OpenAI Solves Navier-Stokes Amid Scooping Allegations — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI announced Tuesday that an unreleased AI model solved the Navier-Stokes Millennium Prize problem in 88 hours using a swarm of roughly 10,000 AI agents, calling the result a "milestone."
- Tristan Buckmaster, an NYU mathematics professor, published findings on a related problem one day earlier with Anthropic researcher Levent Alpöge (acting independently) and alleged OpenAI pressured him to drop Alpöge as coauthor and credit OpenAI's internal model instead.
- Buckmaster said an OpenAI researcher told him "If you don't want me to be nice, then I don't have to be nice" when he threatened to go public with his concerns.
- OpenAI denied accessing Buckmaster's specific user data on Codex but acknowledged it could not rule out indirect influence from de-identified usage data informing its models.
- OpenAI researcher Sébastien Bubeck said the company began working on the problem after spotting Millennium Prize rumors on Twitter; the effort cost millions of dollars, though OpenAI will not claim the $1 million prize.
- Mathematicians from Brown, Carnegie Mellon, Oxford, and elsewhere warned the episode could erode the informal norms of trust and openness underpinning their field, with one Carnegie Mellon professor calling "the thought that AI systems might steal ideas from our queries" chilling.
Why it matters: OpenAI demonstrated its AI's frontier mathematical capability but appears to have alienated the academic community it sought to impress — multiple mathematicians warned the episode could push their field toward greater secrecy and undermine the informal norms of trust on which cutting-edge research depends.
Ask SkimNews
