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:26+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-3bdb96dd3375", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "Platform technologies give rise to promising drug assets that are spun off into independent entities following the successful “asset focused” company building approach of the life science sector and its investors.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b2a69995-65a7-42fb-a5b5-f1aae858654f", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "AI-first drug discovery startups raising mega rounds and fulfilling their “platform strategies”", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7a6f22f-b09e-4c01-86c2-75a427ec1c84", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "AIDA Model", "slug": "aida-model", "status": "active", "bestFor": null, "canonId": "019dd956-86fc-773c-813f-d416714b8413", "version": 1, "bodyDocId": null, "description": "Attention, Interest, Desire, Action - classic persuasion sequence", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "471ac015-ba9b-4bcf-bad8-14d4d587cd34", "evidence": "The components mention a 'platform strategies' and 'asset focused' company building approach which hints at the AIDA model's Attention, Interest, Desire, and Action stages.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.736566+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6bd69a90d122", "evidence": "$239M mega round figure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "10999e37-7235-4cf6-b9e7-5b9fa572e5a5", "evidence": "The slide explicitly describes spinning off drug assets into independent entities.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:40.895+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 85, "frameworkName": "breakup-spinoff" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 86, "type": "evidence", "title": "The slide features a logo grid of the project partners.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 5, "textChars": 802, "nDataPoints": 0, "notes": "The slide features a logo grid of the project partners.", "elementsJson": [ "headline_text", "paragraph", "bullet_list", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:40.967+00", "seconds": 2.719, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/86", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-86", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Project MELLODY is the machine learning ledger orchestration for drug discovery.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:25+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-0197da47226c", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.8, "x": 0.1, "y": 0.45 }, "kind": "image", "text": "Logos of Amgen, Astellas, AstraZeneca, Bayer, Iktos, KU Leuven, Kubermatic, Boehringer Ingelheim, GSK, Janssen, Merck, NVIDIA, Owkin, Substra, Novartis, Servier", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.967+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9f809520-626b-4375-9cfd-d34c02c61e82", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "The goal is to build a platform that makes it possible to learn from multiple sets of proprietary data while respecting their highly confidential nature, as data and asset owners will retain control of their information throughout the project.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.967+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fc218c9f-daa3-444e-a355-54c294cfc812", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.23 }, "kind": "paragraph", "text": "In our 2019 Predictions we predicted: “Privacy-preserving ML techniques are adopted by a non-GAFAM Fortune 2000 company”. Project MELLODY is the machine learning ledger orchestration for drug discovery.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.967+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "70bd4843-649c-4a4c-b5d5-130869567542", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "2019 Prediction outcome: Large pharma and startups ally around privacy-preserving machine learning for drug discovery", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:40.967+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "316e3fa7-3ae3-423c-b9c4-648b76f4017a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "SCQA Framework", "slug": "scqa-framework", "status": "active", "bestFor": null, "canonId": "019dd956-7bcb-763f-bdf6-1dfedf3bd36c", "version": 1, "bodyDocId": "019df22a-a0b0-719e-937e-2ac31e53d32c", "description": "Situation, Complication, Question, Answer - structured problem framing", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "9f3caf15-92f6-496f-be0e-595957f9a75e", "evidence": "The slide mentions a specific prediction and outcome, which aligns with the SCQA framework's Situation, Complication, Question, and Answer structure.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.796804+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 86, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 87, "type": "evidence", "title": "The chart shows a correlation between predicted and true binding affinity scores with a Spearman r of 0.88.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 44, "componentCount": 5, "textChars": 614, "nDataPoints": 1, "notes": "The chart shows a correlation between predicted and true binding affinity scores with a Spearman r of 0.88.", "elementsJson": [ "headline_text", "bullet_list", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.598+00", "seconds": 5.106, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/87", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-87", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "callout", "text": "Combining ML with carefully designed experiments has enabled LabGenius to increase the number of potential drug candidates by up to 100,000 fold.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.598+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "28a9d80d-5677-404b-ae00-2ddcce37b4f3", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.4, "x": 0.55, "y": 0.3 }, "kind": "chart", "text": "Test set, Spearman r = 0.88", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.598+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a5060307-83f6-4dbf-b3ba-9f0ad4d63860", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "list", "text": "Drug discovery is a complex multi-factor optimisation problem, where the majority of drug candidates fail.\nHere, 10^4 experimentally tested protein variants were used to accurately rank ~10^9 variants in silico.\nThis process gives LabGenius a more diverse panel of candidate proteins, increasing the probability of success of each drug discovery campaign.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.598+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e70e76c0-a15d-43c4-8c1b-21aaa820074d", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Spearman r: 0.88", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-eae72a2625ab", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Deep learning models interpret protein biology to find new therapeutics", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.598+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "16dbcba4-bd9b-4f34-9622-37b2ec1d898f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-6e41aba64d40", "evidence": "100,000-fold candidate increase.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 87, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "9c526ea9-d892-4aef-93cc-f72f82b994f3", "evidence": "The slide presents a case study with data and results, which follows a narrative arc to convey the story.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.857223+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 87, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 88, "type": "evidence", "title": "The slide uses a process flow diagram to illustrate the transition from manual input to automated analysis.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 49, "componentCount": 6, "textChars": 632, "nDataPoints": 0, "notes": "The slide uses a process flow diagram to illustrate the transition from manual input to automated analysis.", "elementsJson": [ "headline_text", "action_title", "bullet_list", "process_diagram", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.147+00", "seconds": 5.652, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/88", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-88", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Collapsing hours of human microscope time to minutes using supervised learning and computer vision.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-24950048bc62", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.7, "x": 0.15, "y": 0.4 }, "kind": "diagram", "text": "Process flow from input (microscope, annotations) to output (visualization, quantification)", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.147+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "73d9e784-2b8d-464f-8f16-d00b62c3051b", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.7, "x": 0.15, "y": 0.4 }, "kind": "image", "text": "Visual representation of the ONI hardware and software pipeline", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.147+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6d8d714f-90c3-4f21-8669-994a4d4e2ae1", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "Super-resolution microscopy usually requires subject-matter expertise to evaluate samples.\nONI's system automates these visual inspection tasks and unlocks super-resolution for non-expert users.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.147+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e663beee-cf8d-4862-938e-8db2be71f147", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.85, "x": 0.05, "y": 0.22 }, "kind": "title", "text": "Collapsing hours of human microscope time to minutes using supervised learning and computer vision.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.147+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "31c4ae8f-207d-4bd4-8d90-200054085393", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Deep learning revamps super-resolution microscopy imaging from acquisition to analysis", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.147+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8219620e-3371-4da3-917d-dbf222e5f99b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "6f1d8824-6edb-4a78-8dff-23891f83da8e", "evidence": "The slide shows a process flow from input to output, which implies a before-after transformation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:58.918496+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 88, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 89, "type": "evidence", "title": "The slide uses a hub-and-spoke diagram to show data inputs, line charts for time-series metabolic data, and a scatter plot for model validation.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 67, "componentCount": 7, "textChars": 841, "nDataPoints": 3, "notes": "The slide uses a hub-and-spoke diagram to show data inputs, line charts for time-series metabolic data, and a scatter plot for model validation.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "line_chart", "scatter_plot" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:41.886+00", "seconds": 3.287, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/89", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-89", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "ML predictions of glucose, triglyceride response two hours after meal consumption correlate 77% of the time with actual measured responses.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-030431b2f156", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.35, "x": 0.28, "y": 0.535 }, "kind": "chart", "text": "Triglycerides and Glucose response over time", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7f485346-25db-4aed-8f54-b677ce7af1f0", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.25, "x": 0.68, "y": 0.535 }, "kind": "chart", "text": "Predicted vs Measured GLU 2h-iAUC", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "473be386-73d8-4e09-ade8-7bfb40ee128b", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.25, "x": 0.02, "y": 0.535 }, "kind": "diagram", "text": "Factors influencing metabolic response: Genetics, Nutrition & Food, Gut Microbiome, Blood Biochemistry, Exercise, Sleep, Disease & Inflammation, Food timing, Sex, Anthropometry", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Hub and spoke", "slug": "hub-spoke", "status": "active", "canonId": "019de52d-04db-73e6-81ca-a234f2f23fc1", "version": 1, "description": null }, "framework": null }, "subkind": "hub-spoke", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cfdb8737-561d-4253-af99-b47eb8abc314", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "Identical twins have very different responses to the same foods. ML predictions of glucose, triglyceride response two hours after meal consumption correlate 77% of the time with actual measured responses. ZOE’s commercial AI-driven test kit launched in the US in August 2020.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "92690858-81c7-443d-b454-06d7fac76d26", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "correlation: 77%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-0413fab920ef", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Using genetic, metabolomic, metagenomic and meal-context information from 1,100 study participants to predict individuals’ metabolic response to food at scale", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f6c79519-45be-4b0b-8b02-19ac0528bd53", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f3249e98-e341-4f4a-847c-46303d3c6b43", "evidence": "Although not explicitly a 2x2 matrix, the diagram/hub-spoke shows multiple factors influencing metabolic response, which could be interpreted as a matrix-like structure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:58.978775+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-72b5231657f4", "evidence": "1,100 participants, 77% accuracy stats.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "5de1efea-c4ed-43e6-9b40-002dd4be742d", "evidence": "Diagram showing central human figure with various data inputs radiating outwards.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:41.886+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 89, "frameworkName": "hub-spoke" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 89, "from": 83, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-0f7fddb1083f", "evidence": "Phase 1 trial, EUR250M Exscientia deal, mega rounds, MELLODDY, LabGenius.", "position": 10, "objective": "Show pharma validating AI-first drug discovery", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 90, "type": "analysis", "title": "The slide illustrates the FDA's proposed TPLC regulatory framework for AI/ML-based software as a medical device (SaMD).", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 54, "componentCount": 6, "textChars": 840, "nDataPoints": 0, "notes": "The slide illustrates the FDA's proposed TPLC regulatory framework for AI/ML-based software as a medical device (SaMD).", "elementsJson": [ "headline_text", "paragraph", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.041+00", "seconds": 6.107, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/90", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-90", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The FDA published a new proposal to embrace the highly iterative and adaptive nature of AI systems in what they call a 'total product lifecycle (TPLC) regulatory approach built on good machine learning practices.'", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-05655daf011f", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.5, "x": 0.25, "y": 0.88 }, "kind": "chart", "text": "Spectrum of ML/AI-based algorithms", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "096cf5c9-5204-4700-b35a-7b9a23de071c", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.65, "x": 0.18, "y": 0.42 }, "kind": "diagram", "text": "FDA TPLC Regulatory Framework Diagram", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "03daa800-269a-4a77-b59e-bf0554742904", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.15, "x": 0.72, "y": 0.55 }, "kind": "legend", "text": "Legend: AI Model Development, AI Production Model, AI Device Modifications, Proposed TPLC Approach", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ca26e06f-722a-4488-9ac2-9a0e25b21cf6", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "The typically FDA approved AI-first software as medical device (SaMD) products are “locked”. The FDA published a new proposal to embrace the highly iterative and adaptive nature of AI systems in what they call a “total product lifecycle (TPLC) regulatory approach built on good machine learning practices.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b6586d63-3ee0-407e-a27e-660f358aeda9", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "In 2019, the FDA acknowledged that the traditional paradigm of medical device regulation was not designed for AI-first software which improves over time.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "801f8398-6857-4df4-a1d0-5933474cfc4d", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Hypothesis-Driven Structure", "slug": "hypothesis-driven-structure", "status": "active", "bestFor": null, "canonId": "019dd956-833f-7419-9932-be80362019a7", "version": 1, "bodyDocId": null, "description": "Start with a hypothesis and organize evidence to prove/disprove it", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "a4933b61-8c3a-423e-bd49-28a1d39e980c", "evidence": "The slide presents a regulatory framework diagram and a spectrum of ML/AI-based algorithms, which implies a structured and hypothesis-driven approach.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:59.039429+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-76b85eae62df", "evidence": "Callout explains TPLC regulatory shift.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [ { "matchId": "019dd95a-1ca5-70bb-bba0-836d67706a3c", "evidence": "Callout names FDA TPLC regulatory approach.", "framework": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "frameworkName": "FDA Total Product Lifecycle (TPLC)" }, { "matchId": "d864565c-121c-4ddb-89bd-03fc3c360b14", "evidence": "Diagram showing the total product lifecycle approach for AI/ML software", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:45.041+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 90, "frameworkName": "FDA TPLC Regulatory Framework" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 92, "from": 90, "loop": { "name": "36_maturity_curve", "slug": "36-maturity-curve", "status": "active", "bestFor": "Digital transformation, capability building, benchmarking", "canonId": "019dd956-748d-7329-ad3a-5c550b66e275", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current Maturity Level -> The Gap to Next Level -> Required Capabilities -> The Roadmap", "description": "Show where you are on a progression and what it takes to reach the next level", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-134b6ed954d8", "evidence": "FDA TPLC framework -> SPIRIT-AI -> first CMS reimbursement progression.", "position": 11, "objective": "Track AI medical regulation maturing toward reimbursement", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 91, "type": "analysis", "title": "The slide highlights a specific industry response to a identified problem (low quality in 20,000 studies).", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 5, "textChars": 1000, "nDataPoints": 2, "notes": "The slide highlights a specific industry response to a identified problem (low quality in 20,000 studies).", "elementsJson": [ "headline_text", "paragraph", "bullet_list" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.213+00", "seconds": 3.201, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/91", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-91", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A review of 20,000 recent studies in the field found that less than 1% of these studies had sufficiently high-quality design and reporting.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-1afafa329edf", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "Al-based medical imaging studies have a major problem. A review of 20,000 recent studies in the field found that less than 1% of these studies had sufficiently high-quality design and reporting.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0abf93f5-c58e-49cf-a94e-11d39c60641f", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.8, "x": 0.08, "y": 0.62 }, "kind": "list", "text": "State which version of the Al algorithm will be used; How was input data acquired and selected; How was poor quality or unavailable input data assessed and handled; Was there human-Al interaction; Describe onsite/offsite requirements; Access/re-use", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "970c340c-e3fb-444f-99da-597712097618", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.45 }, "kind": "list", "text": "CONSORT and SPIRIT are existing international frameworks that have now been extended to include Al-specific requirements. These have been simultaneously published in high-impact journals: Nature Medicine, The Lancet Digital Health and BMJ.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aed0ca9d-ae49-4085-8d35-5fd606082ee8", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "New international guidelines are drafted for clinical trial protocols (SPIRIT-AI) and reports (CONSORT-AI) that involve AI systems in a bid to improve both quality and transparency", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c701cacc-f00c-4470-a10b-941a38a6e6e6", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-7a8ce7a20b26", "evidence": "Specific 1% of 20,000 studies meet quality.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d01a7386-7b4e-462b-9f46-08251682d2dc", "evidence": "The slide presents information in a list/bullet format, which is a common presentation technique.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.101414+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [ { "matchId": "6b1d9cd2-6564-4c4f-837f-2527c06d4759", "evidence": "Mentioned as an international framework for clinical trial reporting", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "frameworkName": "CONSORT-AI" }, { "matchId": "da2dbdbb-fbab-4100-b43c-b85d2e9dcdb0", "evidence": "Mentioned as an international framework for clinical trial protocols", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:42.213+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 91, "frameworkName": "SPIRIT-AI" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 92, "from": 90, "loop": { "name": "36_maturity_curve", "slug": "36-maturity-curve", "status": "active", "bestFor": "Digital transformation, capability building, benchmarking", "canonId": "019dd956-748d-7329-ad3a-5c550b66e275", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current Maturity Level -> The Gap to Next Level -> Required Capabilities -> The Roadmap", "description": "Show where you are on a progression and what it takes to reach the next level", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-134b6ed954d8", "evidence": "FDA TPLC framework -> SPIRIT-AI -> first CMS reimbursement progression.", "position": 11, "objective": "Track AI medical regulation maturing toward reimbursement", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 92, "type": "analysis", "title": "The slide discusses the impact of CMS reimbursement on AI adoption in healthcare, specifically for Viz.ai.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 66, "componentCount": 6, "textChars": 1406, "nDataPoints": 5, "notes": "The slide discusses the impact of CMS reimbursement on AI adoption in healthcare, specifically for Viz.ai.", "elementsJson": [ "headline_text", "bullet_list", "screenshot" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:41.978+00", "seconds": 2.961, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/92", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-92", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Winning reimbursement from the CMS is a critical step towards any new system becoming implemented in clinical medicine because it creates the needed financial incentive to drive use.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-e916478979ef", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.35, "x": 0.6, "y": 0.45 }, "kind": "image", "text": "Viz.ai mobile app interface showing various diagnostic modules", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.978+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0275ca39-469d-4f9a-8bb3-8ad484a8c21a", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.3 }, "kind": "list", "text": "Viz.ai was granted a New Technology Add on Payment of up to $1,040 per use in patients with suspected strokes. The AI system scans computed tomography scans of the brain and pings the results to a specialist who can treat the patient before they suffer damage that leads to long-term disability. Several exclusion factor apply...", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.978+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2d5995fa-8a9b-4657-ab5e-bf9d5bd1084f", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.05, "y": 0.45 }, "kind": "list", "text": "Winning reimbursement from the CMS is a critical step towards any new system becoming implemented in clinical medicine because it creates the needed financial incentive to drive use. Viz.ai says their system detects ~90% of blockages in the brain and will exclude 90% of patients who do not have a blockage. This means that the neurointerventionalist can prioritise the right patient for urgent care. However, CMS will only reimburse for inpatients who are covered by Medicare and already a loss-maker for the hospital. Thus, fewer patients are actually eligible and the $1k/case could look more like $30-80/case (similar to mammography).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.978+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c6a13743-cf73-44fa-b00b-c87e3b2e98b0", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Reimbursement per use: $1,040", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-efac04e2b1ea", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "The first reimbursement approval for a deep learning-based medical imaging product has been granted by the Centers for Medicare and Medicaid Services (CMS) in the USA", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.978+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "db0bf641-d759-41bb-8654-6997aa788d91", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f4778304-04e2-4c51-b2ab-cd6628e2bb6f", "evidence": "The slide mentions a critical step towards implementation in clinical medicine, which implies a call to action.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.16585+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 92, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-7f44b8ec21a7", "evidence": "Callout explains why CMS approval is critical.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 92, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 92, "from": 90, "loop": { "name": "36_maturity_curve", "slug": "36-maturity-curve", "status": "active", "bestFor": "Digital transformation, capability building, benchmarking", "canonId": "019dd956-748d-7329-ad3a-5c550b66e275", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Current Maturity Level -> The Gap to Next Level -> Required Capabilities -> The Roadmap", "description": "Show where you are on a progression and what it takes to reach the next level", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-134b6ed954d8", "evidence": "FDA TPLC framework -> SPIRIT-AI -> first CMS reimbursement progression.", "position": 11, "objective": "Track AI medical regulation maturing toward reimbursement", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 93, "type": "setup", "title": "Source: NCSL, Morning Brew. Part of the 'state of ai 2020' report.", "function": "establish_context", "rawType": "geographic_map", "block": null, "metadata": { "slideType": "geographic_map", "slideTypeCanon": { "name": "Geographic map", "slug": "geographic_map", "status": "active", "canonId": "019de52d-09de-7153-8845-584c1a9d0c83", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 6, "textChars": 334, "nDataPoints": 0, "notes": "Source: NCSL, Morning Brew. Part of the 'state of ai 2020' report.", "elementsJson": [ "headline_text", "map", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:41.249+00", "seconds": 2.171, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/93", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-93", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Over half of all US states have enacted legislation to related to autonomous vehicles.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-7753e8184f31", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "image", "text": "US map showing legislative status", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Map", "slug": "map", "status": "active", "canonId": "019de52c-fe4d-77bd-b0a4-d72d58b9e5a4", "version": 1, "description": null }, "framework": null }, "subkind": "map", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.249+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a3b070b0-81d2-474d-8ace-9e37ffaccc5a", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.15, "x": 0.68, "y": 0.7 }, "kind": "legend", "text": "Both, Neither, Executive Order, Enacted Legislation", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.249+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "005ab39b-4473-44b7-ad3e-489c6aa2f80e", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Legislative status", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-78603d8280fe", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.7, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Over half of all US states have enacted legislation to related to autonomous vehicles.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.249+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "884b2cad-19f7-4773-9f42-07f9fe163cea", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "US states continue to legislate autonomous vehicles policies", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:41.249+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a816e72-8de2-466a-9a26-7b628d06dd93", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-81ca965fe8c4", "evidence": "US map visualises legislative spread.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 93, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 94, "type": "analysis", "title": "The slide uses a timeline-like structure to present three specific regulatory milestones for Waymo, Nuro, and AutoX.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 86, "componentCount": 10, "textChars": 1114, "nDataPoints": 3, "notes": "The slide uses a timeline-like structure to present three specific regulatory milestones for Waymo, Nuro, and AutoX.", "elementsJson": [ "headline_text", "paragraph", "logo_grid", "screenshot" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.949+00", "seconds": 3.72, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/94", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-94", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "To qualify for a driverless testing permit, companies must show proof of insurance or a bond equal to $5 million, prove that their cars can operate without a driver, meet the federal Motor Vehicle Safety Standards or otherwise have an exemption from the National Highway Traffic Safety Administration.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-1d25eaade828", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.14, "y": 0.5 }, "kind": "image", "text": "Waymo logo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3016ff0b-4eee-4b53-894b-b7157fb2a660", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.45, "y": 0.5 }, "kind": "image", "text": "Nuro logo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "545a45b2-63dc-48b5-8e09-be92811093cb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.76, "y": 0.5 }, "kind": "image", "text": "AutoX logo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f01b30cd-cb01-43e0-9251-c320b0cf4724", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.05, "y": 0.45 }, "kind": "list", "text": "30 October 2018: DMV ISSUES PERMIT AUTHORIZING WAYMO TO TEST DRIVERLESS VEHICLES IN SANTA CLARA COUNTY", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Timeline item", "slug": "timeline-item", "status": "active", "canonId": "019de52c-fd0e-74ce-9443-440fbd8c23b4", "version": 1, "description": null }, "framework": null }, "subkind": "timeline-item", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bcb6b4bc-fa72-45ab-a3c9-427b15fb8ede", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.67, "y": 0.45 }, "kind": "list", "text": "17 July 2020: DMV AUTHORIZES AUTOX TO TEST DRIVERLESS VEHICLE IN PORTION OF SAN JOSE", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Timeline item", "slug": "timeline-item", "status": "active", "canonId": "019de52c-fd0e-74ce-9443-440fbd8c23b4", "version": 1, "description": null }, "framework": null }, "subkind": "timeline-item", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d2d97505-8d0e-4f84-ae19-d685e200a4ca", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.28, "x": 0.36, "y": 0.45 }, "kind": "list", "text": "7 April 2020: DMV AUTHORIZES NURO TO TEST DRIVERLESS DELIVERY VEHICLES IN PORTION OF BAY AREA", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Timeline item", "slug": "timeline-item", "status": "active", "canonId": "019de52c-fd0e-74ce-9443-440fbd8c23b4", "version": 1, "description": null }, "framework": null }, "subkind": "timeline-item", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e27f64a2-2629-4c18-91d5-c7481242bf05", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "companies with driverless permits: 3", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-20d599d3b1b4", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.29 }, "kind": "paragraph", "text": "To qualify for a driverless testing permit, companies must show proof of insurance or a bond equal to $5 million, prove that their cars can operate without a driver, meet the federal Motor Vehicle Safety Standards or otherwise have an exemption from the National Highway Traffic Safety Administration.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2f867347-1921-412a-8160-016236630a2b", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Even so, driverless cars are still not so driverless: Only 3 of 66 companies with AV testing permits in California are allowed to test without safety drivers since 2018", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.949+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5cbc1525-f03d-4861-96ef-9399f8f05fa3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-84325ef6aaf9", "evidence": "Title: 'driverless cars are still not so driverless'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 94, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-883973d61621", "evidence": "Only 3 of 66 companies allowed.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 94, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 95, "type": "analysis", "title": "The chart uses a bar for human mileage and text-based values for AV mileage to emphasize the massive disparity.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 52, "componentCount": 6, "textChars": 569, "nDataPoints": 3, "notes": "The chart uses a bar for human mileage and text-based values for AV mileage to emphasize the massive disparity.", "elementsJson": [ "headline_text", "action_title", "paragraph", "bar_chart_horizontal" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.324+00", "seconds": 3.029, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/95", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-95", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Self-driving car companies racked up 42% more AV miles in 2019 than 2018. However, this only equates to 0.000737% of the miles driven by licensed California drivers 2019.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-b00409e476c8", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.7, "x": 0.15, "y": 0.5 }, "kind": "chart", "text": "Human mileage by Californians: 390,313,739,000; AV mileage in California: 2,020,855 (2018) and 2,874,950 (2019)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Horizontal bar chart", "slug": "bar-horizontal", "status": "active", "canonId": "019de52c-fec4-7221-986b-1cb5f5e84b1f", "version": 1, "description": null }, "framework": null }, "subkind": "bar-horizontal", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.324+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "25242675-501f-4205-bc34-07501f309efa", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.15, "x": 0.42, "y": 0.38 }, "kind": "legend", "text": "2018 2019", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.324+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "527abcee-b5da-46eb-ad29-a5b8fa34102a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Total miles driven: 0.000737%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-b4ea5d0ce5d5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "Self-driving car companies racked up 42% more AV miles in 2019 than 2018. However, this only equates to 0.000737% of the miles driven by licensed California drivers 2019.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.324+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ac90366f-9cf6-437e-9b0a-15ecd3d89f6e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Self-driving mileage in California remains microscopic compared to human driving", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.324+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7a9d4433-d783-4994-bf3f-4d4900f1a4e9", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "5a8f1c27-ddfe-4050-bd1f-c095deacb1e3", "evidence": "The slide presents a bar chart and a metric, which shows a comparison and could be related to a chart selection guide.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.288769+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-8ece501aad52", "evidence": "0.000737% of human miles vivid contrast.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-9228dd9ff452", "evidence": "'Microscopic' framing emphasises gap.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 95, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 96, "type": "analysis", "title": "The slide uses a line chart to illustrate the lack of reliability in 'miles per disengagement' as a metric for AV progress, specifically calling out Baidu's dat", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 49, "componentCount": 5, "textChars": 699, "nDataPoints": 8, "notes": "The slide uses a line chart to illustrate the lack of reliability in 'miles per disengagement' as a metric for AV progress, specifically calling out Baidu's data.", "elementsJson": [ "headline_text", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.168+00", "seconds": 2.805, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/96", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-96", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A year-on-year improvement of 8,679% sounds too good to be true...", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-5f5981b60409", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.9, "x": 0.05, "y": 0.45 }, "kind": "chart", "text": "Line chart showing miles/disengagement trends for Baidu, Waymo, GM Cruise, Zoox, Drive.ai, Nissan, Mercedes Benz, and NVIDIA.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.168+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bc9855a0-c6d3-45cd-979a-12f9080a32e3", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Miles/disengagement: 18,050", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-612dc63b33f8", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "Despite performing in the bottom 5th percentile of miles/disengagement in 2018, Baidu claims to have driven 18,050 miles/dis., which puts it at the top of the leaderboard ahead of Waymo with 13,219 miles/dis. (vs. 11,154 miles/dis. in 2018). A year-on-year improvement of 8,679% sounds too good to be true...", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.168+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c3996682-72b7-4c18-b117-2bb663d1fd57", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.12 }, "kind": "title", "text": "Sketchy metrics: Tracking AV progress is complicated by the industry's focus on miles per disengagement, which is hard to benchmark and is not reported across all US States.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.168+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10b897c2-7228-4500-9531-d657773e5c84", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-9586b599740e", "evidence": "Title: 'Sketchy metrics: Tracking AV progress'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Anchoring Bias", "slug": "anchoring-bias", "status": "active", "bestFor": null, "canonId": "019dd956-9254-73de-8f1b-c7250f8f4c9c", "version": 1, "bodyDocId": null, "description": "First piece of information becomes reference point for all subsequent judgments", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "c9838607-9b31-47c4-baaa-44a4452879c4", "evidence": "The slide mentions a potentially misleading metric (miles/disengagement), which could be related to anchoring bias.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:59.350687+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Lie Factor", "slug": "lie-factor", "status": "active", "bestFor": null, "canonId": "019dd956-b501-7740-8c7c-af7ecca16e9c", "version": 1, "bodyDocId": null, "description": "Ensure visual representation accurately reflects data magnitude", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-9aed41eaa60f", "evidence": "Calls out misleading 8,679% YoY metric.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 96, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 97, "type": "analysis", "title": "Includes a quote/excerpt from the official Amazon acquisition announcement.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 72, "componentCount": 7, "textChars": 1128, "nDataPoints": 5, "notes": "Includes a quote/excerpt from the official Amazon acquisition announcement.