{ "schemaVersion": "deck-intelligence.v1", "id": "019dd923-5de0-76bd-a168-6159b949e2c2", "slug": "ff5d549e5027972a", "title": "GPT-3 and the actuarial landscape", "subtitle": "Oliver Wyman", "author": "Oliver Wyman", "pageCount": 46, "kind": "consulting-deck", "orientation": "landscape", "aspectRatio": 1.778, "document": { "id": "019dd923-5de0-76bd-a168-6159b949e2c2", "slug": "ff5d549e5027972a", "title": "GPT-3 and the actuarial landscape", "rawTitle": "GPT-3 and the actuarial landscape", "authorId": "OliverWyman", "authorName": "Oliver Wyman", "authorDisplay": "Oliver Wyman", "kind": { "slug": "consulting-deck", "label": "Consulting deck" }, "sourceType": { "slug": "strategy_consulting", "label": "Strategy consulting" }, "fileName": "PI-9_GPT-3_and_the_Actuarial_Landscape.pdf", "sourceUrl": "https://www.casact.org/sites/default/files/2023-04/PI-9_GPT-3_and_the_Actuarial_Landscape.pdf", "sourceDate": "2023-03-15 00:00:00", "presentationDate": null, "ingestedAt": "2026-04-24 07:54:33+00", "status": "ready", "pageCount": 46, "orientation": "landscape", "aspectRatio": 1.778, "targetCompany": null, "targetTicker": null, "metadata": { "local_root": "/tmp/legacy-images/ff5d549e5027972a", "origin_corpus": "consulting" } }, "coverage": { "pageCount": 46, "pagesIndexed": 46, "pagesWithImages": 46, "pagesWithMetadata": 46, "imageCoveragePct": 100, "metadataCoveragePct": 100, "componentCount": 215, "chartCount": 8, "metricSlides": 6, "toolMatches": 20, "documentToolMatches": 0, "frameworkMatches": 9, "arcMatches": 1, "beatMatches": 4, "loopMatches": 3, "patternMatches": 0, "storymakersMatch": false, "scoreAvailable": true, "reviewAvailable": true, "activistThesisAvailable": false, "pitchdeckMetadataAvailable": false, "pitchdeckProfileAvailable": false }, "blocks": [], "matches": { "arcs": [ { "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null, "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "evidence": "The deck follows a sparkline narrative arc, starting with an introduction to NLP and GPT-3, then exploring its applications in the actuarial field, and finally discussing the potential for innovation and efficiency gains.", "isPrimary": true, "confidence": 0.8, "extraction": { "at": "2026-07-16 21:13:57.518364+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "beats": [ { "to": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null, "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 }, { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null, "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 }, { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null, "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 }, { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null, "categoryName": "Narrative & Transformation", "categorySlug": "narrative-transformation" }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null, "categoryName": "Vision", "categorySlug": "vision" }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } }, { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null, "categoryName": "Consulting Framework", "categorySlug": "consulting-framework" }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } }, { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null, "categoryName": "Consulting Framework", "categorySlug": "consulting-framework" }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "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": "7bad7c1e-b23c-4a12-9c41-3cba4bfe92b4", "evidence": "This slide is a cover slide with a title, subtitle, and logo.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.460822+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": "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": "c21d97a6-e100-4801-b696-091ce7ab5b21", "evidence": "This slide serves as a section divider with a clear title.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.647747+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 5, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "54364cf1-7f0e-459a-b426-d35c0f081849", "evidence": "The slide includes a table with data on model parameters, dataset size, and compute.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:12.758566+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": "54a791e6-cd3c-43fb-bf22-f8db918faf60", "evidence": "This slide serves as another section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.83357+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": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b962e2f0-303b-4ffd-9526-d349dd5206e5", "evidence": "The slide includes a table with data on model structure, cost function, and formulas for β coefficients.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:12.943561+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": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b1b04942-ed59-4484-971b-e188ec8f7035", "evidence": "The slide includes a table with data on model structure, cost function, and formulas for β coefficients.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.019465+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "256d7a7d-040b-4045-a3cd-ebf95508ccf0", "evidence": "list/bullet: FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11; Intercept β0 = (Σ(y) – β1 * ΣX) / N = 5.15", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.903883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 12, "whyItWorks": null, "antipattern": null, "cardinality": "common", "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": "1e467481-c211-429a-b52f-baf034baf485", "evidence": "chart/scatter: Scatter plot showing data points and a linear regression line with equation Y = 5.15 * X + 10.11 and Cost = 10.7476", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-17 03:18:13.866836+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 12, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "34995123-1480-4322-86d0-a39bdafc6215", "evidence": "table/data: MODEL STRUCTURE: Y = β0 + β1 * X; COST FUNCTION: Cost = Σ(predicted – actual)^2 = 10.7476; FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11, Intercept β0 =", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.976822+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "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": "c15f7299-65e2-4f99-9845-bbcbff197126", "evidence": "table/data: Comparison matrix of Linear Regression, Multivariate Linear Regression, and Logistic Regression across Model Structure, Cost Function, Prediction range, and Examples.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:14.203697+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ebee8e72-944d-46f3-91f6-b29da64231de", "evidence": "table/data: Multi-Label Classification Training Data Example", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:14.315695+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": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "57f02c71-2623-466e-916c-1da08c36bc8b", "evidence": "table/data: Token | Token ID | One-Hot Encoded Vector (50,000 dimensions) | Word Embedding Vector (fewer dimensions)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.635024+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 24, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "e414b55d-69f6-40fe-9e80-db774518598c", "evidence": "table/data: Token | Word Embedding | Positional Encoding | Resulting vectors", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.712812+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null }, { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b9dcd898-138d-46cd-8c19-53e067709c38", "evidence": "table/data: Table showing Preceding Tokens, Key, Query, Value, Unnormalized Weights, Normalized Weights, and New Representation of Peru.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.830907+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "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": "742de825-4d99-439c-bab2-df5f9c084e90", "evidence": "The slide contains a list of bullet points describing various insurance processes that can be improved with GPT.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.430366+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 37, "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": "b841dd85-8031-426d-86c8-723e98dfe11a", "evidence": "The slide contains a list of bullet points describing limitations of CHATGPT and GENERAL LIMITATIONS OF LLMS.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.507405+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 38, "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": "570c0a7e-650c-46de-8b07-546c0c0b5b44", "evidence": "The slide is titled '4 AI: SOFTWARE 2.0' and has a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.646187+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 39, "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": "74601f38-4dc3-4c84-b0a2-76c2f6f07293", "evidence": "The slide contains a list of bullet points describing best practices for software development.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.758102+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "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": "6488c285-416d-4605-9649-4a371a3eabe5", "evidence": "The slide is titled '5 RECAP' and has a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.832513+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 42, "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": "9633a882-7206-4807-b084-84a73b747035", "evidence": "The slide is titled 'QUESTIONS' and appears to be a call to action for questions.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.945107+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 44, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [ { "matchId": "77dd797f-b84f-4ccf-a242-7450cf78361c", "evidence": "The slide uses a vertical timeline to organize historical developments in NLP.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 6, "frameworkName": "timeline" }, { "matchId": "609c641c-4d13-406b-800a-458f34a452be", "evidence": "Comparison of statistical modeling techniques", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "frameworkName": "analytical-method" }, { "matchId": "cdcc384a-3615-407a-85ff-deaac36a84e8", "evidence": "Diagram of layers and nodes", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:13.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "frameworkName": "Neural Network Architecture" }, { "matchId": "1a2fca73-7c91-4ef9-b512-093b62a109a0", "evidence": "The diagram explicitly maps the transformation of input tokens into Key, Query, and Value vectors via matrix multiplication.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:17.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "frameworkName": "Transformer Attention Mechanism" }, { "matchId": "dbe2bd7c-8449-4fae-b722-9d5ba2d2048a", "evidence": "The slide explicitly describes the calculation of keys, queries, and values in self-attention.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "frameworkName": "Self-Attention Mechanism" }, { "matchId": "15706509-49fb-4f41-ab06-c77e0b95db21", "evidence": "Visual representation of the Transformer model components (Attention, Feed Forward, Linear, Softmax)", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:18.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "frameworkName": "Transformer Architecture" }, { "matchId": "1334b54e-a4dc-4b92-99be-84fcb1c0f34e", "evidence": "The slide compares four distinct methods of model adaptation based on demonstration count.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "frameworkName": "learning-paradigm-comparison" }, { "matchId": "ec38b1fa-3616-4085-bbaf-531d2e0dd27a", "evidence": "4-step sequential process", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "frameworkName": "process" }, { "matchId": "5721622c-9046-40e6-b139-91a7f9f3b1ec", "evidence": "The slide explicitly lists six core software engineering principles applied to actuarial work.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "frameworkName": "Software Development Lifecycle (SDLC) Best Practices" } ], "patterns": [] }, "storymakers": null, "score": { "backend": "codex", "scoredAt": "2026-05-02 18:13:46.267+00", "subScores": { "scqa_arc": 45, "action_titles": 48, "mece_structure": 59, "closing_strength": 47, "evidence_quality": 67, "clarity_of_thesis": 44, "production_quality": 70, "visual_storytelling": 73 }, "totalScore": 57, "coveragePct": 100, "explanations": { "scqa_arc": "The deck has Situation in slides 6-7 and a late Answer in slides 40-43, but the middle largely reads as a technical walkthrough rather than a tight Situation-Complication-Question-Answer arc.", "action_titles": "Headline quality is mixed: several slides are declarative, but many are topic labels, for example slide 9, \"MACHINE LEARNING 101: GRADIENT DESCENT\".", "mece_structure": "The five-part agenda is logical, but the long ML/GPT primer overlaps across slides 9-32 and leaves adoption, governance, and implementation implications underdeveloped.", "closing_strength": "Slides 41-43 provide a recap and broad best-practice direction, but the deck closes with Q&A and branding rather than a concrete recommendation, owner, timeline, or next step.", "evidence_quality": "Technical claims are often supported by tables, diagrams, screenshots, and examples, but strategic claims about actuarial leadership and transformation are mostly asserted on slides 40-43.", "clarity_of_thesis": "The first five slides establish topic and agenda, but no declarative thesis appears until much later when slides 40-43 suggest actuaries should lead Software 2.0 work.", "production_quality": "Branding, section dividers, visual consistency, and footnote discipline are solid, but repeated topic titles and dense primer slides reduce Storymakers polish.", "visual_storytelling": "The deck uses appropriate tools such as timelines, process diagrams, comparison tables, screenshots, grids, and a quote slide, though several dense technical slides dilute the message." }, "slidesAnalyzed": 46 }, "review": { "backend": null, "verdict": "A competent educational tutorial on GPT-3 with a strong but late-arriving actuarial thesis — useful as a teaching artifact, weak as a Storymakers exemplar because the recommendation is buried at p.40 of 46 and nine consecutive slides share one topic title.", "reviewedAt": "2026-04-25 05:54:15+00", "slidesSeen": 46, "suggestions": [ "Lead with a one-slide answer-first thesis after the cover ('Actuaries should lead the Software 2.0 transition — here is what GPT-3 changes and where to start') so the technical build serves the argument rather than precedes it", "Compress the gradient-descent explainer (p.9-17) to 2-3 slides with progressive action titles ('Models learn by minimizing error → gradient descent finds coefficients → the model now remembers the data') instead of nine slides sharing one label", "Replace 'LIMITATIONS' (p.38), 'NEURAL NETWORKS' (p.21), and 'WHERE TO START?' (p.41) with declarative titles that state the punchline (e.g. 