Drama swirls around OpenAI’s legendary mathematical milestone

OpenAI announced it has solved the Navier-Stokes problem, one of the seven Millennium Prize Problems in mathematics, using an internal AI model more powerful than GPT-6 Astra. The problem, unsolved for about 90 years, relates to fluid dynamics and carries a $1 million prize. However, the announcement was immediately followed by controversy. New York University professor Tristan Buckmaster, in partnership with Anthropic researcher Levent Alpöge, had published related findings just one day prior. Buckmaster expressed concern that OpenAI may have accessed their work drafts, which were stored in Codex, a tool developed by OpenAI. He questioned whether the model had been trained on or had access to their sessions. OpenAI responded that no specific user data was accessed to solve the problem but added it could not rule out that de-identified data from product usage might have helped improve its models. An OpenAI technical staff member stated their proofs differed significantly from the researchers' work. Buckmaster countered that OpenAI appeared to admit using training data from a period after their result was found. OpenAI has stated it does not plan to claim the $1 million prize.

(Source：The Verge)