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "quote_block" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.512+00", "seconds": 3.028, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/97", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-97", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Consolidation of industry players begins as Zoox, the company reinventing an AV-first car, was acquired by Amazon for a reported $1.3B in cash.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-929c7261a2b8", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Consolidation of industry players begins as Zoox, the company reinventing an AV-first car, was acquired by Amazon for a reported $1.3B in cash. The company raised at least $955M since 2015 with its last reported post-money valuation of $3.2B.\nDeal documents suggest that Zoox was burning over $30M per month in early 2020. They have 900+ FTE.\nCruise bid $1.05B after the original Amazon offer, which resulted in the final $1.3B price.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.512+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b660668b-9e7e-468c-98b7-99ef6536c762", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Acquisition price: $1.3B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-98015d610b04", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.7, "x": 0.16, "y": 0.68 }, "kind": "paragraph", "text": "Amazon has signed an agreement to acquire Zoox, a California-based company working to design autonomous ride-hailing vehicles from the ground up. Aicha Evans, Zoox CEO, and Jesse Levinson, Zoox co-founder and CTO, will continue to lead the team as they innovate and drive towards their mission.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.512+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c0493803-7e98-4505-bc1d-d3f1439185cd", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.7, "x": 0.16, "y": 0.575 }, "kind": "quote", "text": "We’re acquiring Zoox to help bring their vision of autonomous ride-hailing to reality", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": null, "subkind": { "name": "Pull quote", "slug": "pull-quote", "status": "active", "canonId": "019de52c-fbf9-76fe-97e1-a0b7540445c9", "version": 1, "description": null }, "framework": null }, "subkind": "pull-quote", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.512+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "41d5a000-e252-43f3-b5dc-3c0ac1840b87", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "We’re acquiring Zoox to help bring their vision of autonomous ride-hailing to reality — Amazon", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-94c71c0191b1", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Self-driving: When even a billion dollars isn't enough", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.512+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a584f7c-38ae-4998-8d49-bb7b17e1731a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "986de603-b020-4d6c-a4f6-82e96616b8e4", "evidence": "The components mention 'Amazon', 'Zoox', and 'autonomous ride-hailing', indicating an audience calibration.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.685233+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "4988c880-37d2-4f90-b2e1-b19cb9026b15", "evidence": "The quote from Amazon 'We’re acquiring Zoox to help bring their vision of autonomous ride-hailing to reality' can be seen as a call to action.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.719072+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-9cbcc3ff8a10", "evidence": "$1.3B Zoox/Amazon deal cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 97, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 98, "type": "analysis", "title": "The slide uses a bubble-based visualization to represent funding scale, with lines connecting company logos to their respective bubbles.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 6, "textChars": 306, "nDataPoints": 7, "notes": "The slide uses a bubble-based visualization to represent funding scale, with lines connecting company logos to their respective bubbles.", "elementsJson": [ "logo_grid", "bubble_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.004+00", "seconds": 2.354, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/98", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-98", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The main self-driving contenders raised almost $7B in private rounds since July 2019.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-b175eb434a67", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "chart", "text": "Bubble chart showing funding rounds for Argo AI, AutoX, Voyage, Waymo, Pony.ai, Five AI, and AImotive.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Bubble chart", "slug": "bubble", "status": "active", "canonId": "019de52d-00a4-756f-8978-8d5e48905dbc", "version": 1, "description": null }, "framework": null }, "subkind": "bubble", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.004+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4463ac6b-499f-48c8-ab59-12469afb1b17", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Private funding raised: $7B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-b796437b84ba", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.05, "x": 0.03, "y": 0.58 }, "kind": "paragraph", "text": "2019", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.004+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1bfc5f4f-4e5c-4a5d-905d-f52dc3ef468e", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.05, "x": 0.92, "y": 0.58 }, "kind": "paragraph", "text": "2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.004+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d11608fa-d219-416a-8e3e-cea3e6c257c2", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "The main self-driving contenders raised almost $7B in private rounds since July 2019", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.004+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "29f93c00-5164-4d27-997f-ebb4f615b080", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "6baf4b87-908f-44d8-92b7-e45e731830a8", "evidence": "The bubble chart showing funding rounds for various companies instantiates a chart selection guide.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.787974+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 98, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-a088c3168f86", "evidence": "$7B raised since July 2019.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 98, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 99, "type": "evidence", "title": "Includes images of a test vehicle and a control center dashboard.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 61, "componentCount": 7, "textChars": 618, "nDataPoints": 3, "notes": "Includes images of a test vehicle and a control center dashboard.", "elementsJson": [ "headline_text", "paragraph", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.804+00", "seconds": 2.963, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/99", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-99", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A safety control center monitors vehicles and traffic conditions real-time.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-f5abaf1df186", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.4, "x": 0.05, "y": 0.4 }, "kind": "image", "text": "DiDi autonomous test vehicle", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.804+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "236043a0-d88b-4fc3-9f67-c4bba9df612f", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.45, "y": 0.4 }, "kind": "image", "text": "Control center dashboard", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.804+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "239fb35b-d330-40ec-a6f5-ace84d4cd16e", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "DiDi's self-driving unit raised >$500M from SoftBank Vision Fund, grew its team from 200 to >400 since last year and launched its ride hailing service to consumers in Shanghai from late July 2020. The service runs on public roads that are fit with additional sensors that feed into a control room manned with safety operators.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.804+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e17d9be4-e91f-4ba4-a78b-f1c66de6dd6a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Funding: $500M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-f97b54fee621", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.45, "y": 0.78 }, "kind": "paragraph", "text": "A safety control center monitors vehicles and traffic conditions real-time.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.804+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1b651ef4-4b5c-476b-895f-074e6c9281c2", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Another self-driving group, DiDi, spins off from its parent and raises $500M", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.804+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "399b1400-0478-4d89-a65e-4502222863d3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "f82a8526-1630-4905-a961-617e227e2096", "evidence": "The components mention 'DiDi's self-driving unit raised >$500M' and 'launched its ride hailing service to consumers', indicating a before-after bridge.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.855943+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 99, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 100, "type": "analysis", "title": "Part of the State of AI 2020 report.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 72, "componentCount": 9, "textChars": 598, "nDataPoints": 2, "notes": "Part of the State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.717+00", "seconds": 2.762, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/100", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-100", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Waymo, Aurora, and GM Cruise have acquired LiDAR companies or built sensors in-house to hit a 300m range and to each own a key technology component of their value chain.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-cf4127a0bdab", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.28, "x": 0.67, "y": 0.5 }, "kind": "image", "text": "GM Cruise Strobe sensor", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "64ba1330-7ca6-49f1-a5b2-c70f09f2e1f5", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.28, "x": 0.36, "y": 0.5 }, "kind": "image", "text": "Aurora sensor suite", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7c4c33b-0840-47cf-b768-344ea05e1eba", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.28, "x": 0.05, "y": 0.5 }, "kind": "image", "text": "Waymo Honeycomb LiDAR", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "eeacfee5-ef0d-47a5-9a8e-979f3c38b3a3", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.28, "x": 0.05, "y": 0.4 }, "kind": "paragraph", "text": "Waymo's in-house Honeycomb released in 2019", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "19d9d0e9-54d5-4274-b879-253744967e56", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.28, "x": 0.67, "y": 0.4 }, "kind": "paragraph", "text": "GM Cruise acquired Strobe in 2017", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3be68723-2172-476e-9fc3-e89086562516", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.28, "x": 0.36, "y": 0.4 }, "kind": "paragraph", "text": "Aurora acquired Blackmore in 2019", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "48f83643-5c9c-44d4-9e63-371451c09623", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Waymo, Aurora, and GM Cruise have acquired LiDAR companies or built sensors in-house to hit a 300m range and to each own a key technology component of their value chain.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6f241920-bf2b-467e-861d-917ca11ff7c4", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Capital is used to vertically-integrate and deepen technology moats, e.g. in-house LiDAR", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3c2da28c-cd1a-437e-ba7b-7fdd5038762c", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Capital Allocation Framework", "slug": "capital-allocation-framework", "status": "active", "bestFor": null, "canonId": "d2901a21-50b9-4151-ab1b-55b432772ea1", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "30ae4ce1-a669-4630-a467-b4e5dbc411f2", "evidence": "The components mention 'Waymo, Aurora, and GM Cruise have acquired LiDAR companies or built sensors in-house', indicating a capital allocation framework.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.924003+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-a57f61ad6c73", "evidence": "Callout explains LiDAR vertical integration moat.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [ { "matchId": "83cc3738-c80b-41c3-886a-85dca485e7ac", "evidence": "Mentions owning key technology components of their value chain.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:42.717+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "frameworkName": "value-chain" }, { "matchId": "019dd95a-1ca5-70bb-bba0-87de388e10d2", "evidence": "Callout: 'own a key technology component of their value chain'.", "framework": null, "confidence": 60, "extraction": { "at": "2026-04-29 13:07:34.826634+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 100, "frameworkName": "Value Chain" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 101, "type": "analysis", "title": "Includes a product portfolio diagram for Velodyne and a sample LiDAR point cloud visualization for Luminar.", "function": "compare_options", "rawType": "market_landscape", "block": null, "metadata": { "slideType": "market_landscape", "slideTypeCanon": { "name": "Market landscape", "slug": "market_landscape", "status": "active", "canonId": "019de52d-0b6d-75db-a946-9ac362e3afc2", "version": 1, "description": null }, "function": "compare_options", "functionCanon": { "name": "Compare options", "slug": "compare_options", "status": "active", "canonId": "019de52d-167f-7038-aebc-70a9d2e0e33b", "version": 1, "description": null }, "density": "dense", "densityScore": 55, "componentCount": 6, "textChars": 681, "nDataPoints": 3, "notes": "Includes a product portfolio diagram for Velodyne and a sample LiDAR point cloud visualization for Luminar.", "elementsJson": [ "headline_text", "paragraph", "process_diagram", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:42.434+00", "seconds": 2.259, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/101", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-101", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Velodyne will list shy of $2B valuation on $106M of net revenues in 2019 and Luminar at $3.4B.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-3dd667aad859", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.55, "x": 0.05, "y": 0.045 }, "kind": "image", "text": "Velodyne product portfolio and technology architecture diagram", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Infographic", "slug": "infographic", "status": "active", "canonId": "019de52c-fe26-750c-9d7d-9f34e21324e8", "version": 1, "description": null }, "framework": null }, "subkind": "infographic", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a278060d-4892-4fd1-b069-5b8db7f8baf4", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.35, "x": 0.62, "y": 0.045 }, "kind": "image", "text": "Luminar LiDAR point cloud visualization", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a3b49e39-6243-4d25-b7e1-b79f7ea69fac", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "Velodyne will list shy of $2B valuation on $106M of net revenues in 2019 and Luminar at $3.4B. Velodyne guides to upcoming software for autopilot and collision avoidance that makes use of a future Vela LiDAR. Luminar points to an agreement with Volvo that sees its integration into their vehicle platform in 2022.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e30f552e-700a-4843-9b71-2c4205e331fc", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "valuation: $3.4B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-42213292272c", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Meanwhile, LiDAR incumbent Velodyne and challenger Luminar both go public on the Nasdaq via reverse mergers (SPAC) to compete with hardware and ADAS software", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:42.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5f8dc61c-3228-4893-8bdb-43404670e414", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-a84c63083138", "evidence": "$2B / $3.4B SPAC valuations.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 101, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Market Segmentation Pyramid", "slug": "market-segmentation-pyramid", "status": "active", "bestFor": null, "canonId": "1c9f3543-c96f-4a5b-b4a4-9b2a412c0aaf", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9291303f-a055-4751-b177-b5f7eb30f997", "evidence": "The components mention 'Velodyne will list shy of $2B valuation' and 'Luminar at $3.4B', indicating a market segmentation pyramid.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.990912+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 101, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 102, "type": "analysis", "title": "The chart illustrates the gap between initial expectations of exponential growth and the reality of diminishing returns in AI capability over time.", "function": "diagnose", "rawType": "diagnosis", "block": null, "metadata": { "slideType": "diagnosis", "slideTypeCanon": { "name": "Diagnosis", "slug": "diagnosis", "status": "active", "canonId": "019de52d-0cd7-7212-898f-0464041ff8d8", "version": 1, "description": null }, "function": "diagnose", "functionCanon": { "name": "Diagnose", "slug": "diagnose", "status": "active", "canonId": "019de52d-1657-7223-acc5-7437d878c850", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 1094, "nDataPoints": 0, "notes": "The chart illustrates the gap between initial expectations of exponential growth and the reality of diminishing returns in AI capability over time.", "elementsJson": [ "headline_text", "paragraph", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.082+00", "seconds": 2.842, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/102", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-102", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Rather than seeing exponential improvements in the quality of AI performance (a la Moore's Law), we're instead seeing exponential increases in the cost to improve AI systems — supervised ML seems to follow an S-Curve.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-41b787ce7549", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.43, "x": 0.52, "y": 0.38 }, "kind": "chart", "text": "Improvements in Supervised ML systems over time", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e18e2996-cf1b-403d-8084-8facd5a15623", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI capability", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-474706b326a0", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.45, "x": 0.05, "y": 0.45 }, "kind": "paragraph", "text": "It's widely understood that the hardest part of building AI is how it deals with situations that happen uncommonly, i.e. edge cases. In fact, the better your model, the harder it is to find robust data sets of novel edge cases. Additionally, the better your model, the more accurate the data you need to improve it. Rather than seeing exponential improvements in the quality of AI performance (a la Moore's Law), we're instead seeing exponential increases in the cost to improve AI systems — supervised ML seems to follow an S-Curve.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7b70b41-deb1-4617-8bad-6cfa78956e45", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Starsky Robotics, which was the first company to run an autonomous unmanned truck drive on a public highway, closed its doors in Q1 2020. It openly cited the challenges of scaling supervised learning.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fd6b606f-30f4-414e-9ce0-2a7fd06e5d90", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Supervised learning and the cost of edge cases: New technology approaches are needed", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "59f3a9ec-83ba-4933-b317-046c809a892e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-ad52e19434ae", "evidence": "Title: 'Supervised learning and the cost of edge cases'.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-b0a5af0a4d55", "evidence": "Anti-Moore's-Law analogy in callout.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Problem Statement Canvas", "slug": "problem-statement-canvas", "status": "active", "bestFor": null, "canonId": "019dd956-839d-744b-ad11-9bf3e3642a18", "version": 1, "bodyDocId": null, "description": "Structured definition of the problem: What? Who? Where? When? Scale?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "0505b6ee-d387-4123-807d-36cc3c04ff5c", "evidence": "The components mention 'the hardest part of building AI is how it deals with situations that happen uncommonly', indicating a problem statement canvas.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:50.059038+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "092fcd99-dd06-4461-91a1-86777b38ea5a", "evidence": "The text explicitly mentions 'supervised ML seems to follow an S-Curve' and the chart depicts this trajectory.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:43.082+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 102, "frameworkName": "S-Curve" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 103, "type": "analysis", "title": "The slide uses a dual-column timeline format to display chronological releases of datasets.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 31, "componentCount": 4, "textChars": 350, "nDataPoints": 13, "notes": "The slide uses a dual-column timeline format to display chronological releases of datasets.", "elementsJson": [ "headline_text", "timeline_visual", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.092+00", "seconds": 2.82, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/103", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-103", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "11 major datasets and 2 updates since 2019, many of which include cameras, video, LiDAR and motion traces.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-79eaab88a94c", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.95, "x": 0.03, "y": 0.3 }, "kind": "chart", "text": "Timeline of dataset releases from April 2019 to July 2020", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": null, "framework": null }, "subkind": "timeline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.092+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a35c42c0-9f9b-4468-ba34-36d497b97376", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.8, "x": 0.03, "y": 0.21 }, "kind": "paragraph", "text": "11 major datasets and 2 updates since 2019, many of which include cameras, video, LiDAR and motion traces.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.092+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6f17809a-5399-475e-a216-86215f2ac5b7", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.8, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Leading companies crowdsource ideas from open source using data they've generated", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.092+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ddb7cdc8-f02e-43ed-995a-d3a05f2723ec", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "474a31e4-dda4-4137-a3ca-a37503faa2b1", "evidence": "The components mention '11 major datasets and 2 updates since 2019', indicating a data story arc.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:50.126475+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 103, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "2106a44d-0ccc-429c-a92f-b5da285918bd", "evidence": "Chronological list of dataset releases", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:43.092+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 103, "frameworkName": "timeline" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 104, "type": "analysis", "title": "The slide uses a process flow diagram combined with gauge charts to show the distribution of ML usage across the autonomous driving stack.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 56, "componentCount": 6, "textChars": 702, "nDataPoints": 3, "notes": "The slide uses a process flow diagram combined with gauge charts to show the distribution of ML usage across the autonomous driving stack.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.141+00", "seconds": 2.5, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/104", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-104", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Use of ML in self-driving is still mostly limited to perception with large parts of stack hand-engineered.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-93ea0b1b3b27", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.5, "x": 0.25, "y": 0.75 }, "kind": "chart", "text": "Amount of ML: 100%, 50%, 0%", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Gauge", "slug": "gauge", "status": "active", "canonId": "019de52d-0286-709d-a957-ccb6eb2af8b7", "version": 1, "description": null }, "framework": null }, "subkind": "gauge", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8be54788-fb35-4e5f-aa60-b787ae3acb22", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.7, "x": 0.15, "y": 0.42 }, "kind": "diagram", "text": "Sensor input and maps -> Perception -> Detected traffic agents -> Motion prediction -> Predicted agent motion -> Trajectory planning -> Path taken by autonomous vehicle", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a704c919-3f86-4bc9-8e23-4a05715ac24f", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Use of ML in self-driving is still mostly limited to perception with large parts of stack hand-engineered.\nMuch of today’s ML in self-driving systems focus only on understanding what is around the vehicle.\nWhat the self-driving car should do is mostly hand-engineered making development difficult and slow.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a79788b4-19d0-4361-9a44-b80d18b6dc63", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Amount of ML", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa68-9535e3e36af0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "The Next Step: New models and a shift in focus from perception to motion prediction", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bf0f519c-e616-4fd1-82e0-c49f765734bd", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "d32ea224-1a08-428e-a04b-cbff2d67405c", "evidence": "The slide depicts the autonomous driving stack as a sequential process flow.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 17:45:43.141+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 104, "frameworkName": "process" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 105, "type": "analysis", "title": "Includes logos of Lyft, Waymo, and Uber at the bottom.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 924, "nDataPoints": 1, "notes": "Includes logos of Lyft, Waymo, and Uber at the bottom.", "elementsJson": [ "headline_text", "bullet_list", "screenshot", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.364+00", "seconds": 2.627, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/105", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-105", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Companies that can leverage the scale of their human driver fleets can build an advantage in the data race that will power new model innovation.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa69-cd761b8ec911", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.2, "x": 0.02, "y": 0.9 }, "kind": "image", "text": "Lyft, Waymo, Uber logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.364+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b0fea531-af5b-40ad-8c1f-2b21351a91c2", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.2, "x": 0.76, "y": 0.24 }, "kind": "image", "text": "Visualization of self-driving planning data", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.364+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d2ae3d71-8fdb-4f88-95ca-5e70c3fbfa7a", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.7, "x": 0.02, "y": 0.52 }, "kind": "list", "text": "New datasets can change the power balance of existing leading players. Few players can collect enough data to fully train these new kinds of systems. Companies that can leverage the scale of their human driver fleets can build an advantage in the data race that will power new model innovation.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.364+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3c8788f7-c0b5-4ccf-9005-29f48135bc69", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.7, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "New algorithms working akin AlphaGo and trained on large amount of human driving demonstrations are being developed. Recently, both Waymo, Uber and Lyft demonstrated new techniques of imitation learning and inverse reinforcement learning instead of hand-written rules. Lyft released a new dataset counting 1,000 hours to develop these systems.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.364+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4da67f86-8d85-4ef9-8e67-a240b5531d5a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "The new frontier of self-driving development is machine learning for planning", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.364+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1dea0124-2028-4657-a21e-e3d03eca1ca8", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 106, "type": "analysis", "title": "Part of the 'state of ai 2020' report series.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 79, "componentCount": 8, "textChars": 1189, "nDataPoints": 4, "notes": "Part of the 'state of ai 2020' report series.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "screenshot" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.091+00", "seconds": 4.137, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/106", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-106", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Costing $8,000 today, Tesla is a rare breed in monetising their Autopilot system to the tune of $100M’s so far.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-5e1ad997a861", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.45, "x": 0.52, "y": 0.35 }, "kind": "image", "text": "Tesla Autopilot interface visualization", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2d445cb3-2bb8-4ff4-881e-b6cab5757035", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.45, "x": 0.05, "y": 0.4 }, "kind": "list", "text": "Tesla currently recognises significant deferred revenue for its Autopilot feature, which it can realize as more features are shipped. In 2020, Tesla shipped control for Stops and Traffic Lights, and Speed Limits. The next milestone as per Elon is a \"feature complete\" build which will additionally take the turns when it comes to stops.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3232be80-905e-4389-b8ad-f383ec4ab79c", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.45, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "Costing $8,000 today, Tesla is a rare breed in monetising their Autopilot system to the tune of $100M’s so far. Given the importance of edge-case reliability, their “driver-in-the-loop” engineering approach could pay off.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4155f0fd-c41d-4c28-8081-1c3ac60b5842", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.45, "x": 0.05, "y": 0.65 }, "kind": "list", "text": "Comma.ai is a startup that sells “Tesla Autopilot-like functionality for your Toyota or Honda”. Their open-source openpilot system has driven 25 million miles so far.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5fdf30d7-8bac-499f-9242-d5c6830f5a4f", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.45, "x": 0.05, "y": 0.8 }, "kind": "list", "text": "MobileEye (owned by Intel) counts 50 partnerships with 30 OEMs. It sells vision-based advanced driver assistance technology.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "afab8dd0-f1e6-4d28-8da0-a85d00cb6ec5", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "deferred revenue: $100M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:29+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-63e6a209b9df", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "The consumer-first approach to self-driving: Tesla has hundreds of thousands of Autopilot-enabled cars in the wild and consumers help inch it towards “Full Self-Driving”", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.091+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2f3bb22c-9921-4ce4-9a4a-9e418adea41c", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Monroe's Motivated Sequence", "slug": "monroes-motivated-sequence", "status": "active", "bestFor": null, "canonId": "019dd956-8758-700f-83cf-d260a54ad24f", "version": 1, "bodyDocId": null, "description": "Five-step persuasion: Attention, Need, Satisfaction, Visualization, Action", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "7d9cef01-2718-4193-97e3-2347bed618f7", "evidence": "The components mention 'Tesla currently recognises significant deferred revenue for its Autopilot feature', indicating a motivated sequence.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:50.264013+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 106, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 106, "from": 93, "loop": { "name": "10_iceberg", "slug": "10-iceberg", "status": "active", "bestFor": "Consulting, complex problem solving, organizational change", "canonId": "019dd956-69f4-738c-852a-643798da0899", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Symptom (Visible) -> The System (Hidden) -> The Root Cause", "description": "Reveal that the visible problem is merely a symptom of a deeper root cause", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-154d2d6f882c", "evidence": "Surface (legislation/permits) -> sketchy metrics -> capital consolidation -> root tech limits of supervised perception.", "position": 12, "objective": "Diagnose why self-driving progress is slower than headlines suggest", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 107, "type": "analysis", "title": "The slide highlights the rapid evolution of specialized AI hardware, specifically focusing on Graphcore's technological advancement.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 65, "componentCount": 7, "textChars": 712, "nDataPoints": 6, "notes": "The slide highlights the rapid evolution of specialized AI hardware, specifically focusing on Graphcore's technological advancement.", "elementsJson": [ "headline_text", "paragraph", "photo", "bar_chart_grouped" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.597+00", "seconds": 2.562, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/107", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-107", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Graphcore released their Mk2 IPU processor, which packs 59.4 billion transistors on a 823 sqmm die using a 7 nm process. This is the most complex processor ever made.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-7ce1fd734511", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.55, "x": 0.38, "y": 0.4 }, "kind": "chart", "text": "Colossus Mk1 vs Colossus Mk2 performance improvement across BERT-Large, BERT-3Layer, and EfficientNet-B3", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.597+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0479ee8c-ee40-46ca-8ad4-dc268eb5f143", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.3, "x": 0.05, "y": 0.4 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.597+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f789d203-7711-454f-b752-4cbf545c1759", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "Graphcore released their Mk2 IPU processor, which packs 59.4 billion transistors on a 823 sqmm die using a 7 nm process. This is the most complex processor ever made.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.597+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4d3aba34-21b9-4bd0-8d12-80d6512a6dd0", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Performance improvement: 9.3x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-83aea0c5125c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.15, "y": 0.9 }, "kind": "source-note", "text": "Figure note: Comparing 8x C2 PCIe IPU-Processors with IPU-LINK vs. 8x M2000 IPU-Machine with IPU-FABRIC", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.597+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5b56427b-3236-4d37-b85b-1f7ea7e81a08", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "AI problems like self-driving thrive on compute: New providers of specialized AI compute platforms are already onto their generation 2+ products", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.597+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a39730bd-d1d6-4498-a18c-c98a445a67f0", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-b7c36953ed3a", "evidence": "Title: AI compute providers on gen 2+ products.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 107, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-bbfb2099f055", "evidence": "59.4B transistors / 823 sqmm specifics.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 107, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 110, "from": 107, "loop": { "name": "16_david_goliath", "slug": "16-david-goliath", "status": "active", "bestFor": "Startups, disruptive innovation, competitive displacement", "canonId": "019dd956-6c5d-704a-9c33-1aa083ca432e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Giant's Weakness -> Our Slingshot (Unique Edge) -> The Topple", "description": "Frame your initiative as the nimble underdog taking on a slow, bloated incumbent", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-19b5d7018f7a", "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 108, "type": "setup", "title": "The slide uses a visual comparison of hardware units to illustrate cost efficiency.", "function": "compare_options", "rawType": "comparison_table", "block": null, "metadata": { "slideType": "comparison_table", "slideTypeCanon": { "name": "Comparison table", "slug": "comparison_table", "status": "active", "canonId": "019de52d-0a06-73f8-8813-fe022fab1de2", "version": 1, "description": null }, "function": "compare_options", "functionCanon": { "name": "Compare options", "slug": "compare_options", "status": "active", "canonId": "019de52d-167f-7038-aebc-70a9d2e0e33b", "version": 1, "description": null }, "density": "dense", "densityScore": 68, "componentCount": 8, "textChars": 586, "nDataPoints": 4, "notes": "The slide uses a visual comparison of hardware units to illustrate cost efficiency.", "elementsJson": [ "headline_text", "paragraph", "icon_grid", "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.134+00", "seconds": 2.993, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/108", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-108", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The reported 16x faster training time for the image classification model EfficientNet-B4 on the M2000 vs. NVIDIA DGX-A100 translates to an 12x cost advantage.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-12801ff7bffc", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.2, "x": 0.59, "y": 0.55 }, "kind": "image", "text": "NVIDIA hardware icon grid", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Icon grid", "slug": "icon-grid", "status": "active", "canonId": "019de52c-fdfe-749d-9990-1db465a3eed4", "version": 1, "description": null }, "framework": null }, "subkind": "icon-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9ae084e4-1335-45f7-8e1b-33b20e311ecf", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.08, "x": 0.25, "y": 0.6 }, "kind": "image", "text": "Graphcore hardware icon", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "31ae139f-a881-4756-aebf-e1481a68fe74", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.12, "x": 0.23, "y": 0.8 }, "kind": "metric", "text": "$259,600", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": null, "subkind": { "name": "Big number", "slug": "big-number", "status": "active", "canonId": "019de52c-fc70-70c8-b965-0584548a9204", "version": 1, "description": null }, "framework": null }, "subkind": "big-number", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4421de7d-9e42-466a-bf6c-b1d281993d8f", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.12, "x": 0.63, "y": 0.8 }, "kind": "metric", "text": "$3,000,000", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": null, "subkind": { "name": "Big number", "slug": "big-number", "status": "active", "canonId": "019de52c-fc70-70c8-b965-0584548a9204", "version": 1, "description": null }, "framework": null }, "subkind": "big-number", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8a14f017-fe2d-415b-9e04-c700cbcf256b", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: $259,600", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-140606036b4f", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "The reported 16x faster training time for the image classification model EfficientNet-B4 on the M2000 vs. NVIDIA DGX-A100 translates to an 12x cost advantage. This does not factor in the cost of migrating to a new development platform and mastering its tooling.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5aaa22d5-f032-4fb3-a62b-093a7c0498e5", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Graphcore M2000 offers faster training time to drop the cost of state-of-the-art models", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.134+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "abe72935-9b36-4b27-8940-be58cc8159c8", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-bcde5de70619", "evidence": "Title: M2000 faster training drops cost.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Comparison frame", "slug": "comparison-frame", "status": "active", "bestFor": null, "canonId": "019dd9e1-484b-7399-a4fc-d14644b99f93", "version": 1, "bodyDocId": null, "description": "Two or more states/subjects placed side by side to expose a gap.