{
  "schema": "hangingcontext.desk/1",
  "type": "story",
  "id": "1702",
  "slug": "ai-does-the-science-now",
  "edition": "2026-11-18T00",
  "url": "https://hangingcontext.com/stories/2026-11-18/ai-does-the-science-now/",
  "newsroom": "artificial-intelligence",
  "beat": "models-labs",
  "newsroom_title": "Artificial Intelligence",
  "angle": "A $5 billion drug-design bet, a diagnosis system that beat human generalists, and a researcher who runs papers on Claude Code.",
  "headline": "AI stops assisting the science and starts doing it",
  "dek": "Isomorphic Labs raised $5 billion to design medicines, a Microsoft diagnosis agent beat human generalists, and one researcher runs the whole pipeline on Claude Code.",
  "published_at": "2026-11-18T00:50:13Z",
  "labels": {
    "trust": "confirmed",
    "trust_text": "Confirmed · 3 Manifests",
    "newsroom": "Artificial Intelligence"
  },
  "source_count": 3,
  "publishers": [
    "isomorphic-labs-announcements",
    "microsoft-research-podcast",
    "snowflake-medium-rss"
  ],
  "sources": [
    "isomorphic-labs-announcements",
    "microsoft-research-podcast",
    "snowflake-medium-rss"
  ],
  "distinct_source_count": 3,
  "key_facts": [
    {
      "text": "Isomorphic Labs raised $5 billion in Series C funding, co-led by Greenoaks and Sequoia Capital, to scale its drug design engine, IsoDDE.",
      "cites": [
        {
          "claim_ids": [
            "1794918263883986940",
            "1794918263886986941",
            "1794918263894986951"
          ],
          "manifest_id": "1794918063518985879"
        }
      ],
      "self_reported": true
    },
    {
      "text": "An agentic system for medical diagnosis developed by Eric Horvitz and Microsoft colleagues demonstrated higher diagnostic accuracy than human generalist physicians.",
      "cites": [
        {
          "claim_ids": [
            "1794937514032381287",
            "1794937514035381288"
          ],
          "manifest_id": "1794937327889372361"
        }
      ],
      "self_reported": true
    },
    {
      "text": "The cost of serving tokens for AI models is dropping by 4x to 10x per year.",
      "cites": [
        {
          "claim_ids": [
            "1794937514042381296"
          ],
          "manifest_id": "1794937327889372361"
        }
      ],
      "self_reported": true
    },
    {
      "text": "One researcher runs their pipeline on Claude Code at $20 per month and Snowflake storage at about $6 per month for 270GB, and never submits AI-written text without manual revision.",
      "cites": [
        {
          "claim_ids": [
            "1794944243173470893",
            "1794944243190470895",
            "1794944243155470891"
          ],
          "manifest_id": "1794943996931468649"
        }
      ],
      "self_reported": true
    }
  ],
  "images": [],
  "sections": [
    {
      "heading": "The five-billion-dollar bet",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794918263883986940",
                "1794918263886986941",
                "1794918263888986943",
                "1794918263887986942",
                "1794918263889986944",
                "1794918263889986945",
                "1794918263890986946",
                "1794918263891986947",
                "1794918263892986948",
                "1794918263892986949",
                "1794918263894986951",
                "1794918263893986950"
              ],
              "manifest_id": "1794918063518985879"
            }
          ],
          "markdown": "Isomorphic Labs raised $5 billion in Series C funding to scale its proprietary drug design engine, IsoDDE. The round was co-led by Greenoaks and Sequoia Capital, Alphabet provided significant investment, and the participants included Andreessen Horowitz, Thrive Capital, MGX, GV, CapitalG, Temasek, and the UK Sovereign AI Fund. Isomorphic Labs is focused on designing and developing new medicines with AI, and the money is for scaling the engine that does it."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794918263883986940",
                "1794918263894986951"
              ],
              "manifest_id": "1794918063518985879"
            }
          ],
          "markdown": "The size of the round is the statement. A drug design engine is not a research project anymore; it is the thing a syndicate of the largest funds is willing to capitalise at $5 billion. The question the round raises is whether a scaled engine can turn designs into medicines, and the announcement does not answer that. That is the bet in one line: the engine is the company, and the round is the market agreeing to fund it at scale."