'Hallucination and IP risk make GPT unsafe for unsupervised actuarial work today')", "Convert the closing 'QUESTIONS' slide into an explicit call-to-action with 2-3 next steps for the audience (pilot use cases, skills to build, governance to set up) before opening Q&A" ], "closingScore": 58, "openingScore": 38, "topStrengths": [ "Clean section-divider spine with five numbered pillars — easy to navigate and table-of-contents friendly", "Several insight-bearing action titles in the technical core, e.g. p.32 explicitly explains the GPT→ChatGPT transition rather than labeling it", "Strong rhetorical pivot slide on p.40 ('SOFTWARE IS EATING THE WORLD, AI IS EATING SOFTWARE') with a clear thesis callout that actuaries should lead", "Callout sentences (p.17, p.23, p.25, p.28, p.43) carry the argument explicitly instead of relying on visuals alone" ], "topWeaknesses": [ "Same title 'MACHINE LEARNING 101: GRADIENT DESCENT' repeated across nine consecutive slides (p.9-17) — zero progressive disclosure of insight in the titles", "No SCQA opening: thesis ('actuaries should lead Software 2.0') is buried until p.40 of 46, so the deck reads as exposition with a recommendation tacked on", "Closing devolves into 'QUESTIONS' (p.44) → disclaimer → cover slide; no concrete next steps, owners, or commitments", "Sections 1-2 (intro + building blocks) consume 27 of 46 pages — analytically heavy, narratively thin, and disproportionate to the 'so what' acts" ], "narrativeScore": 52, "pillarCritique": "Five numbered dividers (Intro to NLP / Building Blocks / Actuarial Applications / AI: Software 2.0 / Recap) give the deck a visible spine, but they are a teaching outline, not MECE strategic pillars — Section 4 'Software 2.0' is a thesis act sitting awkwardly alongside three explanatory acts.", "closingCritique": "p.43 RECAP delivers a real takeaway ('Actuaries are well suited to lead this work but need to modernize their skillset'), which is the strongest moment of the deck — but it is followed by a generic 'QUESTIONS' slide, a legal disclaimer, and a re-cover, so there is no concrete next-step or call-to-action.", "openingCritique": "The first five slides are cover/confidentiality/speakers/overview/section-divider — no hook, no thesis, no stakes for actuaries. The reader has to wait until p.40 ('AI is eating software... actuaries are ideally suited to lead this work') to learn why the deck exists.", "extractionSeconds": 75.199844, "narrativeCritique": "Reads as an educational tutorial (NLP intro → GPT internals → applications → Software 2.0 → recap) rather than a SCQA arc; the Complication and Answer ('actuaries should lead this work') only surface at slides 40-43, after ~25 pages of pure exposition. Setup and analysis are present, but tension is missing and the resolution is stapled on.", "titleQualityScore": 55, "titleQualityCritique": "Strong action titles do appear ('ACCESS TO POWERFUL RESOURCES ENABLE LARGE LANGUAGE MODELS' p.7, 'ADDING REINFORCEMENT LEARNING LAYERS AND A MODERATION API ENABLES THE TRANSITION FROM GPT TO CHATGPT' p.32, 'GPT-ENABLED TOOLS CAN HELP ACTUARIES EXECUTE THEIR WORK' p.34-36), but they are diluted by nine consecutive 'MACHINE LEARNING 101: GRADIENT DESCENT' slides (p.9-17) and pure-topic labels like 'LIMITATIONS' (p.38), 'NEURAL NETWORKS' (p.21), 'WHERE TO START?' (p.41)." }, "activistThesis": null, "pitchdeck": { "metadata": null, "profile": null }, "slides": [ { "page": 1, "type": "setup", "function": "front_matter", "imagePath": "https://imgproxy.kitesheet.com/uBbwUCNWKJzqxzHIJXa_giTngI2rEZStRI51H3a9_jA/rs:fit:1200:1200:0/q:82/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9mZjVkNTQ5ZTUwMjc5NzJhL3AwMDEuanBn", "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": "sparse", "densityScore": 21, "componentCount": 5, "textChars": 151, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "subtitle_text" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:09.974+00", "seconds": 2.088, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/1", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-1", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.03, "w": 0.15, "x": 0.05, "y": 0.92 }, "kind": "disclaimer", "text": "A business of Marsh McLennan", "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 18:38:09.974+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "21efa5dd-dc39-48c4-83e9-3112fb9c08df", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.2, "x": 0.05, "y": 0.05 }, "kind": "image", "text": "OliverWyman", "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 18:38:09.974+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "99ac87b3-0197-4455-b9ea-5ebbabb30d0a", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.2, "x": 0.05, "y": 0.82 }, "kind": "paragraph", "text": "CAS RPM Seminar\nMarch 2023", "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 18:38:09.974+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f0586130-caac-4e5b-a18c-955c953e5974", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.7, "x": 0.05, "y": 0.45 }, "kind": "title", "text": "GPT-3 AND THE ACTUARIAL LANDSCAPE", "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 18:38:09.974+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9d4bba49-0f03-41f6-ad82-f1e88acda94e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.05, "y": 0.7 }, "kind": "title", "text": "An Overview of Large Language Models and Applications", "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 18:38:09.974+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f8b1684a-7095-4aeb-802e-0049d26cee6e", "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": "7bad7c1e-b23c-4a12-9c41-3cba4bfe92b4", "evidence": "This slide is a cover slide with a title, subtitle, and logo.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.460822+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": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 } ], "loops": [], "thumbSrc": "https://imgproxy.kitesheet.com/39Cfd-a4tJ7UC5xDvJVW57HlG6I3Fxl4P8evj3LYBr0/rs:fit:480:480:0/q:78/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9mZjVkNTQ5ZTUwMjc5NzJhL3AwMDEuanBn", "imagePathAlt": "https://imgproxy.kitesheet.com/00HFoSdzoPk-mfJ_Bs6IhVpwOskMf2OLCQJ3WdMTshM/rs:fit:1200:1200:0/q:82/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9mZjVkNTQ5ZTUwMjc5NzJhL3AwMDEucG5n", "thumbSrcAlt": "https://imgproxy.kitesheet.com/Vqne87KUZ7P_B5g6zJPxLNtlEPY5hHQecFGngbl1XUc/rs:fit:480:480:0/q:78/f:webp/czM6Ly9raXRlc2hlZXQvY29ycHVzL3NsaWRlcy9mZjVkNTQ5ZTUwMjc5NzJhL3AwMDEucG5n" }, { "page": 2, "type": "setup", "function": "other", "rawType": "disclaimer", "block": null, "metadata": { "slideType": "disclaimer", "slideTypeCanon": { "name": "Disclaimer", "slug": "disclaimer", "status": "active", "canonId": "019de52d-0801-727b-b98a-5ed095aeee78", "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": 31, "componentCount": 3, "textChars": 656, "nDataPoints": 0, "notes": null, "elementsJson": [ "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:26.457+00", "seconds": 18.545, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/2", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-2", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.138, "w": 0.791, "x": 0.037, "y": 0.441 }, "kind": "disclaimer", "text": "Our clients' industries are extremely competitive, and the maintenance of confidentiality with respect to our clients' plans and data is critical. Oliver Wyman rigorously applies internal confidentiality practices to protect the confidentiality of all client information. Similarly, our industry is very competitive. We view our approaches and insights as proprietary and therefore look to our clients to protect our interests in our proposals, presentations, methodologies, and analytical techniques. Under no circumstances should this material be shared with any third party without the prior written consent of Oliver Wyman.", "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 18:38:26.457+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "14cd7894-bfa8-4a36-90e7-c3d8962c0348", "parentComponentId": null }, { "bbox": { "h": 0.018, "w": 0.085, "x": 0.037, "y": 0.598 }, "kind": "paragraph", "text": "© Oliver Wyman", "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 18:38:26.457+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "62c75029-1233-4069-9d84-47ed5e7476f3", "parentComponentId": null }, { "bbox": { "h": 0.033, "w": 0.204, "x": 0.037, "y": 0.383 }, "kind": "title", "text": "CONFIDENTIALITY", "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 18:38:26.457+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5761fedb-62c5-4425-8d56-0001a08bffcf", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 3, "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": "dense", "densityScore": 51, "componentCount": 7, "textChars": 463, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "headshot", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:11.277+00", "seconds": 3.364, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/3", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-3", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.25, "w": 0.15, "x": 0.73, "y": 0.21 }, "kind": "image", "text": "Sabrina Tan", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1169f61b-c84b-4217-ac08-c1245622bb9e", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.15, "x": 0.06, "y": 0.21 }, "kind": "image", "text": "Olivier Brown", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3f58714f-bef5-45c9-9e92-c2d8818fe441", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.15, "x": 0.4, "y": 0.21 }, "kind": "image", "text": "Hugo Latendresse", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f9c38572-0799-4cc2-bc5d-ad23a382d8dd", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.28, "x": 0.69, "y": 0.45 }, "kind": "paragraph", "text": "Sabrina Tan, ACAS. P&C Insurance Practice: Consultant. Sabrina Tan is a Consultant in Oliver Wyman’s Actuarial practice.", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4a100dbe-3878-43d6-8bd3-558d7312ccec", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.28, "x": 0.36, "y": 0.45 }, "kind": "paragraph", "text": "Hugo Latendresse, FCAS. P&C Insurance Practice: Senior Manager. Hugo Latendresse is a Senior Manager in Oliver Wyman’s Actuarial practice.", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a81111e7-edbf-488f-87fe-0ae027d72382", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.28, "x": 0.037, "y": 0.45 }, "kind": "paragraph", "text": "Olivier Brown, FCAS, MAAA. P&C Insurance Practice: Principal. Olivier Brown leads P&C Actuarial Innovation within Oliver Wyman’s Actuarial practice.", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f7b9b97d-0fab-483b-a990-9643081aa6bd", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.25, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "MEET THE SPEAKERS", "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 18:38:11.277+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ccbd089-4bde-4159-a8a4-b48110732cdc", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 4, "type": "setup", "function": "summarize", "rawType": "agenda", "block": null, "metadata": { "slideType": "agenda", "slideTypeCanon": { "name": "Agenda", "slug": "agenda", "status": "active", "canonId": "019de52d-061b-7470-b138-3ffa5e612cce", "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": 6, "componentCount": 2, "textChars": 132, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "numbered_list", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:10.009+00", "seconds": 1.889, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/4", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-4", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.6, "w": 0.6, "x": 0.35, "y": 0.25 }, "kind": "list", "text": "1. Intro to Natural Language Processing\n2. The Building Blocks of GPT\n3. Actuarial Applications\n4. AI: Software 2.0\n5. Recap", "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 18:38:10.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "72892ad5-e1fd-4839-84a4-42114c3b2421", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.12, "x": 0.358, "y": 0.05 }, "kind": "title", "text": "OVERVIEW", "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 18:38:10.009+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7285eda7-bb6e-4a1f-b566-75498f9d1080", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 5, "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": 2, "textChars": 38, "nDataPoints": 0, "notes": null, "elementsJson": [ "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:10.587+00", "seconds": 2.161, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/5", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-5", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 1, "w": 0.25, "x": 0.75, "y": 0 }, "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 18:38:10.587+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3386f628-165d-4d9d-aa92-063f481bccde", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.6, "x": 0.05, "y": 0.35 }, "kind": "title", "text": "1 INTRO TO NATURAL LANGUAGE PROCESSING", "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 18:38:10.587+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c9008693-f18b-4263-a1f0-c2e3735b7df2", "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": "c21d97a6-e100-4801-b696-091ce7ab5b21", "evidence": "This slide serves as a section divider with a clear title.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.647747+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 5, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 5, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 1, "name": "What Is", "beatId": "bfc10720-c292-42d4-a612-64530096e5a4", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-1", "status": "active", "canonId": "019dd9b8-25d5-7468-8649-454886d75343", "version": 1, "description": null }, "evidence": "The deck introduces GPT-3 and NLP, providing context and background information.", "position": 0, "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": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 6, "type": "setup", "function": "present_framework", "rawType": "timeline", "block": null, "metadata": { "slideType": "timeline", "slideTypeCanon": { "name": "Timeline", "slug": "timeline", "status": "active", "canonId": "019de52d-0851-75eb-b758-ea2965e7739a", "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": 64, "componentCount": 7, "textChars": 928, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "bullet_list", "timeline_visual", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:11.302+00", "seconds": 2.852, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/6", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-6", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.2, "x": 0.4, "y": 0.2 }, "kind": "diagram", "text": null, "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "925eb9f4-dfea-4ce2-a314-66cff35f8c9d", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.35, "x": 0.6, "y": 0.4 }, "kind": "list", "text": "Use of neural networks for language modeling: Initial uses of neural networks for next word prediction; First representations of words with dense vectors called word embeddings.", "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0aeb5d20-190f-4f61-9a8a-3b7388abdfaa", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.4, "x": 0.03, "y": 0.17 }, "kind": "list", "text": "Rule-based systems to simple statistical models: First application of NLP was for machine translation; Initial rule-based models required significant manual coding; Machine learning and statistical models replaced hard-coded rules.", "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "53332d91-7911-4827-baf7-46bcf0db0347", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.35, "x": 0.6, "y": 0.75 }, "kind": "list", "text": "Generative pre-trained transformers (GPT): OpenAI releases first version of GPT language model (2018); ChatGPT, fined-tuned on GPT-3.5, launched 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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5e8da071-d3f8-4e01-b66d-906f45ab6708", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.4, "x": 0.03, "y": 0.58 }, "kind": "list", "text": "Attention, transformers, and large language models: The attention mechanism along with transformer architecture enable state-of-the-art performance; Ability to consider context in texts increased the ability to produce human-like texts.", "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f4679484-341f-434e-afc5-727516076cdb", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.4, "x": 0.03, "y": 0.92 }, "kind": "source-note", "text": "Source: https://medium.com/nlplanet/a-brief-timeline-of-nlp-bc45b640f07d", "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e70dbcfd-5e2d-4b26-b0ae-ed896f3e5f85", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.75, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "RECENT KEY INNOVATIONS HAVE ACCELERATED ADVANCEMENTS IN NLP", "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 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5b9e951e-311a-479b-a945-30852cc128b9", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "77dd797f-b84f-4ccf-a242-7450cf78361c", "evidence": "The slide uses a vertical timeline to organize historical developments in NLP.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:11.302+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 6, "frameworkName": "timeline" } ], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 7, "type": "analysis", "title": "The slide highlights the convergence of data availability, hardware, software frameworks, and algorithmic innovation.", "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": 674, "nDataPoints": 15, "notes": "The slide highlights the convergence of data availability, hardware, software frameworks, and algorithmic innovation.", "elementsJson": [ "headline_text", "bullet_list", "data_table", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:16.932+00", "seconds": 7.975, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/7", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-7", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.55, "w": 0.35, "x": 0.14, "y": 0.3 }, "kind": "list", "text": "Increasing availability of text data from the internet\nDevelopment of powerful computational resources (GPUs and TPUs)\nFrameworks for developing neural networks (TensorFlow and PyTorch)\nAdvances in ML algorithms (transformers and attention)", "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 18:38:16.932+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5c261785-044a-4b58-884b-721ccfed0f7b", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Number of parameters: 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": "2026-04-24 13:39:16+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ea-7ca91ed2afc2", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.45, "x": 0.03, "y": 0.2 }, "kind": "paragraph", "text": "NLP has achieved groundbreaking results through LLMs, enabled by various modern 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 18:38:16.932+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4a086350-ee04-411d-a93c-6ce66a4ea7e5", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.3, "x": 0.65, "y": 0.95 }, "kind": "source-note", "text": "Source: https://huggingface.co/transformers/v2.4.0/pretrained_models.html", "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 18:38:16.932+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fa195aad-25cd-468a-b28c-ec4d7756292a", "parentComponentId": null }, { "bbox": { "h": 0.7, "w": 0.45, "x": 0.52, "y": 0.2 }, "kind": "table", "text": "Model | Parameters | Dataset Size | Compute\nBERT | 110M | 16GB | -\nGPT | 117M | 40GB | -\nROBERTA | 125M | 160GB | -\nGPT-2 | 1.5B | 800GB | -\nGPT-3 | 175B | 45TB | 3,600+ GPU days, 330+ MWH", "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 18:38:16.932+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e11ccc85-3ae3-405e-95c0-877052190329", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.75, "x": 0.03, "y": 0.04 }, "kind": "title", "text": "ACCESS TO POWERFUL RESOURCES ENABLE LARGE LANGUAGE 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 18:38:16.932+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8309b040-a0da-425c-afc8-a7abf4834014", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Number of parameters", "numberRaw": "175B", "numberKind": "plain", "actionTitle": "ACCESS TO POWERFUL RESOURCES ENABLE LARGE LANGUAGE MODELS", "calloutText": null, "numberScale": "b", "numberValue": 175, "metricFamily": "volume_count", "numberCurrency": null } ], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "54364cf1-7f0e-459a-b426-d35c0f081849", "evidence": "The slide includes a table with data on model parameters, dataset size, and compute.