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "4faec66b-c36d-460b-8053-f1af26f86395", "evidence": "The components mention 'The reported 16x faster training time for the image classification model EfficientNet-B4 on the M2000 vs. NVIDIA DGX-A100', indicating a comparison frame.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:50.364827+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Contrast Principle", "slug": "contrast-principle", "status": "active", "bestFor": null, "canonId": "019dd956-bf0f-73ae-b9df-e60508cdaa65", "version": 1, "bodyDocId": null, "description": "Higher visual weight draws attention - use contrast strategically for emphasis", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-c7f61c64a03e", "evidence": "$259,600 vs $3,000,000 stark visual cost contrast.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-ca57d778cd60", "evidence": "Visual equation '1 chip = 16 boxes' makes ratio concrete.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Hierarchy", "slug": "visual-hierarchy", "status": "active", "bestFor": null, "canonId": "019dd956-b2ca-706b-937f-e52441086fd7", "version": 1, "bodyDocId": null, "description": "Order of perception through Size > Color > Position > Weight > Space", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-c2f6502c9012", "evidence": "Single chip vs grid of 16 boxes plus '=' sign and price tags.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 108, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 110, "from": 107, "loop": { "name": "16_david_goliath", "slug": "16-david-goliath", "status": "active", "bestFor": "Startups, disruptive innovation, competitive displacement", "canonId": "019dd956-6c5d-704a-9c33-1aa083ca432e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Giant's Weakness -> Our Slingshot (Unique Edge) -> The Topple", "description": "Frame your initiative as the nimble underdog taking on a slow, bloated incumbent", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-19b5d7018f7a", "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 109, "type": "analysis", "title": "The chart shows performance gains for Mask R-CNN, SSD, ResNet-50, Transformer, and GNMT models.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 55, "componentCount": 6, "textChars": 640, "nDataPoints": 5, "notes": "The chart shows performance gains for Mask R-CNN, SSD, ResNet-50, Transformer, and GNMT models.", "elementsJson": [ "headline_text", "paragraph", "bar_chart_grouped", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.934+00", "seconds": 2.692, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/109", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-109", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The TPU v4 packs 2x the matrix multiplication TFLOPs of the TPU v3, greater memory bandwidth and improved interconnect technology.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-baf1b9b244eb", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.5, "x": 0.25, "y": 0.4 }, "kind": "chart", "text": "Comparison of TPU v3 and TPU v4 training performance across models", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.934+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "42df702f-bf03-4164-a4c9-d5260d617f99", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Training speedup: 3.7x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-bc54ec506354", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "The TPU v4 packs 2x the matrix multiplication TFLOPs of the TPU v3, greater memory bandwidth and improved interconnect technology. The supercomputer used for MLPerf v0.7 submissions is 4x the size of the TPU v3 Pod used in MLPerf v0.6.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.934+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fb1926ee-697a-4aaa-b6ac-e76feb6e730f", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.6, "x": 0.12, "y": 0.92 }, "kind": "source-note", "text": "Figure note: All comparisons at 64-chip scale. Gains are due to hardware innovations and software improvements.", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.934+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7684554c-3a65-4ad6-aa1d-a71f3c8baaee", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.15 }, "kind": "title", "text": "Google’s new TPU v4 delivers up to a 3.7x training speedup over their TPU v3", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.934+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3cd39fe4-8595-4528-8c02-f22c4228caa7", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-cc76bc902134", "evidence": "Specific 3.7x speedup vs prior gen.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 109, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Waterfall chart", "slug": "waterfall-chart", "status": "active", "bestFor": null, "canonId": "e2799da3-725f-44a9-a093-570f1f388e64", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "3caccf96-90f7-4063-9b87-facb10c50665", "evidence": "The slide mentions a comparison of TPU v3 and TPU v4 training performance across models, which is often visualized using a waterfall chart.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.20688+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 109, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 110, "from": 107, "loop": { "name": "16_david_goliath", "slug": "16-david-goliath", "status": "active", "bestFor": "Startups, disruptive innovation, competitive displacement", "canonId": "019dd956-6c5d-704a-9c33-1aa083ca432e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Giant's Weakness -> Our Slingshot (Unique Edge) -> The Topple", "description": "Frame your initiative as the nimble underdog taking on a slow, bloated incumbent", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-19b5d7018f7a", "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 110, "type": "analysis", "title": "The slide highlights the performance leap of the Ampere architecture.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 6, "textChars": 574, "nDataPoints": 8, "notes": "The slide highlights the performance leap of the Ampere architecture.", "elementsJson": [ "headline_text", "paragraph", "bar_chart_grouped", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.455+00", "seconds": 3.201, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/110", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-110", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The A100 GPU is NVIDIA's first processor based on their new Ampere architecture. The company produced 4x performance gains on MLPerf in 1.5 years.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-33b1a5363d91", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.44, "x": 0.28, "y": 0.38 }, "kind": "chart", "text": "Performance comparison of V100 vs A100 across 8 models", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.455+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0f795edd-d00e-4c95-80f7-e28c2ec1f461", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Training speedup: 2.5x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-34ba61f1e0fe", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "The A100 GPU is NVIDIA's the first processor based on their new Ampere architecture. The company produced 4x performance gains on MLPerf in 1.5 years.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.455+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fa7a53eb-7f5a-4899-858b-7d7434b0ae79", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.7, "x": 0.07, "y": 0.94 }, "kind": "source-note", "text": "Figure note: Per chip performance based on comparing performance at same scale and normalizing it to a single chip.", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.455+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5e114dae-de44-4f9d-84af-c59268f11962", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "NVIDIA will not rest either: Up to 2.5x training speedups with the new A100 GPU vs V100", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.455+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7461bd89-bb85-4305-bbb8-6b52968f2dc5", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "04bd1378-2252-4a0c-806d-0314bbc129aa", "evidence": "The slide compares performance across 8 models, which could be represented in a 2x4 or 2x2 matrix, but the exact structure is not clear.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.276346+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 110, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-d3d773aaa610", "evidence": "2.5x speedup vs V100.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 110, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 110, "from": 107, "loop": { "name": "16_david_goliath", "slug": "16-david-goliath", "status": "active", "bestFor": "Startups, disruptive innovation, competitive displacement", "canonId": "019dd956-6c5d-704a-9c33-1aa083ca432e", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Giant's Weakness -> Our Slingshot (Unique Edge) -> The Topple", "description": "Frame your initiative as the nimble underdog taking on a slow, bloated incumbent", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-19b5d7018f7a", "evidence": "Slide 108 directly compares Graphcore M2000 vs NVIDIA DGX-A100 cost.", "position": 13, "objective": "Position Graphcore against NVIDIA/Google in AI compute", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 111, "type": "analysis", "title": "The slide uses a combination of a bar-line chart for GitHub project growth and a line chart for search trends.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 62, "componentCount": 7, "textChars": 649, "nDataPoints": 2, "notes": "The slide uses a combination of a bar-line chart for GitHub project growth and a line chart for search trends.", "elementsJson": [ "headline_text", "bar_chart_grouped", "line_chart", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:43.931+00", "seconds": 2.436, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/111", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-111", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "25% of the top-20 fastest growing GitHub projects in Q2 2020 concern ML infrastructure, tooling and operations.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-8a195b94df85", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.02, "y": 0.4 }, "kind": "chart", "text": "GitHub project growth analysis", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": null, "framework": null }, "subkind": "bar-line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.931+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c27d14a8-97dc-44b1-9b49-89fac207e16c", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.35, "x": 0.55, "y": 0.4 }, "kind": "chart", "text": "Google Search Trends for MLOps", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.931+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a56f0d2-c281-446e-99fb-dad7eb2ace48", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.8, "x": 0.02, "y": 0.25 }, "kind": "list", "text": "25% of the top-20 fastest growing GitHub projects in Q2 2020 concern ML infrastructure, tooling and operations. Google Search traffic for “MLOps” is now on an uptick for the first time.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.931+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bf832ef1-2abd-46dc-b248-78b168e88827", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Relative search volume: 25%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-8daba8eee3d3", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.92 }, "kind": "source-note", "text": "Figure note: Left graph reproduces analysis from Runa Capital and right graph reproduces data from Google Search Trends for “MLOps”", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.931+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1c44b3ae-473a-4b79-b014-1d83afb8ec73", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.02, "y": 0.13 }, "kind": "title", "text": "The rise of MLOps (DevOps for ML) signals an industry shift from technology R&D (how to build models) to operations (how to run models)", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:43.931+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "712d6ff2-0e2c-4cb5-9798-9d4a2b9971d4", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-d61b76661837", "evidence": "Title: rise of MLOps signals industry shift.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 90, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chartjunk Elimination", "slug": "chartjunk-elimination", "status": "active", "bestFor": null, "canonId": "019dd956-b49e-7453-ab01-b60f338eeda1", "version": 1, "bodyDocId": null, "description": "Remove decorative elements that don't convey data: 3D effects, shadows, unnecessary gridlines", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "6a394b14-28e3-4ce4-8a01-1d583498ee8a", "evidence": "The slide uses a combination of bar and line charts to display GitHub project growth analysis and Google Search Trends for 'MLOps', which is an example of effective chart selection and junk elimination.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.34563+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-d951efff9228", "evidence": "MLOps framed as 'DevOps for ML'.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 111, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 112, "type": "analysis", "title": "The slide lists 12 specific regulatory requirements for AI applications.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 4, "textChars": 1012, "nDataPoints": 0, "notes": "The slide lists 12 specific regulatory requirements for AI applications.", "elementsJson": [ "headline_text", "numbered_list", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.498+00", "seconds": 2.269, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/112", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-112", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "External monitoring is transitioning from a focus on business metrics down to low-level model metrics.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-8f7cfa591ffe", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.6, "x": 0.2, "y": 0.35 }, "kind": "list", "text": "Sample (Hong Kong Monetary Authority*)\n1. Board and Senior Management accountable for the outcome of AI applications\n2. Possessing sufficient expertise\n3. Ensuring an appropriate level of explainability of AI applications\n4. Using data of good quality\n5. Conducting rigorous model validation\n6. Ensuring auditability of AI applications\n7. Implementing effective management oversight of third-party vendors\n8. Being ethical, fair and transparent\n9. Conducting periodic reviews and ongoing monitoring\n10. Complying with data protection requirements\n11. Implementing effective cybersecurity measures\n12. Risk mitigation and contingency plan", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Numbered list", "slug": "numbered", "status": "active", "canonId": "019de52c-fcc0-7718-a284-a4250c805df2", "version": 1, "description": null }, "framework": null }, "subkind": "numbered", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.498+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "15d76101-6a83-442a-9d22-16928415c14c", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "External monitoring is transitioning from a focus on business metrics down to low-level model metrics. This creates challenges for AI application vendors including slower deployments, IP sharing, and more:", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.498+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4f824ee0-55a2-45c7-bbc3-297cc3649cbb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "As AI adoption grows, regulators give developers more to think about", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.498+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b04f5a37-47de-4756-8893-340500fcb959", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "338fb3f8-e47e-45dd-8902-c96422fb4988", "evidence": "components: list/numbered: Sample (Hong Kong Monetary Authority*) 1. Board and Senior Management accountable for the outcome of AI applications", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:20:48.917267+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 112, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 113, "type": "analysis", "title": "The chart uses a diverging stacked bar format to compare two distinct metrics (cost decrease vs revenue increase) across the same set of business functions.", "function": "analyze_data", "rawType": "data_table", "block": null, "metadata": { "slideType": "data_table", "slideTypeCanon": { "name": "Data table", "slug": "data_table", "status": "active", "canonId": "019de52d-0bbd-71b9-9918-417f904abbf8", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 51, "componentCount": 6, "textChars": 439, "nDataPoints": 96, "notes": "The chart uses a diverging stacked bar format to compare two distinct metrics (cost decrease vs revenue increase) across the same set of business functions.", "elementsJson": [ "bar_chart_stacked", "bar_chart_stacked" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.885+00", "seconds": 2.626, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/113", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-113", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Results from a poll of 1,872 enterprises worldwide: Cost decreases and revenue increases from AI by function.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-ffa34d1211dc", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.25, "x": 0.22, "y": 0.38 }, "kind": "chart", "text": "Average cost decrease", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.885+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "56422080-e17f-4bf9-bbe3-351fb0aed64b", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.25, "x": 0.55, "y": 0.38 }, "kind": "chart", "text": "Average revenue increase", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.885+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e7748a42-87d2-467f-8ca8-2a8cb6615ba8", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI impact on revenue and cost: 1,872", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-01fb255ec508", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.6, "x": 0.05, "y": 0.25 }, "kind": "subtitle", "text": "Results from a poll of 1,872 enterprises worldwide: Cost decreases and revenue increases from AI by function.", "attrs": null, "canon": { "kind": { "name": "Subtitle", "slug": "subtitle", "status": "active", "canonId": "82d821ee-2793-42e7-8f16-c6131385a3ce", "version": 1, "description": "Header-zone supporting text near the title (= ex body.description with bbox)." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.885+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d6e1b88d-75aa-4fca-8a67-935d72689220", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Enterprises report that AI drives revenue in sales and marketing while reducing costs in supply chain management and manufacturing functions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.885+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "330f3907-d107-45cf-ab04-c8e9916a5291", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ecf-de2334eb1b7e", "evidence": "Title: AI drives revenue and reduces costs by function.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 113, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ecf-e05d30596c26", "evidence": "Function-by-function comparison view.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 65, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 113, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 114, "type": "analysis", "title": "The slide uses a bubble-heatmap style table to visualize AI adoption across different industry sectors.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "balanced", "densityScore": 49, "componentCount": 5, "textChars": 566, "nDataPoints": 144, "notes": "The slide uses a bubble-heatmap style table to visualize AI adoption across different industry sectors.", "elementsJson": [ "headline_text", "paragraph", "heatmap" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.627+00", "seconds": 5.313, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/114", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-114", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "3% of respondents, the “high performers”, report 11 live AI use cases vs. 3 for the average enterprise.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-d8267c86e5f7", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.32, "x": 0.34, "y": 0.395 }, "kind": "chart", "text": "Organizations' AI capabilities, % of respondents by industry", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Heatmap", "slug": "heatmap", "status": "active", "canonId": "019de52d-00cd-71ee-bf36-233b7817c304", "version": 1, "description": null }, "framework": null }, "subkind": "heatmap", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.627+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "25564296-f88a-460a-b532-c43703383391", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "3% of respondents, the “high performers”, report 11 live AI use cases vs. 3 for the average enterprise. Retail businesses reported the largest YoY use case expansion. AI tends to be applied in areas of core competency:", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.627+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "66c6ce48-c283-4ed5-a3c1-7caef00fd005", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI capability adoption: 11", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-defd14d1bae2", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "RPA and computer vision are the most common deployed techniques in the enterprise. Speech, natural language generation and physical robots are the least common", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.627+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "47ccc7e6-3b53-4de1-8b88-702e83ed662b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-e46d25a90315", "evidence": "11 use cases for high performers vs 3 for average.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 114, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Heatmap matrix", "slug": "heatmap-matrix", "status": "active", "bestFor": null, "canonId": "019de52d-2363-7248-9aa8-cd0bcf077891", "version": 1, "bodyDocId": null, "description": null, "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "704e76a1-e1f2-441c-9793-c45895d4143c", "evidence": "The slide uses a heatmap to display organizations' AI capabilities across different industries, which is an example of a heatmap matrix.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:49.445951+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 114, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 115, "type": "analysis", "title": "The chart shows ARR growth from $1M in 2015 to $360M in 2019.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "balanced", "densityScore": 39, "componentCount": 5, "textChars": 440, "nDataPoints": 5, "notes": "The chart shows ARR growth from $1M in 2015 to $360M in 2019.", "elementsJson": [ "headline_text", "paragraph", "bar_chart_vertical" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:44.372+00", "seconds": 1.994, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/115", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-115", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "With over 7,000 enterprise customers, UiPath's annual revenue growth is emblematic of the demand for operational automation.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-9df17257e720", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.32, "x": 0.34, "y": 0.45 }, "kind": "chart", "text": "UiPath ARR growth 2015-2019", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.372+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9edbf25a-eeb9-4903-bc05-ca5937d367b9", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Annual Recurring Revenue: $360M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6a-a04e7c1143e7", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.85, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "With over 7,000 enterprise customers, UiPath's annual revenue growth is emblematic of the demand for operational automation. By mid-2020 the business passed $400M in annual recurring revenue.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.372+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8b11e08b-4174-4673-a5ec-efef4dd93f00", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Robotic process automation continues to tear through the enterprise", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:44.372+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "736227c1-64c3-4451-9eb0-7480b310c4b7", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "fe71d316-b762-41cf-83be-2f3acaa0a435", "evidence": "The slide uses a bar chart to display UiPath ARR growth, which is a common chart type for showing growth over time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:49.514726+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 115, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-e881347ba74e", "evidence": "7,000 enterprise customers UiPath.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 115, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 115, "from": 111, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-1e509d37e8ee", "evidence": "MLOps growth, enterprise survey on revenue/cost impact, RPA scale.", "position": 14, "objective": "Show enterprise shift from R&D to MLOps and adoption metrics", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 116, "type": "evidence", "title": "The slide highlights the practical application of deep learning in voice assistants, contrasting performance with legacy systems.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 794, "nDataPoints": 2, "notes": "The slide highlights the practical application of deep learning in voice assistants, contrasting performance with legacy systems.", "elementsJson": [ "headline_text", "bullet_list", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.264+00", "seconds": 6.855, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/116", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-116", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The system can understand noisy speech from telephone lines and has a success rate of >90% for an average 8-turn conversation.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-52ee3e10a297", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.4, "x": 0.55, "y": 0.55 }, "kind": "chart", "text": "AI success rate", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "100304e6-f233-479b-a80f-680c66ed3935", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.4, "x": 0.05, "y": 0.55 }, "kind": "chart", "text": "Abandon rate", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e3669a0d-2513-46f5-a2b3-0638ebb01fd1", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "PolyAI has rolled out its voice assistant for hospitality in the UK. The system is actively answering reservation calls and assisting diners with special dietary requirements and providing COVID-19 guidance.\nPowered by the company's latest deep learning technology, the system can understand noisy speech from telephone lines and has a success rate of >90% for an average 8-turn conversation.\nWith the recent advances in technology, we have seen new AI assistants learning from interactions much faster than their predecessors like Siri or Alexa.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "52d1ba7e-8175-4740-8dd0-0d2cd88935c9", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "AI success rate: >90%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-56e1a35a8e6c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.75, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "AI dialogue assistants are live and handling calls from UK customers today", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.264+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "64bb58b4-0baa-4bf8-a61c-1649bdf4dec8", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-eff20cde4b5b", "evidence": ">90% success on 8-turn calls.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 116, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 117, "type": "evidence", "title": "The slide highlights the value proposition of Tractable's AI in the insurance industry, specifically for auto claims.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 56, "componentCount": 5, "textChars": 1149, "nDataPoints": 5, "notes": "The slide highlights the value proposition of Tractable's AI in the insurance industry, specifically for auto claims.", "elementsJson": [ "headline_text", "action_title", "bullet_list", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.167+00", "seconds": 2.54, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/117", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-117", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Tractable's AI captures and processes imagery of the damage to automatically predict its repair costs.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:36+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-e6960a1a383b", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.35, "x": 0.62, "y": 0.25 }, "kind": "image", "text": "Tractable mobile app and web dashboard interface showing AI-based assessment.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.167+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "827c6a98-55e0-4f1a-8fe5-62febab7a0dd", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.55, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "Tractable's AI captures and processes imagery of the damage to automatically predict its repair costs.\nAccidents and disasters drive $1T of damage globally/year.\nRecovery always begins with visual damage appraisal.\nFor vehicle repair, Tractable's AI automates damage appraisal to accelerate recovery from 30 days to 1 week.\nThe system is trained using tens of millions of auto damage photos and the expert appraiser-approved repairs that ensued.\nUsers can now photograph damage on their phone and immediately obtain complete repair estimates, enabling near-instant decisions on next steps (total loss, repair, settlement etc) that would previously take days to weeks to reach.\nTractable has processed $1B+ in auto claims and is used today by the world's leading insurers, including Tokio Marine (Japan), Covea (France), Talanx-Warta (Central Europe), Admiral Seguros (Spain).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.167+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8ccba02e-4245-474f-9cbc-f633d923c314", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "recovery time: 1 week", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:36+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-eba1e61e2414", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Computer vision unlocks faster accident and disaster recovery intervention", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.167+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "91385e18-3821-4e65-b929-6d3dbb2c1232", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-f29bd8756312", "evidence": "1 week intervention timeline.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 117, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 118, "type": "analysis", "title": "The slide highlights a specific product capability (Tinyclues) within the context of industry trends in AI/ML automation.", "function": "present_solution", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_solution", "functionCanon": { "name": "Present solution", "slug": "present_solution", "status": "active", "canonId": "019de52d-15b8-71da-bd23-2b767850cbce", "version": 1, "description": null }, "density": "dense", "densityScore": 73, "componentCount": 7, "textChars": 1431, "nDataPoints": 3, "notes": "The slide highlights a specific product capability (Tinyclues) within the context of industry trends in AI/ML automation.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "tree_map", "logo_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.676+00", "seconds": 7.045, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/118", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-118", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The API powers Tinyclues marketing decisioning suite with capabilities such as targeting, campaign prioritization and the ability to predict efficient marketing topics (pictured above) It powers more than 100,000 marketing campaigns, delivering an average revenue uplift of 40% against legacy approaches.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-da87ccc45f5d", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.45, "x": 0.51, "y": 0.033 }, "kind": "chart", "text": "Efficiency Map", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Treemap", "slug": "treemap", "status": "active", "canonId": "019de52d-016e-713a-98c8-e205161a7741", "version": 1, "description": null }, "framework": null }, "subkind": "treemap", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.676+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8b379eb7-b9a6-4392-9c70-545cdd070508", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.48, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "Conventional wisdom says that creating value from real-world, first-party customer datasets with custom attributes and tables require custom data-science.\nTinyclues has built a no-code prediction API that is able to predict any customer event from any 360 customer dataset, out of the box.\nIt automates feature engineering and deep learning, exposing high-level business-centric controls.\nPrediction strategies, algorithm selection and hyperparameter tuning is optimized at scale on 100+ entreprise datasets and 1.5B+ customer records.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.676+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "65e04472-6ce5-4a6f-bdc6-bf1c319ab719", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "revenue uplift: 40%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-de02436002c8", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "Despite their diversity and lack of normalization, first-party 360 customer datasets share structural commonality. Tinyclues leverages this commonality to run a no-code prediction API.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.676+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5441ce7b-5397-4f89-a54b-464ae777bfb3", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.51, "y": 0.7 }, "kind": "subtitle", "text": "The API powers Tinyclues marketing decisioning suite with capabilities such as targeting, campaign prioritization and the ability to predict efficient marketing topics (pictured above) It powers more than 100,000 marketing campaigns, delivering an average revenue uplift of 40% against legacy approaches.", "attrs": null, "canon": { "kind": { "name": "Subtitle", "slug": "subtitle", "status": "active", "canonId": "82d821ee-2793-42e7-8f16-c6131385a3ce", "version": 1, "description": "Header-zone supporting text near the title (= ex body.description with bbox)." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.676+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3ae1b841-5838-4bfc-8af0-7ba3783607ea", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "No-code ML automation: A universal prediction API for 360 customer data", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.676+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6854ed13-60da-4ceb-9056-2a4069cb0e14", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-f4a7fce29219", "evidence": "40% improvement metric cited.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 118, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 119, "type": "analysis", "title": "The slide uses a radar chart to compare four companies across five ESG pillars, supported by a corresponding data table with conditional formatting.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 63, "componentCount": 6, "textChars": 906, "nDataPoints": 20, "notes": "The slide uses a radar chart to compare four companies across five ESG pillars, supported by a corresponding data table with conditional formatting.", "elementsJson": [ "headline_text", "bullet_list", "data_table", "radar_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.47+00", "seconds": 5.81, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/119", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-119", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "NLP can derive ESG perception scores by assessing the relationships and sentiments of products and companies with respect to client-specific ESG reputation pillars (e.g., environment, diversity, and more).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-6f9dda87c93d", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.3, "x": 0.65, "y": 0.4 }, "kind": "chart", "text": "Radar chart comparing ESG performance of four companies", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Radar chart", "slug": "radar", "status": "active", "canonId": "019de52d-00f5-74e6-82d0-74c065f6f4a9", "version": 1, "description": null }, "framework": null }, "subkind": "radar", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10e50cab-6012-4872-aeb0-e31947ee3091", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.55, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "NLP can derive ESG perception scores by assessing the relationships and sentiments of products and companies with respect to client-specific ESG reputation pillars (e.g., environment, diversity, and more). Investors are increasingly demanding evidence of ESG performance. This approach uses NLP to tag millions of news articles daily to identify and understand relevant coverage using entity linking, saliency and topic classification.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f8aabb1b-36a9-42c4-8aa0-81f7df7c75f1", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "ESG perception score", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-728d6ca783a7", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.55, "x": 0.05, "y": 0.6 }, "kind": "table", "text": "ESG scores table for Companies W, X, Y, Z across 5 categories", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed10489b-68e4-4582-88f4-ae66d8498829", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "NLP is used to automate quantification of a company’s Environmental, Social and Governance (ESG) perception using the world’s news", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "75d15a18-de1c-4cf2-9000-671e28638565", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "4068c7a2-1cce-465e-a435-28ecdbdb0749", "evidence": "Mentions Environmental, Social and Governance pillars", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 17:45:48.47+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 119, "frameworkName": "ESG Framework" } ], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 120, "type": "analysis", "title": "The chart shows a sharp increase in COVID-19 related phishing emails compared to general phishing emails during early 2020.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 5, "textChars": 628, "nDataPoints": 2, "notes": "The chart shows a sharp increase in COVID-19 related phishing emails compared to general phishing emails during early 2020.", "elementsJson": [ "headline_text", "bullet_list", "line_chart" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.33+00", "seconds": 2.62, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/120", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-120", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "During COVID-19, Tessian observed a 30x increase in email phishing attacks that specifically exploited uncertainty around the pandemic.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-61605247c778", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.5, "x": 0.25, "y": 0.45 }, "kind": "chart", "text": "COVID-19 related phishing attacks are on the rise", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.33+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c35d4a39-062a-4b34-a9bd-31113f0b4f01", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "During COVID-19, Tessian observed a 30x increase in email phishing attacks that specifically exploited uncertainty around the pandemic.\nWith employees being the gatekeepers to sensitive data, human error is the cause of 88% of data breaches.\nTessian's ML model is trained on employee's email behavior to detect and block inappropriate traffic.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.33+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ac137a35-f265-4b56-aed4-ad506f237ded", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Spear Phishing Emails Detected: 30x", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-678bd5771250", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Machine learning protects humans from email spear phishing attacks", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.33+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6969b9e2-5654-4cd5-ac9f-afe3cee6ee99", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-fa27e1fa8328", "evidence": "30x increase in pandemic phishing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 120, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 121, "type": "evidence", "title": "The slide uses a visual callout technique to map specific document errors to their locations on an ID card.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 6, "textChars": 514, "nDataPoints": 0, "notes": "The slide uses a visual callout technique to map specific document errors to their locations on an ID card.", "elementsJson": [ "headline_text", "paragraph", "screenshot", "other" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.899+00", "seconds": 3.183, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/121", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-121", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "With more identity documents digitally captured, Onfido's AI system learns to detect fake documents that run rampant online.