OpenAI says internal model solved 100+ open math problems, forms advisory group with top mathematicians
机器之心 · wechat · 2026-09-22
OpenAI announced a partnership with the Advisory Group on Mathematics and AI (AGMAI), hosted by the Institute for Advanced Study, to address math results being produced faster than peer review can absorb.
- OpenAI says a new internal model trained since Aug 28 has solved 100+ long-standing open problems, beyond its Navier–Stokes work (paper and Lean proof published; it won't claim the Millennium Prize)—though the claims await independent mathematical review
- The nine founding members include Timothy Gowers, Martin Hairer, Edward Witten, Ravi Vakil, and Melanie Matchett Wood; members take no AI-company pay, publish recommendations openly, serve beyond OpenAI, but hold no veto over publication
- Context: a Sept 11 open letter signed by multiple Fields Medalists criticized AI companies for treating open problems as benchmarks, warning results could outpace understanding and hurt research ecosystems
- OpenAI's announcement directly responded to the letter, acknowledging the need for more careful engagement with the math community
More from AGI Musings
- Researcher willing to bet $100K the intelligence explosion won't happen next year — NathanpmYoung · 2026-11-12
- ICLR reviewer: the worst parts of my job are reading and answering AI slop — chhaviyadav_ · 2026-11-12
- Fermi paradox twist: advanced civilizations may go digital after developing AI — VraserX · 2026-11-12
- Mark Cuban: AI could make healthcare costs 'drop like a rock' — alvelda · 2026-11-12
- Researcher's paradox: if AI makes everyone unemployed, who pays for tokens? — prajdabre · 2026-11-12
- Dear ImGui Creator ocornut: AI Empowers Mediocre Makers to Produce More Mediocrity, Faster — rms80 · 2026-11-12