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:12.758566+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": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "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": 5, "componentCount": 3, "textChars": 27, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:13.348+00", "seconds": 4.332, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/8", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-8", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 1, "w": 0.254, "x": 0.746, "y": 0 }, "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 18:38:13.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1e91a734-ffd7-45c7-aed0-f7cec2d880de", "parentComponentId": null }, { "bbox": { "h": 0.063, "w": 0.503, "x": 0.037, "y": 0.642 }, "kind": "title", "text": "THE BUILDING BLOCKS OF GPT", "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 18:38:13.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "51617933-534b-4404-b1ef-749f598fd782", "parentComponentId": null }, { "bbox": { "h": 0.235, "w": 0.098, "x": 0.037, "y": 0.378 }, "kind": "title", "text": "2", "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 18:38:13.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9aa0fa90-72ba-49e5-b7ad-a28616b0c89c", "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": "54a791e6-cd3c-43fb-bf22-f8db918faf60", "evidence": "This slide serves as another section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:12.83357+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": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 9, "type": "appendix", "title": "The slide serves as an educational visual aid for a machine learning presentation.", "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": "sparse", "densityScore": 13, "componentCount": 2, "textChars": 65, "nDataPoints": 80, "notes": "The slide serves as an educational visual aid for a machine learning presentation.", "elementsJson": [ "headline_text", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:11.306+00", "seconds": 2.14, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/9", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-9", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.6, "x": 0.36, "y": 0.138 }, "kind": "chart", "text": "Scatter plot of data points", "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 18:38:11.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d62d0c0f-5e79-4bed-8811-84784fd3984b", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.25, "x": 0.04, "y": 0.4 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:11.306+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8a351e79-53c2-4da1-8d4f-69f8f7aacf96", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 10, "type": "appendix", "title": "The slide provides the standard OLS (Ordinary Least Squares) formulas for linear regression, though the title mentions 'Gradient Descent' which is an alternativ", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "balanced", "densityScore": 31, "componentCount": 5, "textChars": 375, "nDataPoints": 0, "notes": "The slide provides the standard OLS (Ordinary Least Squares) formulas for linear regression, though the title mentions 'Gradient Descent' which is an alternative optimization method not explicitly shown in the formulas.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:14.925+00", "seconds": 5.71, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/10", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-10", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.4, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot showing a positive linear correlation between two variables.", "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 18:38:14.925+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8414d5bd-165f-4aa4-83d9-a2dfbf1919a2", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.1, "x": 0.03, "y": 0.95 }, "kind": "disclaimer", "text": "© Oliver Wyman", "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 18:38:14.925+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2865c7bf-98ec-4f54-9af8-c70a6fefb6f5", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.03, "y": 0.275 }, "kind": "table", "text": "Model Structure: Y = β0 + β1 * X; Cost Function: Cost = Σ(predicted - actual)^2; Formulas for β Coefficients: Slope β1 = (n * Σ(x*y) - Σ(x) * Σ(y)) / (n * Σ(x^2) - (Σ(x))^2), Intercept β0 = (Σ(y) - β1 * Σ(x)) / N", "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 18:38:14.925+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "589155f1-6664-4c61-9d2f-d5bbf3778211", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:14.925+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "eab21531-c0fb-4477-91ab-3b2adf31531e", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.3, "x": 0.03, "y": 0.2 }, "kind": "title", "text": "Linear Regression Model Using Formulas", "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 18:38:14.925+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8b8a15a0-e44b-411d-9ef4-fff29668c6c0", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b962e2f0-303b-4ffd-9526-d349dd5206e5", "evidence": "The slide includes a table with data on model structure, cost function, and formulas for β coefficients.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:12.943561+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": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [ { "to": 10, "from": 5, "loop": { "name": "11_golden_circle", "slug": "11-golden-circle", "status": "active", "bestFor": "Visionary leadership, brand positioning, mission statements", "canonId": "019dd956-6a50-7202-882d-505fcb9922ed", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Why (Belief) -> The How (Process) -> The What (Result)", "description": "Invert the typical pitch by starting with why you exist, rather than what you do", "familyLabel": null }, "matchId": "604ad8ff-b2cb-4f7f-bd6f-eb583fd391a1", "evidence": "The deck uses the golden circle framework to explain the what, how, and why of GPT-3 and NLP.", "position": 0, "objective": "Why should we care about GPT-3 and NLP?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.69431+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 11, "type": "appendix", "title": "The slide uses a standard linear regression example to illustrate basic ML concepts.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 57, "componentCount": 8, "textChars": 449, "nDataPoints": 0, "notes": "The slide uses a standard linear regression example to illustrate basic ML concepts.", "elementsJson": [ "paragraph", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:16.696+00", "seconds": 7.223, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/11", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-11", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.75, "w": 0.42, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot showing data points and a linear regression line with equation Y = 5.15 * X + 10.11 and Cost = 10.7476", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "970283f4-f888-4e7b-ad1e-026b1929265a", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.08, "x": 0.037, "y": 0.95 }, "kind": "disclaimer", "text": "© Oliver Wyman", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "75add1b6-58c8-42fa-a577-e18635a51f6f", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.48, "x": 0.037, "y": 0.44 }, "kind": "list", "text": "COST FUNCTION: Cost = Σ(predicted – actual)^2 = 10.7476", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "63f36491-8510-4e55-9337-303fe7386c1d", "parentComponentId": null }, { "bbox": { "h": 0.12, "w": 0.48, "x": 0.037, "y": 0.27 }, "kind": "list", "text": "MODEL STRUCTURE: Y = β0 + β1 * X", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6b467bae-b12a-4b43-bb77-04d105d23d54", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.48, "x": 0.037, "y": 0.61 }, "kind": "list", "text": "FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11; Intercept β0 = (Σ(y) - β1 * Σx) / N = 5.15", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f1d16b56-e017-4a8e-a2d6-577f1b460090", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: 10.7476", "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": "2026-04-24 13:39:16+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82e9-f4cd22f4a01c", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.28, "x": 0.037, "y": 0.205 }, "kind": "paragraph", "text": "Linear Regression Model Using Formulas", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c65b996e-5f3a-4c60-991c-b39c8c89911b", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.48, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:16.696+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4f8804c3-a8e8-4c33-baaa-702767c50aaa", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Cost", "numberRaw": "10.7476", "numberKind": "plain", "actionTitle": null, "calloutText": null, "numberScale": null, "numberValue": 10.7476, "metricFamily": "cost_savings", "numberCurrency": null } ], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b1b04942-ed59-4484-971b-e188ec8f7035", "evidence": "The slide includes a table with data on model structure, cost function, and formulas for β coefficients.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.019465+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 11, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 12, "type": "setup", "title": "The slide explains the derivation of beta coefficients for a simple linear regression model.", "function": "present_framework", "rawType": "appendix_methodology", "block": null, "metadata": { "slideType": "appendix_methodology", "slideTypeCanon": { "name": "Appendix methodology", "slug": "appendix_methodology", "status": "active", "canonId": "019de52d-07d9-726c-8599-167f2d2ffaa7", "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": 8, "textChars": 472, "nDataPoints": 0, "notes": "The slide explains the derivation of beta coefficients for a simple linear regression model.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.534+00", "seconds": 2.785, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/12", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-12", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.4, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot showing data points and a linear regression line with equation Y = 5.15 * X + 10.11 and Cost = 10.7476", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2a0e18fe-b0e8-4d4a-b498-51ad706ce2eb", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.48, "x": 0.037, "y": 0.275 }, "kind": "list", "text": "MODEL STRUCTURE: Y = β0 + β1 * X", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0fb67319-c005-424b-b262-194defefe875", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.48, "x": 0.037, "y": 0.44 }, "kind": "list", "text": "COST FUNCTION: Cost = Σ(predicted – actual)^2 = 10.7476", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "30b17311-8eda-4845-b588-d2a67f67c0aa", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.48, "x": 0.037, "y": 0.825 }, "kind": "list", "text": "RESULTING MODEL: Y = 10.11 * X + 5.15", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "387ab37c-4d30-47d0-b5a9-45106344b73d", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.48, "x": 0.037, "y": 0.615 }, "kind": "list", "text": "FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11; Intercept β0 = (Σ(y) – β1 * ΣX) / N = 5.15", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "95703873-c52d-4764-82c3-7e425a722591", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: 10.7476", "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": "2026-04-24 13:39:18+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-3117bd7c9ce0", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.28, "x": 0.037, "y": 0.205 }, "kind": "paragraph", "text": "Linear Regression Model Using Formulas", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d5a96725-e2c4-4532-9f35-22a09c5696d9", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.48, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:12.534+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d18cd98e-f794-4325-a644-2e8253aa6052", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Cost", "numberRaw": "10.7476", "numberKind": "plain", "actionTitle": null, "calloutText": null, "numberScale": null, "numberValue": 10.7476, "metricFamily": "cost_savings", "numberCurrency": null } ], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "256d7a7d-040b-4045-a3cd-ebf95508ccf0", "evidence": "list/bullet: FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11; Intercept β0 = (Σ(y) – β1 * ΣX) / N = 5.15", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.903883+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 12, "whyItWorks": null, "antipattern": null, "cardinality": "common", "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": "1e467481-c211-429a-b52f-baf034baf485", "evidence": "chart/scatter: Scatter plot showing data points and a linear regression line with equation Y = 5.15 * X + 10.11 and Cost = 10.7476", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.8, "extraction": { "at": "2026-07-17 03:18:13.866836+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 12, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 13, "type": "setup", "title": "The slide sets up a transition from closed-form solutions to iterative optimization.", "function": "transition", "rawType": "transition", "block": null, "metadata": { "slideType": "transition", "slideTypeCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-0693-7766-9f05-887f41f8e41e", "version": 1, "description": null }, "function": "transition", "functionCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-1365-7149-92aa-66df859e75fd", "version": 1, "description": null }, "density": "balanced", "densityScore": 31, "componentCount": 4, "textChars": 407, "nDataPoints": 4, "notes": "The slide sets up a transition from closed-form solutions to iterative optimization.", "elementsJson": [ "headline_text", "process_diagram", "callout_box" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.852+00", "seconds": 3.02, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/13", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-13", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.28, "w": 0.46, "x": 0.51, "y": 0.615 }, "kind": "callout", "text": "Without those formulas, How can we find the coefficients?", "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 18:38:12.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "add9fdc5-f7bb-4c74-99d7-6c8c1e15f388", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.25, "x": 0.03, "y": 0.21 }, "kind": "paragraph", "text": "Linear Regression Model Using Formulas", "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 18:38:12.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "31676cf5-4f78-47f1-8aba-cbedbba212ed", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.48, "x": 0.03, "y": 0.27 }, "kind": "table", "text": "MODEL STRUCTURE: Y = β0 + β1 * X; COST FUNCTION: Cost = Σ(predicted – actual)^2 = 10.7476; FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11, Intercept β0 = (Σ(y) – β1 * ΣX) / N = 5.15; RESULTING MODEL: Y = 10.11 * X + 5.15", "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 18:38:12.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7d61d827-2146-425f-a605-71ddf7f17e63", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:12.852+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "abae0ea8-b76e-42dd-ba70-5fc92176691f", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "34995123-1480-4322-86d0-a39bdafc6215", "evidence": "table/data: MODEL STRUCTURE: Y = β0 + β1 * X; COST FUNCTION: Cost = Σ(predicted – actual)^2 = 10.7476; FORMULAS FOR β COEFFICIENTS: Slope β1 = (n * Σ(x*y) – Σ(x) * Σ(y)) / (n * Σ(x^2) – (Σ(x))^2) = 10.11, Intercept β0 =", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:13.976822+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 13, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 14, "type": "appendix", "title": "The slide contrasts the iterative nature of gradient descent (no closed-form formula for coefficients) with the model structure and cost function.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 18, "componentCount": 4, "textChars": 226, "nDataPoints": 0, "notes": "The slide contrasts the iterative nature of gradient descent (no closed-form formula for coefficients) with the model structure and cost function.