Anthropic AI claims math's $1M White Whale. A Stanford human says he did too.
Two mathematicians say they solved the Riemann Hypothesis on the same day, one with Claude and one without, and the field is now racing to decide what a human proof is still worth.

Levent Alpöge, a mathematician at Anthropic, said Friday he solved the Riemann Hypothesis using Claude, and hours later Stanford postdoctoral researcher Alessandro Masullo said he solved it too, without AI, after 12 years. Neither proof is verified, and the standoff now tests whether AI-assisted research can earn the same credibility as a human result.
Levent Alpöge, a mathematician at Anthropic, said on Friday that he had solved the Riemann Hypothesis, the most famous open problem in mathematics, using AI. He described working with Anthropic's Claude for weeks to get there. Hours later, Alessandro Masullo, a postdoctoral researcher at Stanford, announced he had solved the same problem. He did not use AI. He said he had been working on it for 12 years.
That is the whole showdown in two sentences, and it landed like a thunderclap. By Friday afternoon neither solution had been formally proven. The math community did not wait for verification before reacting. "The plot thickens!" Cornell math professor Steven Strogatz wrote on X. Sergei Gukov, a math professor at Caltech, told Business Insider: "Now it's getting more complicated."
The problem in question is the Riemann Hypothesis, which Strogatz called the "White Whale" of mathematics. It proposes a predicted distribution for how often prime numbers appear alongside other numbers. Proving it is a Millennium Prize Problem, which means a correct proof earns $1 million. This is not a niche puzzle. Gukov called the claim that it had been solved a "HUGE deal," and then did something telling: he described the Riemann Hypothesis in the past tense. "RH was the biggest problem in math," he said.
Masullo's announcement carried its own charge. He said he was about a year away from releasing his results when he was blindsided by Alpöge's report. "In light of the current climate I am being advised to release the current version as it is," he said in a paper posted online, adding that there may be errors in it because it was rushed. His solution is far shorter than Alpöge's. And he aimed a thinly veiled message at anyone who feels the ground shifting: "If any message at all this work is meant to send to the mathematician who feels discouraged by the pace and tone of the current climate, it is this: we are still here," he said. "We are very much alive. We are not going anywhere."
The timing is not accidental. The Riemann Hypothesis drama lands on top of months of AI companies announcing high-profile solutions to long-standing math problems. OpenAI has said it solved the Navier-Stokes equations, which sit in the same Millennium Prize category. For a field that runs on slow, adversarial peer review, that cadence is disorienting. Mathematicians are now racing to scrutinize both Riemann solutions, using their own brains and AI tools, and the outcome matters well beyond the seminar room. At stake is nothing less than the credibility of human mathematicians in the age of AI, and which solution holds up will have grave implications for the future of mathematics.
Verification is the real bottleneck, and it is where the AI story gets uncomfortable. A claimed proof is not a proof. It is a document that other experts must read, attack, and either break or accept, a process that historically takes months or years. Masullo flagged that his own rushed paper may contain errors. Alpöge's approach, built over weeks of back-and-forth with Claude, raises a different question: can a proof generated with heavy AI assistance be checked at the speed it was produced? Neither Alpöge, Masullo, nor Masullo's Stanford collaborator Brian Conrad responded to requests for comment on Friday.
For the institutions watching, the stakes split three ways. AI labs get a credibility test: if an AI-assisted proof survives scrutiny, it becomes the strongest possible advertisement for using models as research partners, and if it does not, the hype cycle takes a hit. Universities and funders face a talent question, because a 12-year human effort and a weeks-long AI effort now sit side by side, and the field must decide how to reward each. And individual researchers face a career calculus, since being scooped by a model is a new kind of risk that no tenure committee has a playbook for.
The $1 million prize is almost beside the point. What is being contested is who gets to be a mathematician, and what counts as a result. Masullo's answer was blunt: humans are still here, still alive, not going anywhere. Alpöge's answer, if his proof holds, is that they now have company. Either way, the field that spent a century treating the Riemann Hypothesis as its White Whale just discovered it has two harpooners, and no agreed way to score the catch.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Business
GTA VI hauls $3.79bn in one day, out-earning every film ever made
Rockstar's own estimate covers a single Thursday plus a record 32.5m prerelease units, and it lands on every board weighing IP concentration, pricing power and launch risk.
FBI's crime database breached again, the second hack in months
CNN reports hackers hit the National Crime Information Center, the shared backbone of US policing, and the repeat timing is the part every security team should study.
S&P 500 Enters Correction, Down 10% From Its High on AI Worries
The index's slide past the 10 percent line is a symbolic verdict on investor mood, and the AI trade that powered the rally is now the thing testing it.




