OpenAI’s Remarkable Math Achievement Sparks Controversy

OpenAI has reportedly cracked a significant math challenge that has stumped experts for nearly 90 years. In a recent blog post, the organization revealed it has found a solution to the Navier-Stokes problem, which deals with the dynamics of fluid flow, utilizing a powerful internal AI model that surpasses the capabilities of the newly launched GPT-6 Astra, along with 10,000 concurrent agents. The Navier-Stokes problem is one of the seven Millennium Prize Problems, each offering a $1 million reward for a solution.

The training of this internal AI model began on August 28th, showcasing remarkable performance in various benchmarks, including mathematics. This breakthrough is a major milestone for the math community, but it has sparked some controversy.

Just a day before OpenAI’s announcement, Tristan Buckmaster, a mathematics professor at New York University, published findings on a related issue in collaboration with Levent Alpöge, a researcher at Anthropic. Buckmaster stated that he reached out to OpenAI after discovering they were aware of their progress. However, he was surprised to find that OpenAI had developed a proof for the Navier-Stokes equation using a method he and Alpöge had been exploring with OpenAI’s Codex and Anthropic’s Claude.

In his remarks, Buckmaster expressed concerns about whether OpenAI had accessed their Codex data to aid in reaching the Navier-Stokes solution. He noted, “I asked whether the model had been trained on, or had access to, our sessions in Codex, where we had been compiling all our drafts for this project. I was told the model did not look up user data. I asked again about training, and I did not receive a response.”

In response to these allegations, OpenAI’s announcement emphasized that “no specific user data was accessed in order to solve this problem.” They also mentioned that “while unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.” When asked for clarification, OpenAI directed inquiries to its statement on X, which aligns with its blog post.

Sébastien Bubeck, a technical staff member at OpenAI, reiterated that they had not seen any of Buckmaster and Alpöge’s work until it was made public. He noted that their proofs differ significantly, including the specific results achieved.

In response, Buckmaster took to Mastodon, asserting that OpenAI is “openly admitting they used training data from a period after we found our result.” OpenAI has stated that it does not intend to claim the $1 million prize.