After months of bombarding mathematicians with a never-ending stream of AI-generated results, OpenAI CEO Sam Altman says he sympathizes with researchers struggling to make sense of it all. But he has no plans to stop – and hints that the company’s stockpile of discoveries could deliver much more.
“I think on the Internet, the world deserves to understand how the universe works,” Altman said during an episode of the TV show. Cheeky pint podcast released Thursday. “But I have empathy for the method.”
Altman argued that the public’s right to new scientific knowledge “outweighs” the disruption caused by OpenAI’s releases, while acknowledging that the company needs to improve how it communicates its findings. “I am unequivocal on this point: humanity deserves fundamental knowledge, and we should not settle for one major fundamental scientific discovery,” he said. “The world deserves this.”
Mathematicians speaking to The edge have repeatedly criticized OpenAI for the way it announces its results, citing inadequate attribution, little context, and insufficient verification. The researchers say this has forced them to do the painstaking work of verifying and making sense of their findings, even as the company continues to move forward. This month alone, OpenAI has published hundreds of mathematical results, often accompanied by brief papers that researchers have described as poorly written and difficult to follow.
Yet OpenAI is now being criticized for doing precisely what Altman says it shouldn’t do: sitting on potentially significant discoveries. Mathematicians and scientists speaking to The edge This week, they expressed frustration with what they see as research control by OpenAI, following the announcement of three major advances in mathematics, physics and computer science. Details of one of the claimed results – involving Yang-Mills theory, the mathematical foundations of modern particle physics – have not been made public. OpenAI says this is due to recommendations from a hand-picked group of experts it consulted on how best to release it. The practice goes directly against the advice of a previous group of mathematicians consulted by the company, but is part of what the company sees as a new approach to communicating its results.
On the podcast, Altman hints that OpenAI also relies on far more mathematical discoveries than he has made public. Asked if the company had more than the hundreds of proofs already released, Altman said, “We do. Of course.”
Altman also acknowledged that OpenAI’s search for mathematical problems had “gone out of control” for a short period of time, but said this was deliberate and necessary to probe the limits of its AI models. He described an internal Slack channel called “Model Name-Hold My Beer,” where employees “just solved the hardest problems possible” to see how the models would work.
Altman also estimated that OpenAI spent on the order of hundreds of millions of dollars solving math problems, including about $10 million – “maybe a little more” – for its Navier-Stokes solution, one of the legendary Millennium Prize math problems. The spending provides a glimpse into the vast resources AI labs devote to solving scientific problems, sparking concern among researchers that they could be excluded from their own field.
It seems mathematicians and other researchers hoping for a break will be out of luck.
Gn bussni