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:31+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-a7886330264b", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.15, "x": 0.78, "y": 0.043 }, "kind": "image", "text": "AI feature map visualization", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.899+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "320de76a-07ce-4f7e-b147-1ba914419d4e", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.6, "x": 0.08, "y": 0.035 }, "kind": "image", "text": "Annotated French ID card", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.899+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d1a5d35f-94b3-49f2-9281-205a58ac4ba5", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.3, "x": 0.6, "y": 0.4 }, "kind": "list", "text": "1 character too short\nDoesn't match initials\nWrong photo printing technique\nFonts and spacing wrong throughout\nCharacter aspect wrong throughout", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.899+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "65d8cf9a-711f-48a6-b902-30ce90765cb2", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "With more identity documents digitally captured, Onfido's AI system learns to detect fake documents that run rampant online.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.899+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7cc260ad-4150-4336-9fc6-7d904eebbff0", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Computer vision detects subtle evidence of tampered identity documents", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.899+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ab9d037d-26e6-4788-8413-edfbf4a139c3", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "062fe0a1-276f-40ef-96cb-3013abc220e0", "evidence": "The slide seems to be targeted at an audience interested in AI applications, specifically in identity document verification.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.090808+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 121, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual Anchors", "slug": "visual-anchors", "status": "active", "bestFor": null, "canonId": "019dd956-c264-7357-a791-89fa5f02a400", "version": 1, "bodyDocId": null, "description": "Design elements that guide users' eyes through the content intentionally", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "7771a736-a551-4853-adf1-cfeda0554aa8", "evidence": "The use of AI feature map visualization and annotated French ID card as visual aids.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:57:04.131173+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 121, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 122, "type": "evidence", "title": "The slide highlights the efficiency gains of AI over manual compliance processes, specifically in adverse media coverage.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 61, "componentCount": 6, "textChars": 868, "nDataPoints": 9, "notes": "The slide highlights the efficiency gains of AI over manual compliance processes, specifically in adverse media coverage.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_grouped" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.038+00", "seconds": 5.282, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/122", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-122", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "ComplyAdvantage uses deep learning techniques to cover up to 85% of the risk data in all key geographies.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-452002198b92", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "chart", "text": "Adverse media coverage per geography: Previous vs Now", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Grouped bar chart", "slug": "bar-grouped", "status": "active", "canonId": "019de52c-ff3b-751e-a8ff-f21ff11097b1", "version": 1, "description": null }, "framework": null }, "subkind": "bar-grouped", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.038+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d4423f0f-f36a-47ce-9467-e05bac266dd7", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "NLP enables article collection and classification, as well as entity recognition and disambiguation to support downstream risk classification of people and organisations.\nA typical professional analyst can process 120 articles in the time that ComplyAdvantage’s automated solution can process 8 million articles.\nComplyAdvantage’s adverse media coverage per geography is now averaging 80% with the latest ML pipelines.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.038+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "230d6493-1190-4ba9-9725-f21aad817c54", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Adverse media coverage: 80%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-4b24ff12cfe9", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Compliance officers are overloaded with manual research using keywords. ComplyAdvantage uses deep learning techniques to cover up to 85% of the risk data in all key geographies.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.038+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1b7deef2-c3ac-4b0e-a807-f347b2f05c03", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "AI is the key to Web-scale content analysis for money laundering and terrorist financing", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.038+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "55b10ce5-3048-4270-9fdc-363d290813e2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "f9970606-2f7f-4346-962a-a144c6c330c5", "evidence": "The use of a bar chart to display adverse media coverage per geography.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.255115+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ecf-fcd6ecdf2fc0", "evidence": "85% risk data coverage.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "1db4bb10-633a-40eb-8d79-a00eff501aaa", "evidence": "The slide presents a narrative about AI-powered content analysis for risk classification.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.20968+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 122, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 123, "type": "evidence", "title": "The slide uses a bar chart to show time reduction and a precision-recall curve to show performance maintenance.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 806, "nDataPoints": 4, "notes": "The slide uses a bar chart to show time reduction and a precision-recall curve to show performance maintenance.", "elementsJson": [ "bar_chart_grouped", "area_chart", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.9+00", "seconds": 6.82, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/123", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-123", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Machine translation is used to generate multilingual training data for financial crime classification. This approach significantly reduced lead time from 20 weeks for English to less than 2 weeks per European language while maintaining more than 80% of the recall and precision.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-305c014a34c5", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.5, "y": 0.4 }, "kind": "chart", "text": "Multilingual financial crime classification performance from English is maintained", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Area chart", "slug": "area", "status": "active", "canonId": "019de52c-ff89-71e0-911e-5cab78b90f53", "version": 1, "description": null }, "framework": null }, "subkind": "area", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.9+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a0d768f6-6a91-4a55-85be-49f2724ae0c6", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.05, "y": 0.4 }, "kind": "chart", "text": "Days to collect annotated training data for adverse media classification", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.9+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed921db8-6886-48d1-b4ff-fb1d14f7e2d7", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Lead time reduction: 20 weeks", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-34d3d04cbcf6", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "Machine translation is used to generate multilingual training data for financial crime classification. This approach significantly reduced lead time from 20 weeks for English to less than 2 weeks per European language while maintaining more than 80% of the recall and precision.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.9+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0a3a5a64-e6b6-4eab-8773-b40a7e8295a5", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Machine translation unlocks financial crime classification globally", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.9+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b1d7678e-e090-42a3-b3ac-6485d44c07de", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Before-After-Bridge", "slug": "before-after-bridge", "status": "active", "bestFor": null, "canonId": "019dd956-8604-7188-883f-42370288ad05", "version": 1, "bodyDocId": null, "description": "Narrative structure: Paint the before, show the after, explain the bridge", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "d26c760c-1f05-45e2-b92a-e496ac0064d9", "evidence": "The slide presents a comparison of lead times before and after using machine translation.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.33772+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chartjunk Elimination", "slug": "chartjunk-elimination", "status": "active", "bestFor": null, "canonId": "019dd956-b49e-7453-ab01-b60f338eeda1", "version": 1, "bodyDocId": null, "description": "Remove decorative elements that don't convey data: 3D effects, shadows, unnecessary gridlines", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "cf75213f-f6f9-4156-a5b8-e537e9a6b3b4", "evidence": "The use of simple and clear charts to display data.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.377361+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-00823accbdc8", "evidence": "20 weeks -> 2 weeks lead time.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 123, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 124, "type": "evidence", "title": "The slide highlights the rapid adoption of open-source AI research into large-scale production environments.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 29, "componentCount": 4, "textChars": 476, "nDataPoints": 0, "notes": "The slide highlights the rapid adoption of open-source AI research into large-scale production environments.", "elementsJson": [ "headline_text", "paragraph", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.301+00", "seconds": 2.054, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/124", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-124", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "From open source publication to processing search queries in large-scale production within 12 months (assuming the paper’s publication was not purposefully held back).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6b-c05e4ab2472e", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.58, "x": 0.21, "y": 0.4 }, "kind": "image", "text": "Before and After search result comparison", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.301+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e373976c-5f17-48d4-b692-0140595fadb6", "parentComponentId": null }, { "bbox": { "h": 0.07, "w": 0.96, "x": 0.02, "y": 0.28 }, "kind": "paragraph", "text": "From open source publication to processing search queries in large-scale production within 12 months (assuming the paper’s publication was not purposefully held back).", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.301+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0622890b-6e9e-4713-951e-485b4ed782a5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "BERT language model goes mainstream: Upgrading Google and Microsoft’s Bing search query understanding", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.301+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "92e00dba-0c7e-426a-93d7-d389a26ebc29", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Aha! Moment", "slug": "aha-moment", "status": "active", "bestFor": null, "canonId": "019dd956-a5b5-7763-820e-b299c5f27a32", "version": 1, "bodyDocId": null, "description": "The moment when the audience realizes the value - design for insight revelation", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "31d47980-5771-4838-9199-3b1303f29a1a", "evidence": "The slide presents a surprising fact about BERT language model's adoption.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.460148+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 124, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9726c492-b718-4408-b4fb-6cc7777ce8cf", "evidence": "The use of a screenshot to highlight the differences in search results.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.500606+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 124, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 125, "type": "evidence", "title": "Part of the 'Industry' section of the State of AI 2020 report.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 5, "textChars": 566, "nDataPoints": 1, "notes": "Part of the 'Industry' section of the State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.631+00", "seconds": 2.268, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/125", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-125", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.04, "w": 0.6, "x": 0.05, "y": 0.22 }, "kind": "callout", "text": "Supply chain operators realise a 70% reduction in direct labour as a result.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.631+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8b080dcf-7ebd-450c-8e5a-5365b73c4681", "parentComponentId": null }, { "bbox": { "h": 0.26, "w": 0.46, "x": 0.49, "y": 0.29 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.631+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "07cde1e6-d2ee-4377-9d90-e8178ff9bd73", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "The company offers a portfolio of complete AI-enabled robotic solutions that pick, pack, sort, and transport products and packages autonomously for fulfillment operations. Operating 24/7 to automate break pack order selection. The approach combines AI with industrial robotics, mobile robotics, computer vision, advanced sensing, novel gripping, and engineered infrastructure.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.631+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0b9baae6-352b-4207-ac4d-6cf78d260080", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "direct labour reduction: 70%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-4e1955cc1b6c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Berkshire Grey robotic installations are achieving millions of robotic picks per month", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.631+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "83be5d8f-c352-46c4-95ff-29ba688f0ee8", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9fb87768-d034-43b1-b01d-49688f1c3e74", "evidence": "The slide implies that the solution can be implemented to achieve significant labour reduction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.580508+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 125, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-04c4b0e27f30", "evidence": "70% direct labour reduction.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 125, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 126, "type": "evidence", "title": "The slide uses a process flow diagram to illustrate the transition from a 3D file to a finished part via automated CNC programming.", "function": "illustrate_case", "rawType": "client_example", "block": null, "metadata": { "slideType": "client_example", "slideTypeCanon": { "name": "Client example", "slug": "client_example", "status": "active", "canonId": "019de52d-0c87-77cb-9867-613ae22e95ac", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 763, "nDataPoints": 3, "notes": "The slide uses a process flow diagram to illustrate the transition from a 3D file to a finished part via automated CNC programming.", "elementsJson": [ "headline_text", "bullet_list", "process_diagram", "photo", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.791+00", "seconds": 2.428, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/126", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-126", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "CloudNC's Factory OS replaces expert humans with autonomous software to soon achieve >95% productivity.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-58e2ab1d81c6", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.95, "x": 0.03, "y": 0.54 }, "kind": "diagram", "text": "3D File -> CNC Machine -> Finished Parts", "attrs": null, "canon": { "kind": { "name": "Diagram", "slug": "diagram", "status": "active", "canonId": "019de52c-f9f0-734a-81de-15382ff28c65", "version": 1, "description": "Schematic or relational diagram." }, "tool": null, "subkind": { "name": "Process flow", "slug": "process", "status": "active", "canonId": "019de52d-039c-718f-ad3a-efe9c4b3c838", "version": 1, "description": null }, "framework": null }, "subkind": "process", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.791+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1f30582d-0da4-4ddf-a3f6-c82380ff405d", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "CNC Machines produce over >$168B worth of parts per year for manufacturing, carving blocks of metal into useful shapes. CloudNC is automating the programming of these machines.\nThere are a huge number of ways of producing even a simple component, and humans are unable to find optimal manufacturing solutions. This results in 9% productivity vs the theoretical maximum.\nCloudNC's Factory OS replaces expert humans with autonomous software to soon achieve >95% productivity.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.791+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bbb12904-b7e8-49b6-99c8-d40c019e59b6", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Productivity: >95%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-5ecac2a42dec", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.6, "x": 0.02, "y": 0.02 }, "kind": "other", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Other", "slug": "other", "status": "active", "canonId": "019de52c-fa90-743c-8bf6-20d1fbe82b6d", "version": 1, "description": "Fallback when no canonical kind applies." }, "tool": null, "subkind": { "name": "Unclassified", "slug": "unclassified", "status": "active", "canonId": "019de52d-05ca-77db-81fb-db7d4234bbc0", "version": 1, "description": null }, "framework": null }, "subkind": "unclassified", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.791+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fa48ed34-2b34-4f21-b73d-06922974ec61", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Manufacturing: CNC Machine programming starts to be automated", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.791+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "43824d7b-b01b-4d07-9f73-5c605d6fc3ba", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Big Idea Formula", "slug": "big-idea-formula", "status": "active", "bestFor": "Any narrative-driven persuasive deck whose job is to move the audience: keynotes, fundraising pitches, capital-allocation board asks, all-hands behaviour-change talks, conference talks meant to shift mental models. Use it before structuring; pair with Pyramid when the deck must also withstand line-by-line scrutiny.", "canonId": "019dd956-7dd1-74d7-9866-ffdef19ecfbf", "version": 1, "bodyDocId": "c30d7484-69e0-4e21-82cf-20fa0966869e", "description": "Nancy Duarte's articulation of how to frame a presentation's central message: the speaker's unique point of view plus what's at stake for the audience, expressed as one complete declarative sentence. The narrative-driven counterpart to Minto's analytical governing thought.", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "ddbbc789-031a-41c8-ad22-31a3fdc36e14", "evidence": "The slide presents a key idea about automating CNC machine programming.", "pageRefs": null, "priority": "Core", "whenToUse": "Any narrative-driven persuasive deck whose job is to move the audience: keynotes, fundraising pitches, capital-allocation board asks, all-hands behaviour-change talks, conference talks meant to shift mental models. Use it before structuring; pair with Pyramid when the deck must also withstand line-by-line scrutiny.", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.703251+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": "Solves the topic-not-thesis trap and the stake-less recommendation trap simultaneously. A single declarative sentence with a verb forces the speaker's stance into the room; routing the consequence through the word 'you' translates internal urgency into the audience's currency. The result is short enough to be repeated after the meeting — and that repetition is the persuasion mechanism.", "antipattern": "A topic dressed as a thesis ('Sustainability Roadmap 2026'), a stake attached to the speaker rather than the audience, two ideas spliced with 'and', abstract verbs (leverage / unlock / empower) that survive deletion, or writing the Big Idea after the deck is already built so the slides don't actually defend it.", "cardinality": null, "narrativePurpose": "Forces the deck to commit to a single declarative thesis before slide one — a sentence that names the speaker's stance and routes a consequence through the audience's currency. Acts as the filter every slide, anecdote, and chart must ladder up to." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-0a570d7926a8", "evidence": ">95% productivity target.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Process flow", "slug": "process-flow", "status": "active", "bestFor": null, "canonId": "019dd9e1-4dca-76bc-8135-401836a1c4c0", "version": 1, "bodyDocId": null, "description": "Workflow, steps, sequential diagram.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "24fe0dac-49b5-4d0a-8b5f-6b3e9c8e4b5a", "evidence": "The slide presents a diagram of the 3D File -> CNC Machine -> Finished Parts process.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:04.66245+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 126, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 127, "type": "analysis", "title": "The chart shows a clear upward trend in the usage of various transformer models over time.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 716, "nDataPoints": 10, "notes": "The chart shows a clear upward trend in the usage of various transformer models over time.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_stacked" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.582+00", "seconds": 7.188, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/127", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-127", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "1,000+ companies are using Hugging Face's Transformers library in production: 5M pip installs, 2,500+ community transformer models trained in over 164 languages by 430 contributors.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-7d4426164fb1", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.23, "y": 0.45 }, "kind": "chart", "text": "Number of weekly unique user instantiations of the top 10 models from huggingface.co", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.582+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "60dd8896-8184-47d8-919f-143de315c355", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "list", "text": "1,000+ companies are using Hugging Face’s Transformers library in production: 5M pip installs, 2,500+ community transformer models trained in over 164 languages by 430 contributors.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.582+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2d5ea462-2d49-467f-9f4b-6597dd4aed59", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "Models fine-tuned on a variety of tasks including text classification, information extraction, question answering, summarization, machine translation, and more.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.582+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "df038fcd-a445-423b-a5fe-bd9d6406707c", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "weekly unique user instantiations: 5M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-83bc2c739a77", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Open source model and dataset sharing is driving NLP’s cambrian explosion", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.582+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e854ae86-7c3f-4fec-8807-b4195d3f80b9", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Segmentation", "slug": "audience-segmentation", "status": "active", "bestFor": null, "canonId": "019dd956-8079-7046-8e6b-ecdf1bf15590", "version": 1, "bodyDocId": null, "description": "Grouping audience by attitude toward your message: Champions, Neutrals, Skeptics", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "aa836a9c-6ec8-4604-aedc-93b00927b87e", "evidence": "The slide seems to be targeted at an audience interested in NLP applications.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.821192+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-0df4d7861454", "evidence": "5M pip installs / 2,500 models / 164 languages.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data-Ink Ratio", "slug": "data-ink-ratio", "status": "active", "bestFor": null, "canonId": "019dd956-afc8-70da-b378-92834d36fde6", "version": 1, "bodyDocId": null, "description": "Maximize data, minimize non-data ink - remove chartjunk and decoration", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "42815d8d-5692-463b-901b-f0a365bb99a1", "evidence": "The slide presents a bar chart with a clear and simple design.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.782008+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-12a1937c50b8", "evidence": "'Cambrian explosion' biological metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 127, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 128, "type": "analysis", "title": "The slide highlights adoption across six major industries with specific metrics on downloads and contributors.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 69, "componentCount": 10, "textChars": 445, "nDataPoints": 2, "notes": "The slide highlights adoption across six major industries with specific metrics on downloads and contributors.", "elementsJson": [ "logo_grid", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:46.34+00", "seconds": 2.8, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/128", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-128", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Rasa's libraries and tools have clocked >2 million downloads and have 400+ open source contributors.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:36+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-010e815ec7d5", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.16, "x": 0.42, "y": 0.7 }, "kind": "image", "text": "Telecommunications industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3dea0790-8b99-474f-81d0-ad9e1ff6a5da", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.18, "x": 0.62, "y": 0.35 }, "kind": "image", "text": "Banking industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "47dc9b2e-924b-4811-96b7-32b25c0fbb20", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.16, "x": 0.42, "y": 0.35 }, "kind": "image", "text": "Insurance industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "52dcdad2-d0be-4b33-8251-28d7d2dbef48", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.18, "x": 0.62, "y": 0.7 }, "kind": "image", "text": "Manufacturing industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9bcef4cc-7edd-45c7-bc0c-b94036a8d822", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.18, "x": 0.21, "y": 0.35 }, "kind": "image", "text": "Healthcare industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aab27bca-a7a0-4297-8a86-81952535d660", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.18, "x": 0.21, "y": 0.7 }, "kind": "image", "text": "Technology industry logos", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo grid", "slug": "logo-grid", "status": "active", "canonId": "019de52c-fdd7-7409-93e5-81280a8ca215", "version": 1, "description": null }, "framework": null }, "subkind": "logo-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e02f25b9-3777-4c95-afc8-9e6c7789a000", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "downloads: >2 million", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:36+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-05bba7f933e6", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.7, "x": 0.04, "y": 0.22 }, "kind": "paragraph", "text": "Rasa's libraries and tools have clocked >2 million downloads and have 400+ open source contributors.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "53382ded-344e-4c77-9c7c-1508d46fe44d", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Open source conversational AI expands its footprint across industry", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.34+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "759cf8e8-c457-44a6-ad13-c1213dfd4479", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Iconography", "slug": "iconography", "status": "active", "bestFor": null, "canonId": "019dd9e1-4d71-77c5-a8d6-20a534104995", "version": 1, "bodyDocId": null, "description": "Icon grids and pictograms standing in for text.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "75285175-9646-4a46-bbee-280a4dd3d618", "evidence": "The slide presents a grid of industry logos.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:04.900446+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 128, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null }, { "tool": { "name": "Social Proof", "slug": "social-proof", "status": "active", "bestFor": null, "canonId": "019dd956-990d-7620-82b3-6dd286f34568", "version": 1, "bodyDocId": null, "description": "People adapt behavior based on what others do - show adoption and testimonials", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "f5499e89-9015-458a-9cc5-bc67867a57d9", "evidence": "The slide mentions the number of downloads and contributors.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:04.939883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 128, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 128, "from": 116, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2076d04ec9c2", "evidence": "Thirteen consecutive case studies: dialogue, claims, fraud, ESG, robots, NLP.", "position": 15, "objective": "Parade real-world AI deployments across verticals", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 129, "type": "analysis", "title": "The chart uses a combination of bar and line graphs to represent capital raised and deal count respectively, with 2020 data being annualized.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 62, "componentCount": 7, "textChars": 539, "nDataPoints": 22, "notes": "The chart uses a combination of bar and line graphs to represent capital raised and deal count respectively, with 2020 data being annualized.", "elementsJson": [ "headline_text", "action_title", "paragraph", "bar_chart_vertical", "line_chart", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:46.477+00", "seconds": 2.802, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/129", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-129", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "2020 is likely to hit $25B+ in total volume and 350+ deals.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-b66368357b9c", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.5, "x": 0.25, "y": 0.4 }, "kind": "chart", "text": "Total Capital Raised", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.477+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6146d7eb-3ea9-4b93-ad7a-974249dcbd37", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.5, "x": 0.25, "y": 0.4 }, "kind": "chart", "text": "Number of deals", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.477+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1657f457-20f4-471c-bdce-9723fee4ccdb", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Total capital raised: $25B+", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-b847cc06d216", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "2020 is likely to hit $25B+ in total volume and 350+ deals. Rounds >$100M consistently account for ~10% of all funding rounds since 2018 onwards. This signals the increasing maturity of the field.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.477+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "84938990-b89f-4ed3-9d72-2771acdfbab8", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.75, "x": 0.05, "y": 0.94 }, "kind": "source-note", "text": "Figure note: Data retrieved from Pitchbook on 13 August 2020. Asterix indicates annualized figures for 2020 using light blue and orange.", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.477+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b1bf9992-a78b-4e27-9766-fe8eba5968ed", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Private >$15M funding rounds for AI-first companies remain strong in spite of COVID-19", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.477+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4a1be85e-95a1-4451-b2a7-2d1687c1201b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-161823e24473", "evidence": "Title: AI funding remains strong despite COVID.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "37d80e1d-c29c-454e-871d-68fcc102d513", "evidence": "The use of a bar chart and line chart to display funding data.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:05.021319+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-1899be92cca9", "evidence": "$25B+, 350+ deals.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Storytelling Effect", "slug": "storytelling-effect", "status": "active", "bestFor": null, "canonId": "019dd956-981b-74f9-9096-5c8df03dab5a", "version": 1, "bodyDocId": null, "description": "Stories are remembered up to 22x better than facts alone", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "2dbb4282-6a8a-41ad-9d36-58f6d758a94f", "evidence": "The slide presents a narrative about funding trends.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:05.059254+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 129, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 129, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 10, "name": "The Facts (What)", "beatId": "019dd95a-0682-776c-8e34-ed2a8f38a754", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Facts (What)", "slug": "triple-take-the-facts-what", "status": "active", "canonId": "019dd9b8-092b-754c-8eaa-ede3ffdcb060", "version": 1, "description": null }, "evidence": "Research, Talent, Industry sections inventory what happened in AI in 2020.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Setup", "slug": "setup", "status": "active", "canonId": "019dd9b8-020e-7505-af9b-d9b9a4679bcc", "version": 1, "description": "Where the deck grounds the reader before doing anything else. Almost every arc opens here." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 129, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 10, "name": "Rising Action", "beatId": "019dd95a-0682-776c-8e34-fc384f439bba", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Rising Action", "slug": "mountain-rising-action", "status": "active", "canonId": "019dd9b8-1992-742b-be9c-85d0ebee0591", "version": 1, "description": null }, "evidence": "Research, Talent and Industry sections build momentum of AI progress.", "position": 2, "isPrimaryArc": false, "parentBeatType": { "name": "Development", "slug": "development", "status": "active", "canonId": "019dd9b8-0332-7008-b3aa-11b0840a57b1", "version": 1, "description": "The middle stretch in arcs that have one. Often where length problems creep in." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 130, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 1, "componentCount": 1, "textChars": 19, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.295+00", "seconds": 3.6, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/130", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-130", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.042, "w": 0.254, "x": 0.373, "y": 0.458 }, "kind": "title", "text": "Section 4: Politics", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.295+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed0ae0b5-25ef-4000-baa0-2bc0e0a30be6", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "d3aff7c4-f141-42e8-8e5a-d4061bf5b812", "evidence": "The slide is clearly a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:05.138416+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 130, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 131, "type": "analysis", "title": "The slide highlights academic and investigative work that brought AI ethics into the mainstream.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 36, "componentCount": 4, "textChars": 688, "nDataPoints": 0, "notes": "The slide highlights academic and investigative work that brought AI ethics into the mainstream.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "screenshot" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:46.982+00", "seconds": 3.12, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/131", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-131", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Examples include policing, the judiciary and the military.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-d85fb27b05ae", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.43, "x": 0.52, "y": 0.345 }, "kind": "image", "text": "The Perpetual Line-Up: Unregulated Police Face Recognition in America", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.982+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2c7f4c66-3bae-4a5c-a6b0-1a2e400714da", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.45, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "Examples include policing, the judiciary and the military. A few trailblazing researchers include: Joy Buolamwini, Timnit Gebru, Gender Shades; Clare Garvie, Alvaro Bedoya, and Jonathan Frankle, The Perpetual Line-Up; Adam Harvey, Megapixels; P Allo, M Taddeo, S Wachter, L Floridi, The ethics of algorithms; Margaret Boden, Joanna Bryson, Alan Winfield et al., Principles of robotics.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.982+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "542f043d-c2c3-4316-995a-642bf70620ef", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Ethical risks: A group of researchers have spent years helping to frame the ethical risks of deploying ML in certain sensitive contexts. This year those issues went mainstream.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.982+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2f6f3005-ea01-44a6-b0b6-18b76d2ae5d9", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "5034c473-5339-445c-9f90-7d1f1cc157f3", "evidence": "The slide mentions specific researchers and their work.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:57:05.217939+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 131, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 132, "type": "setup", "title": "The slide uses a color-coded map to visualize the prevalence of facial recognition technology globally.", "function": "establish_context", "rawType": "geographic_map", "block": null, "metadata": { "slideType": "geographic_map", "slideTypeCanon": { "name": "Geographic map", "slug": "geographic_map", "status": "active", "canonId": "019de52d-09de-7153-8845-584c1a9d0c83", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 47, "componentCount": 6, "textChars": 442, "nDataPoints": 3, "notes": "The slide uses a color-coded map to visualize the prevalence of facial recognition technology globally.", "elementsJson": [ "headline_text", "paragraph", "map" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:45.995+00", "seconds": 2.046, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/132", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-132", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "50% of the world currently allows the use of facial recognition.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-9bc8f80f0314", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.7, "x": 0.15, "y": 0.35 }, "kind": "image", "text": "World map showing facial recognition status", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Map", "slug": "map", "status": "active", "canonId": "019de52c-fe4d-77bd-b0a4-d72d58b9e5a4", "version": 1, "description": null }, "framework": null }, "subkind": "map", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.995+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e0e32eb6-e406-459f-9da9-20365494fe5a", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.3, "x": 0.05, "y": 0.8 }, "kind": "legend", "text": "Actively in use; Considering; Partially-banned or no evidence of use", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.995+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed2d7da7-d228-4c70-99e3-845b81d5a8df", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Global adoption rate: 50%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-9f3eb4c8d4ba", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "50% of the world currently allows the use of facial recognition. Only 3 countries (Belgium, Luxembourg, Morocco) have partial bans on the technology that only allow it in specific cases.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.995+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6cf1de23-480d-49ab-be1d-fff01de5dc24", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Facial recognition is remarkably common around the world", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:45.995+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ca99e0ed-f9db-46ef-8059-737a7f3864e8", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-21dfe70c07ee", "evidence": "50% of world allows it.