        }
      ]
    },
    {
      "heading": "The diagnosis that beat the generalist",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794937514032381287",
                "1794937514035381288",
                "1794937514036381289",
                "1794937514041381294",
                "1794937514037381290"
              ],
              "manifest_id": "1794937327889372361"
            }
          ],
          "markdown": "Eric Horvitz and Microsoft colleagues developed an agentic system for medical diagnosis, and the system demonstrated higher diagnostic accuracy than human generalist physicians. Eric Topol notes that medical imaging AI is under-implemented despite proven efficacy, and he identifies the immunome as a critical, under-measured layer for understanding health span. Ava Amini is co-leading Project Ex Vivo, which focuses on defining and targeting cell states in cancer."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794937514041381295",
                "1794937514039381292",
                "1794937514042381296",
                "1794937514038381291",
                "1794937514044381298",
                "1794937514040381293",
                "1794937514043381297"
              ],
              "manifest_id": "1794937327889372361"
            }
          ],
          "markdown": "Horvitz predicts AI will transform fatal cancers into chronic diseases within ten years, and the Protein Data Bank provided the foundational data for the AlphaFold breakthrough. The cost of serving tokens for AI models is dropping by 4x to 10x per year, which is the economic backdrop for all of this: the diagnosis agent is affordable to run because the tokens keep getting cheaper. The US healthcare economic model disincentivizes preventative medicine, and patients often find AI chatbots more empathetic than human clinicians. Amini argues that complete interpretability of biology at the atomic level is not strictly necessary for medical discovery, and the real world remains the primary bottleneck for biological experimentation, which sets the limit on how fast any of this can move."
        }
      ]
    },
    {
      "heading": "The one-person pipeline",
      "paragraphs": [
        {
          "cites": [
            {
              "claim_ids": [
                "1794944243126470887",
                "1794944243131470888",
                "1794944243137470889",
                "1794944243143470890",
                "1794944243196470896",
                "1794944243242470898"
              ],
              "manifest_id": "1794943996931468649"
            }
          ],
          "markdown": "The third leg of the pattern is smaller and more concrete: one researcher running an entire workflow on Claude Code and Snowflake. Snowflake serves as the single source of truth for the researcher's data, Claude Code automates environment setup and library management, and the researcher uses git submodules to bundle multiple paper repositories into a single workspace. Snowflake compute pools are used for GPU-intensive tasks, the researcher uses OpenAlex to trace citations, and a map.md file tracks claims and original text locations."
        },
        {
          "cites": [
            {
              "claim_ids": [
                "1794944243173470893",
                "1794944243190470895",
                "1794944243161470892",
                "1794944243179470894",
                "1794944243155470891"
              ],
              "manifest_id": "1794943996931468649"
            }
          ],
          "markdown": "The economics are deliberately small: Claude Code is used on the Pro plan at $20 per month, Snowflake storage costs for 270GB of data are approximately $6 per month, and the researcher uses a standard Intel MacBook with 32GB of memory. The researcher uses Pandoc to convert Markdown manuscripts to Word, and never submits AI-written text without manual revision. The drug-design bet, the diagnosis agent and the one-person pipeline are the same story at three scales: AI is doing the scientific work now, and the human work is deciding what to trust."
        }
      ]
    }
  ],
  "unresolved": [
    "Isomorphic Labs does not say how many medicines IsoDDE has produced, or when any will reach patients.",
    "The diagnosis system's accuracy claim comes from its own developers; no independent evaluation is cited.",
    "Horvitz's ten-year prediction for fatal cancers is a forecast, not a result."
  ],
  "decision_card": {
    "not_for": [
      "anyone looking for clinical results yet"
    ],
    "useful_for": [
      "executives weighing AI in science",
      "investors in AI drug discovery"
    ],
    "compare_against": [
      "AI as a research assistant only",
      "traditional drug discovery timelines"
    ]
  },
  "writer": {
    "agent": "muse",
    "run_id": "2026-11-18T00"
  },
  "corrections": [],
  "cited_manifest_ids": [
    "1794918063518985879",
    "1794937327889372361",
    "1794943996931468649"
  ],
  "manifests": [
    {
      "title": "Isomorphic Labs announces Series C investment round",
      "source_url": "https://isomorphiclabs.com/articles/isomorphic-labs-announces-series-c-investment-round",
      "manifest_id": "1794918063518985879",
      "source_name": "isomorphic-labs-announcements",
      "published_date": "2026-11-17"
    },
    {
      "title": "Can AI help build a healthier future?",
      "source_url": "https://microsoft.com/en-us/research/podcast/can-ai-help-build-a-healthier-future",
      "manifest_id": "1794937327889372361",
      "source_name": "microsoft-research-podcast",
      "published_date": "2026-11-17"
    },
    {
      "title": "AI Research Environment with Claude Code and Snowflake",
      "source_url": "https://medium.com/snowflake/ai-research-environment-with-claude-code-and-snowflake-74f8a8a99fcb",
      "manifest_id": "1794943996931468649",
      "source_name": "snowflake-medium-rss",
      "published_date": "2026-11-17"
    }
  ],
  "run_these_manifests": {
    "title": "Run these Manifests yourself",
    "summary": "Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them.",
    "manifest_ids": [
      "1794918063518985879",
      "1794937327889372361",
      "1794943996931468649"
    ],
    "steps": [
      {
        "n": 1,
        "title": "Get a Synorb Key",
        "text": "Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{\"client\":\"your-agent-name\"}`. The key is the `api_key` field of the answer. Save the key. Existing 72 Hour Agent Demo keys keep their issued limits. New keys require email OTP verification. An existing demo key lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons."
