{
  "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": 81,
  "pageCount": 313,
  "prevPage": 80,
  "nextPage": 82,
  "slideType": "case_study",
  "function": "illustrate_case",
  "density": "dense",
  "nDataPoints": 0,
  "notes": "The slide uses a diagram to explain the EMMA model architecture, showing inputs (text/vision) feeding into the model and outputs (reasoning/planning).",
  "elementsJson": [
    "headline_text",
    "paragraph",
    "bullet_list",
    "flowchart"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-0d0f98caffe1/81",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1#slide-81",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "Waymo's EMMA is an end-to-end multimodal model that reimagines autonomous driving as a vision-language problem.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a822-9b705ff604cf",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.41,
        "x": 0.56,
        "y": 0.41
      },
      "kind": "diagram",
      "text": "EMMA model architecture diagram showing inputs (text, vision), the core model, and outputs (reasoning, planning, spatial reasoning, road comprehension, scene understanding).",
      "attrs": null,
      "subkind": "process",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "50939596-c74e-4101-b7a8-4c2be182348a",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.52,
        "x": 0.03,
        "y": 0.41
      },
      "kind": "list",
      "text": "EMMA achieves high performance on public datasets such as nuScenes and the Waymo Open Motion Dataset, particularly excelling in motion planning and 3D object detection using camera inputs alone.\nA key feature is its CoT reasoning, which enhances decision-making transparency by prompting the model to explain its decisions sequentially, integrating world knowledge. This approach produces outputs such as future vehicle trajectories and object detection estimates in a readable, interpretable format.\nAlthough promising, EMMA is limited by only processing a few frames at a time, not using accurate 3D sensing modalities like LiDAR, and being computationally expensive.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "a2683640-961a-4cf7-941b-30de7fa0ada6",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.94,
        "x": 0.03,
        "y": 0.21
      },
      "kind": "paragraph",
      "text": "Waymo's EMMA is an end-to-end multimodal model that reimagines autonomous driving as a vision-language problem. By mapping camera inputs directly to driving-specific outputs (such as trajectories and road graph elements) and representing them in natural language, EMMA leverages the reasoning and world knowledge of LLMs to achieve SOTA results while offering human-readable rationales.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "d2840085-57d7-4dd0-af71-e5a974b47231",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.04,
        "w": 0.65,
        "x": 0.03,
        "y": 0.14
      },
      "kind": "title",
      "text": "Reading the road: processing driving tasks in a unified language space",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "a42d492b-4e7f-4572-814b-cfa814f76258",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [],
  "frameworks": [
    {
      "name": "chain-of-thought",
      "slug": null,
      "matchId": "74feb4e1-496c-49d6-8bdf-b79387c2eeb6",
      "evidence": "Mentions 'CoT reasoning' as a key feature for decision-making transparency.",
      "confidence": 1
    }
  ],
  "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
}