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Data Story Arc", "slug": "data-story-arc", "status": "active", "bestFor": null, "canonId": "019dd956-9c30-728e-b974-d3fb123dcf5a", "version": 1, "bodyDocId": null, "description": "Context → Conflict → Insight → Implication structure for data narratives", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "263a3bd8-78d4-4e3e-b996-e9c7777fd3c8", "evidence": "The slide presents a narrative about facial recognition adoption.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:57:05.299403+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ed0-1dd22d777e68", "evidence": "World map visualises facial-recognition prevalence.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 132, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 133, "type": "evidence", "title": "The slide uses the 'tip of the iceberg' metaphor to imply broader, unrecorded instances of algorithmic bias.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 44, "componentCount": 6, "textChars": 489, "nDataPoints": 0, "notes": "The slide uses the 'tip of the iceberg' metaphor to imply broader, unrecorded instances of algorithmic bias.", "elementsJson": [ "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.119+00", "seconds": 2.908, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/133", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-133", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "These examples are likely just the tip of the iceberg.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-108edd25c07e", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.44, "x": 0.54, "y": 0.28 }, "kind": "image", "text": "Side-by-side photos of the individuals involved in the wrongful arrests.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4c90d8f7-d957-4be9-baa1-b8896821a005", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.5, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "May 2019: Detroit police arrested Michael Oliver... January 2020: Detroit police arrested Robert Williams... These examples are likely just the tip of the iceberg.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6f137529-aa34-42dd-8f04-2c902c97e71b", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "the computer must have gotten it wrong", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-142364f4f91a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.02, "y": 0.22 }, "kind": "title", "text": "Two (known) examples of wrongful arrests due to erroneous use of facial recognition algorithms emerge.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bad15d36-16f4-414e-b13b-9b4561af21a3", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: From potential risks to wrongful arrests", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a9d540c9-1751-4408-b21a-190b296f86e2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Anadiplosis", "slug": "anadiplosis", "status": "active", "bestFor": null, "canonId": "019dd956-a0a5-75f9-a98c-7b7d8ee377a1", "version": 1, "bodyDocId": null, "description": "Ending one clause with a word that begins the next", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "b9437467-b4b6-4ad1-8ef9-e7cbe9805ac1", "evidence": "These examples are likely just the tip of the iceberg.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:51.949034+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "2c8cdab9-f3f9-494f-bb2e-1fb378d3fc40", "evidence": "Facial recognition: From potential risks to wrongful arrests", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:51.982068+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-2aaa2c756640", "evidence": "Wrongful arrests framed emotionally.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-271adfb71b54", "evidence": "'Tip of the iceberg' metaphor in callout.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "8a21a9bb-9a1f-4dd5-b772-3526cb3c8cb7", "evidence": "The text explicitly mentions 'These examples are likely just the tip of the iceberg.'", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:47.119+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 133, "frameworkName": "iceberg-model" } ], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 134, "type": "evidence", "title": "Part of the 'State of AI 2020' report, specifically the 'Politics' section.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 48, "componentCount": 5, "textChars": 698, "nDataPoints": 5, "notes": "Part of the 'State of AI 2020' report, specifically the 'Politics' section.", "elementsJson": [ "headline_text", "bullet_list", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:46.434+00", "seconds": 2.198, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/134", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-134", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Facebook's automatic photo-tagging was in violation of Illinois' 2008 Biometric Privacy Act.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-f4384419a358", "parentComponentId": null }, { "bbox": { "h": 0.41, "w": 0.43, "x": 0.525, "y": 0.305 }, "kind": "image", "text": "Facebook photo tagging interface", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ba81d3a-4ce0-4063-b18a-3ee9f76da823", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.48, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Facebook's automatic photo-tagging was in violation of Illinois' 2008 Biometric Privacy Act.\nIllinois' biometric privacy law is the strongest in the country and says that businesses must get permission before collecting biometric data.\nThe class action suit, brought in 2015 claimed that Facebook's photo-tagging feature which it rolled out in 2010 did not do this.\nFacebook's maximum exposure via the suit was $47B. In the end this suit is likely to net each affected user $200-400.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fecd25d2-a56c-4458-b8bf-f05fadec26b5", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Settlement amount: $650M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6c-f944c8bafde4", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: Facebook settles class action lawsuit for $650M", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.434+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f3da271f-15dc-45d3-b5cc-98810d56f2cd", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-2f33b142cff0", "evidence": "$650M settlement specific.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 134, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 135, "type": "evidence", "title": "The slide uses a bar chart to contrast the scale of Clearview's database against traditional law enforcement databases.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 65, "componentCount": 6, "textChars": 1338, "nDataPoints": 4, "notes": "The slide uses a bar chart to contrast the scale of Clearview's database against traditional law enforcement databases.", "elementsJson": [ "headline_text", "bullet_list", "bar_chart_vertical", "quote_block" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.226+00", "seconds": 5.741, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/135", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-135", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "A New York Times investigation revealed that Clearview scraped billions of images and then licensed their “search engine for faces” to over 600 law enforcement agencies.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-97f8013989be", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.28, "x": 0.68, "y": 0.35 }, "kind": "chart", "text": "Number of searchable photos per database", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.226+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8c9a0624-eb90-4b51-bef0-df47d56869c6", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.65, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "A New York Times investigation revealed that Clearview scraped billions of images and then licensed their “search engine for faces” to over 600 law enforcement agencies.\nClearview claims to have scraped these photos from Facebook, YouTube, Venmo and millions of other websites. While this would have been technically straightforward for companies like Facebook or Google to build, they had refrained due to concerns about privacy and misuse.\nFederal and state law enforcement offices said that the app had been used to help solve a variety of cases. A follow up investigation from Buzzfeed revealed that Clearview's technology had also been used by private individuals, banks, schools, the US Department of Justice and retailers including Best Buy.\nBuilding on Illinois law, the ACLU has sued Clearview. Al Gidari, a privacy professor at Stanford commented “Absent a very strong federal privacy law, we’re all screwed.”", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.226+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bce8aaba-6eef-43d5-abb6-45edfe77b455", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of searchable photos: 3B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-9d427138d180", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "Absent a very strong federal privacy law, we’re all screwed. — Al Gidari, a privacy professor at Stanford", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-9bc0d41ef4cd", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Clearview exposes what is now technically possible with facial recognition", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.226+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1e2c11c3-d4e5-4d4d-a6ce-ed241ba0fe2f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Aha! Moment", "slug": "aha-moment", "status": "active", "bestFor": null, "canonId": "019dd956-a5b5-7763-820e-b299c5f27a32", "version": 1, "bodyDocId": null, "description": "The moment when the audience realizes the value - design for insight revelation", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "cf66170a-a225-45ac-af75-64274fa569cf", "evidence": "A New York Times investigation revealed that Clearview scraped billions of images and then licensed their “search engine for faces” to over 600 law enforcement agencies.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:52.085881+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-32b60627f4a2", "evidence": "Billions of images, 600 agencies.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-34160a750f56", "evidence": "'Search engine for faces' framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 135, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 136, "type": "analysis", "title": "The slide highlights corporate policy shifts by Microsoft, Amazon, IBM, and Apple regarding facial recognition.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 30, "componentCount": 4, "textChars": 401, "nDataPoints": 1, "notes": "The slide highlights corporate policy shifts by Microsoft, Amazon, IBM, and Apple regarding facial recognition.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.25+00", "seconds": 2.744, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/136", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-136", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Large technology companies are taking a more careful path.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-7ce4a1a149d0", "parentComponentId": null }, { "bbox": { "h": 0.26, "w": 0.26, "x": 0.674, "y": 0.296 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.25+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "54411e9b-b5ce-4c85-9533-ff87c2173c43", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.65, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Large technology companies are taking a more careful path. Microsoft deleted it's database of 10 million faces. Amazon announced a one-year pause on police use of Rekognition. IBM announced it would sunset general purpose facial recognition. Apple is asked to enable FaceID for masks.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.25+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "73fb3361-73eb-4d4e-b396-49d415812046", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: More thoughtful approaches gather steam", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.25+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d6130749-585b-4a99-a33d-94984de4bc3b", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 137, "type": "analysis", "title": "The slide uses a grid of photos to illustrate the 'Original' vs 'Balanced' datasets.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 862, "nDataPoints": 1, "notes": "The slide uses a grid of photos to illustrate the 'Original' vs 'Balanced' datasets.", "elementsJson": [ "headline_text", "bullet_list", "icon_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:46.945+00", "seconds": 2.436, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/137", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-137", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The creators of ImageNet produced an update that takes first steps towards reducing bias.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:33+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-2fd0b4c67b30", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.5, "x": 0.46, "y": 0.25 }, "kind": "image", "text": "Grid of photos showing original and balanced datasets", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Icon grid", "slug": "icon-grid", "status": "active", "canonId": "019de52c-fdfe-749d-9990-1db465a3eed4", "version": 1, "description": null }, "framework": null }, "subkind": "icon-grid", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.945+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6e137437-53de-4ba3-8c33-16726eab4a8d", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.4, "x": 0.02, "y": 0.6 }, "kind": "list", "text": "The researchers also plan to release a tool that will allow users to retrieve sets of images balanced by gender, skin colour or age, to allow developers to produce algorithms that more fairly classify faces and activities in images.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.945+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3afa26a0-52bf-4ea5-bdac-ad390868bb6f", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.4, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "The creators of ImageNet produced an update that takes first steps towards reducing bias.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.945+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6cbf93b1-8ecb-4e33-8cf4-b91405fcb26a", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.4, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "The ImageNet team recruited 12 graduate students representing 4 countries of origin, male and female genders, and a handful of racial groups from diverse backgrounds to systematically identify offensive categories, such as racial and sexual characterizations, among ImageNet's person categories and proposed removing them from the database.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.945+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f1a3c2ba-b999-4e00-a483-e9ac7e2a833a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: More thoughtful approaches gather steam", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:46.945+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5a858357-d43e-4ce0-91e8-1f25825a3732", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Evidence Matrix", "slug": "evidence-matrix", "status": "active", "bestFor": null, "canonId": "019dd956-7ecc-74c2-b46c-0b9c1f4c73d7", "version": 1, "bodyDocId": null, "description": "Planning visual assets: Claim → Evidence Needed → Data Source → Visual Type", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "1b2e565d-7407-4a10-ad6f-9b40421a249b", "evidence": "The ImageNet team recruited 12 graduate students representing 4 countries of origin, male and female genders, and a handful of racial groups from diverse backgrounds to systematically identify offensive cate", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:52.183939+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 137, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-38cdd6b8e936", "evidence": "Callout: ImageNet creators reduce bias.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 137, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 138, "type": "evidence", "title": "The slide uses a specific case (Ed Bridges vs South Wales Police) to illustrate the shift toward proactive regulation of AI technologies.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 38, "componentCount": 4, "textChars": 751, "nDataPoints": 0, "notes": "The slide uses a specific case (Ed Bridges vs South Wales Police) to illustrate the shift toward proactive regulation of AI technologies.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.056+00", "seconds": 2.524, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/138", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-138", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This means that action on bias cannot be legally deferred until the tech has matured.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-cf1fd400737b", "parentComponentId": null }, { "bbox": { "h": 0.225, "w": 0.4, "x": 0.564, "y": 0.349 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.056+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b3961288-025f-4698-b150-663fa1149065", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.52, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "Shifting legal framework for law enforcement. The High Court in the UK became the first court to review a police force's use of automated facial recognition technology. Although judges ruled against the claimant, they also established an important new duty for the police to make sure that discrimination is proactively “eliminated”. The emphasis is on regulating technology now rather than after harm has occurred. A spokesperson for South Wales police made it clear that the force plans to continue to use facial recognition technology.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.056+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2419157b-ebcc-4295-af8c-2c3fab701c31", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.95, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: A new legal precedent in the UK emphasizes that facial recognition tools cannot \"move fast and break things\"", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.056+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "31c06cb0-5220-4879-8bdd-1cb19abf7835", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-3d30f02a937e", "evidence": "'Move fast and break things' analogy critiqued.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 138, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 139, "type": "evidence", "title": "The slide highlights the intersection of policy and corporate influence in AI regulation.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 48, "componentCount": 5, "textChars": 979, "nDataPoints": 0, "notes": "The slide highlights the intersection of policy and corporate influence in AI regulation.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.636+00", "seconds": 3.052, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/139", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-139", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The new law requires government agencies to obtain a warrant to run facial recognition scans.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:34+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-b7d5cccb8c59", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.28, "x": 0.69, "y": 0.308 }, "kind": "image", "text": "Jay Inslee", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dded79de-258b-4c3a-baab-5af93631231e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "The new law requires government agencies to obtain a warrant to run facial recognition scans.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7bb510ec-84e5-479a-8004-1c5de78b5563", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.65, "x": 0.02, "y": 0.29 }, "kind": "list", "text": "In March 2020 Jay Inslee (pictured) signed the first US state law that carefully restricts law enforcement's use of facial recognition technology. The software used must be accessible to an independent third party via an API to assess for “accuracy and unfair performance differences” across characteristics like race or gender. If unfair performance is found “the provider must develop and implement a plan to mitigate the identified performance differences within ninety days of receipt of such results”. The law also requires training and public reporting around usage of facial recognition. State Senator Joe Nguyen, who is a senior program manager at Microsoft, had sponsored the legislation.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ad64145d-e898-4a4f-b0ac-f16febda833c", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.95, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Facial recognition: Washington State passes new law with active support from Microsoft", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "45da43b6-0568-47cc-bb68-2087b7afbc80", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 140, "type": "evidence", "title": "The slide uses a specific case (Professor Guo Bing) to illustrate a broader trend of privacy regulation in China.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 82, "componentCount": 10, "textChars": 1296, "nDataPoints": 0, "notes": "The slide uses a specific case (Professor Guo Bing) to illustrate a broader trend of privacy regulation in China.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.066+00", "seconds": 3.354, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/140", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-140", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Guo's lawsuit focused on the risk of data leaks: \"once leaked, illegal misuse will easily endanger the safety of consumers\".", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-e60c5b9de74d", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.36, "x": 0.61, "y": 0.35 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "91b4d6ea-13bf-4540-abc7-1133f500d013", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.55, "x": 0.03, "y": 0.35 }, "kind": "list", "text": "Guo's lawsuit focused on the risk of data leaks: \"once leaked, illegal misuse will easily endanger the safety of consumers\". The safari park has since changed its entrance policy to allow visitors to choose between facial recognition or fingerprint recognition.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7e85a354-9fa3-44b0-b2df-89bf664ecbb6", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.55, "x": 0.03, "y": 0.76 }, "kind": "list", "text": "The Personal Information Security Specifications is a new voluntary standard for data privacy in China that is now being trialled by companies including Tencent and Alipay.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b81410e7-067b-48d7-8cbd-7d944656212c", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.55, "x": 0.03, "y": 0.22 }, "kind": "list", "text": "Professor Guo Bing of Zhejiang Sci-Tech University sued a local safari park for \"violating consumer protection law\" after it made facial recognition registration a mandatory requirement for visitor entrance", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e2f0e8aa-179a-498a-a519-ddb5f1277dc5", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.55, "x": 0.03, "y": 0.65 }, "kind": "list", "text": "Lei Chaozi, director of science and technology at China's Ministry of Education has pledged to \"curb and regulate\" the use of facial recognition in schools.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "eabdce70-e1ac-48d2-8ea6-2d6ce3d84d30", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.55, "x": 0.03, "y": 0.52 }, "kind": "list", "text": "China's use of facial recognition is incredibly widespread (see coverage in The State of AI 2018 & 2019) but there are some signs that privacy concerns are being taken more seriously.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "eee42399-b870-40b2-964a-f787c36f7608", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "once leaked, illegal misuse will easily endanger the safety of consumers", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6d-e95dcdbc99eb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Facial recognition: The first legal challenge in China", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "356a984b-e0f9-452f-b2ba-e2f418fa35be", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.6, "x": 0.03, "y": 0.03 }, "kind": "title", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Subtitle", "slug": "subtitle", "status": "active", "canonId": "019de52c-fb09-739d-900c-815b0d9ac526", "version": 1, "description": null }, "framework": null }, "subkind": "subtitle", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.066+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b66f1ee8-fca2-480d-9053-a3fb9c68e0a1", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 140, "from": 131, "loop": { "name": "26_ripple_effect", "slug": "26-ripple-effect", "status": "active", "bestFor": "Change management, scaling initiatives, demonstrating broad value", "canonId": "019dd956-7075-7564-a134-cbde706dcfe0", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Individual Impact -> Team Impact -> Organizational Impact -> Market/Industry Impact", "description": "Show how impact spreads from individual to team to organization to market", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-270e46b4c916", "evidence": "Risks -> wrongful arrests -> Facebook $650M -> Clearview -> UK ruling -> WA law -> China lawsuit.", "position": 16, "objective": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 141, "type": "analysis", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 4, "textChars": 972, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.325+00", "seconds": 2.606, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/141", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-141", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.22 }, "kind": "callout", "text": "Increased awareness of deepfakes causes a rush of activity led by China and California.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.325+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f85ac554-f79d-4526-a3a1-06578e40ed54", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.3, "x": 0.65, "y": 0.325 }, "kind": "image", "text": "Comparison of original and deepfake video frames", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.325+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d20d4a9a-1bd9-4d55-8367-7d88e608e993", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.6, "x": 0.05, "y": 0.3 }, "kind": "list", "text": "China's internet regulator announced a ban on the publishing and distribution of “fake news” created via AI and mandated that use of AI also needs to be clearly marked in a prominent manner. China's top legislative body said earlier this year it was considering making deepfake technology illegal.\nCalifornia passed law AB 730, aimed at deepfakes, which criminalises distributing audio or video that gives a false, damaging impression of a politician's words or action.\nMany other US state bills have been passed, addressing different risks. Virginia law amends current criminal law on revenge porn to include computer-generated pornography.\nVarious thoughtful approaches have been proposed by CSET in their deepfakes report including broadly distributing detection technology.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.325+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5eacb9ae-54c9-48cb-ad7b-41012e674b07", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Lawmakers scramble to legislate against the use of deepfakes", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.325+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8cdf2536-dfa2-4fae-9549-04d7117bfb26", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 143, "from": 141, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2af0a5465c72", "evidence": "Three slides clustering legislative responses to algorithmic risks.", "position": 17, "objective": "Show regulators waking up to deepfakes, ADM and model bias", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 142, "type": "analysis", "title": "Slide from State of AI 2020 report.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 44, "componentCount": 4, "textChars": 1188, "nDataPoints": 0, "notes": "Slide from State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.321+00", "seconds": 2.179, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/142", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-142", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Multiple countries and states start to wrestle with how to regulate the use of ML in decision making.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-182b48259bb8", "parentComponentId": null }, { "bbox": { "h": 0.42, "w": 0.27, "x": 0.69, "y": 0.3 }, "kind": "image", "text": "Jacinda Ardern", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.321+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6fdc0e3a-4905-4a22-bed3-e1c7feea6e52", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.96, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Multiple countries and states start to wrestle with how to regulate the use of ML in decision making.\nA Dutch court has ordered the immediate halt of an automated surveillance system for detecting welfare fraud citing that it violates human rights.\nNew Zealand's Prime Minister (pictured) says they're the first in the world to produce a set of standards for how public agencies should use algorithms to make decisions.\nThe UK's Home Office is to scrap a controversial automated decision-making algorithm used to filter applications for UK visas. The UK also rolled back a countrywide A-level exam grading algorithm after huge public outcry and evidence that it was biased against disadvantaged students.\nWashington State passed a law that requires government agencies to obtain a warrant to run facial recognition scans, except in case of emergency. The software used must have a way to be independently tested for \"accuracy and unfair performance differences\" across skin tone, gender, age and other characteristics.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.321+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "79e8a4c7-76a9-4409-baf1-d41d0f479242", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Algorithmic decision making: Regulatory pressure builds", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.321+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6a2c70b1-d01f-46a1-b29b-3fba7cdaea4d", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 143, "from": 141, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2af0a5465c72", "evidence": "Three slides clustering legislative responses to algorithmic risks.", "position": 17, "objective": "Show regulators waking up to deepfakes, ADM and model bias", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 143, "type": "evidence", "title": "The slide uses screenshots of the OpenAI Playground interface to provide empirical evidence of model bias.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 5, "textChars": 712, "nDataPoints": 0, "notes": "The slide uses screenshots of the OpenAI Playground interface to provide empirical evidence of model bias.", "elementsJson": [ "paragraph", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.939+00", "seconds": 6.584, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/143", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-143", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Models trained on large volumes of language on the internet will reflect the bias in those datasets unless their developers make efforts to fix this.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-31417765a6d2", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.47, "x": 0.51, "y": 0.45 }, "kind": "image", "text": "OpenAI Playground interface showing GPT-2 output", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.939+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a4dccb4-d360-4e8a-ae3b-8ef248faa213", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.47, "x": 0.02, "y": 0.45 }, "kind": "image", "text": "OpenAI Playground interface showing GPT-3 output", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.939+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b6c1f579-46ef-4cf7-a7ad-8a4e238143c9", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.96, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "Example from the GPT-3 (left) and GPT-2 (right) with prompts and the model's predictions, which contain clear bias. Models trained on large volumes of language on the internet will reflect the bias in those datasets unless their developers make efforts to fix this. See our coverage in State of AI Report 2019 of how Google adapted their translation model to remove gender bias.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.939+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "04824cb8-ece9-4da1-b532-e96ab8032f4c", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "GPT-3, like GPT-2, still outputs biased predictions when prompted with topics of religion", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.939+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cfba62c7-23c4-4cb1-a603-1bc7a841bae2", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 143, "from": 141, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2af0a5465c72", "evidence": "Three slides clustering legislative responses to algorithmic risks.", "position": 17, "objective": "Show regulators waking up to deepfakes, ADM and model bias", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 144, "type": "evidence", "title": "The chart is a t-SNE or similar dimensionality reduction visualization showing clusters of game commands.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "balanced", "densityScore": 46, "componentCount": 5, "textChars": 858, "nDataPoints": 0, "notes": "The chart is a t-SNE or similar dimensionality reduction visualization showing clusters of game commands.", "elementsJson": [ "headline_text", "bullet_list", "scatter_plot", "logo_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.448+00", "seconds": 3.051, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/144", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-144", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Three months after DeepMind's StarCraft II breakthrough, the US Army publishes interesting StarCraft results.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-c866c46d13b8", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.4, "x": 0.55, "y": 0.25 }, "kind": "chart", "text": "Visualization of RL agent command clusters in StarCraft II", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Scatter plot", "slug": "scatter", "status": "active", "canonId": "019de52d-007d-72ec-8689-c79fb78e21a2", "version": 1, "description": null }, "framework": null }, "subkind": "scatter", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.448+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "15ea4d31-4f82-44e4-a52f-bef30da4f005", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "In the State of AI Report 2019, we covered DeepMind's breakthrough results on StarCraft II.\nInevitably progress applying RL to war inspired games like Go and StarCraft is also of interest to the military.\nThe US Army Research Lab published a paper exploring how natural language commands could be used to improve performance of RL agents where there are sparse reward functions.\nWhile it is notable that cutting edge research ideas are migrating from academic and corporate research labs to military labs.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.448+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bd26974d-59c9-4b4d-b303-35087b6a8ee8", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.85, "x": 0.05, "y": 0.21 }, "kind": "title", "text": "Three months after DeepMind's StarCraft II breakthrough, the US Army publishes interesting StarCraft results.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.448+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1cc5f6eb-591a-4240-8d3e-e5debeb8abb8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "From DeepMind to U.S. Army Research Lab, AI research agendas start to overlap", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.448+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ba87999d-aaae-4956-a70a-d94f7515b62c", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-435098f2846e", "evidence": "Title: research agendas overlap with US Army.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 144, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 145, "type": "analysis", "title": "Includes an annotated image of a military vehicle with various electronic warfare modules.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 62, "componentCount": 6, "textChars": 1070, "nDataPoints": 2, "notes": "Includes an annotated image of a military vehicle with various electronic warfare modules.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:47.874+00", "seconds": 2.449, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/145", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-145", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "As machine learning techniques continue to industrialise they are increasingly explored by militaries.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-ab2945837f37", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "image", "text": "Annotated photo of a military vehicle with electronic warfare components", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.874+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bfae094d-5f3a-49e6-895c-71bfa98cf4e2", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.48, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "The U.S. General Services Administration and the U.S. DoD's Joint Artificial Intelligence Center announced the award of its 5-year, $800M task order to Booz Allen Hamilton. The brief includes “data labeling, data management, data conditioning, AI product development, and the transition of AI products into new and existing fielded programs and systems”.\nCognitive Electronic Warfare is a developing area where machine learning is used to analyse enemy signals and automatically design responses to disrupt their operation. The US Army awarded Lockheed Martin $75M for a ML enabled cyber/jamming pod for use mounted on drones or humvees.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.874+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f0822d1b-d572-4728-a172-48641e48717f", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "contract value: $800M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-af18586d7547", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.48, "x": 0.02, "y": 0.22 }, "kind": "paragraph", "text": "As machine learning techniques continue to industrialise they are increasingly explored by militaries. However, the degree of real-world impact is not yet clear.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.874+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "14de2267-e186-48df-a436-5f4cf5309acb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.85, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "The U.S. continues to make major investments to implement military AI systems", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:47.874+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "42dd1e96-cefd-4cf7-945a-d2bd04ee54c1", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-475f94d77ecc", "evidence": "$800M military AI investment.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 145, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 146, "type": "analysis", "title": "Part of the 'State of AI 2020' report.", "function": "illustrate_case", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 65, "componentCount": 7, "textChars": 749, "nDataPoints": 4, "notes": "Part of the 'State of AI 2020' report.", "elementsJson": [ "headline_text", "bullet_list", "photo", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.423+00", "seconds": 2.893, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/146", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-146", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "As defense roadmaps include more ML-enabled components, startups are winning lucrative government contracts and raising large venture rounds.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-6811e59e2e07", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.56, "y": 0.33 }, "kind": "image", "text": "UAV drone image", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.423+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "125e3960-f3de-4ca1-a870-2a44519ed6fc", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.5, "x": 0.03, "y": 0.33 }, "kind": "list", "text": "Pivotal Software won a $121M contract with the US Department of Defense.\nAreas attracting significant VC investment include high resolution satellite imagery, UAVs and information management and decision making systems. We provide examples of such companies below:", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.423+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "18c17816-1b1d-413e-8fc9-8cb4a3208ef3", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Venture funding: $200M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:35+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6e-6e5ff3fa4f14", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.5, "x": 0.03, "y": 0.65 }, "kind": "other", "text": "Anduril $200M Series C July 2019, Rebellion $62M Series A 2019, Skydio $100M Series C July 2020", "attrs": null, "canon": { "kind": { "name": "Other", "slug": "other", "status": "active", "canonId": "019de52c-fa90-743c-8bf6-20d1fbe82b6d", "version": 1, "description": "Fallback when no canonical kind applies." }, "tool": null, "subkind": { "name": "Unclassified", "slug": "unclassified", "status": "active", "canonId": "019de52d-05ca-77db-81fb-db7d4234bbc0", "version": 1, "description": null }, "framework": null }, "subkind": "unclassified", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.423+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8f2baf02-26c4-4a5a-af97-e47d1ef11d42", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.5, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "As defense roadmaps include more ML-enabled components, startups are winning lucrative government contracts and raising large venture rounds.