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.718+00", "seconds": 2.57, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/14", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-14", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.75, "w": 0.4, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot of data points", "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 18:38:12.718+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e88805f8-bbe6-485e-bef2-4057f21fa941", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.33, "x": 0.037, "y": 0.205 }, "kind": "paragraph", "text": "Linear Regression Model Using Gradient Descent", "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 18:38:12.718+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "75807346-ed6c-49d6-b186-e94654c5c267", "parentComponentId": null }, { "bbox": { "h": 0.48, "w": 0.44, "x": 0.037, "y": 0.27 }, "kind": "table", "text": "Model Structure: Y = β0 + β1 * X; Cost Function: Cost = Σ(predicted – actual)^2; Formulas for β Coefficients: None!", "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 18:38:12.718+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "25d5816f-918c-4e9d-a7a9-eae28a0bc1be", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.48, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:12.718+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "493d9a11-05d5-4c34-8799-99c4c918dd81", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 15, "type": "appendix", "title": "The slide illustrates the iterative nature of gradient descent by showing a poor initial fit (red line) for the data points.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "balanced", "densityScore": 36, "componentCount": 6, "textChars": 310, "nDataPoints": 0, "notes": "The slide illustrates the iterative nature of gradient descent by showing a poor initial fit (red line) for the data points.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.433+00", "seconds": 2.226, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/15", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-15", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.4, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot with regression line showing Y = 20.00 * X + -5.00 and Cost = 226.5655", "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 18:38:12.433+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "13d59a9f-4d48-4013-927e-b7c1a601ca2a", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: 226.5655", "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": "2026-04-24 13:39:18+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-78b3db48f256", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.35, "x": 0.03, "y": 0.21 }, "kind": "paragraph", "text": "Linear Regression Model Using Gradient Descent", "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 18:38:12.433+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10b0dc6e-133f-4500-8c10-2a1f1e8be0a2", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.1, "x": 0.03, "y": 0.95 }, "kind": "source-note", "text": "© Oliver Wyman", "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 18:38:12.433+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "83761592-97c3-483c-8bb1-d970c381ecd1", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.45, "x": 0.03, "y": 0.27 }, "kind": "table", "text": "MODEL STRUCTURE: Y = β0 + β1 * X; COST FUNCTION: Cost = Σ(predicted – actual)^2; FORMULAS FOR β COEFFICIENTS: None!", "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 18:38:12.433+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9c9816ec-6f45-4f20-92ca-514b1c663801", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:12.433+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f4c283cd-ce18-424d-9839-8e8e0e754013", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Cost", "numberRaw": "226.5655", "numberKind": "plain", "actionTitle": null, "calloutText": null, "numberScale": null, "numberValue": 226.5655, "metricFamily": "cost_savings", "numberCurrency": null } ], "tools": [], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 16, "type": "appendix", "title": "The slide illustrates the iterative nature of gradient descent by showing a poor initial fit (red line) for the data points.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "balanced", "densityScore": 37, "componentCount": 6, "textChars": 328, "nDataPoints": 0, "notes": "The slide illustrates the iterative nature of gradient descent by showing a poor initial fit (red line) for the data points.", "elementsJson": [ "headline_text", "paragraph", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:13.269+00", "seconds": 3.06, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/16", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-16", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.4, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot showing data points and a regression line with Y = 20.72 * X + -2.86 and Cost = 138.6722", "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 18:38:13.269+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2ac75cdf-335b-492a-abaf-81befc54285c", "parentComponentId": null }, { "bbox": { "h": 0.5, "w": 0.45, "x": 0.03, "y": 0.25 }, "kind": "list", "text": "MODEL STRUCTURE: Y = β0 + β1 * X; COST FUNCTION: Cost = Σ(predicted – actual)^2; FORMULAS FOR β COEFFICIENTS: None!", "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": "Comparison list", "slug": "comparison", "status": "active", "canonId": "019de52c-fce7-733e-99d3-030a104727bf", "version": 1, "description": null }, "framework": null }, "subkind": "comparison", "groupRole": null, "confidence": null, "extraction": { "at": "2026-05-02 18:38:13.269+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c2aec4d2-c112-4312-92da-5668e0e912d3", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: 138.6722", "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": "2026-04-24 13:39:18+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-8cf683ec4e35", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.35, "x": 0.03, "y": 0.2 }, "kind": "paragraph", "text": "Linear Regression Model Using Gradient Descent", "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 18:38:13.269+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6b28a5c0-ba9f-4411-b230-ee8445c493ac", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.1, "x": 0.03, "y": 0.95 }, "kind": "source-note", "text": "© Oliver Wyman", "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 18:38:13.269+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "885f828a-f322-4555-8d85-6072304c824a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:13.269+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b913d88d-c41f-4d80-bbab-33e11aca69d2", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Cost", "numberRaw": "138.6722", "numberKind": "plain", "actionTitle": null, "calloutText": null, "numberScale": null, "numberValue": 138.6722, "metricFamily": "cost_savings", "numberCurrency": null } ], "tools": [], "frameworks": [], "arcBeats": [ { "to": 16, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 6, "name": "What Is", "beatId": "d061c184-c7cd-4a92-b451-664f35ad62f3", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Is", "slug": "sparkline-what-is-3", "status": "active", "canonId": "019dd9b8-2701-72ba-a477-4121f4fb5d33", "version": 1, "description": null }, "evidence": "The deck delves deeper into the technical aspects of GPT-3 and NLP.", "position": 1, "isPrimaryArc": true, "parentBeatType": { "name": "Complication", "slug": "complication", "status": "active", "canonId": "019dd9b8-0275-7709-b2d6-4f77c3fc7579", "version": 1, "description": "The point in the narrative where the audience should feel discomfort. Required for argumentative arcs." }, "alignedBlockIds": null, "matchConfidence": 0.8 } ], "loops": [], "locked": true }, { "page": 17, "type": "appendix", "title": "The slide uses a structured layout to define model components (structure, cost function, coefficients, resulting model) and shows the resulting regression line ", "function": "present_framework", "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": "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": 7, "textChars": 442, "nDataPoints": 50, "notes": "The slide uses a structured layout to define model components (structure, cost function, coefficients, resulting model) and shows the resulting regression line on a scatter plot.", "elementsJson": [ "headline_text", "data_table", "scatter_plot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.907+00", "seconds": 5.648, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/17", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-17", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Through the two coefficients, the model “remembers” the data.", "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": "2026-04-24 13:39:18+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-50f334316b17", "parentComponentId": null }, { "bbox": { "h": 0.72, "w": 0.42, "x": 0.55, "y": 0.2 }, "kind": "chart", "text": "Scatter plot showing data points and a fitted linear regression line.", "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 18:38:15.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a71c925c-aa0c-41af-9770-09353c43dcff", "parentComponentId": null }, { "bbox": { "h": 0.63, "w": 0.48, "x": 0.037, "y": 0.27 }, "kind": "list", "text": "Model Structure: Y = β0 + β1 * X; Cost Function: Cost = Σ(predicted – actual)^2; Formulas for β Coefficients: None!; Resulting Model: Y = 10.11 * X + 5.15", "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 18:38:15.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0313a077-8eab-423f-a9d9-d504e68dfd0f", "parentComponentId": null }, { "bbox": null, "kind": "metric", "text": "Cost: 10.7476", "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": "2026-04-24 13:39:18+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-57c1c906e890", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.26, "x": 0.037, "y": 0.915 }, "kind": "paragraph", "text": "Through the two coefficients, the model “remembers” the data.", "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 18:38:15.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "00828475-af3d-479b-9fcb-84c0764ba51f", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.33, "x": 0.037, "y": 0.205 }, "kind": "paragraph", "text": "Linear Regression Model Using Gradient Descent", "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 18:38:15.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c9b6e490-1540-4a3e-86a6-4601de89038e", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.48, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "MACHINE LEARNING 101: GRADIENT DESCENT", "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 18:38:15.907+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "aa8e72e5-0c59-49d8-8a83-0c0a3d067066", "parentComponentId": null } ], "metrics": [ { "metricRaw": "Cost", "numberRaw": "10.7476", "numberKind": "plain", "actionTitle": null, "calloutText": "Through the two coefficients, the model “remembers” the data.", "numberScale": null, "numberValue": 10.7476, "metricFamily": "cost_savings", "numberCurrency": null } ], "tools": [], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 18, "type": "setup", "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": "sparse", "densityScore": 8, "componentCount": 2, "textChars": 194, "nDataPoints": 0, "notes": null, "elementsJson": [ "comparison_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.38+00", "seconds": 2.113, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/18", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-18", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.92, "x": 0.037, "y": 0.18 }, "kind": "table", "text": "Comparison matrix of Linear Regression, Multivariate Linear Regression, and Logistic Regression across Model Structure, Cost Function, Prediction range, and Examples.", "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 18:38:12.38+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e1d34f96-f0e0-48aa-bab6-9f518b794339", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.36, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "REGRESSION VS CLASSIFICATION", "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 18:38:12.38+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8274b945-93d4-4d39-8cf5-b410353ab584", "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": "c15f7299-65e2-4f99-9845-bbcbff197126", "evidence": "table/data: Comparison matrix of Linear Regression, Multivariate Linear Regression, and Logistic Regression across Model Structure, Cost Function, Prediction range, and Examples.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:14.203697+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 18, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 19, "type": "setup", "title": "The slide uses a node-link diagram to illustrate the mathematical structure of three different machine learning models.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 17, "componentCount": 4, "textChars": 208, "nDataPoints": 0, "notes": "The slide uses a node-link diagram to illustrate the mathematical structure of three different machine learning models.", "elementsJson": [ "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.39+00", "seconds": 4.947, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/19", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-19", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.6, "w": 0.3, "x": 0.35, "y": 0.2 }, "kind": "diagram", "text": "Multivariate Linear Regression model structure and input-output 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 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0970df7e-805f-4d46-919a-fba17ddb6253", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.3, "x": 0.037, "y": 0.2 }, "kind": "diagram", "text": "Linear Regression model structure and input-output 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 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3bf9de4c-fbe5-43c4-9ec2-0a39ac4b4e00", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.3, "x": 0.67, "y": 0.2 }, "kind": "diagram", "text": "Logistic Regression model structure and input-output 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 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "caa8df28-faf7-4475-9a8c-a9f723bb7619", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.35, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "REGRESSION VS CLASSIFICATION", "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 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "19223c26-8500-4ab4-afd0-6a4d563a6f79", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "609c641c-4d13-406b-800a-458f34a452be", "evidence": "Comparison of statistical modeling techniques", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 18:38:15.39+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 19, "frameworkName": "analytical-method" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 20, "type": "appendix", "title": "The slide illustrates the difference between a single sigmoid output node (binary/single-label) and multiple sigmoid output nodes (multi-label/multi-class) in a", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 15, "componentCount": 4, "textChars": 169, "nDataPoints": 0, "notes": "The slide illustrates the difference between a single sigmoid output node (binary/single-label) and multiple sigmoid output nodes (multi-label/multi-class) in a neural network context.", "elementsJson": [ "headline_text", "data_table", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:12.754+00", "seconds": 2.258, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/20", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-20", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.6, "w": 0.28, "x": 0.68, "y": 0.25 }, "kind": "diagram", "text": "Multi-Classification Model Structure", "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 18:38:12.754+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2e418853-b60d-4b60-8b31-6726acdef577", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.25, "x": 0.05, "y": 0.25 }, "kind": "diagram", "text": "Single-Label Classification Model Structure", "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 18:38:12.754+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cef7e88e-7449-45a5-ab35-2ca12938b421", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.3, "x": 0.35, "y": 0.25 }, "kind": "table", "text": "Multi-Label Classification Training Data Example", "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 18:38:12.754+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "92050abd-a786-4cb9-a271-7831d0adf005", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "SINGLE-LABEL VS MULTI-LABEL CLASSIFICATION", "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 18:38:12.754+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "95bf430f-2f21-499b-884b-b4f9938e739a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "ebee8e72-944d-46f3-91f6-b29da64231de", "evidence": "table/data: Multi-Label Classification Training Data Example", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.7, "extraction": { "at": "2026-07-17 03:18:14.315695+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 20, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 21, "type": "appendix", "title": "The slide illustrates a standard feed-forward neural network with an input layer, two hidden layers, and an output layer, specifically applied to the Iris datas", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 21, "componentCount": 4, "textChars": 303, "nDataPoints": 0, "notes": "The slide illustrates a standard feed-forward neural network with an input layer, two hidden layers, and an output layer, specifically applied to the Iris dataset classification problem.", "elementsJson": [ "headline_text", "process_diagram", "callout_box" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:13.386+00", "seconds": 2.716, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/21", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-21", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The model adjusts those parameters to minimize the cost 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": "2026-04-24 13:39:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-d6119ba4ba64", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.6, "x": 0.05, "y": 0.25 }, "kind": "diagram", "text": "Multi-Layer Perceptron neural network 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 18:38:13.