{
  "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": 51,
  "pageCount": 163,
  "prevPage": 50,
  "nextPage": 52,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 0,
  "notes": "The slide highlights the shift from pure generation to instruction-based editing using models like InstructPix2Pix and Imagen Editor.",
  "elementsJson": [
    "headline_text",
    "paragraph",
    "bullet_list",
    "photo"
  ],
  "metadataConfidence": 0.95,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-06b04d219fea/51",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-06b04d219fea#slide-51",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "This year has seen new methods enabling co-pilot style capability for image generation and editing.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a225-1baf09de3671",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.4,
        "x": 0.57,
        "y": 0.35
      },
      "kind": "image",
      "text": "Visual examples of image editing: sunflowers to roses, Eiffel tower with fireworks, street scene with snow, character style transfer.",
      "attrs": null,
      "subkind": "illustration",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "6174d7f5-eac5-4227-85e0-831694db29e7",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.55,
        "x": 0.02,
        "y": 0.35
      },
      "kind": "list",
      "text": "InstructPix2Pix, leverages pre-trained GPT3 and StableDiffusion to generate a large dataset of {input image, text instruction, generated image} triplets to train a supervised conditional diffusion model. Editing then happens in a feed-forward way without any per image fine tuning/inversion, enabling modifications in seconds.\nMasked inpainting methods such as Imagen Editor require providing the model with an overlay or “mask” to indicate the region to modify, alongside text instructions.\nBuilding on these approaches, startups such as Genmo AI’s “Chat” provide a co-pilot style interface for image generation with text-guided semantic editing.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "7f4d9764-c174-49e5-8b8d-7d5062b39da0",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.96,
        "x": 0.02,
        "y": 0.2
      },
      "kind": "paragraph",
      "text": "Last year saw the emergence of a host of text-image generation models: DALLE-2, Imagen, Parti, Midjourney, Stability and more. But controlling the generation requires experimenting extensively with prompts and custom syntax. This year has seen new methods enabling co-pilot style capability for image generation and editing.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "99ec6405-111e-44e9-9fc0-7eee5c56cdcd",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.65,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Instruction based editing assistants for text-image generation",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "fe5b57d1-975a-4857-8aa3-844b1dee7497",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "List presentation",
      "slug": "list-presentation",
      "agent": null,
      "layer": "slide",
      "matchId": "8f137a32-f7e2-4197-851a-e0c87db77fce",
      "evidence": "list/bullet: InstructPix2Pix, leverages pre-trained GPT3 and StableDiffusion to generate a large dataset of {input image, text instruction, generated image} triplets to train a supervised conditional diffusion model.",
      "confidence": 0.8
    }
  ],
  "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": 58,
      "from": 50,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3d-f70f6435b722",
      "evidence": "Text-to-video, instruction editing, Gaussian splatting, NeRFs, depth, SAM, DINOv2, weather, music.",
      "position": 6,
      "objective": "Generative media and visual foundation models proliferate",
      "structure": "Evidence A -> Evidence B -> Evidence C -> Pattern/Conclusion",
      "confidence": 72,
      "description": "Group multiple pieces of evidence that together point to a pattern or conclusion"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}