OpenAI claims Navier-Stokes breakthrough, faces data-theft accusations — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI announced it solved the Navier-Stokes problem — one of seven Millennium Prize Problems, each carrying a $1 million reward — using an internal AI model more powerful than GPT-6 Astra alongside 10,000 concurrent agents, with training that began on August 28.
- Tristan Buckmaster, an NYU mathematics professor, published findings on a related problem the day before OpenAI's announcement in partnership with Levent Alpöge, a researcher at Anthropic, prompting concerns OpenAI may have trained on their shared Codex sessions.
- OpenAI stated "no specific user data was accessed" to solve the problem, but added it "cannot rule out" that de-identified data derived from Buckmaster and Alpöge's product usage helped improve its models.
- Sebastien Bubeck, a member of OpenAI's technical staff, claimed the company's proofs "differ significantly" from Buckmaster and Alpöge's and that OpenAI didn't see their work until public release — a characterization Buckmaster rejected on Mastodon, accusing OpenAI of "openly admitting they used training data."
- OpenAI says it does not plan to claim the $1 million Millennium Prize tied to solving the problem.
- Buckmaster's research collaboration used both OpenAI's Codex and Anthropic's Claude, putting a frontier-lab product at the center of a dispute between two competing AI companies.
Why it matters: The Navier-Stokes announcement is a landmark mathematical claim, but OpenAI's hedged admission that de-identified user data *may* have influenced training — combined with Buckmaster's public accusation — puts a spotlight on whether frontier AI models are inadvertently absorbing researchers' unpublished work and presenting it as original breakthroughs, a question with direct implications for academic integrity and AI training-data governance.
Ask SkimNews
