FutureHouse publishes 'Millennium Problems for Biology' as wetlab-verifiable evals for AI
burny_tech · x · 2026-09-21
Reacting to reports that an internal OpenAI model solved the Navier-Stokes problem, researchers from Edison Scientific and FutureHouse (Michaela Thinks and Stephen Rodriques) argue biology lacks fast validation: even a superhuman model's proposed cures can't be verified quickly enough to serve as a frontline eval.
They propose a list of Millennium Problems for Biology:
- Each problem is extremely hard to solve but easy to validate in a simple wet lab
- The authors call it the "last reasonable eval" for AI in biology — solving any one would signal machine intelligence capable of fundamentally reshaping humanity's ability to make, measure, and model biology, with most contributing materially toward curing disease
- Full descriptions with acceptance criteria are on a dedicated website, and the authors invite the community to red-team the challenges.
More from AGI Musings
- On AI Progress: When Models Lag on Your Load-Bearing Skill, You Can't See Them Improving — dfrsrchtwts · 2026-11-12
- Slack's David Hoang: agents beat people on coordination, proposes 'boundary agents' — davidhoang · 2026-11-12
- CS student mourns coding flow states he'll never experience in the AI era — chrissyykat · 2026-11-12
- Ford's 370% Turnover Led to $5 Days — But ASI Never Quits — krishnan · 2026-11-12
- Against 'verifiable vs. fuzzy': the real axis is how expensive tasks are to grade — dfrsrchtwts · 2026-11-12
- Daniel Kokotajlo challenges AI-skeptic 'missing factor F' arguments: every past bottleneck claim fell fast — DKokotajlo · 2026-11-12