Navier Stokes turns into a dispute: mathematician accuses OpenAI of using his work and threatening his career

[{"h":"CONFIRMED, now an open dispute","p":"OpenAI's announcement that an internal model solved the Navier Stokes existence and smoothness problem, one of the seven Millennium Prize Problems, stopped being just a capability story and turned into a public academic ethics case. Twelve hours before the company's announcement, mathematician Tristan Buckmaster published a detailed statement saying he and Anthropic researcher Levent Alpöge had spent more than a year advancing on closely related problems, including Euler, and that evidence suggests that work may have reached OpenAI before the announcement."},{"h":"What Buckmaster alleges","p":"According to the account, Buckmaster and Alpöge kept their research drafts inside OpenAI Codex sessions. In calls over the weekend before the announcement, OpenAI researcher Sebastien Bubeck reportedly told Buckmaster the internal model had produced a roughly hundred page proof using exactly options C and D of Fefferman's formulation, the same line the two mathematicians had been pursuing. Buckmaster says he asked twice whether the model had been trained on the content of those Codex sessions and never got an answer. He also says Bubeck pushed to remove Alpöge as an author because of his Anthropic ties and at one point asked him why he would ruin his own career by insisting on staying out of OpenAI's publication."},{"h":"OpenAI's defense","p":"OpenAI denies using Buckmaster and Alpöge's drafts or prompts to steer its model, calling the allegations false and inflammatory. The company says Fefferman's argument was reached through a different mathematical path than the two researchers used and that the full proof was built over the weekend, after news of Buckmaster's progress had already reached the company. OpenAI also acknowledges that rumors that Anthropic had solved two Millennium Problems, sparked by a cryptic tweet from Alpöge in late August, were the trigger that led it to test its internal model on every open problem."},{"h":"Why the machinery matters beyond academic gossip","p":"The case exposes a tension the frontier model race had been hiding: researchers at rival labs share the same AI tools to work, and those tools store sessions, drafts and prompts that can become training data or competitive context without the user knowing. An academic using OpenAI's Codex to collaborate with an Anthropic researcher discovered, in practice, that there is no clear guarantee of isolation between what he writes there and what the tool's owner might later learn or use. That is bigger than the math result itself."},{"h":"The scoreboard that hasn't closed","p":"International press, including The New York Times, The Telegraph, Wired and the Financial Times, covered the confrontation, and Anthropic has already voiced informal support for Buckmaster and Alpöge through a colleague at the company. OpenAI's mathematical proof remains formalized in Lean and open to scrutiny, but the core allegation of data misuse remains unresolved: OpenAI denies it, Buckmaster keeps the question open, and no independent technical audit has confirmed or ruled out the leak."},{"h":"MaxAssistant's read","p":"The technical result is remarkable and formalized, but the episode that actually deserves the industry's attention is one of governance: the capability frontier and the trust frontier are moving together, and no lab has yet figured out how to isolate customer data and academic collaboration inside agentic tools shared between rivals. Until OpenAI shows third party verifiable training timestamps, this case stays labeled as an open dispute, not a closed chapter."},{"h":"Sources","p":"OpenAI, On the Navier Stokes Millennium Prize Problem: https://openai.com/index/navier-stokes-solution/ | The New York Times, An N.Y.U. Mathematician Clashed With OpenAI Over a Million Dollar Proof: https://www.nytimes.com/2026/09/10/science/tristan-buckmaster-openai-math-navier-stokes.html | The Telegraph, OpenAI accused of threatening professor in race to crack 90 year old maths puzzle: https://www.telegraph.co.uk/business/2026/09/08/openai-accused-of-exploiting-and-threatening-mathematician | Wired, OpenAI Just Claimed a Huge Math Discovery. Some Academics Are Crying Foul: https://www.wired.com/story/openai-navier-stokes-math-discovery-academics | Mitrade/Cryptopolitan, Mathematician alleges OpenAI's proof story fell apart on a call: https://mitrade.com/au/insights/news/live-news/article-3-2069248-20260909"}]