When Machines Solve the Hardest Math Problems, What Happens to the Mathematicians?

Steven Strogatz cried when he talked about this. That should tell you something.

The Cornell mathematician has spent his career chasing proofs, the kind of intellectual mountain climbing that takes years and sometimes a lifetime. This past week, he watched two AI labs race through problems that had stumped human experts for decades, and he's still trying to figure out what it means for the people, not just the field, left in the wake.

On Tuesday, OpenAI announced its systems had used tens of thousands of AI agents to solve a 90 year old math problem tied to a $1 million prize: the Navier-Stokes existence and smoothness question, a deeply technical puzzle about whether certain equations governing fluid flow always produce sensible solutions. The claimed solution still needs independent verification. It builds on groundwork laid by Spanish mathematicians Diego Córdoba and Luis Martínez-Zoroa. But the announcement quickly soured. Mathematician Tristan Buckmaster says OpenAI learned of unpublished work he had done with Anthropic researcher Levent Alpöge on closely related problems, then rushed to publish first, and even tried to shape who received credit.

Strogatz doesn't mince words about what he thinks is driving the speed. "This is a marketing device for them to prove how good their machines are," he told WIRED. "Nobody cares about the Navier-Stokes singularity problem; only a tiny subset of pure mathematicians care about that. It doesn't affect anybody except it's maybe worth a trillion dollars for OpenAI to show they're better than Anthropic." He's careful to note that the Navier-Stokes problem, for all its fame, has essentially no bearing on the work of a civil engineer or an aerodynamicist. Its value is almost entirely symbolic, a trophy in a competition between corporate labs ahead of blockbuster IPOs.

The credit dispute matters beyond bragging rights, too. Strogatz says he'd like to see Córdoba and Martínez-Zoroa rewarded for their foundational work, but he's also sympathetic to Buckmaster, who posted his own solution to three closely related problems just days before OpenAI's announcement. "They were on the trail, and I think they would've gotten there, but we don't know," Strogatz says. He describes Buckmaster as unusually gracious under the circumstances, someone handling a real loss with dignity most people wouldn't manage.

This isn't an isolated incident. Last week, Anthropic said its Claude model had proved 29,500 small theorems while formalizing an existing proof of Fermat's Last Theorem, a project other mathematicians had spent years on already. In August, OpenAI claimed progress on 10 other long-standing math problems. The pace is accelerating fast enough that Strogatz predicts 2026 will be remembered as either a miraculous year for mathematics or a horrifying one. Possibly both.

The Cost to the Climbers

For Strogatz's collaborator Alex Townsend, the disruption isn't theoretical. Townsend recently used ChatGPT to help solve a decades-old numerical linear algebra problem, and he's candid that without the tool, the sheer volume of work required would have made the project unfeasible. The technology helped him do better research. It also gutted something he valued about doing that research in the first place.

"I actually feel kind of upset that I've dedicated 15 years of my life to research mathematics, and at a point in my career where I'm very productive and at my peak strength as a mathematician, that peak skill is no longer there," Townsend says. "Something is able to surpass me." He describes the shift from feeling like a pioneer at the frontier of knowledge to feeling like an operator managing an AI agent. It's a different job. He says plainly that he feels threatened by it.

Strogatz reaches for a horror movie to describe the sensation, something unseen and not fully understood creeping closer with each passing week. "But we're not at the end of the movie yet," he says. "Instinctively, I'm really terrified."

Not everyone will lose the same things. Strogatz thinks proof digestion, the work of translating a machine-generated proof into language humans can actually understand and appreciate, might remain a human specialty for a while longer, though he suspects even that will eventually be automated. He also points to applied math, and fields like economics or sociology, as more resistant territory simply because they're messier, less reducible to clean formal problems. And there's a harder question lurking underneath all of it: who decides which mathematical questions are worth asking at all. Machines haven't shown any sign of aesthetic judgment yet, Strogatz says, but he doesn't see why they couldn't eventually be trained on it too.

There's also a case to be made that this is progress, not loss. Strogatz is willing to make it himself. If AI opens mathematics to people who never had the decade of formal training that gatekept the field before, that's democratization, not devastation. "I don't want to be in the ivory tower and say: 'Oh no, it's gross. I'm 67. I put in a lifetime of training to be able to do this stuff. I don't want you guys getting in here without putting in the work,'" he says.

But he also worries about what happens to the motivation that drove people like him and Townsend into the field in the first place, the thrill of being first to the summit. If machines always get there before humans do, that particular game ends. Mathematicians might still do math the way amateurs still play tennis or chess, for love rather than for being the best. Strogatz isn't sure that's sustainable, though. If AI does all the frontier work, he asks, why would anyone keep paying humans to do it too?

Why It Matters

Strogatz frames what's happening in math as a preview, not an isolated event. Pure mathematics might be a low-stakes proving ground compared to medicine, law, or national security, but the pattern, human expertise losing its footing to systems most people don't fully understand, is the same pattern likely to show up everywhere else. "Are we the canary in the coal mine for what's going to face humanity?" he asks. It's not a rhetorical flourish. It's the question he says he can't stop sitting with.