{
  "docId": "019dd923-5e88-73ef-bd5d-01cf0d8a8fbc",
  "docSlug": "cc8183ec02431b7a",
  "documentTitle": "2022 Air Street Capital The State of AI Report 2022",
  "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": 37,
  "pageCount": 114,
  "prevPage": 36,
  "nextPage": 38,
  "slideType": "client_example",
  "function": "illustrate_case",
  "density": "dense",
  "nDataPoints": 4,
  "notes": "Includes visual diagrams of the SayCan decision process and success metrics.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "other"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc/37",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-01cf0d8a8fbc#slide-37",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "Researchers tested SayCan on 101 instructions from 7 types of language instructions. It was successful in planning and execution 84% and 74% of the time respectively.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a217-35bb597c80ca",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.2,
        "x": 0.78,
        "y": 0.35
      },
      "kind": "image",
      "text": "Diagrams showing the SayCan decision tree and task execution examples.",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "dcc37f80-e288-428e-a1a5-925072bc3e92",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.55,
        "w": 0.75,
        "x": 0.02,
        "y": 0.35
      },
      "kind": "list",
      "text": "Given an ambiguous instruction “I spilled my drink, can you help?”, a carefully prompt-engineered LLM (e.g. Google’s PaLM) can devise a sequence of abstract steps to pick up and bring you a sponge. But any given skill (e.g. pick up, put down) needs to be doable by the robot in concordance with its environment (e.g. robot sees a sponge).\nTo incentivise the LLM to output feasible instructions, SayCan maximises the likelihood of an instruction being successfully executed by the robot.\nAssume the robot can execute a set of skills. Then, for any given instruction and state, the system selects the skill that maximizes: the probability of a given completion (restricted to the set of available skills) times the probability of success given the completion and the current state. The system is trained using reinforcement learning.\nResearchers tested SayCan on 101 instructions from 7 types of language instructions. It was successful in planning and execution 84% and 74% of the time respectively.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "078eeff4-9f1f-418b-8c47-4c46b5acf853",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Success rate: 84%",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-47c4-7719-a217-3997c61ed9fa",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.75,
        "x": 0.02,
        "y": 0.21
      },
      "kind": "paragraph",
      "text": "Thanks to their large range of capabilities, LLMs could in principle enable robots to perform any task by explaining its steps in natural language. But LLMs have little contextual knowledge of the robot's environment and its abilities, making their explanations generally infeasible for the robot. PaLM-SayCan solves this.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "2a2ffe69-430a-4836-97ee-7b3c6c9f6624",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.65,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "LLMs empower robots to execute diverse and ambiguous instructions",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "1065913c-a6d0-4532-b335-69a545d8aea0",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Concrete Language",
      "slug": "concrete-language",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "019dd95a-1055-74e0-a956-14cee2bfa007",
      "evidence": "Cites 84%/74% success on 101 instructions.",
      "confidence": 75
    },
    {
      "name": "Problem-Agitate-Solve (PAS)",
      "slug": "problem-agitate-solve-pas",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "cfa7a66f-2d27-49b1-bf18-2410c8a5f789",
      "evidence": "The slide presents a problem (ambiguous instructions), agitates it ( highlighting the difficulty), and then presents a solution (LLMs empowering robots).",
      "confidence": 0.7
    }
  ],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 80,
      "from": 5,
      "beatId": "019dd95a-0682-776c-8e35-2557e3799e96",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Definitions, exec summary, Research and Industry sections inventory the state of AI.",
      "position": 1,
      "confidence": 78,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 93,
      "from": 11,
      "beatId": "019dd95a-0682-776c-8e35-35abf646aa8b",
      "arcName": "The Mountain",
      "arcSlug": "mountain",
      "beatName": "Rising Action",
      "beatSlug": "mountain-rising-action",
      "evidence": "Research, Industry, Politics sections accumulate signals.",
      "position": 2,
      "confidence": 55,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [
    {
      "to": 43,
      "from": 37,
      "name": "Pattern Hunter",
      "slug": "02-pattern-hunter",
      "bestFor": "Time-pressed audiences, building consensus, when data is strong",
      "matchId": "019dd95a-07fe-70ce-8d3d-afe811e71ad6",
      "evidence": "SayCan, ConvNets vs ViT, vision-language unification, Gato, IRIS, NeRFs as parallel evidence.",
      "position": 6,
      "objective": "Show transformers becoming the universal cross-modality architecture",
      "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
}