{
  "docId": "019dd923-5e88-73ef-bd5d-0856e1444fb9",
  "docSlug": "8757f1b44ef7f176",
  "documentTitle": "2024 Air Street Capital The State of AI Report 2024",
  "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": 44,
  "pageCount": 213,
  "prevPage": 43,
  "nextPage": 45,
  "slideType": "client_example",
  "function": "illustrate_case",
  "density": "overcrowded",
  "nDataPoints": 4,
  "notes": "Part of the State of AI 2024 report.",
  "elementsJson": [
    "bullet_list",
    "photo"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-0856e1444fb9/44",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9#slide-44",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "The core of Movie Gen is a 30B video generation and a 13B audio generation model, capable of producing 16-second videos at 16 frames per second and 45-second audio clips respectively.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c3-7407-8d11-a1a507b3e958",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.3,
        "w": 0.32,
        "x": 0.66,
        "y": 0.55
      },
      "kind": "image",
      "text": "AI-generated video still of a person performing fire poi on a beach.",
      "attrs": null,
      "subkind": "photo",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "508f3de9-d9b2-44b4-8a30-906e0a696acb",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.65,
        "x": 0.02,
        "y": 0.5
      },
      "kind": "list",
      "text": "Movie Gen's video editing capabilities combine advanced image editing techniques with video generation, allowing for both localized edits and global changes while preserving original content.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "16e5b5a4-6f64-49bf-a09b-355d28a083f6",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.65,
        "x": 0.02,
        "y": 0.42
      },
      "kind": "list",
      "text": "These models leverage joint optimization techniques for text-to-image and text-to-video tasks, as well as novel audio extension methods for generating coherent audio for videos of arbitrary lengths.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "2791dffe-b180-4502-ad8b-7196dde751f2",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.65,
        "x": 0.02,
        "y": 0.68
      },
      "kind": "list",
      "text": "Meta used A/B human evaluation comparisons to demonstrate positive net win rates against competing industry models across their four main capabilities. The researchers say they intend to make the model available in future, but don't commit to a timeline or release strategy.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "5888d629-eaf1-4521-a725-ecddbb397b94",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.65,
        "x": 0.02,
        "y": 0.3
      },
      "kind": "list",
      "text": "The core of Movie Gen is a 30B video generation and a 13B audio generation model, capable of producing 16-second videos at 16 frames per second and 45-second audio clips respectively.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "63bef952-3981-4f7a-afee-55984ed9de7d",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.06,
        "w": 0.65,
        "x": 0.02,
        "y": 0.62
      },
      "kind": "list",
      "text": "The models were trained on a combination of licensed and publicly available datasets.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "6a0cdbf2-0926-427b-b4ba-169a9ab75349",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.08,
        "w": 0.95,
        "x": 0.02,
        "y": 0.21
      },
      "kind": "paragraph",
      "text": "Keeping the gated approach of other labs, Meta has brought together its work on different modalities via the Make-A-Scene and Llama families to build Movie Gen.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "488157c4-e80b-45c2-a1b1-71ec86dadc92",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.5,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Meta goes even further, throwing audio into the mix",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "64040484-e547-4b76-a9b9-7188830a8ccb",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 153,
      "from": 9,
      "beatId": "019dd95a-0682-776c-8e35-5f2398b8d1d0",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Exec summary + Research + Industry sections inventory the year",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 85,
      "from": 12,
      "beatId": "019dd95a-0682-776c-8e35-6d9d8ee06d0a",
      "arcName": "Voyage and Return",
      "arcSlug": "voyage-return",
      "beatName": "The Unknown",
      "beatSlug": "voyage-return-the-unknown",
      "evidence": "Research section explores frontier model uncharted territory",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Complication",
      "parentBeatSlug": "complication"
    }
  ],
  "loops": [
    {
      "to": 44,
      "from": 40,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3e-3752140f3c2d",
      "evidence": "VLMs, diffusion image, Stable Video Diffusion, gated text-to-video, Movie Gen audio",
      "position": 6,
      "objective": "Map progress in multimodal vision/video/audio generation",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 76,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}