{
  "docId": "019dd923-5e88-73ef-bd5d-0d0f98caffe1",
  "docSlug": "5df6bd1b0447b5f6",
  "documentTitle": "2025 Air Street Capital The State of AI Report 2025",
  "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": 80,
  "pageCount": 313,
  "prevPage": 79,
  "nextPage": 81,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 0,
  "notes": "Discusses Molmo-Act (AI2), Gemini Robotics 1.5 (GDM), and MIT's 'Teaching LLMs to Plan' research.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "process_diagram"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-0d0f98caffe1/80",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1#slide-80",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "The 'Chain-of-Action' pattern - explicit intermediate plans before low-level control - is becoming a standard for embodied reasoning.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a822-7e9565a8d178",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.32,
        "x": 0.65,
        "y": 0.4
      },
      "kind": "image",
      "text": "Molmo-Act architecture diagram",
      "attrs": null,
      "subkind": "infographic",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "4804a2f8-a380-4c17-a42e-d051dc83e915",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.6,
        "x": 0.04,
        "y": 0.42
      },
      "kind": "list",
      "text": "Molmo-Act (AI2). From a high-level command, the model emits intermediate visual/geometry artifacts (e.g., depth/trajectory sketches) that a separate decoder turns into continuous motor commands. It makes behavior easier to inspect and debug complex manipulation tasks such as pick-and-place or dishwasher loading.\nGemini Robotics 1.5 (GDM): Uses the same plan-then-act architecture with ER 1.5 (the high-level planner) generating structured action plans for Robotics 1.5 to executes via a visuomotor policy.\nTeaching LLMs to Plan (MIT): Parallel work in language introduces explicit plan tokens before final answers, improving long-horizon reliability and giving auditors something to inspect, an LLM analogue of Chain-of-Action.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "42f66722-1119-4b00-90aa-4d0d69a4d482",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.95,
        "x": 0.02,
        "y": 0.21
      },
      "kind": "paragraph",
      "text": "The “Chain-of-Action” pattern - explicit intermediate plans before low-level control - is becoming a standard for embodied reasoning. First shown by AI2’s Molmo-Act in 2025, and rapidly adopted by Gemini Robotics 1.5, this approach mirrors Chain-of-Thought in LLMs, boosting both interpretability and long-horizon reliability in real-world robotics.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "7b0aa88d-ae4a-405b-8495-52dfcd498431",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.5,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Emergent reasoning moves into the physical world",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "7fb47744-4b1b-4496-8f76-5de57d7802f5",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [],
  "frameworks": [
    {
      "name": "Chain-of-Thought",
      "slug": null,
      "matchId": "3ae93a04-4af9-4b40-8404-182578837831",
      "evidence": "Mentioned as the conceptual mirror for the Chain-of-Action pattern in robotics.",
      "confidence": 0.9
    }
  ],
  "arcBeats": [
    {
      "to": 189,
      "from": 12,
      "beatId": "019dd95a-0682-776c-8e35-9f380673831f",
      "arcName": "The Triple Take",
      "arcSlug": "triple-take",
      "beatName": "The Facts (What)",
      "beatSlug": "triple-take-the-facts-what",
      "evidence": "Sections 1-2 lay out research findings and industry data with charts and case studies.",
      "position": 1,
      "confidence": 70,
      "parentBeatName": "Setup",
      "parentBeatSlug": "setup"
    },
    {
      "to": 89,
      "from": 12,
      "beatId": "019dd95a-0682-776c-8e35-ac68557d958b",
      "arcName": "The Onion",
      "arcSlug": "onion",
      "beatName": "First Layer",
      "beatSlug": "onion-first-layer",
      "evidence": "Research section unpacks technical layer beneath the headlines.",
      "position": 2,
      "confidence": 45,
      "parentBeatName": "Development",
      "parentBeatSlug": "development"
    }
  ],
  "loops": [],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}