Story
Oct 31, 2026
Altman Won’t Slow OpenAI’s Math Rush Despite Researchers’ Alarm
Altman frames OpenAI’s mathematical push as a public good that must not be held back; researchers see a powerful company moving too quickly, forcing the academic community to verify and absorb results on its terms.
OpenAI’s latest clash with the mathematics community began as its models started producing headline-grabbing results on problems long treated as the discipline’s hardest tests. The company drew particular attention after saying its systems had found a singularity in the Navier-Stokes equations, a 90-year-old question tied to one of the Millennium Prize Problems.1
The backlash came quickly. OpenAI withdrew from a Caltech “math hackathon” after mathematicians warned the event could harm their community, while 28 Fields Medal winners argued that using mathematical problem-solving as an AI benchmark was “detrimental to the science of mathematics, and to the mathematical community.”1
Researchers’ concern is not simply that AI can solve problems. They have criticized OpenAI’s rollouts as thinly explained, inadequately attributed and insufficiently verified—leaving mathematicians to do the laborious checking after announcements land. This month, the company released hundreds of results, often through brief writeups researchers said were difficult to follow.2
Altman, speaking on the Cheeky Pint podcast, said he understands the unease but rejects a pause. “I think, on net, the world deserves to understand how the universe works,” he said. “But I do have empathy for the method.”2 He later made the trade-off explicit: “Humanity deserves fundamental knowledge, and we should not be sitting on a major fundamental science discovery.”2
That position is complicated by OpenAI’s own restraint. The company has withheld details of a claimed Yang-Mills result after consulting a selected panel of experts, despite criticism that this amounts to the very gatekeeping Altman condemns.2
Altman said OpenAI has spent hundreds of millions of dollars on mathematics, including roughly $10 million on Navier-Stokes, and confirmed it has more discoveries in reserve: “We do. Of course.”2 For him, the task is to make the process more sensitive. For critics, the unanswered question is whether a well-funded AI lab can race ahead without remaking mathematics around itself.