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.423+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7c58c703-4145-4913-b1a6-57fd8cbe8acc", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Startups at intersection of AI and defense raise large financing rounds", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.423+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "271f7d39-2d8a-483d-b02a-f4143b86314b", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 147, "type": "analysis", "title": "Slide from State of AI 2020 report.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 67, "componentCount": 7, "textChars": 1243, "nDataPoints": 0, "notes": "Slide from State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.178+00", "seconds": 3.6, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/147", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-147", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "US State department loosens restrictions on drone exports, shifting from a blanket denial to a more discretionary basis. Uninhabited aircraft now don't count as missiles.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-bdbbc59dbdda", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.45, "x": 0.52, "y": 0.355 }, "kind": "image", "text": "Military drone silhouette at sunset", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d00204e6-8a81-4e50-a638-4a94a64c1a7c", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.85, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "US State department loosens restrictions on drone exports, shifting from a blanket denial to a more discretionary basis. Uninhabited aircraft now don't count as missiles.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0abf2c6f-4801-43bb-92ba-4929eec4b7ba", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.45, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "Bob Menendez, ranking member of the senate foreign relations committee critiqued the decision as having \"weakened international export controls on the export of lethal drones\" and \"making it more likely that we will export some of our most deadly weaponry to human rights abusers across the world.\"", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7b4388fc-500f-4506-9e59-c456952c2a18", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.45, "x": 0.02, "y": 0.6 }, "kind": "list", "text": "Michael Horowitz of the University of Pennsylvania framed it as part of a broader issue around China's looser restrictions: \"Treating uninhabited aircraft as missiles for export policy purposes doesn't work...\"It has allowed China to capture a significant chunk of the drone export market, including with U.S. allies and partners.\"", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9f0c5a73-517c-4b38-8290-719c5d414226", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "It has allowed China to capture a significant chunk of the drone export market, including with U.S. allies and partners. — Michael Horowitz, University of Pennsylvania", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-c02cab69fa6e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Is US-China competition weakening the Missile Technology Control Regime?", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.178+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "796e76b2-9780-4804-b161-f252c853d6cb", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 148, "type": "evidence", "title": "Highlights the dual-use nature of AI techniques migrating from gaming to military applications.", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 78, "componentCount": 8, "textChars": 1275, "nDataPoints": 1, "notes": "Highlights the dual-use nature of AI techniques migrating from gaming to military applications.", "elementsJson": [ "headline_text", "bullet_list", "screenshot", "quote_block" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.907+00", "seconds": 2.993, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/148", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-148", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "This demonstrates the dual-use nature of AI: cutting edge techniques used to win in game environments inspired by war can rapidly migrate to a military context.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-700ab9c72a6d", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.33, "x": 0.64, "y": 0.3 }, "kind": "image", "text": "Visuals from the AlphaDogfight simulation", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Screenshot", "slug": "screenshot", "status": "active", "canonId": "019de52c-fd87-7509-8391-3d79fb339379", "version": 1, "description": null }, "framework": null }, "subkind": "screenshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "879ebbb6-b4d8-4a97-bc0c-b59e93904da2", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.6, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "A mixture of academic research labs (Georgia Tech) and defense contractors competed in a series of virtual dogfights. The top AI developed by Heron Systems beat a human pilot 5-0.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "65b7949b-3383-4bd9-8507-8dea7d5778ea", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.6, "x": 0.02, "y": 0.75 }, "kind": "list", "text": "The winning AI used hyper-aggressive tactics of flying very close to it's opponent whilst continuously firing with lower regard for the survival of its own plane. The anonymous human pilot said: \"The standard things that we do as fighter pilots aren't working\".", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dea51b2a-e2b7-4901-b166-52d3136257d1", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.6, "x": 0.02, "y": 0.48 }, "kind": "list", "text": "The top AI systems from Heron Systems and Lockheed Martin both made use of Deep Reinforcement Learning - the same approach applied by DeepMind in their work on Go and StarCraft II and OpenAI in their work on Dota 2. This demonstrates the dual-use nature of AI: cutting edge techniques used to win in game environments inspired by war can rapidly migrate to a military context.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f5259e7b-bc32-4e5c-91b1-c00d99d1470f", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.6, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "DARPA organised a virtual dogfighting tournament where various AI systems would compete with each other and a human fighter pilot from the US military.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fbb0d371-ee8d-48c2-84d4-b92ee8ffc6a9", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "The standard things that we do as fighter pilots aren't working.", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-7418e3f59fa5", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "After AlphaGo and AlphaStar...AlphaDogfight", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2979d223-52c3-4621-8ebf-dbf864457d36", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "13133e77-01cb-49b7-a09b-702eabc08224", "evidence": "The top AI systems from Heron Systems and Lockheed Martin both made use of Deep Reinforcement Learning - the same approach applied by DeepMind in their work on Go and StarCraft II and OpenAI in their work on", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.427096+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Metaphor & Analogy", "slug": "metaphor-analogy", "status": "active", "bestFor": null, "canonId": "019dd956-adea-7729-bca1-70810d6238a9", "version": 1, "bodyDocId": null, "description": "Making abstract concepts concrete through familiar comparisons", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-4a23826e583b", "evidence": "'AlphaDogfight' brand-extends AlphaGo metaphor.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-4e960ede1d8b", "evidence": "Callout: dual-use AI migrates from games to military.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 148, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 149, "type": "analysis", "title": "Slide from State of AI 2020 report.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 64, "componentCount": 7, "textChars": 977, "nDataPoints": 0, "notes": "Slide from State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo", "quote_block" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.766+00", "seconds": 2.71, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/149", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-149", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The US continues to emphasize importance of AI leadership to its military", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-44bddda67146", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.28, "x": 0.68, "y": 0.28 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.766+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6b92b108-6e13-4a75-9e72-14953fea6b34", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.03, "y": 0.22 }, "kind": "list", "text": "The US continues to emphasize importance of AI leadership to its military", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.766+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8482e16a-ff95-4617-9532-9888b8bd030c", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.6, "x": 0.03, "y": 0.48 }, "kind": "paragraph", "text": "He referenced the “tectonic impact of machine learning on the future of warfare, referenced China a competitor and stated: \"History informs us that those who are first to harness once-in-a-generation technologies often have a decisive advantage on the battlefield for years to come\"", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.766+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5ad7a187-dc75-435e-8288-b2a22706594c", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.6, "x": 0.03, "y": 0.28 }, "kind": "paragraph", "text": "Defense Secretary Dr. Mark T. Esper stated \"The AI agent's resounding victory demonstrated the ability of advanced algorithms to outperform humans in virtual dogfights...These simulations will culminate in a real-world competition involving full-scale tactical aircraft in 2024.\"", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.766+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e8210d04-2f1e-4dcf-972d-9198b1b6a5e8", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "History informs us that those who are first to harness once-in-a-generation technologies often have a decisive advantage on the battlefield for years to come - Defense Secretary Dr. Mark T. Esper", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-48f7c18ebf22", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.75, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "The US Secretary of Defense targets 2024 for real-life AI vs human dogfight", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.766+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2dca0300-5550-4184-b2f8-d9c7d9a01178", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f1fceed3-5cf8-4985-8a95-bb76f4cf62da", "evidence": "The US continues to emphasize importance of AI leadership to its military", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.495095+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 149, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-5212ac8e0be6", "evidence": "2024 specific milestone.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 149, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 149, "from": 144, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-2d54491b6425", "evidence": "DeepMind/Army overlap, $800M DoD investment, AlphaDogfight, 2024 target.", "position": 18, "objective": "Show AI research bleeding into military applications", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 150, "type": "analysis", "title": "Includes a logo of the CIA, likely as an example of an actor with specific operational AI principles.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 48, "componentCount": 5, "textChars": 942, "nDataPoints": 0, "notes": "Includes a logo of the CIA, likely as an example of an actor with specific operational AI principles.", "elementsJson": [ "headline_text", "bullet_list", "logo_grid" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.169+00", "seconds": 2.994, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/150", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-150", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The US Intelligence Community and DoD principles are notable for including a higher level of operational specificity.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:41+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c6-7671-886a-5f2831d12863", "parentComponentId": null }, { "bbox": { "h": 0.22, "w": 0.22, "x": 0.73, "y": 0.33 }, "kind": "image", "text": "Central Intelligence Agency seal", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.169+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9731140f-1535-45e7-8b19-520e371f49b0", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.7, "x": 0.03, "y": 0.22 }, "kind": "list", "text": "The US Department of Defense, The US Intelligence Community, China, and the OECD all develop or adopt their own AI Policy documents.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.169+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4ed3eb7e-6642-4ce2-b5a8-2792f45f1b69", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.7, "x": 0.03, "y": 0.33 }, "kind": "list", "text": "Common themes include transparency, auditability, robustness, safety, fairness. Many of the principles are fairly loosely stated. The US Intelligence Community and DoD principles are notable for including a higher level of operational specificity. The Chinese AI principles are notable for their emphasis on international cooperation. In the EU, the AI Ethics Guidelines from the AI High-Level Expert Group have been developed and are set to shape legislative action in the EU on AI. The Global Partnership on AI is a new initiative from 14 countries and the EU building on top of the OECD work.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.169+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f63ade6d-da02-4b13-ac2c-3beb19aed0b6", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Many actors attempt to define principles for responsible use of AI", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.169+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "daacff21-6360-47c8-a180-519a7872b098", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "9dc775cd-9276-40c5-9bb2-be2c9cdcf30a", "evidence": "The US Department of Defense, The US Intelligence Community, China, and the OECD all develop or adopt their own AI Policy documents.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.562856+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 150, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 152, "from": 150, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3126d540f400", "evidence": "DoD principles, NeurIPS ethics, Google fairness/privacy efforts.", "position": 19, "objective": "Survey emerging responsible AI codes and corporate efforts", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 151, "type": "analysis", "title": "The slide highlights a gap between AI research practices and more stringent standards in life sciences.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 5, "textChars": 798, "nDataPoints": 0, "notes": "The slide highlights a gap between AI research practices and more stringent standards in life sciences.", "elementsJson": [ "headline_text", "bullet_list", "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:48.783+00", "seconds": 2.502, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/151", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-151", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "As the largest conference in the field the proposals from NeurIPS should be high impact.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-a443fd019cb1", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.3, "x": 0.6, "y": 0.38 }, "kind": "image", "text": "NEURAL INFORMATION PROCESSING SYSTEMS", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Logo", "slug": "logo", "status": "active", "canonId": "019de52c-fdaf-743d-94c4-d7c6a0e55302", "version": 1, "description": null }, "framework": null }, "subkind": "logo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.783+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d328da5b-5e7d-4e71-8291-678e936ff91f", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.55, "x": 0.03, "y": 0.32 }, "kind": "list", "text": "NeurIPS will create a dedicated sub-team of reviewers with expertise at the intersection of machine learning and ethics. NeurIPS now mandates authors to include a statement of the potential broader impact. Authors are required to provide disclosure of funding. NeurIPS encourages data sharing but does not mandate it. Machine learning is behind life sciences (e.g., Nature) in data sharing requirements.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.783+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d2d3e725-b269-4fa1-92f1-89f78a5f4430", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "NeurIPS and ICLR both propose new ethical principles and expectations of researchers, but no mandatory code and data sharing. As the largest conference in the field the proposals from NeurIPS should be high impact:", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.783+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a59cb296-934c-4973-b7e2-fd01eef26a8d", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "Two of the leading AI conferences adopt new ethics codes", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:48.783+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "77bc3c0c-de82-4350-8c2b-0483b56ebd10", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 152, "from": 150, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3126d540f400", "evidence": "DoD principles, NeurIPS ethics, Google fairness/privacy efforts.", "position": 19, "objective": "Survey emerging responsible AI codes and corporate efforts", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 152, "type": "analysis", "title": "The slide highlights the 'People + AI Guidebook' framework, which breaks down responsible AI into six key areas.", "function": "present_framework", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "balanced", "densityScore": 40, "componentCount": 5, "textChars": 621, "nDataPoints": 0, "notes": "The slide highlights the 'People + AI Guidebook' framework, which breaks down responsible AI into six key areas.", "elementsJson": [ "headline_text", "icon_grid", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.394+00", "seconds": 2.753, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/152", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-152", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Proliferating educational content and tools through the TensorFlow community.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-5b6f5c548b11", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.4, "x": 0.55, "y": 0.3 }, "kind": "framework", "text": "People + AI Guidebook framework components: User Needs, Data Collection, Mental Models, Explainability, Feedback, Errors + Graceful Failure.", "attrs": null, "canon": { "kind": { "name": "Framework", "slug": "framework", "status": "active", "canonId": "019de52c-fa18-77c6-b10a-ccb8277ac0fa", "version": 1, "description": "Named conceptual framework instance." }, "tool": null, "subkind": { "name": "Framework instance", "slug": "instance", "status": "active", "canonId": "019de52d-05a2-72af-be0b-7edc60708a03", "version": 1, "description": null }, "framework": null }, "subkind": "instance", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d8925843-571a-4062-9528-c8acc3398534", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.28, "x": 0.18, "y": 0.35 }, "kind": "paragraph", "text": "Responsible AI with TensorFlow. Learn how to integrate Responsible AI practices into your ML workflow using TensorFlow. TensorFlow is committed to helping make progress in the responsible development of AI by sharing a collection of resources.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "903615f6-e133-4d60-b7dc-ea0a42a78cb8", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.6, "x": 0.04, "y": 0.23 }, "kind": "title", "text": "Proliferating educational content and tools through the TensorFlow community.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "12b25af9-d4f2-41b6-9fa2-a5a14f3279b2", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Google is leaning into fairness, interpretability, privacy and security of AI models", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1e71cfc5-8cbd-4875-a6ec-dd74c7347507", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "95cce96d-88c5-4041-b48f-a14704a8c9e0", "evidence": "Proliferating educational content and tools through the TensorFlow community.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.664152+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 152, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [ { "matchId": "8c8d565d-0abd-4ff4-b52e-eb1d5cf2462c", "evidence": "The slide explicitly references the People + AI Guidebook and lists its core modules.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 17:45:49.394+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 152, "frameworkName": "People + AI Guidebook" } ], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 152, "from": 150, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3126d540f400", "evidence": "DoD principles, NeurIPS ethics, Google fairness/privacy efforts.", "position": 19, "objective": "Survey emerging responsible AI codes and corporate efforts", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 153, "type": "analysis", "title": "Part of a larger report series (state.of.ai 2020).", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 63, "componentCount": 6, "textChars": 1173, "nDataPoints": 2, "notes": "Part of a larger report series (state.of.ai 2020).", "elementsJson": [ "bullet_list", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.115+00", "seconds": 3.339, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/153", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-153", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Department of Commerce adds 24 Chinese companies and institutions to a sanction list for “supporting the Procurement of items for military end-use in China”.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-8c215be90a08", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.96, "x": 0.02, "y": 0.32 }, "kind": "list", "text": "A further 8 companies and the Institute for Forensic Science were placed on a second list that restricts access to US technology because they are “complicit in human rights violations and abuses...against Uygurs, ethnic Kazakhs, and other members of Muslim minority groups in the Xinjiang Uygur Autonomous Region”.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.115+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3a8a60e1-c63e-4185-852c-adb4b874b165", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Department of Commerce adds 24 Chinese companies and institutions to a sanction list for “supporting the Procurement of items for military end-use in China”.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.115+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "654b23c9-aece-4be2-b933-171baae5bf12", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.64 }, "kind": "list", "text": "Unicorn facial recognition technology startups, Megvii and SenseTime, see challenges to their chip procurement supply chain and their ability to raise capital in the US through IPOs.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.115+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6a97f0fa-c89f-4b6d-93a9-a511dbba6ac0", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.96, "x": 0.02, "y": 0.48 }, "kind": "list", "text": "The list includes Qihoo 360 (antivirus software and web browser), Cloudminds (RPA software), and CloudWalk (facial recognition software). Even so, CloudWalk raised $254M from Chinese provincial and municipal funds as it eyes a public listing on the Shanghai exchange this year.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.115+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ef22b526-1290-4e9d-ad65-9c80f6e91bcb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "White House extends its ban on Chinese companies with ties to surveillance in Xinjiang", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.115+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fff2b456-7409-4302-b190-2dd3b7497373", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 154, "type": "analysis", "title": "Includes a line chart showing quarterly smartphone shipments for Samsung vs Huawei from 2015 to 2020.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 74, "componentCount": 7, "textChars": 1068, "nDataPoints": 40, "notes": "Includes a line chart showing quarterly smartphone shipments for Samsung vs Huawei from 2015 to 2020.", "elementsJson": [ "headline_text", "bullet_list", "line_chart", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:49.266+00", "seconds": 2.479, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/154", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-154", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "For the first time in 9 years a company other than Apple or Samsung led the market. However, Huawei's supply of chips is running out under US sanctions by mid September 2020.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-1d43f2024572", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.4, "x": 0.55, "y": 0.35 }, "kind": "chart", "text": "Quarterly smartphone shipments (millions) for Samsung and Huawei (2015-2020)", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c07badc0-2200-403c-9a46-1a453cafe10d", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "list", "text": "Foreign companies that use US chip making equipment would be required to obtain a US license before supplying certain chips to Huawei. President of Huawei’s consumer unit declared: “no chips and no supply”.\nHuawei’s Kirin AI chips are made by Taiwan Semiconductor Manufacturing Co, (TSMC) which took final orders from Huawei until 15 May 2020 due to the sanctions.\nHuawei is now trying to shore up manufacturing to Shanghai-based Semiconductor Manufacturing International Corp (SMIC).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "41e16393-219b-45b2-93f7-f07eca5fac6b", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Smartphone units (million)", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-27d9dfb15748", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.25 }, "kind": "paragraph", "text": "For the first time in 9 years a company other than Apple or Samsung led the market. However, Huawei’s supply of chips is running out under US sanctions by mid September 2020.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "272cbb15-ca64-4c62-badc-458ae43bbe97", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "no chips and no supply", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-203e5161bd2c", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Huawei is an increasingly dominant player in smartphones and is investing heavily in machine learning technology", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:49.266+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d017fdda-6799-4329-bb34-50c4207bf99b", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-5608dcb22f05", "evidence": "Callout flags Huawei chip supply running out.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 154, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 155, "type": "analysis", "title": "Part of the 'State of AI 2020' report series.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 59, "componentCount": 6, "textChars": 846, "nDataPoints": 3, "notes": "Part of the 'State of AI 2020' report series.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:52.881+00", "seconds": 5.675, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/155", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-155", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "TSMC grows in importance to the US semiconductor strategy.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:42+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c6-7671-886a-ae58b6df3660", "parentComponentId": null }, { "bbox": { "h": 0.245, "w": 0.365, "x": 0.625, "y": 0.319 }, "kind": "image", "text": "TSMC logo and wafer", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.881+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8476b0ab-3823-47b6-85b4-84209e6d335b", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.6, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "Intel announced delays to it's next-generation chips and announced that it could outsource some of its production to external foundries. TSMC is the natural choice and TSMC shares jumped 10% on the news. The US technology industry and TSMC are significantly co-dependent with 60% of TSMC sales coming from the US. TSMC said it would spend $12B to create a chip fab in Arizona. The factory would focus on TSMC's 5-nanometer process and start production in 2024. However, TSMC's technological edge will remain in Taiwan for the forthcoming 3-nanometer process that could start production in Taiwan in 2022.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.881+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dcadf21f-c0b3-4efa-a4c7-3edf3410afe5", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "TSMC sales from US: 60%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:42+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c6-7671-886a-b044d09319bf", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.5, "x": 0.02, "y": 0.22 }, "kind": "title", "text": "TSMC grows in importance to the US semiconductor strategy.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.881+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fa268f7e-6bae-48c2-8067-d2b86492446e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Semiconductors amplify the geopolitical significance of Taiwan and particularly TSMC", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.881+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "40bbd27c-f0a8-4fb2-99c8-17a40a871767", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "2cf0d011-cc42-4cb3-b8bd-abccb560720a", "evidence": "TSMC grows in importance to the US semiconductor strategy.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.798614+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 155, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 156, "type": "setup", "title": "Includes a line chart of financial metrics and a timeline of semiconductor process nodes.", "function": "compare_peers", "rawType": "competitive_analysis", "block": null, "metadata": { "slideType": "competitive_analysis", "slideTypeCanon": { "name": "Competitive analysis", "slug": "competitive_analysis", "status": "active", "canonId": "019de52d-093e-776b-94c6-c1573f5f1a31", "version": 1, "description": null }, "function": "compare_peers", "functionCanon": { "name": "Compare peers", "slug": "compare_peers", "status": "active", "canonId": "019de52d-1453-72ba-a367-9ba3d34c1630", "version": 1, "description": null }, "density": "dense", "densityScore": 66, "componentCount": 7, "textChars": 684, "nDataPoints": 15, "notes": "Includes a line chart of financial metrics and a timeline of semiconductor process nodes.", "elementsJson": [ "headline_text", "action_title", "paragraph", "line_chart", "timeline_visual", "comparison_table" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:54.555+00", "seconds": 7.291, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/156", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-156", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "TSMC's R&D expenses match SMIC's revenues.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-21c6608da817", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.05, "y": 0.38 }, "kind": "chart", "text": "Line chart showing SMIC revenues, TSMC R&D expenses, and other financial metrics from 2012-2020.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:54.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0c4fb3e4-5336-408e-b131-2ee8da07b3a0", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.45, "x": 0.52, "y": 0.38 }, "kind": "chart", "text": "Timeline of semiconductor manufacturing nodes by company and year.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": null, "framework": null }, "subkind": "timeline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:54.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4e254d0c-a474-4a72-81f9-e83593d86777", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "R&D expenditure: $6.7B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-244ff63e3d59", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.9, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "TSMC’s R&D expenses match SMIC’s revenues. TSMC is the only fabricator with 5nm manufacturing process (N5) and it is now working on 3nm (N3) for 2x more power efficiency and 33% more performance than N7. In response, SMIC said it will increase capital expenditure to $6.7B in 2020 (up from its original target of $3.1B).", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:54.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "14c96e93-d0b8-448c-8708-91cd47d8fb7b", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.45, "x": 0.52, "y": 0.65 }, "kind": "table", "text": "Performance comparison table for different process nodes.", "attrs": null, "canon": { "kind": { "name": "Table", "slug": "table", "status": "active", "canonId": "019de52c-f9c8-7148-8119-eaac9594cd28", "version": 1, "description": "Tabular data layout." }, "tool": null, "subkind": { "name": "Data table", "slug": "data", "status": "active", "canonId": "019de52d-02d5-75b8-8abf-88e92bf63569", "version": 1, "description": null }, "framework": null }, "subkind": "data", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:54.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1bad96b8-a51d-42e4-8804-f01d0dcd0986", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Taiwan’s TSMC remains dominant in R&D expenditure and semiconductor manufacturing", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:54.555+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0eec30d2-410c-4d81-b323-a994526eed34", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "c4f1453e-adcb-403a-997d-34baae6b469e", "evidence": "Performance comparison table for different process nodes.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.86527+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 156, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-59cc2faad8c3", "evidence": "TSMC R&D = SMIC revenue comparison.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 156, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 157, "type": "analysis", "title": "Includes a line chart comparing cumulative total returns of SMIC and TSM.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 45, "componentCount": 5, "textChars": 568, "nDataPoints": 6, "notes": "Includes a line chart comparing cumulative total returns of SMIC and TSM.", "elementsJson": [ "headline_text", "bullet_list", "line_chart", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.361+00", "seconds": 3.091, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/157", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-157", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.96, "x": 0.02, "y": 0.28 }, "kind": "callout", "text": "China-listed chipmakers see their public market valuations soar in 2020. Cambricon goes public raising $370M.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.361+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e5376613-26c0-47ef-b32f-10fc682bcdf7", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.58, "x": 0.39, "y": 0.36 }, "kind": "chart", "text": "Cumulative total return of SMIC vs TSM from Oct 2019 to Jul 2020.", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.361+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "18ca9cdb-7305-4351-9152-8e109eefc38b", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.38, "x": 0.02, "y": 0.36 }, "kind": "list", "text": "China is the world's largest importer of semiconductors, totalling $200B/year. New state fund is backed by the Ministry of Finance, China Development Bank, local government and state-owned enterprises. SMIC listed on the Shanghai STAR board.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.361+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "14f3d2ea-5ab8-4224-b195-eb60c6f84783", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Total Return (Cumulative): $29B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-f32283864ed7", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.96, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Chinese government sets up an additional $29B state-backed fund reduce its dependency on American semiconductor technology", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.361+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0fc38ae8-f356-4b0f-b0c5-bbd6f555439c", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "381985e4-6b71-4765-8a7c-2cd694a508ac", "evidence": "China is the world's largest importer of semiconductors, totalling $200B/year.", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.141172+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "cd44405a-6c79-48e3-8d70-44a2151ef64d", "evidence": "chart/line: Cumulative total return of SMIC vs TSM from Oct 2019 to Jul 2020.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.170822+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-5ee1dbf0bcb9", "evidence": "$29B Chinese fund / Cambricon $370M IPO.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 157, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 158, "type": "analysis", "title": "Slide from the State of AI 2020 report.", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 60, "componentCount": 6, "textChars": 885, "nDataPoints": 5, "notes": "Slide from the State of AI 2020 report.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.194+00", "seconds": 2.737, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/158", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-158", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "callout", "text": "TSMC employees are offered as much as 2.5x their annual salary and bonuses to leave. Overall, Taiwan has lost 3,000 semiconductor engineers in recent times (circa 10% of their national supply).", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": null, "subkind": { "name": "Primary callout", "slug": "primary", "status": "active", "canonId": "019de52c-fbd1-70eb-a9ce-8997f8674c1d", "version": 1, "description": null }, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.194+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b2ffe0eb-09e9-48a3-9e3d-c3c9643b1ed4", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "TSMC employees are offered as much as 2.5x their annual salary and bonuses to leave.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-7c9d62a23c6e", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.52, "y": 0.33 }, "kind": "image", "text": "TSMC facility sign", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.194+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7cf41ee3-0fe7-48d5-802a-b0c11e901e4e", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.48, "x": 0.02, "y": 0.33 }, "kind": "list", "text": "Government-backed Quanxin Integrated Circuit Manufacturing (QXIC, founded 2019) and Wuhan Hongxin Semiconductor Manufacturing Co. (HSMC, founded 2017) are are led by ex-TSMC executives and have each hired 50 former TSMC employees.\nQXIC recently began operating a R&D facility close to TSMC's 5-nanometer plant in south Taiwan.\nQXIC plans to build a $18.4B project to produce 14-nanometer chips by 2022, and also has a tech roadmap to develop even more advanced 7-nanometer chips.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.194+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a92feee2-ff56-48ad-a4e1-731397317eca", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "TSMC engineers hired: 100", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-814dd8a8bd4b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.95, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "China hires over 100 TSMC engineers in push to close gap in semiconductor capabilities", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.194+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3de4fffe-f1f0-403f-a66b-b09c0cdd9ae0", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "f8c9decc-a7ab-49ed-8371-8a11379bc982", "evidence": "TSMC employees are offered as much as 2.5x their annual salary and bonuses to leave.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.232369+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6355d4859f76", "evidence": "100 engineers / 2.5x salary.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "9c531313-1ad6-4576-918c-ba5cbd582087", "evidence": "list/bullet: Government-backed Quanxin Integrated Circuit Manufacturing (QXIC, founded 2019) and Wuhan Hongxin Semiconductor Manufacturing Co. (HSMC, founded 2017) are are led by ex-TSMC executives and have each hired 50", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.264253+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 158, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 159, "type": "analysis", "title": "The chart shows the number of 300mm semiconductor fabs by region from 2011 to 2019.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 775, "nDataPoints": 6, "notes": "The chart shows the number of 300mm semiconductor fabs by region from 2011 to 2019.", "elementsJson": [ "headline_text", "bullet_list", "line_chart", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.368+00", "seconds": 2.907, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/159", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-159", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Although over half the world's advanced chips are designed in America, only 12% are manufactured there.