OpenAI chief executive Sam Altman says he sympathizes with mathematicians who have been struggling to keep up with the company's stream of AI-generated results, but he does not plan to slow down. In a Cheeky Pint podcast episode, as reported by The Verge on 30 October 2026, Altman said OpenAI has more discoveries it has not yet made public.
What did Altman say about the criticism?
Altman argued that the public's right to new scientific knowledge outweighs the disruption caused by OpenAI's releases. He also acknowledged that the company needs to improve how it communicates its findings. According to The Verge's account of the podcast, he framed the trade-off in these terms:
"But I do have empathy for the method."
Altman's statement is one of the first times he has directly addressed the company's effect on mathematicians. It does not change the release practice itself. The company's stated position is that the releases will go on.
Why are mathematicians frustrated with the releases?
Mathematicians who spoke to The Verge criticized how OpenAI announces its results. Their complaints centre on three things: inadequate attribution, little context, and insufficient verification. This month alone, OpenAI has released hundreds of mathematical results, often with brief writeups that researchers described as poorly written and hard to follow.
The result is that researchers spend their time checking and interpreting the work. A proof that cannot be quickly understood is hard to use, and a writeup that does not explain its method leaves reviewers guessing what was actually shown.
Is OpenAI holding back findings?
The Verge reports that OpenAI now faces the opposite criticism: that it is withholding significant findings. After announcing three major breakthroughs in mathematics, physics and computer science, the company has not released details of one claimed result. That result involves Yang-Mills theory, the mathematical foundation of modern particle physics.
OpenAI says the limited release follows the advice of a hand-picked group of experts it consulted on how to publish the work. The Verge reports that this runs against the advice of an earlier group of mathematicians the company consulted. OpenAI presents the approach as part of a new way of communicating its findings.
Altman's comments sharpen this tension. He said humanity deserves fundamental knowledge and that the company should not be sitting on a major discovery. When asked whether OpenAI had more than the hundreds of proofs it has already released, he replied: "We do. Of course."
What does the spending and the Navier-Stokes claim mean?
Altman estimated that OpenAI has spent in the order of hundreds of millions of dollars on mathematical problems. He put around $10 million, "maybe a little more," into its Navier-Stokes solution. Navier-Stokes is one of the Millennium Prize problems, which are among mathematics's best-known open challenges.
The source does not say whether the Navier-Stokes result has been independently checked or accepted by the mathematical community. That is the most important open question for readers. The spending figures are also Altman's own estimates, given in a podcast, not audited numbers.
Altman also said the pursuit of mathematical problems got out of control for a short time, and that this was deliberate. He described an internal Slack channel called "Model Name-Hold My Beer," where employees threw the hardest problems they could find at the models to see how they would perform.
Who it is for and where it falls short
Researchers who rely on OpenAI's published results have the most at stake. So do scientists deciding whether AI-generated proofs can be trusted in their own work. The Verge also raises a wider concern that the resources AI labs are putting into science may price some researchers out of their own fields. The report presents this as a concern, not a measured finding.
Several things remain unclear. The source does not give a full list of the released results or say how many have been verified. The Yang-Mills details are not public. Altman's spending estimates and his statement that more discoveries exist are the company's own claims so far.
Bottom line: Treat OpenAI's math results as unverified claims until outside mathematicians have checked them. Expect more releases, since Altman has said more discoveries exist. Wait for the Yang-Mills details and independent review before relying on any of this work.
Source and verification: This article is based on The Verge's report, published 30 October 2026, of Sam Altman's remarks on the Cheeky Pint podcast. The spending figures and the Navier-Stokes details are Altman's estimates as reported by The Verge. No repository or GitHub figures were used.
Related: ChatGPT Shared Link Flaw Let Attackers Read Private Chats
Comments (0)