OpenAI claims AI solved Navier-Stokes equation, academics dispute credit

This digest was compiled by AI from multiple sources — links to the originals are below.
OpenAI said on Monday that its AI model solved the Navier-Stokes equation, a 200-year-old fluid dynamics problem and one of the Clay Millennium Prize challenges worth $1 million. The claim was immediately disputed by mathematician Tristan Buckmaster, who said OpenAI rushed to solve the problem after learning of his and Anthropic researcher Levent Alpöge's progress. OpenAI executives denied accessing the pair's work before it was published.
Key Facts
- OpenAI began training a new AI model with advanced mathematical capabilities on August 28, according to researcher Sebastien Bubeck.
- The company deployed over a thousand AI agents for more than 50 hours to find the solution, at a compute cost in the millions of dollars.
- Tristan Buckmaster of NYU and Levent Alpöge of Anthropic posted documents on Monday claiming key advances relevant to the Navier-Stokes problem.
- Buckmaster claims OpenAI offered him a proposal to publish a paper without Alpöge's name included.
The Navier-Stokes Solution
OpenAI announced that its AI model produced a formalized proof of the Navier-Stokes equation, a problem that has remained unsolved for two centuries. The equation is one of the seven Clay Millennium Prize problems, each carrying a $1 million reward. Sebastien Bubeck, a mathematician and AI researcher at OpenAI, said the company began training the model on August 28. After seeing rumors that Anthropic was making progress, OpenAI dedicated more resources and had over a thousand agents work for more than 50 hours. Mark Chen, head of research at OpenAI, said the compute cost was in the millions of dollars.
Credit Dispute
Tristan Buckmaster, a mathematician at NYU, and Levent Alpöge, a researcher at Anthropic, posted documents on Monday claiming key advances relevant to the Navier-Stokes problem. Buckmaster said he learned last week that OpenAI had become aware of their work and started putting significant resources toward the problem. He claims OpenAI offered several proposals, including one where Buckmaster could publish a paper announcing the solution without Alpöge's name. Buckmaster, Alpöge, and Anthropic did not immediately respond to WIRED's request for comment.
OpenAI Response
In the press briefing, Bubeck and other OpenAI executives denied that the company inspected the pair's Codex prompts to inform their work. Bubeck said, "We, whether it's the researchers or the agents, did not see any of their work until it was released publicly last night." OpenAI noted that solving the problem required considerably more compute than it had previously spent on mathematical problems.