{
  "docId": "019dd923-5e88-73ef-bd5c-f341d4394195",
  "docSlug": "46f66c49fd159048",
  "documentTitle": "2018 Air Street Capital The State of AI Report 2018",
  "authorId": "AirStreetCapital",
  "authorName": "Air Street Capital",
  "documentKindSlug": "consulting-deck",
  "documentKindLabel": "Consulting deck",
  "sourceTypeSlug": "vc_research",
  "sourceTypeLabel": "VC research",
  "presentationDate": null,
  "orientation": "landscape",
  "aspectRatio": 1.777,
  "pageNumber": 11,
  "pageCount": 156,
  "prevPage": 10,
  "nextPage": 12,
  "slideType": "industry_trends",
  "function": "analyze_data",
  "density": "dense",
  "nDataPoints": 15,
  "notes": "The slide uses two charts to demonstrate the scaling efficiency of GPU clusters in AI model training.",
  "elementsJson": [
    "headline_text",
    "bar_chart_vertical",
    "line_chart",
    "callout_box"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5c-f341d4394195/11",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5c-f341d4394195#slide-11",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "The hardware war: More GPUs allows for faster training, as well as bigger (more powerful) models.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a80c-75bd65449c4b",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.5,
        "w": 0.45,
        "x": 0.05,
        "y": 0.3
      },
      "kind": "chart",
      "text": "Training speedup over one GPU (x-fold)",
      "attrs": null,
      "subkind": "bar-vertical",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "0db9fe69-26af-40ee-9239-a8268898d5fe",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.5,
        "w": 0.45,
        "x": 0.52,
        "y": 0.3
      },
      "kind": "chart",
      "text": "32 GPUs = same accuracy, 25x faster",
      "attrs": null,
      "subkind": "line",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "f4bc794e-6cd2-4850-95dd-717e81666b05",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Training speedup: 25x",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a80c-7b579b54ab15",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.95,
        "x": 0.02,
        "y": 0.92
      },
      "kind": "source-note",
      "text": "Baidu Google AI stateof.ai 2018",
      "attrs": null,
      "subkind": null,
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "829dc2e0-20e4-4dc6-8dd8-fb7d8bf16990",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.03,
        "w": 0.85,
        "x": 0.02,
        "y": 0.22
      },
      "kind": "title",
      "text": "The hardware war: More GPUs allows for faster training, as well as bigger (more powerful) models.",
      "attrs": null,
      "subkind": "action-title",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "6d6c5094-80de-43fc-9ce1-50294ea6f4d5",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.35,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "AI hardware as the new frontier",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "0c75a12f-3149-440c-aa5d-17b90c231661",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "019dd95a-0fd5-7148-8eca-5078758dbbcd",
      "evidence": "Headline 'The hardware war: More GPUs allows for faster training'.",
      "confidence": 85
    },
    {
      "name": "Action Titles",
      "slug": "action-titles",
      "agent": "Architect",
      "layer": "slide",
      "matchId": "6a394c96-7517-489c-9aa0-abb752d9c208",
      "evidence": "title/action-title: The hardware war: More GPUs allows for faster training, as well as bigger (more powerful) models.",
      "confidence": 0.8
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 69,
      "from": 4,
      "beatId": "019dd95a-0682-776c-8e34-ad4df4fe3ce7",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions then research breakthroughs (transfer learning, hardware, RL) and talent supply data.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 55,
      "from": 10,
      "beatId": "019dd95a-0682-776c-8e34-be65c7627fe7",
      "arcName": "The Onion",
      "arcSlug": "onion",
      "beatName": "First Layer",
      "beatSlug": "onion-first-layer",
      "evidence": "Hardware, vision, RL, bias - technical research layer.",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 21,
      "from": 10,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3c-4a0086ef4670",
      "evidence": "Twelve contiguous slides on GPU growth, Moore's Law, new architectures, hourly cost, ending in TPUv2 vs V100 cost case.",
      "position": 2,
      "objective": "Stack evidence that AI hardware is the binding constraint",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 78,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}