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1ce0b9a9-6706-44bb-8fc3-e822792d2895", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.28, "x": 0.68, "y": 0.25 }, "kind": "list", "text": "MODEL STRUCTURE: The number of nodes in the hidden layer is chosen by the modeler. COST FUNCTION: Cross-entropy. MODEL PARAMETERS: All beta parameters are initially set a random.", "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 18:38:13.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "53d06eda-4992-4012-ba86-a5ec691c6cf5", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.23, "x": 0.035, "y": 0.05 }, "kind": "title", "text": "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 18:38:13.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "81a33fbc-6509-46dc-a8ee-c18302129906", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "cdcc384a-3615-407a-85ff-deaac36a84e8", "evidence": "Diagram of layers and nodes", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:13.386+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 21, "frameworkName": "Neural Network Architecture" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 22, "type": "setup", "function": "transition", "rawType": "transition", "block": null, "metadata": { "slideType": "transition", "slideTypeCanon": { "name": "Transition", "slug": "transition", "status": "active", "canonId": "019de52d-0693-7766-9f05-887f41f8e41e", "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": 32, "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 18:38:12.867+00", "seconds": 2.064, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/22", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-22", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.2, "w": 0.5, "x": 0.25, "y": 0.4 }, "kind": "title", "text": "BUT WHAT ABOUT PREDICTING WORDS?", "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 18:38:12.867+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c3d8c122-23f1-4a5e-8a9a-2ca095dc7a6f", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 23, "type": "appendix", "title": "Explains the technical process of how GPT models convert text to tokens and predict the next token as a classification task.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 70, "componentCount": 7, "textChars": 1059, "nDataPoints": 2, "notes": "Explains the technical process of how GPT models convert text to tokens and predict the next token as a classification task.", "elementsJson": [ "headline_text", "paragraph", "bullet_list", "process_diagram", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.876+00", "seconds": 7.066, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/23", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-23", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Fundamentally, GPT-3 and ChatGPT are neural networks that constantly give a probability to what should be the next outputted word.", "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": "2026-04-24 13:39:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-f63595690025", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.3, "x": 0.67, "y": 0.39 }, "kind": "diagram", "text": "Example: Tokenization of an Input\nWhat do actuaries do exactly?\n[2061, 466, 43840, 3166, 466, 3446, 30]", "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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9290cae9-da73-4b1e-b1f6-288539282dbb", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.35, "y": 0.39 }, "kind": "list", "text": "Classification Problem\nNext word prediction becomes a classification problem\nInput: series of tokens (a sentence)\nOutput: probability distribution over all tokens\nVocab size of GPT-3 = 50,257\nThe problem becomes a classification problem with 50,257 labels", "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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9a38eb60-14e2-40bd-970a-34314b19aace", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.03, "y": 0.39 }, "kind": "list", "text": "First Step: Tokenization\nFirst step of NLP any model is to convert text into numbers, or \"tokens\".\nGPT-3's tokenizer assign integers to chunks of characters.\nIt's a one-to-one mapping, fixed mapping.\nIn the input layer, \"exactly\" will always be mapped to the number 3446\nIn the output layer, 3446 will always be mapped to \"exactly\"", "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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b0853dbb-07a2-4805-9b79-61e57a48d752", "parentComponentId": null }, { "bbox": { "h": 0.11, "w": 0.94, "x": 0.03, "y": 0.23 }, "kind": "paragraph", "text": "Fundamentally, GPT-3 and ChatGPT are neural networks that constantly give a probability to what should be the next outputted word. That's why ChatGPT types one word at a 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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "696b7558-1292-4474-b823-7cf89af98dd8", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.26, "x": 0.71, "y": 0.91 }, "kind": "source-note", "text": "Source: https://platform.openai.com/tokenizer", "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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "976b34ec-cfdc-43b8-bdfd-816254d7020e", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.28, "x": 0.03, "y": 0.04 }, "kind": "title", "text": "NEXT WORD 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 18:38:17.876+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1b9164b9-d485-4e23-bcb2-feb26825dd09", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 24, "type": "appendix", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 52, "componentCount": 5, "textChars": 756, "nDataPoints": 15, "notes": null, "elementsJson": [ "headline_text", "bullet_list", "data_table" ], "confidence": 0.95, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:14.15+00", "seconds": 2.953, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/24", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-24", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.15, "w": 0.4, "x": 0.05, "y": 0.25 }, "kind": "list", "text": "Token IDs cannot be used as-is.\nWord Embedding: a large vector assigned to each token\nValues in the vector are initially assigned at random", "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 18:38:14.15+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fdd279c9-2f88-4f32-810f-ea56c336cee0", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.3, "x": 0.05, "y": 0.2 }, "kind": "paragraph", "text": "How to quantify meanings of words?", "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 18:38:14.15+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bbca0989-9ded-4e8e-92ad-58f8b05ccc14", "parentComponentId": null }, { "bbox": { "h": 0.25, "w": 0.9, "x": 0.05, "y": 0.65 }, "kind": "table", "text": "Token | Token ID | One-Hot Encoded Vector (50,000 dimensions) | Word Embedding Vector (fewer dimensions)\nround | 35634 | (0, 0, 0, 0, 0, 0, ..., 0, 0, 0, 0, 0, 0, 1, 0, ..., 0, 0, 0, 0) | (0.932, 0.321, 0.456, 0.571, 0.984, ..., 0.654)\nball | 1894 | (0, 0, 0, 0, 0, 0, ..., 0, 1, 0, 0, 0, 0, 0, 0, ..., 0, 0, 0, 0) | (0.524, 0.329, 0.132, 0.134, 0.952, ..., 0.213)\nnet | 3262 | (0, 0, 0, 0, 0, 0, ..., 0, 0, 0, 0, 1, 0, 0, 0, ..., 0, 0, 0, 0) | (0.187, 0.818, 0.118, 0.901, 0.347, ..., 0.221)", "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 18:38:14.15+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bb6fed45-646a-4142-95cb-cc814f50ab89", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.2, "x": 0.05, "y": 0.58 }, "kind": "title", "text": "Word Embedding Examples", "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 18:38:14.15+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f46cd2d5-02e9-48f8-8b17-308851999ea2", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.05 }, "kind": "title", "text": "SUMMARIZING MEANING AND REDUCING DIMENSIONALITY WITH WORD EMBEDDINGS", "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 18:38:14.15+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fc269fe3-cb50-4c11-8de9-66fe8247288a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "57f02c71-2623-466e-916c-1da08c36bc8b", "evidence": "table/data: Token | Token ID | One-Hot Encoded Vector (50,000 dimensions) | Word Embedding Vector (fewer dimensions)", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.635024+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 24, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 25, "type": "appendix", "title": "Explains the mechanism of adding positional encoding vectors to word embedding vectors to preserve sequence information.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 55, "componentCount": 5, "textChars": 920, "nDataPoints": 15, "notes": "Explains the mechanism of adding positional encoding vectors to word embedding vectors to preserve sequence information.", "elementsJson": [ "headline_text", "paragraph", "bullet_list", "data_table" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:18.651+00", "seconds": 7.219, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/25", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-25", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The resulting vectors represent both the meaning and position of tokens.", "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": "2026-04-24 13:39:30+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ed-dcc886ff3f4b", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.6, "x": 0.03, "y": 0.28 }, "kind": "list", "text": "First, obvious reason: the order of the tokens need to be considered. Solution: Positional Encoding (see below)\nSecond, less obvious reason: some words “care” more about each other than others. Solution: Self-Attention (see next slides)", "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 18:38:18.651+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2b82f121-22eb-467f-996a-7bde8a78605e", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.75, "x": 0.03, "y": 0.18 }, "kind": "paragraph", "text": "Network nodes need to consider multiple tokens at once. How to do that? A naïve approach of simply taking an average or a sum of all word embedding vectors would be wrong for two reasons.", "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 18:38:18.651+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3e9c53f8-4743-4ee9-a7ca-500e6de789cb", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.95, "x": 0.03, "y": 0.55 }, "kind": "table", "text": "Token | Word Embedding | Positional Encoding | Resulting vectors\nName | (0.638...) | + (0, 1, 0, 1, ...) | = (0.638, 1.759...)\nthe | (0.655...) | + (0.031, 1.000...) | = (0.686, 1.324...)\ncapital | (0.082...) | + (0.062, 0.998...) | = (0.144, 1.324...)\nof | (0.194...) | + (0.094, 0.995...) | = (0.288, 1.289...)\nPeru | (0.825...) | + (0.125, 0.992...) | = (0.95, 1.935...)", "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 18:38:18.651+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "367147be-9439-42f3-bbb9-6410e595cb56", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.67, "x": 0.03, "y": 0.04 }, "kind": "title", "text": "REPRESENTING ORDER OF WORDS WITH POSITIONAL ENCODING", "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 18:38:18.651+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7c64ca74-d4fa-4e33-8875-470f85c96b5a", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "e414b55d-69f6-40fe-9e80-db774518598c", "evidence": "table/data: Token | Word Embedding | Positional Encoding | Resulting vectors", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.712812+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 25, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [ { "to": 25, "from": 17, "loop": { "name": "40_mece_breakdown", "slug": "40-mece-breakdown", "status": "active", "bestFor": "Problem structuring, ensuring completeness, strategic analysis", "canonId": "019dd956-763c-765e-8b68-6ee3ae1ef184", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Whole -> Category A (distinct) -> Category B (distinct) -> Category C (distinct) -> Complete Coverage", "description": "Divide a complex topic into mutually exclusive, collectively exhaustive categories", "familyLabel": null }, "matchId": "5fc4f6a3-76b0-4e01-8746-40ac791c194d", "evidence": "The deck provides a detailed breakdown of GPT-3's capabilities and applications in the actuarial field.", "position": 1, "objective": "How can GPT-3 be applied to actuarial work?", "confidence": 0.6, "extraction": { "at": "2026-07-16 21:13:57.728379+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 26, "type": "appendix", "title": "The slide uses a simple linguistic example to explain the concept of self-attention in AI models.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 64, "componentCount": 7, "textChars": 927, "nDataPoints": 0, "notes": "The slide uses a simple linguistic example to explain the concept of self-attention in AI models.", "elementsJson": [ "headline_text", "paragraph", "bullet_list" ], "confidence": 0.9, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:14.428+00", "seconds": 2.96, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/26", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-26", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Therefore, we want the next word predictions to highly depend on \"music\" and \"lifting\" and not so much on \"enthusiastic\" and \"apartment\".", "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": "2026-04-24 13:39:19+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82eb-ae535d3ded7a", "parentComponentId": null }, { "bbox": { "h": 0.15, "w": 0.8, "x": 0.037, "y": 0.57 }, "kind": "list", "text": "In the two sentences above:\n- The words \"music\" and \"lifting\" give a lot of meaning to the token \"heavy\", since those tokens help specify the context.\n- The words \"enthusiastic\" and \"apartment\", however are not very useful in finding out what is \"heavy\".", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "177faf72-834c-4907-a7c7-5d4491aca7f9", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.6, "x": 0.037, "y": 0.44 }, "kind": "list", "text": "I can enjoy almost any music genre, but I was never enthusiastic about heavy ___.\nI run instead of lifting, because my apartment building's gym doesn't have heavy ___.", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f1a970d3-8803-4760-92f2-a04672197523", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.037, "y": 0.75 }, "kind": "paragraph", "text": "Therefore, we want the next word predictions to highly depend on \"music\" and \"lifting\" and not so much on \"enthusiastic\" and \"apartment\".", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "705d64e1-bc31-4e66-9084-77bc65a5595c", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.25, "x": 0.037, "y": 0.39 }, "kind": "paragraph", "text": "Motivation for Self-Attention", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9e2197e5-1525-46f9-99ed-1e3c8de75b71", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.92, "x": 0.037, "y": 0.24 }, "kind": "paragraph", "text": "Self-Attention is the mechanism used by transformer models to weigh the importance of difference words in a sentence or piece of text based on their relationships to other words.", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cd896e53-b0d4-4c15-808a-cfd4a81f2a37", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.32, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "ATTENTION IS ALL YOU NEED", "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 18:38:14.428+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "de2b3eb3-0f1c-4134-b270-509b1dd885dd", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 27, "type": "appendix", "title": "This slide illustrates the core linear algebra operations (matrix multiplication) used in the Transformer architecture's attention mechanism.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 7, "componentCount": 2, "textChars": 154, "nDataPoints": 0, "notes": "This slide illustrates the core linear algebra operations (matrix multiplication) used in the Transformer architecture's attention mechanism.", "elementsJson": [ "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.253+00", "seconds": 5.76, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/27", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-27", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.65, "w": 0.92, "x": 0.036, "y": 0.25 }, "kind": "diagram", "text": "Input text to tokenization, encoding, and matrix multiplication into K, Q, V vectors", "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 18:38:17.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "77f3b73d-eb09-4c07-a0b2-487ef8bc5ca3", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.86, "x": 0.036, "y": 0.05 }, "kind": "title", "text": "CREATING KEYS, QUERIES, AND VALUES TO ALLOW SELF-ATTENTION CALCULATION", "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 18:38:17.