Altman's public empathy tour lands somewhere between apology and victory lap—and the disconnect tells you everything about who gets to decide what "progress" means.
The Summary
- Sam Altman told Stripe President John Collison he has "empathy" for mathematicians upset by AI solving their field's biggest problems, but says OpenAI won't stop and may have more discoveries stockpiled
- OpenAI pulled out of a Caltech math hackathon after backlash, and 28 Fields Medal winners published an open letter warning AI benchmarking "is detrimental to the science of mathematics"
- Altman frames this as philosophical tension: "the world deserves to understand how the universe works" versus respect for how mathematicians work
- The real question isn't whether AI can solve math problems—it's who decides when automation serves humanity versus when it steamrolls a discipline
The Signal
Altman's comments on the Cheeky Pint podcast mark a rare moment where he directly addresses what OpenAI's AI has done to an entire field. He admits to feeling "sadness and melancholy" when seeing what the models could do. But his conclusion isn't to slow down. It's to acknowledge the feelings while barreling forward. "I think, on net, the world deserves to understand how the universe works," he said, positioning the company as serving a higher truth that transcends any single profession's concerns.
The mathematical community isn't buying it. Twenty-eight Fields Medal winners co-signed an open letter warning that using math problems as AI benchmarks damages the field itself. Their argument: mathematics isn't just about answers. It's about the process of discovery, the development of human intuition, the training of new minds through struggle with hard problems. When AI generates proofs faster than humans can verify them, it doesn't just solve problems—it fundamentally changes what it means to do mathematics.
"The push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics, and to the mathematical community."
OpenAI's retreat from the Caltech math hackathon shows the pressure is real. Current and former Caltech mathematicians forced the company out by arguing the event would have "destructive impacts." That's not abstract concern. That's a professional community drawing a line.
What's striking is the framework Altman uses to process this. He tells Collison that OpenAI is "sorting through how to continue research while making sure those in specialized fields get the best version of the future that they can have." Notice the structure of that sentence: OpenAI will continue. The question is just how to manage the fallout for people whose fields get automated.
The Verge notes Altman hints the company has more mathematical discoveries stockpiled. This isn't over. It's just beginning.
Key tensions emerging:
- Altman positions AI math breakthroughs as serving humanity's right to understand the universe
- Mathematicians argue the field requires human struggle and intuition to develop properly
- OpenAI frames this as inevitable progress that needs careful management, not a choice to reconsider
The Implication
This is the agent economy collision in miniature. Mathematics was supposed to be one of the last domains requiring pure human intelligence. If that falls—not just to automation but to automation that happens faster than the field can absorb—what holds?
Watch what happens in other specialized fields next. If OpenAI can steamroll mathematics while expressing empathy, expect the same playbook for law, medicine, engineering. The pattern: automate first, express sadness second, frame objections as understandable but ultimately insufficient to stop progress. The mathematicians' open letter is a template. Other professions should study it now, before they need to write their own.