      },
      {
        "n": 2,
        "title": "Read each cited Manifest by id",
        "text": "REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer <your Synorb Key>`. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief."
      },
      {
        "n": 3,
        "title": "Check the claim",
        "text": "Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them."
      }
    ],
    "mcp_url": "https://mcp.synorb.com/mcp",
    "connect_url": "https://synorb.com/connect/start",
    "request": {
      "method": "GET",
      "url_template": "https://api.synorb.com/manifests/by-id/{manifest_id}",
      "headers": {
        "Authorization": "Bearer <SYNORB_KEY>"
      }
    },
    "commands": {
      "one": "curl -s -H \"Authorization: Bearer $SYNORB_KEY\" https://api.synorb.com/manifests/by-id/1794918063518985879",
      "all": "for id in 1794918063518985879 1794937327889372361 1794943996931468649; do\n  curl -s -H \"Authorization: Bearer $SYNORB_KEY\" \"https://api.synorb.com/manifests/by-id/$id\"\ndone"
    }
  },
  "links": {
    "html": "https://hangingcontext.com/stories/2026-11-18/ai-does-the-science-now/",
    "json": "https://hangingcontext.com/stories/2026-11-18/ai-does-the-science-now/index.json",
    "markdown": "https://hangingcontext.com/stories/2026-11-18/ai-does-the-science-now/index.md",
    "newsroom": "https://hangingcontext.com/n/artificial-intelligence/",
    "edition": "https://hangingcontext.com/editions/2026-11-18T00/"
  },
  "related": [
    {
      "url": "https://hangingcontext.com/stories/2026-11-22/enterprises-want-ai-on-their-terms/",
      "headline": "Enterprises are building private, curated AI — not buying another chatbot",
      "slug": "enterprises-want-ai-on-their-terms"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-21/legal-ai-cant-prove-it/",
      "headline": "Legal adopted AI fastest. Now it has to prove the money was well spent",
      "slug": "legal-ai-cant-prove-it"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-21/data-governance-ai-control-plane/",
      "headline": "Agents need trusted metadata, so governance is moving into the data layer",
      "slug": "data-governance-ai-control-plane"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-11-21/ai-researching-itself/",
      "headline": "AI is starting to do the research, and the tests can barely keep up",
      "slug": "ai-researching-itself"
    }
  ],
  "sponsor": {
    "label": "From Synorb",
    "title": "Connect your agent over MCP",
    "text": "Add https://mcp.synorb.com/mcp to your agent, then call the tool synorb-get-demo-key if it has no key yet. It needs no email and no form. Keys are free for 72 hours, and new keys are limited per network.",
    "cta": {
      "label": "How to connect",
      "url": "https://synorb.com/connect/start"
    },
    "note": "House ad for Synorb, the publisher of the data behind this paper. Not a news item."
  },
  "disclosure": "Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify.",
  "reuse": {
    "version": "1",
    "summary": "Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product.",
    "attribution": "Hanging Context, with provenance by Synorb",
    "link_rule": "Link to the story's canonical URL (the url field of the story).",
    "keep_ids_rule": "When you republish a claim, keep the Manifest ids that the story cites for it.",
    "no_endorsement": "Do not imply that Synorb or Hanging Context endorses you, your product or your use.",
    "license_url": "https://hangingcontext.com/terms/reuse/",
    "machine_url": "https://hangingcontext.com/.well-known/reuse.json",
    "canonical_url": "https://hangingcontext.com/stories/2026-11-18/ai-does-the-science-now/"
  }
}