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "11b1f12d-b35e-4abc-8d40-b74524a13e60", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "1a2fca73-7c91-4ef9-b512-093b62a109a0", "evidence": "The diagram explicitly maps the transformation of input tokens into Key, Query, and Value vectors via matrix multiplication.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:17.253+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 27, "frameworkName": "Transformer Attention Mechanism" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 28, "type": "analysis", "title": "The slide explains the mathematical steps of the self-attention mechanism in transformer models.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 50, "componentCount": 5, "textChars": 666, "nDataPoints": 24, "notes": "The slide explains the mathematical steps of the self-attention mechanism in transformer models.", "elementsJson": [ "headline_text", "data_table", "bullet_list", "callout_box", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.384+00", "seconds": 5.725, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/28", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-28", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.2, "w": 0.928, "x": 0.036, "y": 0.715 }, "kind": "callout", "text": "Queries: vector describing what each token cares about\nKeys: vector describing what each token can inform about\nValue: vector describing information each token has to offer", "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 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a03e317f-767e-4233-a179-5b3da74444e9", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "Queries: vector describing what each token cares about; Keys: vector describing what each token can inform about; Value: vector describing information each token has to offer.", "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": "2026-04-24 13:39:30+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ee-09fbfb76d015", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.5, "x": 0.036, "y": 0.59 }, "kind": "list", "text": "Matrices used to obtain keys, queries, and values are common to all tokens. They are initialized at random and trained using gradient descent.", "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 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "779050cd-4623-47a9-9775-33ee477c5ea4", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.928, "x": 0.036, "y": 0.215 }, "kind": "table", "text": "Table showing Preceding Tokens, Key, Query, Value, Unnormalized Weights, Normalized Weights, and New Representation of Peru.", "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 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2c65277b-1b51-4412-ae77-872df2949cbb", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.65, "x": 0.036, "y": 0.05 }, "kind": "title", "text": "COMBINING KEYS, QUERIES, AND VALUES IN SELF-ATTENTION", "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 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bb3d341e-7064-4e30-9386-984b3c684042", "parentComponentId": null } ], "metrics": [], "tools": [ { "tool": { "name": "Table data", "slug": "table-data", "status": "active", "bestFor": null, "canonId": "019dd9e1-4fab-7168-a82f-b847350a006d", "version": 1, "bodyDocId": null, "description": "Tabular layouts of structured data.", "familyLabel": null, "categoryName": null, "categorySlug": null }, "agent": null, "layer": "slide", "agents": null, "matchId": "b9dcd898-138d-46cd-8c19-53e067709c38", "evidence": "table/data: Table showing Preceding Tokens, Key, Query, Value, Unnormalized Weights, Normalized Weights, and New Representation of Peru.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-17 03:18:13.830907+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [ { "matchId": "dbe2bd7c-8449-4fae-b722-9d5ba2d2048a", "evidence": "The slide explicitly describes the calculation of keys, queries, and values in self-attention.", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:17.384+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 28, "frameworkName": "Self-Attention Mechanism" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 29, "type": "evidence", "title": "The slide uses a direct screenshot of a ChatGPT interface to illustrate the capabilities of transformer architecture.", "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": "sparse", "densityScore": 18, "componentCount": 4, "textChars": 229, "nDataPoints": 0, "notes": "The slide uses a direct screenshot of a ChatGPT interface to illustrate the capabilities of transformer architecture.", "elementsJson": [ "headline_text", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:13.655+00", "seconds": 1.971, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/29", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-29", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.48, "w": 0.92, "x": 0.037, "y": 0.32 }, "kind": "image", "text": "ChatGPT interface showing a definition of actuaries", "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 18:38:13.655+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b0f6c9e6-eab7-45ab-b5c9-aee49a503687", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "Actuaries are professionals who use mathematical and statistical techniques to analyze financial risk and uncertainty.", "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": "2026-04-24 13:39:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-8e24e905574e", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.15, "x": 0.037, "y": 0.91 }, "kind": "source-note", "text": "Source: chat.openai.com/chat", "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 18:38:13.655+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2606a81b-7379-436c-a515-c25fb595536a", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "TRANSFORMER ARCHITECTURE EXAMPLE", "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 18:38:13.655+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "70d3de1d-297a-4779-a4e2-b46018ceccdd", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 30, "type": "appendix", "title": "The diagram maps the standard Transformer architecture (Attention is All You Need) to a specific example involving actuarial text processing.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "sparse", "densityScore": 8, "componentCount": 3, "textChars": 97, "nDataPoints": 0, "notes": "The diagram maps the standard Transformer architecture (Attention is All You Need) to a specific example involving actuarial text processing.", "elementsJson": [ "headline_text", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:18.348+00", "seconds": 6.549, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/30", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-30", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.75, "w": 0.92, "x": 0.05, "y": 0.15 }, "kind": "diagram", "text": "Transformer architecture 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 18:38:18.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "61116e11-14b1-4de7-981f-5cb1cd06b9b6", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.15, "x": 0.17, "y": 0.915 }, "kind": "source-note", "text": "Source: arxiv.org/pdf/1706.03762.pdf", "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 18:38:18.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6fdad998-f2c8-4921-ac36-b8aff2e08a84", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.45, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "TRANSFORMER ARCHITECTURE EXAMPLE", "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 18:38:18.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6bb36a6c-d5e6-4673-8a1e-d8bc64bcc5c6", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "15706509-49fb-4f41-ab06-c77e0b95db21", "evidence": "Visual representation of the Transformer model components (Attention, Feed Forward, Linear, Softmax)", "framework": null, "confidence": 1, "extraction": { "at": "2026-05-02 18:38:18.348+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 30, "frameworkName": "Transformer Architecture" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 31, "type": "appendix", "title": "The slide illustrates the reduction in required demonstrations as models move from fine-tuning to zero-shot learning.", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "dense", "densityScore": 63, "componentCount": 8, "textChars": 586, "nDataPoints": 0, "notes": "The slide illustrates the reduction in required demonstrations as models move from fine-tuning to zero-shot learning.", "elementsJson": [ "headline_text", "process_diagram", "paragraph", "callout_box", "footnote" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:20.556+00", "seconds": 8.68, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/31", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-31", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "No demonstrations are allowed – the model is only given a natural language description of the task", "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": "2026-04-24 13:39:20+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-2eca1c13abfe", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.92, "x": 0.036, "y": 0.2 }, "kind": "diagram", "text": "Comparison of Fine-tuning, Few-shot, One-shot, and Zero-shot learning", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "7befe997-487d-4fbe-8a86-1ac92089097b", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.2, "x": 0.505, "y": 0.695 }, "kind": "paragraph", "text": "Same as few-shot but only one demonstration is allowed", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "22d23b7f-489e-4f69-aa12-e3a149741830", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.2, "x": 0.74, "y": 0.695 }, "kind": "paragraph", "text": "No demonstrations are allowed – the model is only given a natural language description of the task", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "74afafd7-8b11-48e7-8c20-ce7bfc2e33d5", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.2, "x": 0.27, "y": 0.695 }, "kind": "paragraph", "text": "Model is given a few demonstrations of the task as conditioning, but no weight updates are allowed", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "bf951d9b-fe93-4af1-b68c-435fd1c75d38", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.2, "x": 0.036, "y": 0.695 }, "kind": "paragraph", "text": "Update weights of pre-trained model by training on a dataset specific to the desired task", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f502fc22-4fdf-4d0a-a51d-710a7df136df", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.2, "x": 0.036, "y": 0.91 }, "kind": "source-note", "text": "Source: https://arxiv.org/pdf/2005.14165.pdf", "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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f83b8797-d4d9-4f68-ba43-4f258aa76526", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.46, "x": 0.036, "y": 0.05 }, "kind": "title", "text": "GPT IS CAPABLE OF ZERO-SHOT 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 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3186b77c-e2b4-482d-82cf-d6b952e87239", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "1334b54e-a4dc-4b92-99be-84fcb1c0f34e", "evidence": "The slide compares four distinct methods of model adaptation based on demonstration count.", "framework": null, "confidence": 0.8, "extraction": { "at": "2026-05-02 18:38:20.556+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 31, "frameworkName": "learning-paradigm-comparison" } ], "arcBeats": [ { "to": 31, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 17, "name": "What Could Be", "beatId": "5259c9cc-c0ba-4428-ac14-7a6ba8d02be1", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "What Could Be", "slug": "sparkline-what-could-be-2", "status": "active", "canonId": "019dd9b8-266a-7046-9259-94dc86700d37", "version": 1, "description": null }, "evidence": "The deck explores the potential applications of GPT-3 in the actuarial field.", "position": 2, "isPrimaryArc": true, "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": 0.8 } ], "loops": [], "locked": true }, { "page": 32, "type": "setup", "title": "The slide uses a circular process diagram to represent a linear 4-step sequence.", "function": "present_framework", "rawType": "process", "block": null, "metadata": { "slideType": "process", "slideTypeCanon": 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": 41, "componentCount": 4, "textChars": 922, "nDataPoints": 0, "notes": "The slide uses a circular process diagram to represent a linear 4-step sequence.", "elementsJson": [ "headline_text", "numbered_list", "process_diagram", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:20.192+00", "seconds": 7.617, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/32", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-32", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.5, "w": 0.3, "x": 0.05, "y": 0.3 }, "kind": "diagram", "text": "Circular process flow with 4 numbered steps", "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 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "de39a026-ea8d-45d1-b595-5c061734bff4", "parentComponentId": null }, { "bbox": { "h": 0.7, "w": 0.6, "x": 0.38, "y": 0.2 }, "kind": "list", "text": "1. Fine-tune GPT-3.5: ChatGPT focused language model that has been fine-tuned on conversational data such as short, informal sentences and specific conversational conventions. 2. Train a reward model: A labeler ranks possible responses to prompts, and this data is used to train a reward model to determine the final response. 3. Use reinforcement learning to optimize reward: An agent learns to choose the best response to a prompt by receiving feedback in the form of the rewards from step 2. 4. Moderation endpoint: A separate language model is used to classify text as whether they violate content policy by being “sexual, hateful, violent, or promoting self-harm”.", "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 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "4a948c91-47e9-44a9-b05b-68d693ecde20", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.3, "x": 0.03, "y": 0.9 }, "kind": "source-note", "text": "Sources: https://openai.com/blog/chatgpt/ https://openai.com/blog/new-and-improved-content-moderation-tooling/", "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 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f695443e-371b-4f43-9166-d78fda925bff", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.9, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "Adding reinforcement learning layers and a moderation API enables the transition from GPT to ChatGPT", "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 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "00188001-c3e6-4633-ab49-5da677617afe", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [ { "matchId": "ec38b1fa-3616-4085-bbaf-531d2e0dd27a", "evidence": "4-step sequential process", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 18:38:20.192+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 32, "frameworkName": "process" } ], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 33, "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": 6, "componentCount": 3, "textChars": 42, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:14.906+00", "seconds": 2.169, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/33", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-33", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 1, "w": 0.25, "x": 0.75, "y": 0 }, "kind": "image", "text": "Keyboard background", "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 18:38:14.906+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0761453d-daaf-44d9-ad1b-32af36f48cc9", "parentComponentId": null }, { "bbox": { "h": 0.08, "w": 0.45, "x": 0.03, "y": 0.63 }, "kind": "title", "text": "ACTUARIAL APPLICATIONS", "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 18:38:14.906+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "264efb9d-0a2b-43a6-9787-0bd35f936856", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.1, "x": 0.03, "y": 0.35 }, "kind": "title", "text": "3", "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 18:38:14.906+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d14a0d06-4f98-4de5-a0c7-9f6cb0b1f458", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 34, "type": "evidence", "title": "The slide shows a side-by-side comparison of a Jupyter notebook environment and the GitHub Copilot suggestion interface.", "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": 25, "componentCount": 5, "textChars": 249, "nDataPoints": 0, "notes": "The slide shows a side-by-side comparison of a Jupyter notebook environment and the GitHub Copilot suggestion interface.", "elementsJson": [ "headline_text", "subtitle_text", "screenshot" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.817+00", "seconds": 2.933, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/34", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-34", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.75, "w": 0.4, "x": 0.56, "y": 0.18 }, "kind": "image", "text": "GitHub Copilot suggestion interface showing code completion options", "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 18:38:15.817+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "0a1f553b-1074-4725-bbbd-03be1325d11f", "parentComponentId": null }, { "bbox": { "h": 0.65, "w": 0.5, "x": 0.03, "y": 0.18 }, "kind": "image", "text": "Jupyter notebook code blocks showing data preparation