The fight over mathematics' most famous unsolved problem has turned into a two-horse race between humans and machines. Within hours of Anthropic mathematician Levent Alpöge announcing that he and Claude had produced a proof of the Riemann Hypothesis, a team at Stanford released a competing draft of the problem's solution written entirely without AI assistance, according to a report from The Decoder.
For more context on this story, see our ongoing breaking AI news coverage.
The dueling manuscripts land less than a day after Alpöge's four-page remark, posted to his personal site, set off the largest mathematics discussion the internet has seen this year. A Hacker News thread linking the paper collected more than 2,750 points within hours, while The New York Times, The Washington Post, New Scientist, NPR and Al Jazeera all carried the story.
What Alpöge Actually Claims
The paper itself is unusually direct. "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge writes. He adds that the method "easily extends much further, resolving the so-called Generalized Riemann Hypothesis," and that he expects it to reach the Extended Riemann Hypothesis "with more time." The Riemann Hypothesis, first posed in 1859, concerns the distribution of prime numbers and is widely considered the most important open problem in pure mathematics.
Rather than a traditional paper, Alpöge's release centers on a machine-checked Lean formalization, distributed as a fourteen-part archive on his website. He acknowledges the full formalization is enormous — a footnote in his remark says it exceeds 250,000 pages of "unreadable" output — but argues that presenting the verified Lean code was preferable to posting what he calls "some massive unreadable travesty of slop."
His description of the collaboration has already become one of the most quoted passages in modern mathematics. "Claude and I have had a great time together with such breadth in recent months, but this time Claude far outpaced me near the finish line," he writes. "No, Claude flew, and danced, and sparkled near that finish line."
The remark includes a development timeline running from early October through November 16, 2026, ending with an entry marked simply "GRH." Alpöge writes that he messaged Anthropic CEO Dario Amodei late Sunday night before announcing, asking about his ability to go public, and says "there was a constraint to respect." He also pre-empts suspicions about the announcement's timing, writing that "there is an obvious cynical view, to do with the timing, but it is false." The announcement arrived just days after Anthropic's record-setting IPO collided with one of Wall Street's worst weeks for technology stocks since 2020, as CNBC reported.
Independent Experts Say the Proof Is Plausible
Despite the unconventional format, early expert reaction has been more receptive than many expected. According to Scientific American, independent mathematicians who examined the work attest that the AI-assisted proof is "almost certainly true" — a striking assessment for a claim that would normally be treated with deep skepticism until a full referee process ran its course.
The machine-checked Lean formalization is central to that reception. Because the code compiles under a proof assistant, the logical skeleton of the argument does not depend on trust in the authors. What remains unverified is whether the formalization genuinely captures the mathematical statements it claims to capture, a subtler question that mathematicians are now parsing in public.
Terry Tao Weighs In
Terence Tao, widely regarded as the world's leading living mathematician, published a blog post on November 20 titled "Unforced Euler Singularity, Unforced Navier-Stokes Singularity, and the Generalized Riemann Hypothesis," examining connections surrounding the claimed result. The post drew hundreds of points on Hacker News within hours. Alpöge's own remark references Navier-Stokes, noting that after studying prior AI-assisted results he "immediately understood Navier-Stokes was going to fall" — a reference to the OpenAI-backed solution of that Millennium Prize Problem earlier this year, which independent mathematicians subsequently verified.
Computer scientist Scott Aaronson of UT Austin also joined the conversation, publishing a post titled "The stochastic parrots have apparently proved the Riemann Hypothesis," signaling that even prominent AI skeptics are treating the claim as serious enough to dissect.
Stanford's Answer: No AI Involved
The most dramatic response came from Stanford. According to The Decoder, the university's group released their own competing draft of a Riemann proof within hours of Anthropic's announcement — one produced, they emphasize, entirely by human mathematicians.
The dueling drafts set up an unprecedented experiment. If both arguments survive scrutiny, mathematics would gain two independent proofs of the same monumental result, one machine-accelerated and one entirely human. If only one holds, the outcome will shape how the field adjudicates AI-assisted mathematics for years.
The episode is the latest escalation in a year-long pattern we have tracked closely: Claude's unexpected breakthrough on a Riemann zeta problem in August, an OpenAI quasi-proof of the Riemann Hypothesis verified by mathematicians in October, and the Clay Mathematics Institute's acknowledgment that the Navier-Stokes problem has "apparently" been settled. Alpöge himself credits researchers at OpenAI for what he calls their good taste in choosing which mathematical results to pursue.
What Comes Next
Alpöge says the argument is "incredibly long at the moment" and that he expects it to be "massively simplified, likely immediately." He has promised a proper written exposition, talks, and collaborations to make the work digestible. Mathematicians, meanwhile, are combing both the Lean code and Stanford's human-written draft, looking for gaps that automated verification might not catch.
For a problem that has resisted capture for 167 years, the endgame may now be a contest not between mathematicians, but between two very different ways of doing mathematics.
Stay Ahead of AI
Get the latest AI news, analysis, and breakthroughs — all in one place. Follow our latest AI developments coverage. Read more AI news →.