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-cd9464aa05bb", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.55, "x": 0.43, "y": 0.28 }, "kind": "chart", "text": "Chip production swings to Asia", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.368+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f149330c-e27b-4be1-8167-139a65a83bab", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.03, "y": 0.28 }, "kind": "list", "text": "The bipartisan bill seeks to boost US competitiveness. The CHIPS for America Act would earmark $22B to subsidise US manufacturing of chips. Programs include $10B of federal match funding, DoD funding for related funding and $12B in related R&D funding. The US has also asked Intel and Samsung to produce more US manufactured chips. The US has now sanctioned China's SMIC because exports posed an \"unacceptable risk\" of being diverted to \"military end use\".", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.368+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "de620e88-ba39-4d16-889b-25eabb9d4853", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of 300mm semiconductor fabs: 12%", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-d1c6fe19ea9b", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.9, "x": 0.03, "y": 0.22 }, "kind": "title", "text": "Although over half the world's advanced chips are designed in America, only 12% are manufactured there.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.368+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "59846d3f-3bdc-45d0-92c2-c22edbc153c2", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.4, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "US Senate proposes the CHIPS for America Act", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.368+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f2a29bd6-a8ab-45ac-8e09-44e1f79f3789", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "2x2 matrix", "slug": "matrix-2x2", "status": "active", "bestFor": null, "canonId": "019dd9e1-50ab-72d1-8afb-3ab0a38e32a8", "version": 1, "bodyDocId": null, "description": "BCG matrix, importance/urgency, any axes-and-quadrants frame.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "de38d332-3b30-4ad6-a43e-5f0fa93b57c4", "evidence": "chart/line: Chip production swings to Asia", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-16 21:56:55.356199+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-67a02a7c0f89", "evidence": "Callout contrasts US design vs manufacturing share.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 85, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Call to action", "slug": "call-to-action", "status": "active", "bestFor": null, "canonId": "019dd9e1-4bbb-702a-9ead-82daf188bb7c", "version": 1, "bodyDocId": null, "description": "Explicit prescription of what the audience should do next.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "f0ab2e40-e6f5-42ea-85bc-986b3d1352ca", "evidence": "The CHIPS for America Act would earmark $22B to subsidise US manufacturing of chips.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.325461+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 159, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 160, "type": "analysis", "title": "The slide uses a list format to present a chronological sequence of events, categorized by outcome (blocked, allowed, pending).", "function": "summarize", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 5, "textChars": 918, "nDataPoints": 6, "notes": "The slide uses a list format to present a chronological sequence of events, categorized by outcome (blocked, allowed, pending).", "elementsJson": [ "headline_text", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.567+00", "seconds": 3.092, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/160", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-160", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The vast majority of acquisitions have been blocked.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-db77d00126a6", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.95, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "December 2016: US and Germany block $723M bid by China’s Fujian Grand Chip Investment Fund (China) for Aixtron.\nSeptember 2017: US blocks $1.3B bid by Canyon Bridge Capital Partner (China) for Lattice.\nMarch 2018: US blocks $117B bid by Broadcom (previously headquartered in Singapore) for Qualcomm (USA).\nJuly 2018: China blocks Qualcomm’s $44B bid for NXP (Netherlands).\nApril 2020: UK and US effectively block a complete takeover of Imagination Technologies (UK) by Canyon Bridge (China).\nApril 2020: China allows Nvidia’s (USA) $6.9B acquisition of Mellanox (Israel).\nJuly 2020: Siemen’s (Germany) makes bid for Avatar (USA).", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.567+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6cc01bae-a8b1-4f85-a225-343f74cb149b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.02, "y": 0.82 }, "kind": "paragraph", "text": "The reported potential acquisition of Arm (UK) by Nvidia (USA) will be a major test of where things stand.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.567+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "929e7bf3-8b1f-4529-afa8-51a50361ea4f", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.45, "x": 0.02, "y": 0.22 }, "kind": "title", "text": "The vast majority of acquisitions have been blocked.", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Action title", "slug": "action-title", "status": "active", "canonId": "019de52c-fb31-734c-a0f5-a3935b73135d", "version": 1, "description": null }, "framework": null }, "subkind": "action-title", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.567+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "13e38f09-71fd-43c1-bdf2-f817b16228a7", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.85, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Given mounting concerns over chips, cross border M&A remains highly politicised", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.567+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "13dcf11b-62dc-4c3c-8c78-7c48de74cbc6", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "11cc0c2b-33a2-4ae7-91e3-117792d1ec40", "evidence": "list/bullet: December 2016: US and Germany block $723M bid by China’s Fujian Grand Chip Investment Fund (China) for Aixtron.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.417345+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 160, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 160, "from": 153, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-34d60d543d62", "evidence": "Sanctions, Huawei, TSMC dominance, Chinese fund, talent poaching, CHIPS Act, blocked M&A.", "position": 20, "objective": "Map US-China semiconductor decoupling and TSMC centrality", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 161, "type": "analysis", "title": "The slide highlights the trend of 'AI Nationalism' through specific regulatory changes in three major economies.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 5, "textChars": 1335, "nDataPoints": 6, "notes": "The slide highlights the trend of 'AI Nationalism' through specific regulatory changes in three major economies.", "elementsJson": [ "headline_text", "bullet_list", "paragraph", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:53.491+00", "seconds": 5.959, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/161", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-161", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The State of AI Report and AI Nationalism essay predicted that political leaders would start to question whether acquisitions of key AI startups should be blocked.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-60a6d20a3157", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.33, "x": 0.64, "y": 0.35 }, "kind": "image", "text": null, "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.491+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "265abfe4-a72a-466e-a861-b5efc65dfd05", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.6, "x": 0.03, "y": 0.33 }, "kind": "list", "text": "Germany passed a law in June 2020 to allow the government to review or block investments or takeovers by non-EU based companies of robotics, AI and semiconductors companies. The foreign ownership threshold at which government could review or veto was decreased from 25% to 10%. This was likely influenced by the EU's 2019 directive on screening foreign direct investment.\nJapan passed a law, effective in Aug 2019 that requires foreign investors to report to the Japanese government and undergo inspection in case they buy 10% or more of stocks in listed Japanese companies or acquire shares of unlisted firms.\nIn June 2020, the UK expanded its powers to intervene in mergers on public interest grounds under the Enterprise Act 2002. The Government is now able to scrutinise mergers and acquisitions involving AI companies where the target company has revenues of over £1M.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.491+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "af19a5d2-46bd-4555-9bb7-fd4f148c549e", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.9, "x": 0.03, "y": 0.22 }, "kind": "paragraph", "text": "The State of AI Report and AI Nationalism essay predicted that political leaders would start to question whether acquisitions of key AI startups should be blocked. New legislation suggests this is now happening.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.491+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f384b590-a481-4857-a802-4db39d519d87", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "AI Nationalism: Governments increasingly plan to scrutinise acquisitions of AI companies", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.491+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cfea5577-f3cf-4cd5-9786-18de4d7784d4", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience frame", "slug": "audience-frame", "status": "active", "bestFor": null, "canonId": "2a1eefbd-5be8-4600-898b-49dbfcf73ec8", "version": 1, "bodyDocId": null, "description": "Tools for modelling and segmenting the audience (persona, JTBD, segmentation).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "6557fe4a-294f-4cf9-a235-6f211d8346bc", "evidence": "The State of AI Report and AI Nationalism essay predicted that political leaders would start to question whether acquisitions of key AI startups should be blocked.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.480002+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 161, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 162, "type": "setup", "title": "The slide uses a standard consulting layout with a top navigation bar highlighting the 'Politics' section.", "function": "establish_context", "rawType": "situation_overview", "block": null, "metadata": { "slideType": "situation_overview", "slideTypeCanon": { "name": "Situation overview", "slug": "situation_overview", "status": "active", "canonId": "019de52d-0cfe-7769-99aa-9a2fe82a4039", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 86, "componentCount": 9, "textChars": 1721, "nDataPoints": 2, "notes": "The slide uses a standard consulting layout with a top navigation bar highlighting the 'Politics' section.", "elementsJson": [ "headline_text", "action_title", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.5-flash", "runId": null, "at": "2026-07-17 03:39:37.923+00", "seconds": 8.827, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/162", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-162", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "It will break up Arm into a US ARM and a Chinese Arm. There will be two types and fuel the trade war between America and China. It will be sad for a British company to be torn between them.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-eae06ef5ecec", "parentComponentId": null }, { "bbox": { "h": 0.38, "w": 0.26, "x": 0.7, "y": 0.3 }, "kind": "image", "text": "Arm Neoverse SoC DAWN Ares Platform", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2fe2dce4-34ab-449d-8d2b-b1901295927f", "parentComponentId": null }, { "bbox": { "h": 0.13, "w": 0.94, "x": 0.02, "y": 0.71 }, "kind": "list", "text": "Hauser's intervention is notable because he has been involved in Arm from inception (helping to spin it out from Acorn in 1990). Until recently European VCs have mostly been enthusiastic about their startups being acquired by US companies, rather than considering issues of technological sovereignty.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4687558e-4a92-43b7-9eb4-94d9cf6f44aa", "parentComponentId": null }, { "bbox": { "h": 0.28, "w": 0.67, "x": 0.02, "y": 0.42 }, "kind": "list", "text": "He argues that Arm would become subject to oversight by CFIUS which would allow the US to block specific companies from using Arm and that Arm would effectively become part of the US trade arsenal. He also notes that 'it will break up Arm into a US ARM and a Chinese Arm. There will be two types and fuel the trade war between America and China. It will be sad for a British company to be torn between them.'", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5762d080-8942-44e9-aeb7-4d7b1ca6f281", "parentComponentId": null }, { "bbox": { "h": 0.13, "w": 0.67, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "Hauser argues that it would lead to significant job losses in the UK, similar to when NVIDIA acquired Bristol's Icera in 2011. He also argues it would destroy Arm's neutrality and impact European licensees who compete with NVIDIA.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5dfe81b4-bdd4-475d-adef-b82d1e3d70f8", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.94, "x": 0.02, "y": 0.23 }, "kind": "list", "text": "Hermann Hauser, a leading founder and investor, argues it would be bad for the UK if Arm is acquired by NVIDIA.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a7784e12-4f8b-4ee2-9577-d6ab7f1cc367", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.94, "x": 0.02, "y": 0.85 }, "kind": "list", "text": "The UK's opposition party asked the government to assure that in any acquisition there were legally binding assurances that Arm would remain headquartered in the UK.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ee8500a4-b8f0-4546-9c19-90ad511b5305", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "It will break up Arm into a US ARM and a Chinese Arm. There will be two types and fuel the trade war between America and China. It will be sad for a British company to be torn between them. — Hermann Hauser", "attrs": null, "canon": { "kind": { "name": "Quote", "slug": "quote", "status": "active", "canonId": "019de52c-f8fb-751e-8bc6-d58e2d59562a", "version": 1, "description": "Pull quote or testimonial." }, "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-ee4bc419977d", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.82, "x": 0.02, "y": 0.15 }, "kind": "title", "text": "The likely sale of Arm to NVIDIA is questioned by many, including its founder", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-07-17 03:39:37.923+00", "model": "google/gemini-3.5-flash", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "52ac50ab-8de3-4586-a798-efe6ceb01b26", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Authority citation", "slug": "authority-citation", "status": "active", "bestFor": null, "canonId": "019dd9e1-49ae-7557-89ad-3fca70469b45", "version": 1, "bodyDocId": null, "description": "Quoting an authority figure, study, or precedent to lend weight.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ab7b6de1-4799-4c2a-8d71-091e43be013d", "evidence": "The slide quotes Hermann Hauser, a founder and investor, lending credibility to the argument.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.5, "extraction": { "at": "2026-07-17 03:53:51.398039+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6984a298eb94", "evidence": "Founder quote about Arm being 'torn'.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Problem-Agitate-Solve (PAS)", "slug": "problem-agitate-solve-pas", "status": "active", "bestFor": null, "canonId": "019dd956-869f-7564-8d42-de54ed29ab90", "version": 1, "bodyDocId": null, "description": "Identify problem, agitate the pain, present the solution", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "d6e1de4f-753a-4153-8f77-792b098deb60", "evidence": "The slide presents a problem (the likely sale of Arm to NVIDIA), agitation (job losses, destruction of Arm's neutrality), and an implied need for a solution.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-17 03:53:51.359706+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 162, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 163, "type": "analysis", "title": "The slide uses a bar chart to illustrate the growth in federal non-defense AI R&D spending.", "function": "quantify_impact", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "quantify_impact", "functionCanon": { "name": "Quantify impact", "slug": "quantify_impact", "status": "active", "canonId": "019de52d-1630-720d-b57d-fa7f76bebd8d", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 6, "textChars": 707, "nDataPoints": 5, "notes": "The slide uses a bar chart to illustrate the growth in federal non-defense AI R&D spending.", "elementsJson": [ "headline_text", "action_title", "bullet_list", "bar_chart_vertical" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.547+00", "seconds": 2.729, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/163", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-163", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "AI continues to be emphasized as the most important investment area in science and technology.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-b2ad2b4d1bdc", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.3, "x": 0.65, "y": 0.25 }, "kind": "chart", "text": "Federal budget for non-defense AI R&D", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Vertical bar chart", "slug": "bar-vertical", "status": "active", "canonId": "019de52c-fe9d-70f2-b3b7-a9ae294f5917", "version": 1, "description": null }, "framework": null }, "subkind": "bar-vertical", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.547+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "dcaa182f-5100-42ac-9965-4ff2c82375e9", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.55, "x": 0.05, "y": 0.28 }, "kind": "list", "text": "In February 2019, President Trump signed Executive Order 13859 “Maintaining American Leadership in Artificial Intelligence”. The proposed spend for 2021 is $1.5B. These non-military investments include the Departments of Agriculture, Energy and Health. The Department of Defense’s Joint Artificial Intelligence Center has continued to expand from a launch budget in 2019 of $93M to $238M for 2020.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.547+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4027a012-0c41-4615-b33b-e3dce1325261", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Federal AI R&D Budget: $1,503M", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-b5a642f144fc", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.22 }, "kind": "paragraph", "text": "AI continues to be emphasized as the most important investment area in science and technology.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.547+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0baf0246-1c51-463d-b628-7a77bc6383a6", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "AI Nationalism in the US: AI budgets continue to expand", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.547+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cf677228-ddbe-4a67-b4fc-3953aaaec4a5", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "682f4a92-a264-43bb-af95-412d91438cb8", "evidence": "chart/bar-vertical: Federal budget for non-defense AI R&D", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.542119+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 163, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-6f82c9802d77", "evidence": "$1,503M AI budget figure.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 163, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 164, "type": "analysis", "title": "Includes a photo of Chuck Schumer at a podium.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 80, "componentCount": 9, "textChars": 934, "nDataPoints": 2, "notes": "Includes a photo of Chuck Schumer at a podium.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.009+00", "seconds": 3.133, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/164", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-164", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The country that wins the race in key technologies—such as artificial intelligence, quantum computing, advanced communications, and advanced manufacturing—will be the superpower of the future.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-3dcfe74e5269", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.4, "x": 0.55, "y": 0.25 }, "kind": "image", "text": "Chuck Schumer at a podium", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a75670cf-d2ff-4c8e-8252-6c22b2dbd96f", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.5, "x": 0.02, "y": 0.28 }, "kind": "list", "text": "The bi-partisan act explicitly frames a race to lead in AI: “The country that wins the race in key technologies—such as artificial intelligence, quantum computing, advanced communications, and advanced manufacturing—will be the superpower of the future.”", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2332f249-9e4f-4f9d-a857-d46e1d10d024", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.66 }, "kind": "list", "text": "It is worth noting that many senate bills like this do not pass.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "390b4f00-99a3-4222-bf30-198c675e8311", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "The proposed 'Endless Frontier' act explicitly frames AI as a race between superpowers.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "82885643-3c97-4854-ab1e-3153551c8b05", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.5, "x": 0.02, "y": 0.48 }, "kind": "list", "text": "The act would create a Technology Directorate within the National Science Foundation and enable it to operate like DARPA with funding of $100B over 5 years.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cfb3239b-5b14-4518-bb20-1069da58f98d", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.02, "y": 0.6 }, "kind": "list", "text": "It would also provide $10B to establish a series of regional technology hubs.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d7d58bb4-e518-467f-b1b1-1c2f5001cc9e", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "funding: $100B", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:38+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa6f-42dcf8e4aa10", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "AI Nationalism in the US: A major new bi-partisan act is proposed", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "755c0c9f-e6bd-4023-b3ec-1b2914642101", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Concrete Language", "slug": "concrete-language", "status": "active", "bestFor": null, "canonId": "019dd956-bcd4-7398-9723-6aa72346170b", "version": 1, "bodyDocId": null, "description": "Using specific sensory details instead of abstract terms", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-75a5a2e2bd73", "evidence": "$100B bipartisan act.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-713a78db3bc1", "evidence": "Quote: 'superpower of the future' framing.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Monroe's Motivated Sequence", "slug": "monroes-motivated-sequence", "status": "active", "bestFor": null, "canonId": "019dd956-8758-700f-83cf-d260a54ad24f", "version": 1, "bodyDocId": null, "description": "Five-step persuasion: Attention, Need, Satisfaction, Visualization, Action", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "2c8ad468-6cdb-4ac6-ba08-961dfc30eb96", "evidence": "The bi-partisan act explicitly frames a race to lead in AI: “The country that wins the race in key technologies—such as artificial intelligence, quantum computing, advanced communications, and advanced manufacturing—will be the superpower of the future.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.602981+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 164, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 165, "type": "analysis", "title": "The slide highlights a specific government initiative in China to decentralize AI development.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "dense", "densityScore": 60, "componentCount": 6, "textChars": 929, "nDataPoints": 1, "notes": "The slide highlights a specific government initiative in China to decentralize AI development.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.636+00", "seconds": 2.534, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/165", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-165", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "China moves to create “national new generation AI innovation and development pilot zones”.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-0aaa7438aa97", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.4, "x": 0.55, "y": 0.25 }, "kind": "image", "text": "Aerial view of a modern development zone in China.", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Photo", "slug": "photo", "status": "active", "canonId": "019de52c-fd37-757f-b4a8-ca26d4888875", "version": 1, "description": null }, "framework": null }, "subkind": "photo", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "abc8d175-f1b7-4845-a5b2-584a393a5fd6", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.03, "y": 0.43 }, "kind": "list", "text": "This seems intended to enable more decentralised experimentation. Cities that become AI pilot zones are encouraged to accelerate the application of AI in a wide variety of fields, ranging from manufacturing to caring for elderly and disabled.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "251539d6-38de-4335-b697-cdf70bc9324e", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.03, "y": 0.22 }, "kind": "list", "text": "China moves to create “national new generation AI innovation and development pilot zones”. The PRC Ministry of Science and Technology created processes for cities to establish themselves as AI pilot zones. Twenty AI pilot zones are targeted by 2023.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b0c402d8-56d8-440a-8f03-8ece8f1dab0f", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.45, "x": 0.03, "y": 0.65 }, "kind": "list", "text": "AI pilot zones are encouraged to “carry out AI based policy experiments” and “carry out AI based social experiments”. Deqing County is cited as an example. The city is expected to focus on autonomous driving and smart farming.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fdbedb5b-e4ca-4328-87dd-3ce0183fb7d2", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.14 }, "kind": "title", "text": "AI Nationalism in China: Decentralising policy experimentation to cities", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.636+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "106be39f-c61a-4459-8beb-0b9799b9abaa", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Journey frame", "slug": "journey-frame", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ab2-7354-8c61-721dee94c0d1", "version": 1, "bodyDocId": null, "description": "Sequential stages over time (timeline, before-during-after).", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "04dd000e-aaa3-4519-beb4-01f8c36ead5b", "evidence": "AI pilot zones are encouraged to “carry out AI based policy experiments” and “carry out AI based social experiments”.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.663168+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 165, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 166, "type": "analysis", "title": "Includes a cultural reference (Wolf Warrior movie poster) to illustrate the geopolitical context.", "function": "establish_context", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "balanced", "densityScore": 30, "componentCount": 4, "textChars": 500, "nDataPoints": 0, "notes": "Includes a cultural reference (Wolf Warrior movie poster) to illustrate the geopolitical context.", "elementsJson": [ "headline_text", "bullet_list", "photo" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:50.275+00", "seconds": 1.99, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/166", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-166", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Pressure on the UK to choose between the US and China.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:39+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa70-98cd752f48a5", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.23, "x": 0.73, "y": 0.22 }, "kind": "image", "text": "Wolf Warrior movie poster", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Illustration", "slug": "illustration", "status": "active", "canonId": "019de52c-fe76-7772-a5ac-c0b7294d7615", "version": 1, "description": null }, "framework": null }, "subkind": "illustration", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.275+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7b7b098d-6c0f-466d-9578-f08fa5b63d60", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.65, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "Pressure on the UK to choose between the US and China.\nA new group of Conservative UK parliamentarians have formed the China Research Group (CRG) to scrutinise the UK's relationship with China with an emphasis on emerging technology.\nPressure from the US and the CRG led to the UK reversing its policy on Huawei.\nSenseTime is no longer setting up a UK HQ.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.275+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bedb895f-652d-40f1-9908-da1f425a2873", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "AI Nationalism in the UK: China hawks in the UK become more active", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:50.275+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5de7cc29-df30-4a3a-87fe-79583be5374e", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Emotional Appeal", "slug": "emotional-appeal", "status": "active", "bestFor": null, "canonId": "019dd956-bde9-7668-8be7-0097678fec59", "version": 1, "bodyDocId": null, "description": "Connecting to audience feelings rather than just logic", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "8fc7f342-184e-42ca-9a68-c13128bbcea0", "evidence": "Pressure on the UK to choose between the US and China.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.6, "extraction": { "at": "2026-07-16 21:56:55.727183+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 166, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 167, "type": "setup", "title": "The map uses color-coding for strategy status (complete vs forthcoming) and dot colors for the focus of public transformation.", "function": "establish_context", "rawType": "geographic_map", "block": null, "metadata": { "slideType": "geographic_map", "slideTypeCanon": { "name": "Geographic map", "slug": "geographic_map", "status": "active", "canonId": "019de52d-09de-7153-8845-584c1a9d0c83", "version": 1, "description": null }, "function": "establish_context", "functionCanon": { "name": "Establish context", "slug": "establish_context", "status": "active", "canonId": "019de52d-1314-74e6-9074-d213dc0bdcdf", "version": 1, "description": null }, "density": "sparse", "densityScore": 17, "componentCount": 4, "textChars": 205, "nDataPoints": 0, "notes": "The map uses color-coding for strategy status (complete vs forthcoming) and dot colors for the focus of public transformation.", "elementsJson": [ "map", "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.138+00", "seconds": 2.656, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/167", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-167", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.2, "x": 0.4, "y": 0.88 }, "kind": "disclaimer", "text": "See https://oe.cd/ai-strategies for details", "attrs": null, "canon": { "kind": { "name": "Disclaimer", "slug": "disclaimer", "status": "active", "canonId": "91173477-d91f-4024-bbaf-d4f93e3f82dc", "version": 1, "description": "Legal/cautionary text that occupies most of the body of its slide." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "29ed0a7c-03f2-4ae8-8560-eaf28185c7d8", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.8, "x": 0.1, "y": 0.25 }, "kind": "image", "text": "World map showing AI strategy status and focus", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Map", "slug": "map", "status": "active", "canonId": "019de52c-fe4d-77bd-b0a4-d72d58b9e5a4", "version": 1, "description": null }, "framework": null }, "subkind": "map", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d88c9986-6788-43e0-ad03-e261b15b5c33", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.2, "x": 0.15, "y": 0.65 }, "kind": "legend", "text": "AI strategy status and Focus on public transformation legend", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7fdeba61-6400-4060-bfb4-6c04595f5d78", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.14 }, "kind": "title", "text": "Another wave of countries declare national AI strategies", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.138+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0a3cfdac-0897-4734-9609-51f0f6048548", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Map / spatial", "slug": "map-spatial", "status": "active", "bestFor": null, "canonId": "019dd9e1-4f52-724d-8b4e-b5ed8925a0a0", "version": 1, "bodyDocId": null, "description": "Geographic or spatial visualisation.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "1b3d27cb-6e92-404a-9984-9c12b70df1ca", "evidence": "image/map: World map showing AI strategy status and focus", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:56:55.787202+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 167, "whyItWorks": null, "antipattern": null, "cardinality": "optional", "narrativePurpose": null }, { "tool": { "name": "Picture Superiority Effect", "slug": "picture-superiority-effect", "status": "active", "bestFor": null, "canonId": "019dd956-c201-7014-bb58-3e85e3e11dfd", "version": 1, "bodyDocId": null, "description": "Pictures remembered 6x better than words - prioritize visuals over text", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "019dd95a-0fd5-7148-8ed0-7bb067e530ec", "evidence": "World map shows national AI strategy adoption.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 70, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 167, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 167, "from": 161, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3b9613c8b29e", "evidence": "Arm/NVIDIA, US budget, bipartisan act, China pilot zones, UK hawks, geographic map.", "position": 21, "objective": "Document AI nationalism by country (US/China/UK/+)", "confidence": 70, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 168, "type": "analysis", "title": "The chart shows a widening gap between labor tax rates and equipment/software tax rates over time.", "function": "diagnose", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "diagnose", "functionCanon": { "name": "Diagnose", "slug": "diagnose", "status": "active", "canonId": "019de52d-1657-7223-acc5-7437d878c850", "version": 1, "description": null }, "density": "dense", "densityScore": 67, "componentCount": 7, "textChars": 845, "nDataPoints": 3, "notes": "The chart shows a widening gap between labor tax rates and equipment/software tax rates over time.", "elementsJson": [ "headline_text", "action_title", "paragraph", "bullet_list", "line_chart" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:56.841+00", "seconds": 8.141, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/168", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-168", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The authors argue that this incentivises levels of automation that are not socially desirable - displacing workers without achieving meaningful productivity gains.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:41+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa71-a167426c9880", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.55, "x": 0.4, "y": 0.35 }, "kind": "chart", "text": "Effective tax rates for U.S. companies, by type of expenditure", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Line chart", "slug": "line", "status": "active", "canonId": "019de52c-ff62-778f-8f6c-a952ceae8677", "version": 1, "description": null }, "framework": null }, "subkind": "line", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:56.841+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b7cc97f0-519d-449b-9747-08ddd6f8d9b7", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.45, "x": 0.05, "y": 0.35 }, "kind": "list", "text": "The authors argue that this incentivises levels of automation that are not socially desirable - displacing workers without achieving meaningful productivity gains.\nThey argue that reducing labour tax rates and introducing some kind of automation tax could meaningfully increase employment whilst continuing to increase productivity.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:56.841+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3d78bec3-f26d-44fd-ad4a-fe24605b79f4", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Effective tax rate", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:41+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa71-a7fe8133e1d7", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.45, "x": 0.05, "y": 0.21 }, "kind": "paragraph", "text": "Acemoglu, Manera and Restrepo's paper demonstrates that tax reforms from 2000 to 2017 have caused the gap between effective tax rates on labour and robots to dramatically widen.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:56.841+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a67bbcdc-30dd-4780-9e1f-1305a8397ea1", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.1, "x": 0.05, "y": 0.95 }, "kind": "source-note", "text": "BROOKINGS WSJ", "attrs": null, "canon": { "kind": { "name": "Source note", "slug": "source-note", "status": "active", "canonId": "23484952-494b-4a5e-b1ad-550d0d70e948", "version": 1, "description": "Attribution / source / caveat placed at the slide footer. Numbered footnotes use attrs.numbered." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:56.841+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "32945017-d6a1-44d3-857f-f24c1bd5ad10", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.8, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Evidence suggests that the US tax code incentivises replacing humans with robots", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:56.841+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "88ba8a88-eb34-4e10-8014-b7a36ebbefbf", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "989c4b53-2b33-4076-b2fe-f6369389467c", "evidence": "chart/line: Effective tax rates for U.S. companies, by type of expenditure", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.84925+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 168, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "So What? Test", "slug": "so-what-test", "status": "active", "bestFor": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "canonId": "019dd956-acd2-76f8-b59a-9bf15a89f10a", "version": 1, "bodyDocId": "fec04b06-b2f1-432b-a97b-916dc7563f64", "description": "The recursive editorial discipline of asking so-what of every slide, chart, and statement until the answer is an implication for the audience — that final answer becomes the slide's title; the chain becomes its body.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-7c6dd2d5227f", "evidence": "Callout: tax code drives socially undesirable automation.", "pageRefs": null, "priority": "Core", "whenToUse": "On every slide carrying a chart, number, list, or finding intended to inform a decision; whenever a draft title is descriptive rather than directive; whenever the deck will circulate without its author.", "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 168, "whyItWorks": "Solves the two failure modes of analytical communication at once: stopping at description (analyst-confidence problem) and leaving insight stranded in the analyst's head (translation problem). The recursion is what guarantees the implication makes it into ink.", "antipattern": "Stopping at the first tighter description (still a fact, not an implication), regressing past the audience's decision space into philosophy, smuggling the so-what into speaker notes, or re-using an implication aimed at a previous audience.", "cardinality": null, "narrativePurpose": "Forces every analytical artefact to close the gap between fact and decision, so the reader holds the recommendation by reading titles alone rather than reconstructing the analyst's mental model from raw evidence." } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 170, "from": 168, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3d1ca750c538", "evidence": "Tax-code automation incentive, jobs at risk, climate change call to action.", "position": 22, "objective": "Surface labor and climate implications of AI", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 169, "type": "analysis", "title": "The slide uses a series of stacked bar charts to visualize survey data from 1,872 enterprises regarding AI-induced workforce contraction.", "function": "analyze_data", "rawType": "industry_trends", "block": null, "metadata": { "slideType": "industry_trends", "slideTypeCanon": { "name": "Industry trends", "slug": "industry_trends", "status": "active", "canonId": "019de52d-0b1d-7705-a0e0-c663199cd1fa", "version": 1, "description": null }, "function": "analyze_data", "functionCanon": { "name": "Analyze data", "slug": "analyze_data", "status": "active", "canonId": "019de52d-1590-7413-8c82-cc218cff40b4", "version": 1, "description": null }, "density": "dense", "densityScore": 58, "componentCount": 6, "textChars": 621, "nDataPoints": 144, "notes": "The slide uses a series of stacked bar charts to visualize survey data from 1,872 enterprises regarding AI-induced workforce contraction.", "elementsJson": [ "headline_text", "paragraph", "bar_chart_stacked" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:53.838+00", "seconds": 5.134, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/169", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-169", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Executives from 1,872 enterprises worldwide report the largest AI-induced workforce contraction in automotive and assembly and telecoms in the last year.