and model fitting", "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 18:38:15.817+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5c76c699-7e03-42fc-b81f-4fbde55a0c0b", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.1, "x": 0.03, "y": 0.96 }, "kind": "source-note", "text": "© Oliver Wyman", "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 18:38:15.817+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "42a53b96-b258-4d17-81b6-5322c2151cf8", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.7, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "GPT-ENABLED TOOLS CAN HELP ACTUARIES EXECUTE THEIR WORK (1/3)", "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 18:38:15.817+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b5a4e5e3-8f4b-473a-89bf-e1c04d5b089b", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.3, "x": 0.03, "y": 0.11 }, "kind": "title", "text": "Fitting a model using GitHub Copilot", "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 18:38:15.817+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c5f704c0-ec26-4583-926f-3b3b72458e6d", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 35, "type": "evidence", "title": "The slide shows a two-column layout: left side details data loading steps, right side provides Python code for plotting histograms with average target overlays.", "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": "sparse", "densityScore": 24, "componentCount": 5, "textChars": 228, "nDataPoints": 0, "notes": "The slide shows a two-column layout: left side details data loading steps, right side provides Python code for plotting histograms with average target overlays.", "elementsJson": [ "headline_text", "subtitle_text", "screenshot", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.518+00", "seconds": 2.528, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/35", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-35", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.3, "w": 0.45, "x": 0.03, "y": 0.2 }, "kind": "image", "text": "ChatGPT response regarding modeling steps", "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 18:38:15.518+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2d27aa74-a9d4-44cd-963c-299ed06dee87", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.45, "x": 0.03, "y": 0.52 }, "kind": "image", "text": "Jupyter notebook snippet for data loading", "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 18:38:15.518+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "71afdd4e-ec8f-4010-b8eb-4f518ebbfecb", "parentComponentId": null }, { "bbox": { "h": 0.72, "w": 0.42, "x": 0.55, "y": 0.2 }, "kind": "image", "text": "ChatGPT response with Python code", "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 18:38:15.518+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "f9786da2-c24b-45e0-9b3c-500347e4f5c9", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.3, "x": 0.03, "y": 0.11 }, "kind": "paragraph", "text": "An entire modeling process using ChatGPT and Copilot", "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 18:38:15.518+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c84e2829-70cb-4abb-b1dd-ae4dab157b50", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.5, "x": 0.03, "y": 0.05 }, "kind": "title", "text": "GPT-ENABLED TOOLS CAN HELP ACTUARIES EXECUTE THEIR WORK (2/3)", "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 18:38:15.518+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1a8643a8-fafa-46c2-80d5-7dc89f4db258", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 36, "type": "evidence", "title": "The slide demonstrates a specific use case for LLMs in technical workflows.", "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": 30, "componentCount": 5, "textChars": 355, "nDataPoints": 0, "notes": "The slide demonstrates a specific use case for LLMs in technical workflows.", "elementsJson": [ "headline_text", "subtitle_text", "screenshot", "process_diagram" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:18.076+00", "seconds": 5.086, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/36", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-36", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "The issue with this code is that `reg:squarederror` should be a string, but it's currently being interpreted as an object.", "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": "2026-04-24 13:39:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-b83c783bb715", "parentComponentId": null }, { "bbox": { "h": 0.32, "w": 0.484, "x": 0.258, "y": 0.45 }, "kind": "image", "text": "ChatGPT response explaining the syntax error and providing the corrected code.", "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 18:38:18.076+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "a1564ef3-6f17-4d21-af2e-3f85eaf0bf8f", "parentComponentId": null }, { "bbox": { "h": 0.105, "w": 0.484, "x": 0.258, "y": 0.208 }, "kind": "image", "text": "Code snippet showing a syntax error in an XGBoost model definition.", "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 18:38:18.076+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "e8a25026-8202-4be2-98eb-adae4cd2de45", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.75, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "GPT-ENABLED TOOLS CAN HELP ACTUARIES EXECUTE THEIR WORK (3/3)", "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 18:38:18.076+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3dc38675-9ef2-44c6-9fe0-3e8c9e381084", "parentComponentId": null }, { "bbox": { "h": 0.03, "w": 0.25, "x": 0.037, "y": 0.105 }, "kind": "title", "text": "Using ChatGPT to debug 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": "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 18:38:18.076+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8fff7485-eb6c-469d-b754-85b9030cc326", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 37, "type": "setup", "function": "present_solution", "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": "present_solution", "functionCanon": { "name": "Present solution", "slug": "present_solution", "status": "active", "canonId": "019de52d-15b8-71da-bd23-2b767850cbce", "version": 1, "description": null }, "density": "dense", "densityScore": 68, "componentCount": 8, "textChars": 802, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "icon_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.948+00", "seconds": 2.921, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/37", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-37", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.7, "w": 0.9, "x": 0.05, "y": 0.2 }, "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": "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d327daa0-36b5-4d5e-ab3e-2a90a852913b", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.65, "y": 0.6 }, "kind": "list", "text": "Actuarial communication and report generation: Generate text to support actuarial analyses and draft reports", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3163ad33-3645-45ac-919a-df328bd9f130", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.05, "y": 0.2 }, "kind": "list", "text": "Webscraping for commercial lines underwriting: Streamline quoting process by automating capture of potential policyholder information", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6703ad0c-6154-40aa-ae47-43b1307db00f", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.65, "y": 0.2 }, "kind": "list", "text": "Summarizing and searching policy contracts: Identify key provisions and search to identify specific clauses or provisions", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6ebeefa3-68c7-48b4-8ce2-539a035fc4de", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.35, "y": 0.2 }, "kind": "list", "text": "Analysis of unstructured claims data: Classify/label unstructured data in claims to gather insights from documents such as medical reports", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ae65d593-71e3-4e3b-b439-95a6be7fce17", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.35, "y": 0.6 }, "kind": "list", "text": "Customer service: Chatbots powered by GPT models can understand natural language and provide personalized responses", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c871d0dc-579a-46da-93a1-cfe5cbbe1065", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.3, "x": 0.05, "y": 0.6 }, "kind": "list", "text": "Fraud detection: Analyze data from alternative sources to identify potential risks through detection of anomalies", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d2d80680-33ab-4243-b9d9-e6ce47f5b264", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.9, "x": 0.05, "y": 0.05 }, "kind": "title", "text": "GPT OPENS THE DOOR TO INNOVATIVE SOLUTIONS FOR SEVERAL INSURANCE PROCESSES", "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 18:38:15.948+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cd9129eb-056e-4d48-a8fa-b291213c4802", "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": "742de825-4d99-439c-bab2-df5f9c084e90", "evidence": "The slide contains a list of bullet points describing various insurance processes that can be improved with GPT.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.430366+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 37, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 38, "type": "analysis", "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": 51, "componentCount": 7, "textChars": 473, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.629+00", "seconds": 2.586, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/38", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-38", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.35, "w": 0.28, "x": 0.037, "y": 0.205 }, "kind": "list", "text": "GPT-SPECIFIC LIMITATIONS: GPT-3 is proprietary; Insurance data privacy; Need for additional layers for output conversion.", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "47171df3-e306-4ac4-93f3-af1d94b9e435", "parentComponentId": null }, { "bbox": { "h": 0.55, "w": 0.28, "x": 0.357, "y": 0.205 }, "kind": "list", "text": "CHATGPT LIMITATIONS: Confidently wrong; Sensitivity to phrasing; Lack of clarifying questions; Potential for harmful/biased output; Supervision needed.", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "5042af3b-6321-4a9f-b846-4eb352faf8be", "parentComponentId": null }, { "bbox": { "h": 0.45, "w": 0.28, "x": 0.677, "y": 0.205 }, "kind": "list", "text": "GENERAL LIMITATIONS OF LLMS: Computationally expensive; Black box/explainability; Bias perpetuation; High technical expertise required.", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "86d6cab8-2fcc-4a03-884a-8c5f077d463c", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.02, "x": 0.95, "y": 0.955 }, "kind": "paragraph", "text": "38", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "16b37988-8253-48d5-af99-a7d2e961583a", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.2, "x": 0.037, "y": 0.915 }, "kind": "paragraph", "text": "Source: https://openai.com/blog/chatgpt", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "38056fdd-aef4-4285-a4dd-0530d981805e", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.08, "x": 0.037, "y": 0.955 }, "kind": "paragraph", "text": "© Oliver Wyman", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c713c94a-b281-442f-8e12-bc38e66e512f", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.17, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "LIMITATIONS", "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 18:38:15.629+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "39aebbc9-6bf3-4013-8dc4-14b1f1a0400d", "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": "b841dd85-8031-426d-86c8-723e98dfe11a", "evidence": "The slide contains a list of bullet points describing limitations of CHATGPT and GENERAL LIMITATIONS OF LLMS.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.507405+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 38, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 39, "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": 5, "componentCount": 3, "textChars": 17, "nDataPoints": 0, "notes": null, "elementsJson": [ "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:19.853+00", "seconds": 6.736, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/39", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-39", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 1, "w": 0.254, "x": 0.746, "y": 0 }, "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 18:38:19.853+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1965cb87-3731-46c0-ac55-dfec7b7e0391", "parentComponentId": null }, { "bbox": { "h": 0.066, "w": 0.315, "x": 0.037, "y": 0.638 }, "kind": "title", "text": "AI: SOFTWARE 2.0", "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 18:38:19.853+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "20420358-6cf0-43d1-8854-b7675ac82592", "parentComponentId": null }, { "bbox": { "h": 0.2, "w": 0.112, "x": 0.037, "y": 0.382 }, "kind": "title", "text": "4", "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 18:38:19.853+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "ff78d67a-1880-4826-99b2-7da4fee3a619", "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": "570c0a7e-650c-46de-8b07-546c0c0b5b44", "evidence": "The slide is titled '4 AI: SOFTWARE 2.0' and has a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.646187+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 39, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [ { "to": 39, "from": 34, "loop": { "name": "47_quick_win_big_bet", "slug": "47-quick-win-big-bet", "status": "active", "bestFor": "Transformation planning, 100-day plans, resource allocation", "canonId": "019dd956-790d-708d-983e-ef9bbc68f741", "version": 1, "bodyDocId": "019df22a-2420-77be-bc41-ded96d08cb21", "structure": "The Full List -> Quick Wins (Low effort, High impact) -> Big Bets (High effort, High impact) -> Sequenced Roadmap", "description": "Separate initiatives into immediate wins and longer-term strategic bets", "familyLabel": null }, "matchId": "84ae78d8-256a-4b4e-9275-bcfe50513445", "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "position": 2, "objective": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?", "confidence": 0.7, "extraction": { "at": "2026-07-16 21:13:57.761457+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": "doc-narrative-v1" } } ], "locked": true }, { "page": 40, "type": "setup", "function": "establish_context", "rawType": "quote_slide", "block": null, "metadata": { "slideType": "quote_slide", "slideTypeCanon": { "name": "Quote slide", "slug": "quote_slide", "status": "active", "canonId": "019de52d-098e-7555-a9fd-4f0af7e80bc2", "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": 76, "componentCount": 8, "textChars": 1644, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "quote_block", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:21.64+00", "seconds": 8.068, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/40", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-40", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.05, "w": 0.1, "x": 0.52, "y": 0.2 }, "kind": "callout", "text": "HOWEVER...", "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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "816a979c-5638-4c56-991c-d1e7e23f7541", "parentComponentId": null }, { "bbox": null, "kind": "callout", "text": "We think actuaries are ideally suited to lead this 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": "2026-04-24 13:39:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-a17026dc74d1", "parentComponentId": null }, { "bbox": { "h": 0.3, "w": 0.4, "x": 0.52, "y": 0.3 }, "kind": "list", "text": "Applying AI to insurance and actuarial problems does not happen automatically.\nThere is a lot of work needed to convert our industry to Software 2.0.\nWho will do this work?