{
  "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": 34,
  "pageCount": 313,
  "prevPage": 33,
  "nextPage": 35,
  "slideType": "industry_trends",
  "function": "summarize",
  "density": "dense",
  "nDataPoints": 6,
  "notes": "The slide highlights the shift toward verifiable AI reasoning in mathematics.",
  "elementsJson": [
    "headline_text",
    "bullet_list",
    "paragraph",
    "photo"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5e88-73ef-bd5d-0d0f98caffe1/34",
  "deckHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1",
  "deckJsonHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1.json",
  "deckAnchorHref": "/decks/019dd923-5e88-73ef-bd5d-0d0f98caffe1#slide-34",
  "components": [
    {
      "bbox": null,
      "kind": "callout",
      "text": "These improvements point to the very real possibility of a non-trivial research-level result in mathematics being proven and formalized by an AI system (with some human supervision involved) within the next year.",
      "attrs": null,
      "subkind": null,
      "toolName": "Visual emphasis",
      "toolSlug": "visual-emphasis",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a81c-b463e42ad9e0",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.35,
        "w": 0.25,
        "x": 0.73,
        "y": 0.42
      },
      "kind": "image",
      "text": null,
      "attrs": null,
      "subkind": "illustration",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "d1f817ea-610e-42df-9773-28ac035354c4",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.45,
        "w": 0.7,
        "x": 0.02,
        "y": 0.35
      },
      "kind": "list",
      "text": "OpenAI: experimental reasoning model reached IMO gold under contest-style conditions (≈35/42, 5/6 problems). Also at the “coding Olympics” (ICPC-style test), GPT-5 solved 12/12 problems (11 on first try).\nDeepMind: after silver in 2024, IMO gold-level performance reported in 2025.\nHarmonic (Aristotle): announced formally verified IMO gold-medal-level results and released verification artifacts.\nGödel-Prover (Princeton/Goedel-LM): open-source prover with 57.6% Pass@32 on miniF2F (+7.6 over prior OS SOTA), 7 problems on PutnamBench, and 29.7k new Lean proofs—fuel for the training flywheel.",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "ae4124bd-6199-4169-b771-7a9d7b222e60",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Pass@32: 57.6%",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-41ac-7180-a81c-b86538303ad0",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.1,
        "w": 0.7,
        "x": 0.02,
        "y": 0.81
      },
      "kind": "paragraph",
      "text": "These improvements point to the very real possibility of a non-trivial research-level result in mathematics being proven and formalized by an AI system (with some human supervision involved) within the next year.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "1de465eb-d4b9-4013-a520-d137dbbb0d87",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.12,
        "w": 0.95,
        "x": 0.02,
        "y": 0.21
      },
      "kind": "paragraph",
      "text": "Math is a verifiable domain: systems can plan, compute, and check every step, and publish artifacts others can audit. So 2025 saw competitive math and formal proof systems jump together: OpenAI, DeepMind and Harmonic hit IMO gold-medal performance, while auto-formalization and open provers set new records.",
      "attrs": null,
      "subkind": "paragraph",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "ee422747-2005-4b29-b05c-e75630241085",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.5,
        "x": 0.02,
        "y": 0.14
      },
      "kind": "title",
      "text": "Ushering in an era of AI-augmented mathematics",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "856c0f3b-2fdf-4042-af58-38b7f537e51d",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [
    {
      "name": "Data Story Arc",
      "slug": "data-story-arc",
      "agent": "Storyteller",
      "layer": "slide",
      "matchId": "23fec0e1-69fc-419b-9cf0-dc79a5a6b71d",
      "evidence": "paragraph/paragraph: Math is a verifiable domain: systems can plan, compute, and check every step, and publish artifacts others can audit.",
      "confidence": 0.5
    }
  ],
  "frameworks": [],
  "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
}