Altman defends OpenAI’s math releases amid researcher criticism

OpenAI CEO Sam Altman says the company will continue releasing AI-generated mathematical findings despite researchers’ struggles to verify and interpret them. Speaking on the Cheeky Pint podcast, he argued that public access to fundamental scientific knowledge outweighs the disruption. He also confirmed that OpenAI has more discoveries than it has publicly disclosed.

Mathematicians have criticized the releases for inadequate attribution, limited context, insufficient verification, and difficult-to-follow explanations. OpenAI has published hundreds of mathematical results this month, leaving researchers to undertake extensive checking while further findings arrive. Altman acknowledged shortcomings in communication and said an intense period of testing difficult problems was a deliberate effort to probe the models’ limits.

The company also faces criticism for withholding details of a claimed breakthrough involving Yang-Mills theory, which underpins modern particle physics. OpenAI says it followed recommendations from a selected expert group, despite conflicting advice from mathematicians it previously consulted.

Altman estimated that OpenAI has spent hundreds of millions of dollars tackling mathematical problems, including around $10 million, possibly more, on its Navier-Stokes solution. That spending has raised concerns among researchers that AI labs’ resources could price them out of their own fields.