\nWe think actuaries are ideally suited to lead this 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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "d6de2153-b714-465d-9658-c9a3a136f7e5", "parentComponentId": null }, { "bbox": { "h": 0.1, "w": 0.2, "x": 0.25, "y": 0.75 }, "kind": "paragraph", "text": "ANDREJ KARPATHY\nFounding Member of OpenAI\nFormer Director of AI at Tesla", "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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "52865bb1-61d8-490e-99ae-e0b9da420d2c", "parentComponentId": null }, { "bbox": { "h": 0.4, "w": 0.4, "x": 0.05, "y": 0.25 }, "kind": "quote", "text": "The “classical stack” of Software 1.0 is what we’re all familiar with — it is written in languages such as Python, C++, etc. It consists of explicit instructions to the computer written by a programmer. By writing each line of code, the programmer identifies a specific point in program space with some desirable behavior. [...] In contrast, Software 2.0 is written in much more abstract, human unfriendly language, such as the weights of a neural network. [...] Software (1.0) is eating the world, and now AI (Software 2.0) is eating software.", "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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "2667618f-2e24-4eb5-af02-c59d5d3862b2", "parentComponentId": null }, { "bbox": null, "kind": "quote", "text": "The “classical stack” of Software 1.0 is what we’re all familiar with — it is written in languages such as Python, C++, etc. It consists of explicit instructions to the computer written by a programmer. By writing each line of code, the programmer identifies a specific point in program space with some desirable behavior. [...] In contrast, Software 2.0 is written in much more abstract, human unfriendly language, such as the weights of a neural network. [...] Software (1.0) is eating the world, and now AI (Software 2.0) is eating software. — Andrej Karpathy, Founding Member of OpenAI, Former Director of AI at Tesla", "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": "2026-04-24 13:39:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-a6e2f16d68fd", "parentComponentId": null }, { "bbox": { "h": 0.02, "w": 0.3, "x": 0.05, "y": 0.92 }, "kind": "source-note", "text": "Source: https://karpathy.medium.com/software-2-0-a64152b37c35", "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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "38f5259c-3d3e-47fd-a900-466b517449ce", "parentComponentId": null }, { "bbox": { "h": 0.05, "w": 0.6, "x": 0.05, "y": 0.05 }, "kind": "title", "text": "SOFTWARE IS EATING THE WORLD, AI IS EATING 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 18:38:21.64+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "6ed84232-0eed-4636-ac20-cf0425121665", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 41, "type": "setup", "function": "present_framework", "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": "present_framework", "functionCanon": { "name": "Present framework", "slug": "present_framework", "status": "active", "canonId": "019de52d-1403-7667-9100-d2ac6709f750", "version": 1, "description": null }, "density": "overcrowded", "densityScore": 83, "componentCount": 10, "textChars": 1430, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "icon_grid", "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.814+00", "seconds": 4.235, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/41", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-41", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Put together, these best practices ensure that code will remain easy to understand and maintain over 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": "2026-04-24 13:39:32+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ee-224aa2af10b4", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.35, "y": 0.203 }, "kind": "list", "text": "Agility: Learn to \"fail fast\" and adapt. It is essential to interact frequently with end users and adjust the trajectory based on their feedback.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "04e869ac-04c6-4141-bf08-c1edbb6cae72", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.35, "y": 0.51 }, "kind": "list", "text": "Modularity: Reduce work duplication. Maximize code understandability and reusability by spreading functionalities into independent components.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "200b7138-3cb0-46c2-94be-735aa20ae876", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.037, "y": 0.203 }, "kind": "list", "text": "Design: \"Simplicity is the ultimate sophistication\". One should fall in love with the problem rather than any given solution. Once the problem is understood, drafts should be presented to users before rushing to the development phase.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "778f491c-d89a-484e-aa40-8571fa45cc6b", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.037, "y": 0.51 }, "kind": "list", "text": "Version Control: Keep track of all changes. Allow collaborative development by tracking changes of individual contributors and setting frameworks for integration.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "81ce64ec-c88e-45ca-9067-3698be370463", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.65, "y": 0.203 }, "kind": "list", "text": "Testing: Foresee bugs and defects before users. Automatic and timely testing of the whole code base for compliance with expected behavior should be in place.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "9be88635-95e2-43b4-81b2-ff0eaddc35a3", "parentComponentId": null }, { "bbox": { "h": 0.35, "w": 0.31, "x": 0.65, "y": 0.51 }, "kind": "list", "text": "Continuous Integration: Scale the collaboration. Frequent integration of all new code that compose the application, leveraging automated testing and building functionalities.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "fad9d50a-797e-43fe-91de-b667fc0adae2", "parentComponentId": null }, { "bbox": { "h": 0.06, "w": 0.92, "x": 0.037, "y": 0.85 }, "kind": "paragraph", "text": "Put together, these best practices ensure that code will remain easy to understand and maintain over time. It makes it easier to implement new functionality and integrate new technologies.", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "10f44433-6083-4204-8b76-ccb47c8b13d8", "parentComponentId": null }, { "bbox": { "h": 0.038, "w": 0.227, "x": 0.037, "y": 0.055 }, "kind": "title", "text": "WHERE TO START?", "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "cf916b2e-dd08-43aa-8747-c4f7802c5b75", "parentComponentId": null }, { "bbox": { "h": 0.033, "w": 0.881, "x": 0.037, "y": 0.108 }, "kind": "title", "text": "Modern software development practices are the foundation; Actuaries can learn a lot from the software 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": "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 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "769a6265-234f-4f36-aa5b-17e90a01ffd2", "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": "74601f38-4dc3-4c84-b0a2-76c2f6f07293", "evidence": "The slide contains a list of bullet points describing best practices for software development.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.758102+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [ { "matchId": "5721622c-9046-40e6-b139-91a7f9f3b1ec", "evidence": "The slide explicitly lists six core software engineering principles applied to actuarial work.", "framework": null, "confidence": 0.9, "extraction": { "at": "2026-05-02 18:38:17.814+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 41, "frameworkName": "Software Development Lifecycle (SDLC) Best Practices" } ], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 42, "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": 2, "textChars": 7, "nDataPoints": 0, "notes": null, "elementsJson": [ "photo" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.223+00", "seconds": 1.589, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/42", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-42", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 1, "w": 0.253, "x": 0.747, "y": 0 }, "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 18:38:15.223+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3bf76b32-b475-4a9c-a79d-6735b11ba515", "parentComponentId": null }, { "bbox": { "h": 0.305, "w": 0.105, "x": 0.037, "y": 0.38 }, "kind": "title", "text": "5 RECAP", "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 18:38:15.223+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "c404e49e-431e-40fe-afce-848b53364cf6", "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": "6488c285-416d-4605-9649-4a371a3eabe5", "evidence": "The slide is titled '5 RECAP' and has a section divider.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.832513+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 42, "whyItWorks": null, "antipattern": null, "cardinality": "common", "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 43, "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": 32, "componentCount": 3, "textChars": 715, "nDataPoints": 0, "notes": null, "elementsJson": [ "headline_text", "bullet_list" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.187+00", "seconds": 3.474, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/43", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-43", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": null, "kind": "callout", "text": "Converting our industry to a Software 2.0 world will require a lot of work. Actuaries are well suited to lead this work but need to modernize their skillset.", "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": "2026-04-24 13:39:21+00", "model": "or:google/gemini-3.1-flash-lite-preview", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "019dd951-d9f1-764a-82ec-d31fdd545d94", "parentComponentId": null }, { "bbox": { "h": 0.6, "w": 0.9, "x": 0.05, "y": 0.2 }, "kind": "list", "text": "We've seen exponential growth in the complexity of machine learning models, which is largely attributable to the use of deep learning techniques.\nTransformer models, including GPTs, have resulted in breakthrough performance on NLP tasks; the process of \"self-attention\" has been pivotal to this breakthrough.\nThese breakthroughs impact all fields of work, including insurance and actuarial work.\nConverting our industry to a Software 2.0 world will require a lot of work. Actuaries are well suited to lead this work but need to modernize their skillset.", "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 18:38:17.187+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "93914a5e-7e0e-4d6f-9b3e-91ad85233042", "parentComponentId": null }, { "bbox": { "h": 0.04, "w": 0.08, "x": 0.037, "y": 0.05 }, "kind": "title", "text": "RECAP", "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 18:38:17.187+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "1cb35bce-e5b3-418a-be22-0ee45ba5a536", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 44, "type": "setup", "function": "state_next_steps", "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": "state_next_steps", "functionCanon": { "name": "State next steps", "slug": "state_next_steps", "status": "active", "canonId": "019de52d-14ca-70ac-b534-971f458cdf29", "version": 1, "description": null }, "density": "sparse", "densityScore": 1, "componentCount": 2, "textChars": 10, "nDataPoints": 0, "notes": null, "elementsJson": [ "other" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:15.54+00", "seconds": 1.625, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/44", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-44", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.093, "w": 0.052, "x": 0.474, "y": 0.404 }, "kind": "image", "text": "?", "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 18:38:15.54+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8e3ba715-8c7a-44c2-ac26-5cdc32641240", "parentComponentId": null }, { "bbox": { "h": 0.056, "w": 0.194, "x": 0.403, "y": 0.528 }, "kind": "title", "text": "QUESTIONS", "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 18:38:15.54+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "3c3d34d3-875f-4b2d-af1b-9f722f49eb8a", "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": "9633a882-7206-4807-b084-84a73b747035", "evidence": "The slide is titled 'QUESTIONS' and appears to be a call to action for questions.", "pageRefs": null, "priority": null, "whenToUse": null, "confidence": 0.9, "extraction": { "at": "2026-07-16 21:34:27.945107+00", "model": "or:meta-llama/llama-4-scout", "runId": null, "seconds": null, "promptVersion": null }, "pageNumber": 44, "whyItWorks": null, "antipattern": null, "cardinality": null, "narrativePurpose": null } ], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 45, "type": "setup", "function": "front_matter", "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": "front_matter", "functionCanon": { "name": "Front matter", "slug": "front_matter", "status": "active", "canonId": "019de52d-133d-73f0-9e26-201eda96b098", "version": 1, "description": null }, "density": "balanced", "densityScore": 40, "componentCount": 2, "textChars": 1853, "nDataPoints": 0, "notes": null, "elementsJson": [ "paragraph" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:17.565+00", "seconds": 3.069, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/45", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-45", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.438, "w": 0.928, "x": 0.036, "y": 0.331 }, "kind": "disclaimer", "text": "This report is for the exclusive use of the Oliver Wyman client named herein. This report is not intended for general circulation or publication, nor is it to be reproduced, quoted, or distributed for any purpose without the prior written permission of Oliver Wyman. There are no third-party beneficiaries with respect to this report, and Oliver Wyman does not accept any liability to any third party.\n\nInformation furnished by others, upon which all or portions of this report are based, is believed to be reliable but has not been independently verified, unless otherwise expressly indicated. Public information and industry and statistical data are from sources we deem to be reliable; however, we make no representation as to the accuracy or completeness of such information. The findings contained in this report may contain predictions based on current data and historical trends. Any such predictions are subject to inherent risks and uncertainties. Oliver Wyman accepts no responsibility for actual results or future events.\n\nThe opinions expressed in this report are valid only for the purpose stated herein and as of the date of this report. No obligation is assumed to revise this report to reflect changes, events, or conditions, which occur subsequent to the date hereof.\n\nAll decisions in connection with the implementation or use of advice or recommendations contained in this report are the sole responsibility of the client. This report does not represent investment advice nor does it provide an opinion regarding the fairness of any transaction to any and all parties. In addition, this report does not represent legal, medical, accounting, safety, or other specialized advice. For any such advice, Oliver Wyman recommends seeking and obtaining advice from a qualified professional.", "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 18:38:17.565+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "b9123033-f5cf-4d66-afc9-03b6e824556d", "parentComponentId": null }, { "bbox": { "h": 0.031, "w": 0.602, "x": 0.036, "y": 0.271 }, "kind": "title", "text": "QUALIFICATIONS, ASSUMPTIONS, AND LIMITING CONDITIONS", "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 18:38:17.565+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "71836dad-2ed6-4ab5-9cc4-891589452d38", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true }, { "page": 46, "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": "sparse", "densityScore": 3, "componentCount": 2, "textChars": 39, "nDataPoints": 0, "notes": null, "elementsJson": [ "logo_grid" ], "confidence": 1, "extraction": { "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "at": "2026-05-02 18:38:16.395+00", "seconds": 1.597, "promptVersion": "v2-canon-conditioned" } }, "links": { "slide": "/slides/019dd923-5de0-76bd-a168-6159b949e2c2/46", "deck": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2", "deckAnchor": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2#slide-46", "deckJson": "/decks/019dd923-5de0-76bd-a168-6159b949e2c2.json" }, "components": [ { "bbox": { "h": 0.074, "w": 0.388, "x": 0.306, "y": 0.463 }, "kind": "image", "text": "OliverWyman", "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 18:38:16.395+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "234efdf7-c4d2-4e90-ac71-d93b8beb972c", "parentComponentId": null }, { "bbox": { "h": 0.023, "w": 0.186, "x": 0.407, "y": 0.907 }, "kind": "paragraph", "text": "A business of Marsh McLennan", "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 18:38:16.395+00", "model": "google/gemini-3.1-flash-lite-preview-20260303", "runId": null, "seconds": null, "promptVersion": null }, "componentId": "8573e234-67e5-48ec-8c21-2c622fc87bcb", "parentComponentId": null } ], "metrics": [], "tools": [], "frameworks": [], "arcBeats": [ { "to": 46, "arc": { "name": "The Sparkline", "slug": "sparkline", "status": "active", "bestFor": "Keynotes, TED-style talks", "canonId": "019dd956-606e-720c-905a-42607ac0bf50", "version": 1, "bodyDocId": "019df22a-2356-73eb-8061-d8c81dc0aa49", "structure": "What Is -> What Could Be -> What Is -> What Could Be -> New Bliss", "description": null, "familyLabel": null }, "from": 32, "name": "New Bliss", "beatId": "28d56390-4e46-4fc5-8628-c83b99b85901", "matchId": "e228d300-7362-48a1-a07a-17da0e1f65e9", "beatType": { "name": "New Bliss", "slug": "sparkline-new-bliss", "status": "active", "canonId": "019dd9b8-2879-72f7-8968-bcfac0921b99", "version": 1, "description": null }, "evidence": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape.", "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": 0.8 } ], "loops": [], "locked": true } ], "arcBeats": [ { "from": 1, "to": 5, "label": "What Is", "description": "The deck introduces GPT-3 and NLP, providing context and background information." }, { "from": 6, "to": 16, "label": "What Is", "description": "The deck delves deeper into the technical aspects of GPT-3 and NLP." }, { "from": 17, "to": 31, "label": "What Could Be", "description": "The deck explores the potential applications of GPT-3 in the actuarial field." }, { "from": 32, "to": 46, "label": "New Bliss", "description": "The deck discusses the potential benefits and opportunities of GPT-3 and other AI solutions in the actuarial landscape." } ], "loops": [ { "from": 5, "to": 10, "label": "Golden Circle", "description": "Why should we care about GPT-3 and NLP?" }, { "from": 17, "to": 25, "label": "Mece Breakdown", "description": "How can GPT-3 be applied to actuarial work?" }, { "from": 34, "to": 39, "label": "Quick Win Big Bet", "description": "What are the potential benefits of GPT-3 and other AI solutions in the actuarial landscape?" } ] }