OpenAI claims AI proof of Navier-Stokes problem after 10,000-agent, 88-hour sprint, accused of scooping a professor

GaryMarcus · x · 2026-09-09

Axios and Wired report OpenAI claims an internal AI model produced a proof resolving the Navier–Stokes Millennium Prize problem, after launching the effort on Sep 1 upon hearing rumors that two Millennium Prize problems had been solved. Axios says 10,000 concurrent agents worked for 88 hours to reach the result.

A widely shared critique argues OpenAI learned a professor was close to solving it after a decade of work and threw massive compute at scooping him — possibly even training on his own work if data sharing wasn't disabled. The controversy: instead of an inspiring 'someone used AI to crack a big problem' story, it reads as a big lab racing an independent scholar.

Related event: OpenAI Claims Navier-Stokes Breakthrough as Credit Dispute Simmers(161 posts)→

Original post →

More from Companies & People

Companies & People channel →