Rumor: OpenAI scanned a math professor's Codex logs over Navier-Stokes proof and spun up 10,000 agents to finish it
ns123abc · x · 2026-09-09
- Sam Altman acknowledged investigating an internet rumor: a math professor was reportedly close to solving Navier-Stokes after a year of using Codex.
- The rumor claims OpenAI reviewed all his N-S-tagged Codex sessions, found the most promising attempt, and spun up 10,000 agents to finish the proof; when discovered, the professor's career was allegedly ruined.
- OpenAI says it "did not see his work" yet "cannot rule out" that de-identified usage data helped improve its models.
- The poster concludes OpenAI "stole it." Treat as unverified AI-community drama.
Related event: OpenAI denies peeking at user chats to steal Navier-Stokes breakthrough(78 posts)→
More from Companies & People
- Yoav Artzi slams NYT's OpenAI coverage: cited prompt evidence was just a replication — yoavartzi · 2026-09-09
- Ex-OpenAI safety VP Brundage: a model 'significantly more capable than GPT-6' already exists — Miles_Brundage · 2026-09-09
- Developer Madison Kanna joins Fireworks AI — Madisonkanna · 2026-09-09
- FLORA opens applications for 5-week AI design Immersion Program, deadline Sept 18 — round · 2026-09-09
- NYU mathematician accuses OpenAI of racing to steal proof with massive compute — Nikvest · 2026-09-09
- Joke: Anthropic reportedly planned its Navier-Stokes reveal right before IPO — IgorCarron · 2026-09-09