{
  "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": 79,
  "pageCount": 213,
  "prevPage": 78,
  "nextPage": 80,
  "slideType": "case_study",
  "function": "illustrate_case",
  "density": "dense",
  "nDataPoints": 1,
  "notes": "The slide highlights a specific technical solution (HumanPlus) to a known industry bottleneck (data for imitation learning).",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "photo"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-0856e1444fb9/79",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-0856e1444fb9#slide-79",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "The system demonstrates impressive results on a variety of tasks, including complex actions like wearing a shoe and walking, using only up to 40 demonstrations.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c3-7407-8d15-738633df171f",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.3,
        "x": 0.68,
        "y": 0.4
      },
      "kind": "image",
      "text": "Visual collage of HumanPlus robot performing tasks (shadowing and autonomous skills).",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "dc5faf43-c698-416b-8f3b-01e3a3df8177",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.65,
        "x": 0.02,
        "y": 0.33
      },
      "kind": "list",
      "text": "HumanPlus is a full-stack system for humanoids to learn from human data. It combines a real-time shadowing system and an imitation learning algorithm.\nThe shadowing system uses a single RGB camera and a low-level policy to allow human operators to control the humanoid's whole body in real-time. This low-level control policy is trained on a large dataset of human motion data in simulation and transfers to the real world without additional training.\nThe imitation learning component enables efficient learning of autonomous skills from shadowing data. It uses binocular egocentric vision and combines action prediction with forward dynamics prediction.\nThe system demonstrates impressive results on a variety of tasks, including complex actions like wearing a shoe and walking, using only up to 40 demonstrations.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "8b80f212-51a6-4e5e-aeae-c96bdeafa37c",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.6,
        "x": 0.02,
        "y": 0.02
      },
      "kind": "other",
      "text": "Introduction | Research | Industry | Politics | Safety | Predictions",
      "attrs": null,
      "subkind": "unclassified",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3293961a-0aea-465c-9a3b-63493c982959",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.1,
        "w": 0.65,
        "x": 0.02,
        "y": 0.22
      },
      "kind": "paragraph",
      "text": "It's challenging to model the intricacies of human behavior with imitation learning, which relies on human demonstrators. While effective, it's difficult to implement at scale. Stanford has some workarounds.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "e7400128-1ffc-429b-b657-7b4171280eb4",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.5,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Can we overcome the data bottleneck for humanoids?",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "1cb200a8-fe00-4067-99bc-d2fd629f8a37",
      "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": 81,
      "from": 72,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3e-431790a1bfd7",
      "evidence": "Wayve LINGO-2, SAM2, DeepMind robotics, LeRobot, diffusion policies, humanoids, Vision Pro",
      "position": 9,
      "objective": "Show robotics renaissance via converging evidence",
      "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
}