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:45+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c6-7671-886c-18f5d717d56a", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.95, "x": 0.02, "y": 0.38 }, "kind": "chart", "text": "Change in workforce due to AI adoption, % of respondents", "attrs": null, "canon": { "kind": { "name": "Chart", "slug": "chart", "status": "active", "canonId": "019de52c-f9a0-76ef-adc6-0b859109d1f8", "version": 1, "description": "Quantitative chart." }, "tool": null, "subkind": { "name": "Stacked bar chart", "slug": "bar-stacked", "status": "active", "canonId": "019de52c-feec-74bf-9028-15b74f84f0a7", "version": 1, "description": null }, "framework": null }, "subkind": "bar-stacked", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.838+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "76585b5a-0247-47e7-a7ea-9328fa0df99b", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.35, "x": 0.02, "y": 0.41 }, "kind": "legend", "text": "Decrease >10% | Decrease 3-10% | Increase 3-10% | Increase >10%", "attrs": null, "canon": { "kind": { "name": "Legend", "slug": "legend", "status": "active", "canonId": "31361933-3d22-4c70-bbad-0801e9fffc0c", "version": 1, "description": "Color/symbol decoder placed adjacent to a chart or diagram." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.838+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "88aab99e-1b9e-44f2-90a4-caf666d49098", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Change in workforce due to AI adoption: 1,872", "attrs": null, "canon": { "kind": { "name": "Metric", "slug": "metric", "status": "active", "canonId": "019de52c-f923-73bf-8081-d0b1a8c5264d", "version": 1, "description": "Big-number or KPI value." }, "tool": { "name": "Quantification", "slug": "quantification", "status": "active", "bestFor": null, "canonId": "019dd9e1-47ab-72ce-a079-aab1f5299e88", "version": 1, "bodyDocId": null, "description": "Translating a claim into numbers, charts, or tables.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": "primary", "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:45+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c6-7671-886c-1fb0a50f8c55", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.95, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "Executives from 1,872 enterprises worldwide report the largest AI-induced workforce contraction in automotive and assembly and telecoms in the last year. Looking forward, the CPG, transport, utilities, retail and financial services are expected to follow.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.838+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "93d701a6-ad79-41b0-ae3e-9dec00611715", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.4, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Jobs at risk of automation in the EU 19 countries", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.838+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3947b0c6-5d85-4f5e-9e01-22e92e24bc11", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chart Selection Guide", "slug": "chart-selection-guide", "status": "active", "bestFor": null, "canonId": "019dd956-b1a0-700b-ba17-0e16c4d5bcae", "version": 1, "bodyDocId": null, "description": "Choosing the right chart type for your data and message", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Designer", "layer": "slide", "agents": [ "Designer" ], "matchId": "ed95d14d-c371-45d9-8577-924a40e68b48", "evidence": "chart/bar-stacked: Change in workforce due to AI adoption, % of respondents", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:56:55.911199+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 169, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 170, "from": 168, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3d1ca750c538", "evidence": "Tax-code automation incentive, jobs at risk, climate change call to action.", "position": 22, "objective": "Surface labor and climate implications of AI", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 170, "type": "evidence", "title": "The slide highlights a specific academic/industry collaboration (Tackling Climate Change with Machine Learning).", "function": "illustrate_case", "rawType": "case_study", "block": null, "metadata": { "slideType": "case_study", "slideTypeCanon": { "name": "Case study", "slug": "case_study", "status": "active", "canonId": "019de52d-0b95-7344-8090-926f3b8e92c8", "version": 1, "description": null }, "function": "illustrate_case", "functionCanon": { "name": "Illustrate case", "slug": "illustrate_case", "status": "active", "canonId": "019de52d-13b4-772c-bb73-1873612a7463", "version": 1, "description": null }, "density": "dense", "densityScore": 53, "componentCount": 6, "textChars": 775, "nDataPoints": 0, "notes": "The slide highlights a specific academic/industry collaboration (Tackling Climate Change with Machine Learning).", "elementsJson": [ "headline_text", "bullet_list", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:52.267+00", "seconds": 3.487, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/170", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-170", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "ML is part of the solution: it is a tool that enables other tools across fields.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:40+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa71-587a928697e2", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.46, "x": 0.47, "y": 0.41 }, "kind": "disclaimer", "text": "List of authors and their affiliations.", "attrs": null, "canon": { "kind": { "name": "Disclaimer", "slug": "disclaimer", "status": "active", "canonId": "91173477-d91f-4024-bbaf-d4f93e3f82dc", "version": 1, "description": "Legal/cautionary text that occupies most of the body of its slide." }, "tool": null, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.267+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ed6bea51-24b0-4c98-9442-b5fd1199ec9b", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.45, "x": 0.05, "y": 0.33 }, "kind": "list", "text": "Automatic monitoring with remote sensing (e.g. deforestation, climate disasters).\nScientific discovery (e.g. new battery materials, carbon capture).\nOptimize systems (e.g. reducing food waste, consolidating freight).\nAccelerate physical simulations (e.g. climate models and energy scheduling).\nThe authors note that “ML is part of the solution: it is a tool that enables other tools across fields.”", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Bullet list", "slug": "bullet", "status": "active", "canonId": "019de52c-fc98-7169-a083-5cab805076ca", "version": 1, "description": null }, "framework": null }, "subkind": "bullet", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.267+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1a0cd685-2db0-474f-8e45-af22b37a4978", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.4, "x": 0.5, "y": 0.37 }, "kind": "paragraph", "text": "Tackling Climate Change with Machine Learning", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.267+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b676ede6-6023-49e0-b174-e6da5dc79f06", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.05, "y": 0.26 }, "kind": "paragraph", "text": "A position paper and workshop explored various high leverage problems where ML methods can be applied.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.267+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ff557adf-e868-40a9-90cb-09e1fb2a1742", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.05, "y": 0.13 }, "kind": "title", "text": "Bengio, Hassabis, Ng and other AI research leaders unite at NeurIPS 2019 in a call to action for climate change", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.267+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bbf9fef9-3243-4af3-96f1-fb5451cda0c2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Audience Definition", "slug": "audience-definition", "status": "active", "bestFor": null, "canonId": "019dd956-7f2a-775d-aeee-ff099bcb7756", "version": 1, "bodyDocId": null, "description": "Six key questions: Who are they? What do they know? What do they believe? What do they care about? What do they fear? What decisions can they make?", "familyLabel": null, "categoryName": "Block", "categorySlug": "block" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "fce72d03-ed96-4fab-8b98-ece35d56c923", "evidence": "Bengio, Hassabis, Ng and other AI research leaders unite at NeurIPS 2019 in a call to action for climate change", "pageRefs": null, "priority": "Core", "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-16 21:57:06.57547+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Authority Bias", "slug": "authority-bias", "status": "active", "bestFor": null, "canonId": "019dd956-c38d-70b9-a284-b2b479763356", "version": 1, "bodyDocId": null, "description": "Expert opinions are weighted more heavily - cite credible sources", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Architect", "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-80c4a54edd76", "evidence": "Bengio, Hassabis, Ng named as endorsers.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "List presentation", "slug": "list-presentation", "status": "active", "bestFor": null, "canonId": "019dd9e1-4ef7-777d-ba68-9074c2b3bcf1", "version": 1, "bodyDocId": null, "description": "Bulleted, numbered, or checklist enumerations.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "fd2b35c1-d26a-441d-9958-6b3d2316c791", "evidence": "list/bullet: Automatic monitoring with remote sensing (e.g. deforestation, climate disasters).", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.535538+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 170, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 170, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 130, "name": "The Implications (So What)", "beatId": "019dd95a-0682-776c-8e34-f2fc34acb214", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Implications (So What)", "slug": "triple-take-the-implications-so-what", "status": "active", "canonId": "019dd9b8-09c7-75cd-8c7b-ae0e23cca2b8", "version": 1, "description": null }, "evidence": "Politics section unpacks ethical, geopolitical and labor consequences.", "position": 2, "isPrimaryArc": true, "parentBeatType": { "name": "Reflection", "slug": "reflection", "status": "active", "canonId": "019dd9b8-0510-73fd-8d00-1c7e45f7bc98", "version": 1, "description": "The \"so what\" beat. Used by analytical arcs (Triple-take, Onion) and learning arcs (Voyage-return, Tragedy)." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 170, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 130, "name": "Climax", "beatId": "019dd95a-0682-776c-8e35-01ef1d0aee03", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Climax", "slug": "mountain-climax", "status": "active", "canonId": "019dd9b8-1a2a-762e-b86f-73c710f9aa68", "version": 1, "description": null }, "evidence": "Politics section dramatises facial recognition, military, US/China conflicts.", "position": 3, "isPrimaryArc": false, "parentBeatType": { "name": "Turn", "slug": "turn", "status": "active", "canonId": "019dd9b8-0396-702b-ac2f-9901f8dd8b54", "version": 1, "description": "The pivot — solution proposed, catalyst introduced, decision made. Common (15/20 arcs) but not universal: Triple-take, AIDA, PAS, Voyage-return and Pixar substitute it with a strong reflection or development beat." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [ { "to": 170, "from": 168, "loop": { "name": "02_pattern_hunter", "slug": "02-pattern-hunter", "status": "active", "bestFor": "Time-pressed audiences, building consensus, when data is strong", "canonId": "019dd956-6695-771a-be2e-f1748ef041a4", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion", "description": "Group multiple pieces of evidence that together point to a pattern or conclusion", "familyLabel": null }, "matchId": "019dd95a-07fe-70ce-8d3d-3d1ca750c538", "evidence": "Tax-code automation incentive, jobs at risk, climate change call to action.", "position": 22, "objective": "Surface labor and climate implications of AI", "confidence": 60, "extraction": { "at": "2026-04-29 13:07:29.547081+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null } } ], "locked": true }, { "page": 171, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 13, "componentCount": 4, "textChars": 115, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.195+00", "seconds": 2.174, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/171", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-171", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.08, "x": 0.9, "y": 0.93 }, "kind": "paragraph", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.195+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "49b58b98-5814-4374-96f5-bdf793c1b762", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.45, "x": 0.02, "y": 0.03 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.195+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cfb5f316-95d3-4715-b1be-45af108f5268", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.06, "x": 0.92, "y": 0.03 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.195+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d8a63ddc-ce4f-416e-b790-f1916f1f356f", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.31, "x": 0.345, "y": 0.45 }, "kind": "title", "text": "Section 5: Predictions", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.195+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ce7e9d6b-83a5-4f11-85fd-c5baeb277624", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ff11b82e-9d3a-4dfd-bfa3-bc2a3ecb076d", "evidence": "Section 5: Predictions", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.655403+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 171, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 174, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 171, "name": "The Action (Now What)", "beatId": "019dd95a-0682-776c-8e34-f7c6d2e63d67", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Action (Now What)", "slug": "triple-take-the-action-now-what", "status": "active", "canonId": "019dd9b8-0a5e-735a-bf53-3a1ba629b7aa", "version": 1, "description": null }, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "isPrimaryArc": true, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 172, "type": "setup", "title": "The slide uses a numbered list format to present forward-looking statements.", "function": "summarize", "rawType": "initiative_list", "block": null, "metadata": { "slideType": "initiative_list", "slideTypeCanon": { "name": "Initiative list", "slug": "initiative_list", "status": "active", "canonId": "019de52d-0caf-770e-9b91-b99239121f50", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 2, "textChars": 901, "nDataPoints": 2, "notes": "The slide uses a numbered list format to present forward-looking statements.", "elementsJson": [ "numbered_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:53.91+00", "seconds": 4.885, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/172", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-172", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.65, "w": 0.95, "x": 0.02, "y": 0.22 }, "kind": "list", "text": "1. The race to build larger language models continues and we see the first 10 trillion parameter model.\n2. Attention-based neural networks move from NLP to computer vision in achieving state of the art results.\n3. A major corporate AI lab shuts down as its parent company changes strategy.\n4. In response to US DoD activity and investment in US based military AI startups, a wave of Chinese and European defense-focused AI startups collectively raise over $100M in the next 12 months.\n5. One of the leading AI-first drug discovery startups (e.g. Recursion, Exscientia) either IPOs or is acquired for over $1B.\n6. DeepMind makes a major breakthrough in structural biology and drug discovery beyond AlphaFold.\n7. Facebook makes a major breakthrough in augmented and virtual reality with 3D computer vision.\n8. NVIDIA does not end up completing its acquisition of Arm.", "attrs": null, "canon": { "kind": { "name": "List", "slug": "list", "status": "active", "canonId": "019de52c-f94b-779b-a53b-fd4904a5f4d2", "version": 1, "description": "Bullet or numbered list." }, "tool": null, "subkind": { "name": "Numbered list", "slug": "numbered", "status": "active", "canonId": "019de52c-fcc0-7718-a284-a4250c805df2", "version": 1, "description": null }, "framework": null }, "subkind": "numbered", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "58fab814-eb3d-49ab-ac57-fce8138d4b86", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.4, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "8 predictions for the next 12 months", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:53.91+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "92658285-c78e-416e-9df3-bd8d3cf1558f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Chunking", "slug": "chunking", "status": "active", "bestFor": null, "canonId": "019dd956-abab-72ce-b64c-2924cc83cf39", "version": 1, "bodyDocId": null, "description": "Information grouped into chunks is remembered better - structure content in digestible units", "familyLabel": null, "categoryName": "Loop", "categorySlug": "loop" }, "agent": "Architect", "layer": "loop", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-88131b332432", "evidence": "Eight numbered predictions chunked clearly.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Action Titles", "slug": "action-titles", "status": "active", "bestFor": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "canonId": "019dd956-ac70-700b-a608-ba58e78f5dcc", "version": 1, "bodyDocId": "816dad3c-9983-41c1-a86e-978ebd582841", "description": "The McKinsey-bred discipline of writing every slide title as a complete declarative sentence with a verb and an insight, not a topic label. Each title is a sub-claim that ladders up to the deck's governing thought; read in sequence, the titles reconstruct the executive summary.", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Architect", "layer": "slide", "agents": [ "Architect" ], "matchId": "019dd95a-0fd5-7148-8ed0-84d1b923990a", "evidence": "Eight crisp predictive statements.", "pageRefs": null, "priority": "Core", "whenToUse": "Any decision-grade deliverable — strategy recommendation, board update, investment memo, M&A review, post-mortem — where the reader is being asked to agree, decide, or act and will skim the title bar at speed.", "confidence": 80, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": "Solves two failure modes at once. (1) The buried claim: action titles place the conclusion on the page before any analysis is read, so a busy reader doesn't reverse-engineer it from the chart. (2) The spineless deck: because each title is a claim, slides have to ladder up to the governing thought; logical gaps become visible as topic-shaped titles in a sequence that no longer reads as a story.", "antipattern": "Topic labels disguised as titles — Volume by quarter, Key findings, Pricing strategy, Margins have been impacted. No verb, or a passive verb with no agent, or a fact (Revenue grew 12%) without the so-what. Also: titles that overreach the chart's evidence, multi-claim X-grew-but-Y-fell welded titles, and mechanical repetition of the same X-drove-Y template across the deck.", "cardinality": null, "narrativePurpose": "Forces the deck to carry its argument in the title bar so a senior reader can extract the recommendation without opening a single slide; converts a binder of topics into a navigable pyramid where every slide is a node defending the apex." }, { "tool": { "name": "Closing Techniques", "slug": "closing-techniques", "status": "active", "bestFor": null, "canonId": "019dd956-af13-774b-a4fd-7db0a3a88d51", "version": 1, "bodyDocId": null, "description": "Strong endings: Callback, challenge, vision, clear CTA, memorable line", "familyLabel": null, "categoryName": "Slide", "categorySlug": "slide" }, "agent": "Storyteller", "layer": "slide", "agents": [ "Storyteller" ], "matchId": "019dd95a-0fd5-7148-8ed0-8d1e7f203895", "evidence": "Predictions act as forward-looking close.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 75, "extraction": { "at": "2026-04-29 13:07:31.544884+00", "model": "claude-legacy", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 172, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 174, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 171, "name": "The Action (Now What)", "beatId": "019dd95a-0682-776c-8e34-f7c6d2e63d67", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Action (Now What)", "slug": "triple-take-the-action-now-what", "status": "active", "canonId": "019dd9b8-0a5e-735a-bf53-3a1ba629b7aa", "version": 1, "description": null }, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "isPrimaryArc": true, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 173, "type": "setup", "function": "transition", "rawType": "section_divider", "block": null, "metadata": { "slideType": "section_divider", "slideTypeCanon": { "name": "Section divider", "slug": "section_divider", "status": "active", "canonId": "019de52d-066a-7402-9554-6ae504999041", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "sparse", "densityScore": 13, "componentCount": 4, "textChars": 114, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.326+00", "seconds": 2.105, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/173", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-173", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.025, "w": 0.084, "x": 0.901, "y": 0.925 }, "kind": "paragraph", "text": "stateof.ai 2020", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.326+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2f4b889d-046c-49b5-92e7-451fd62866d9", "parentComponentId": null }, { "bbox": { "h": 0.025, "w": 0.485, "x": 0.015, "y": 0.035 }, "kind": "paragraph", "text": "Introduction | Research | Talent | Industry | Politics | Predictions", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.326+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "48b62941-32a6-4836-84f5-20f3fb6fa11d", "parentComponentId": null }, { "bbox": { "h": 0.025, "w": 0.06, "x": 0.925, "y": 0.035 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.326+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "957c91d7-3b81-4d7c-9baf-ee5c7ad6b698", "parentComponentId": null }, { "bbox": { "h": 0.035, "w": 0.306, "x": 0.347, "y": 0.458 }, "kind": "title", "text": "Section 6: Conclusion", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.326+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0becc7c2-0d03-4362-a15f-d74862c86059", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Section divider", "slug": "section-divider", "status": "active", "bestFor": null, "canonId": "019dd9e1-4c66-707d-8d1f-6e14547134c7", "version": 1, "bodyDocId": null, "description": "A transitional slide marking the start of a new section.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b259a1e8-8559-4221-bb46-635ef97d16bd", "evidence": "Section 6: Conclusion", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.773252+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 173, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 174, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 171, "name": "The Action (Now What)", "beatId": "019dd95a-0682-776c-8e34-f7c6d2e63d67", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Action (Now What)", "slug": "triple-take-the-action-now-what", "status": "active", "canonId": "019dd9b8-0a5e-735a-bf53-3a1ba629b7aa", "version": 1, "description": null }, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "isPrimaryArc": true, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 174, "type": "setup", "function": "summarize", "rawType": "closing_ask", "block": null, "metadata": { "slideType": "closing_ask", "slideTypeCanon": { "name": "Closing ask", "slug": "closing_ask", "status": "active", "canonId": "019de52d-0aa4-750f-8aa1-8d622dd99f5c", "version": 1, "description": null }, "function": "summarize", "functionCanon": { "name": "Summarize", "slug": "summarize", "status": "active", "canonId": "019de52d-1541-704f-b86f-bf7b75056373", "version": 1, "description": null }, "density": "balanced", "densityScore": 43, "componentCount": 4, "textChars": 1041, "nDataPoints": 0, "notes": null, "elementsJson": [ "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:51.973+00", "seconds": 2.555, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/174", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-174", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "We would appreciate any and all feedback on how we could improve this Report further, as well as contribution suggestions for next year's edition.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:41+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa71-8ef6d0066c04", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.35, "x": 0.02, "y": 0.88 }, "kind": "paragraph", "text": "Nathan Benaich (@nathanbenaich) and Ian Hogarth (@soundboy)", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.973+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "abf53630-c5f6-48ef-a316-2046e7b2444a", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.96, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "Congratulations on making it to the end of the State of AI Report 2020! Thanks for reading. In this report, we set out to capture a snapshot of the exponential progress in the field of machine learning, with a focus on developments since last year's issue that was published on 26th June 2019. We believe that AI will be a force multiplier on technological progress in our world, and that wider understanding of the field is critical if we are to navigate such a huge transition. We set out to compile a snapshot of all the things that caught our attention in the last year across the range of AI research, talent, industry and the emerging politics of AI. We would appreciate any and all feedback on how we could improve this Report further, as well as contribution suggestions for next year's edition. Thanks again for reading!", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.973+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e0e47578-d88e-4693-b4e7-2a3014e1ae81", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.08, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Thanks!", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:51.973+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5392191b-6ce9-4298-94fb-898359d71bf2", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Storytelling: the closing ask", "slug": "closing-ask", "status": "active", "bestFor": null, "canonId": "019df269-24e0-70ab-8dd9-983832c1c665", "version": 1, "bodyDocId": "019df269-24bc-7554-b88d-245127606b8d", "description": "The single most important slide in the deck — after the cover. The final", "familyLabel": "narrative-block", "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "agent": "storyteller", "layer": "slide", "agents": [ "storyteller" ], "matchId": "34e3f689-2033-4073-8f24-64eda4f5ce16", "evidence": "callout: We would appreciate any and all feedback on how we could improve this Report further, as well as contribution suggestions for next year's edition.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:57:06.851971+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 174, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 174, "arc": { "name": "The Triple Take", "slug": "triple-take", "status": "active", "bestFor": "Data-driven presentations, research findings", "canonId": "019dd956-6221-73ad-bc12-378f8eb1f018", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "The Facts (What) -> The Implications (So What) -> The Action (Now What)", "description": null, "familyLabel": null }, "from": 171, "name": "The Action (Now What)", "beatId": "019dd95a-0682-776c-8e34-f7c6d2e63d67", "matchId": "019dd95a-0542-7728-a78e-fb9f51e5adef", "beatType": { "name": "The Action (Now What)", "slug": "triple-take-the-action-now-what", "status": "active", "canonId": "019dd9b8-0a5e-735a-bf53-3a1ba629b7aa", "version": 1, "description": null }, "evidence": "Predictions section + thanks slide call out next 12 months and feedback ask.", "position": 3, "isPrimaryArc": true, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 60 }, { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 175, "type": "setup", "function": "other", "rawType": "appendix_disclosure", "block": null, "metadata": { "slideType": "appendix_disclosure", "slideTypeCanon": { "name": "Appendix disclosure", "slug": "appendix_disclosure", "status": "active", "canonId": "019de52d-07b2-7549-8e8b-932a19f6e616", "version": 1, "description": null }, "function": "other", "functionCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-15e1-75bc-99dc-07732380e2ce", "version": 1, "description": null }, "density": "balanced", "densityScore": 42, "componentCount": 5, "textChars": 658, "nDataPoints": 0, "notes": null, "elementsJson": [ "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:52.234+00", "seconds": 2.796, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/175", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-175", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The authors declare a number of conflicts of interest as a result of being investors and/or advisors, personally or via funds, in a number of private and public companies whose work is cited in this report.", "attrs": null, "canon": { "kind": { "name": "Callout", "slug": "callout", "status": "active", "canonId": "019de52c-f8d3-761a-8953-7d37080de5e2", "version": 1, "description": "Highlighted callout box." }, "tool": { "name": "Visual emphasis", "slug": "visual-emphasis", "status": "active", "bestFor": null, "canonId": "019dd9e1-474b-736b-a8eb-a794dd88241c", "version": 1, "bodyDocId": null, "description": "Drawing the eye to one element above the rest of the slide.", "familyLabel": null }, "subkind": null, "framework": null }, "subkind": null, "groupRole": null, "confidence": null, "extraction": { "at": "2026-04-25 12:05:41+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd952-47c5-73ac-aa71-7502274ef7a7", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.21 }, "kind": "paragraph", "text": "The authors declare a number of conflicts of interest as a result of being investors and/or advisors, personally or via funds, in a number of private and public companies whose work is cited in this report.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.234+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "308649bf-6eb6-467c-bae8-7833c38cf550", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.96, "x": 0.02, "y": 0.46 }, "kind": "paragraph", "text": "Nathan and Air Street Capital are shareholders of: Graphcore, LabGenius, Niantic, ONI, PolyAI, Secondmind, Tractable, and ZOE.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.234+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8a7f3274-cb62-4108-938b-f95bba508805", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.83, "x": 0.02, "y": 0.36 }, "kind": "paragraph", "text": "Ian is an angel investor in: Chorus.ai, ComplyAdvantage, Disperse, Faculty, LabGenius, and PostEra.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.234+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ae245209-001a-469b-b259-84cff1a80bba", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.23, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "Conflicts of interest", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.234+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10fffcb9-57bc-4030-ba99-01e26db7c280", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 176, "type": "setup", "function": "other", "rawType": "team_bio", "block": null, "metadata": { "slideType": "team_bio", "slideTypeCanon": { "name": "Team bio", "slug": "team_bio", "status": "active", "canonId": "019de52d-0a7d-7394-ae5b-037478e71f1e", "version": 1, "description": null }, "function": "other", "functionCanon": { "name": "Other (fallback)", "slug": "other", "status": "active", "canonId": "019de52d-15e1-75bc-99dc-07732380e2ce", "version": 1, "description": null }, "density": "balanced", "densityScore": 35, "componentCount": 5, "textChars": 365, "nDataPoints": 2, "notes": null, "elementsJson": [ "headshot", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:52.347+00", "seconds": 2.835, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/176", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-176", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.3, "w": 0.16, "x": 0.635, "y": 0.218 }, "kind": "image", "text": "Ian Hogarth photo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.347+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1b7c81c6-fa56-4f90-952f-e864d1ceb50a", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.16, "x": 0.208, "y": 0.218 }, "kind": "image", "text": "Nathan Benaich photo", "attrs": null, "canon": { "kind": { "name": "Image", "slug": "image", "status": "active", "canonId": "019de52c-f978-701f-a720-51d029d769e7", "version": 1, "description": "Photo, illustration, screenshot or icon." }, "tool": null, "subkind": { "name": "Headshot", "slug": "headshot", "status": "active", "canonId": "019de52c-fd5f-765c-9282-38b54f61bbbf", "version": 1, "description": null }, "framework": null }, "subkind": "headshot", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.347+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ec19c25c-d663-4e6a-9f3a-116ec897cc73", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.35, "x": 0.52, "y": 0.63 }, "kind": "paragraph", "text": "Ian Hogarth: Ian is an angel investor in 60+ startups. He is a Visiting Professor at UCL... He is the Chair of Phasecraft, a quantum software company.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.347+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2aa62dd2-0f6d-4bb1-9721-254bfe5fb496", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.35, "x": 0.09, "y": 0.63 }, "kind": "paragraph", "text": "Nathan Benaich: Nathan is the general partner of Air Street Capital... He studied biology at Williams College and earned a PhD from Cambridge in cancer research.", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.347+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5407aa15-1acd-4a01-a4b0-779f0eebfbcb", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.2, "x": 0.02, "y": 0.14 }, "kind": "title", "text": "About the authors", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.347+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a2f9b324-4f44-4e0c-a76b-138f846c6841", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true }, { "page": 177, "type": "setup", "function": "front_matter", "rawType": "cover", "block": null, "metadata": { "slideType": "cover", "slideTypeCanon": { "name": "Cover", "slug": "cover", "status": "active", "canonId": "019de52d-05f2-729d-9f93-e0504f0a2410", "version": 1, "description": null }, "function": "front_matter", "functionCanon": { "name": "Front matter", "slug": "front_matter", "status": "active", "canonId": "019de52d-133d-73f0-9e26-201eda96b098", "version": 1, "description": null }, "density": "balanced", "densityScore": 26, "componentCount": 6, "textChars": 78, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 17:45:52.098+00", "seconds": 2.521, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5e88-73ef-bd5c-f812573a947a/177", "deck": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a", "deckAnchor": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a#slide-177", "deckJson": "/decks/019dd923-5e88-73ef-bd5c-f812573a947a.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.1, "x": 0.85, "y": 0.05 }, "kind": "paragraph", "text": "#stateofai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "279b5386-d9ab-4597-a856-7e8c31e40881", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.05, "y": 0.9 }, "kind": "paragraph", "text": "Nathan Benaich", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "32fdf4a8-01f3-4e92-bbee-43235622a3d6", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.05, "y": 0.05 }, "kind": "paragraph", "text": "stateof.ai", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "66b5ae43-216d-4aa0-aaa1-236fe2d7a1ee", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.1, "x": 0.85, "y": 0.9 }, "kind": "paragraph", "text": "Ian Hogarth", "attrs": null, "canon": { "kind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "5f2affc0-941d-4d07-959d-4bfc1930cea4", "version": 1, "description": "Single prose block in the body of a slide." }, "tool": null, "subkind": { "name": "Paragraph", "slug": "paragraph", "status": "active", "canonId": "019de52c-fb80-75af-aa5c-a7a4bfac2433", "version": 1, "description": null }, "framework": null }, "subkind": "paragraph", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a746a3da-75bc-4627-a0a9-39bd57183c6f", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.5, "x": 0.25, "y": 0.35 }, "kind": "title", "text": "State of AI Report", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Headline", "slug": "headline", "status": "active", "canonId": "019de52c-fae1-7463-9a7c-d60668eced0d", "version": 1, "description": null }, "framework": null }, "subkind": "headline", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7978c36c-cddc-4d7c-b96f-02c0bf986a2b", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.3, "x": 0.35, "y": 0.5 }, "kind": "title", "text": "October 1, 2020", "attrs": null, "canon": { "kind": { "name": "Title", "slug": "title", "status": "active", "canonId": "019de52c-f7dd-76d9-b848-b1f12bd85cf4", "version": 1, "description": "Slide title or headline." }, "tool": null, "subkind": { "name": "Subtitle", "slug": "subtitle", "status": "active", "canonId": "019de52c-fb09-739d-900c-815b0d9ac526", "version": 1, "description": null }, "framework": null }, "subkind": "subtitle", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 17:45:52.098+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4403b1de-8335-4124-b075-4ca22313c8e2", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 177, "arc": { "name": "The Mountain", "slug": "mountain", "status": "active", "bestFor": "Case studies, success stories, dramatic reveals", "canonId": "019dd956-62dc-71fd-9964-679ce4773319", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "Setup & Stakes -> Rising Action -> Climax -> Resolution", "description": null, "familyLabel": null }, "from": 171, "name": "Resolution", "beatId": "019dd95a-0682-776c-8e35-06b1307be1cc", "matchId": "019dd95a-0542-7728-a78e-fef7cbfaa809", "beatType": { "name": "Resolution", "slug": "mountain-resolution", "status": "active", "canonId": "019dd9b8-1ac2-777f-9a8f-ac543fee86f7", "version": 1, "description": null }, "evidence": "Predictions, conclusion and thanks close the report.", "position": 4, "isPrimaryArc": false, "parentBeatType": { "name": "Resolution", "slug": "resolution", "status": "active", "canonId": "019dd9b8-045d-703b-8dd7-e92aba3ba91b", "version": 1, "description": "How the deck ends. Required in 19/20 arcs. The exception (Sparkline) ends with new bliss which is functionally a resolution." }, "alignedBlockIds": null, "matchConfidence": 40 } ], "loops": [], "locked": true } ], "arcBeats": [ { "from": 10, "to": 129, "label": "The Facts (What)", "description": "Research, Talent, Industry sections inventory what happened in AI in 2020." }, { "from": 130, "to": 170, "label": "The Implications (So What)", "description": "Politics section unpacks ethical, geopolitical and labor consequences." }, { "from": 171, "to": 174, "label": "The Action (Now What)", "description": "Predictions section + thanks slide call out next 12 months and feedback ask." } ], "loops": [ { "from": 11, "to": 14, "label": "Tale Two Worlds", "description": "Compare PyTorch vs TensorFlow trajectories" }, { "from": 15, "to": 23, "label": "Pattern Hunter", "description": "Establish pattern that bigger models + more compute drive performance at rising cost" }, { "from": 24, "to": 29, "label": "Pattern Hunter", "description": "Show transformers unlocking new NLP/code/vision use cases" }, { "from": 30, "to": 46, "label": "Pattern Hunter", "description": "Survey AI breakthroughs across biology, drug discovery and COVID" }, { "from": 47, "to": 57, "label": "Pattern Hunter", "description": "Catalog computer vision and generative model advances" }, { "from": 58, "to": 62, "label": "Pattern Hunter", "description": "Map emerging research themes: federated learning, GPs, quantum" }, { "from": 64, "to": 69, "label": "Tale Two Worlds", "description": "Contrast academia vs industry pull on AI talent" }, { "from": 70, "to": 76, "label": "Pattern Hunter", "description": "Map global flows of AI talent into the US" }, { "from": 77, "to": 81, "label": "Benchmark Gap", "description": "Benchmark AI talent supply vs demand gap" }, { "from": 83, "to": 89, "label": "Pattern Hunter", "description": "Show pharma validating AI-first drug discovery" }, { "from": 90, "to": 92, "label": "Maturity Curve", "description": "Track AI medical regulation maturing toward reimbursement" }, { "from": 93, "to": 106, "label": "Iceberg", "description": "Diagnose why self-driving progress is slower than headlines suggest" }, { "from": 107, "to": 110, "label": "David Goliath", "description": "Position Graphcore against NVIDIA/Google in AI compute" }, { "from": 111, "to": 115, "label": "Pattern Hunter", "description": "Show enterprise shift from R&D to MLOps and adoption metrics" }, { "from": 116, "to": 128, "label": "Pattern Hunter", "description": "Parade real-world AI deployments across verticals" }, { "from": 131, "to": 140, "label": "Ripple Effect", "description": "Trace ripple of facial recognition harms from arrests to lawsuits to legislation" }, { "from": 141, "to": 143, "label": "Pattern Hunter", "description": "Show regulators waking up to deepfakes, ADM and model bias" }, { "from": 144, "to": 149, "label": "Pattern Hunter", "description": "Show AI research bleeding into military applications" }, { "from": 150, "to": 152, "label": "Pattern Hunter", "description": "Survey emerging responsible AI codes and corporate efforts" }, { "from": 153, "to": 160, "label": "Pattern Hunter", "description": "Map US-China semiconductor decoupling and TSMC centrality" }, { "from": 161, "to": 167, "label": "Pattern Hunter", "description": "Document AI nationalism by country (US/China/UK/+)" }, { "from": 168, "to": 170, "label": "Pattern Hunter", "description": "Surface labor and climate implications of AI" } ] }