{
  "docId": "019dd923-5e88-73ef-bd5d-06b04d219fea",
  "docSlug": "dd91c78f6570bf29",
  "documentTitle": "2023 Air Street Capital The State of AI Report 2023",
  "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": 27,
  "pageCount": 163,
  "prevPage": 26,
  "nextPage": 28,
  "slideType": "case_study",
  "function": "illustrate_case",
  "density": "dense",
  "nDataPoints": 4,
  "notes": "The slide uses a visual comparison (meme vs. truth) to explain the shift from dense models to Mixture of Experts (MoE) architectures.",
  "elementsJson": [
    "headline_text",
    "paragraph",
    "bullet_list",
    "other"
  ],
  "metadataConfidence": 0.95,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-06b04d219fea/27",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea#slide-27",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "If the rumours are to be believed, no fundamental innovation underpins GPT-4's success.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a221-0421f81c8e41",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.28,
        "w": 0.52,
        "x": 0.24,
        "y": 0.63
      },
      "kind": "image",
      "text": "Comparison of GPT-3 and GPT-4 architectures (meme vs truth)",
      "attrs": null,
      "subkind": "illustration",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "181def2b-90aa-4b96-80d3-123d3e0ac88c",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.9,
        "x": 0.05,
        "y": 0.43
      },
      "kind": "list",
      "text": "The tiny corp founder George Hotz presented the most plausible rumour: “Sam Altman won’t tell you that GPT-4 has 220B parameters and is a 16-way mixture model with 8 sets of weights”, and Soumith Chintala, PyTorch co-founder, confirmed. Neither the total size of the model nor using a Mixture of Experts model is unheard of. If the rumours are to be believed, no fundamental innovation underpins GPT-4’s success.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5ddf9329-2f63-49c8-a386-9232d196e36e",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "tokens: 13 trillion",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a221-0d28e7bddf0a",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.2,
        "w": 0.9,
        "x": 0.05,
        "y": 0.22
      },
      "kind": "paragraph",
      "text": "In 2022, we predicted: “A SOTA LM is trained on 10x more data points than Chinchilla, proving data-set scaling vs. parameter scaling”. Although OpenAI didn’t confirm – and we probably won’t know anytime soon – a sort of consensus seems to be reached among experts about leaked information on the model size, architecture, and the dollar cost of GPT-4. GPT-4 was reportedly trained on ~13 trillion tokens, 9.3x more tokens than Chinchilla.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "7842376d-d99a-4f1c-9a93-19a587be5755",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "quote",
      "text": "Sam Altman won't tell you that GPT-4 has 220B parameters and is a 16-way mixture model with 8 sets of weights",
      "attrs": null,
      "subkind": null,
      "toolName": "Authority citation",
      "toolSlug": "authority-citation",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a221-0a0a7608e027",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.8,
        "x": 0.05,
        "y": 0.13
      },
      "kind": "title",
      "text": "2022 Prediction: language models trained on huge amounts of data",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "87d011c6-fab3-4b59-afdf-b21862da7634",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Aha! Moment",
      "slug": "aha-moment",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "7521af47-90ce-4847-b850-9eefeb78b3a7",
      "evidence": "In 2022, we predicted: “A SOTA LM is trained on 10x more data points than Chinchilla, proving data-set scaling vs. parameter scaling”.",
      "confidence": 0.7
    },
    {
      "name": "Call to action",
      "slug": "call-to-action",
      "agent": null,
      "layer": "slide",
      "matchId": "3b649eeb-4c9b-4a0e-af3d-e8672efb91b8",
      "evidence": "If the rumours are to be believed, no fundamental innovation underpins GPT-4's success.",
      "confidence": 0.6
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 120,
      "from": 11,
      "beatId": "019dd95a-0682-776c-8e35-41afd44ef59f",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Research + Industry sections inventory model, compute, funding facts.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 120,
      "from": 11,
      "beatId": "019dd95a-0682-776c-8e35-523bfb7f96e6",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Rising Action",
      "beatSlug": "mountain-rising-action",
      "evidence": "Escalating capabilities, compute concentration and capital flows.",
      "position": 2,
      "confidence": 45,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 31,
      "from": 22,
      "name": "Zoom In",
      "slug": "06-zoom-in",
      "bestFor": "Technical deep-dives, case studies, detailed analysis",
      "matchId": "019dd95a-07fe-70ce-8d3d-ea04ad01cb62",
      "evidence": "Slides drill from emergent-capability mirage into context-length, memory and data-exhaustion details.",
      "position": 3,
      "objective": "Zoom into the data and context-length bottleneck",
      "structure": "The Big Picture -> Key Area of Focus -> Specific Detail -> Implication",
      "confidence": 72,
      "description": "Start broad, then progressively focus on specific details that prove your point"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}