AI explainer
Anthropic mathematician claims Claude proved Riemann hypothesis
Levent Alpöge says he used Claude over six weeks to prove the Riemann hypothesis and its generalised form, but the work has not been independently verified.
What happened
Anthropic mathematician Levent Alpöge claims he solved the Riemann hypothesis with the help of Claude, the company's AI chatbot 1. Alpöge said he used Claude over about six weeks to create a detailed proof of the hypothesis 2. He also says the method resolves the Generalised Riemann Hypothesis 3.
Alpöge messaged Anthropic CEO Dario Amodei last Sunday night about announcing the finding 4. The Riemann hypothesis was first proposed in 1859 5. The Clay Mathematics Institute offers a $1 million bounty for each Millennium Prize problem solved 6.
What is confirmed
Multiple outlets report that Alpöge said he used Claude to prove the Riemann hypothesis 7. They also report his statement that the method resolves the Generalised Riemann Hypothesis 3 and that the work took about six weeks 2.
The claim was made public via a message to CEO Dario Amodei last Sunday night 4. The Riemann hypothesis dates to 1859 5, and the Clay Mathematics Institute administers a $1 million prize for each Millennium Prize problem 6. Separately, OpenAI claimed in September that employees using a non-public version of its AI tools solved part of the Navier-Stokes problem in 88 hours 8.
Where the accounts differ
The ledger contains no direct denial of Alpöge's claim. However, accounts differ on the status of the proof. DW reports that the proof still has to be verified by other mathematicians 9, while TRT World reports that the claim has not been independently verified or reviewed by outside experts 10. The Verge reports that the solution has yet to be verified independently 11.
On the nature of the proof, TechCrunch reports that Ravi Vakil says the General Riemann Hypothesis proof is all human, with no AI 12, and that Vakil posted a proof written by Stanford postdoc Alessandro Masullo 13. Al Jazeera reports that Kevin Buzzard says nobody yet understands what the proof actually shows 14. The Verge reports that Anthropic said in August it was working on the Riemann hypothesis with Claude but did not expect its techniques to prove it 15.
What is not known yet
Several claims rest on single outlets. The Guardian reports that Alpöge posted a note on his personal website saying Claude helped prove the Riemann hypothesis 16, and that he linked to Lean computer code he maintained could verify the findings 17. The Guardian also reports that the note was published one day after Anthropic went public on the Nasdaq 18, and that an independent mathematician in Italy released an incomplete proof he said he had worked on for almost 12 years 19.
Al Jazeera reports that Kevin Buzzard says the proof is 17 million lines of Lean code 20 and that nobody yet understands what it shows 14. DW reports that Martin Bridson called the announcement of historic significance 21 and that Alpöge said the collaboration generated more than 250,000 pages of mathematical material 22. The BBC reports that Terrance Tao said verification is still some way off 23 and that the complete proof is over 250,000 lines long 24. TRT World reports that Ravi Vakil said a postdoctoral researcher in his department independently submitted a proof 25. TechCrunch reports that Alpöge and Tristan Buckmaster released another proof related to Navier-Stokes 26.
Open questions include whether the proof will be accepted by the Clay Mathematics Institute and whether independent verification will confirm the claim.
Why it matters
If confirmed, the result would be the first solution of a Millennium Prize problem since the Poincaré conjecture, which The Verge reports is the only one officially solved in the quarter-century since the prize was established 27. The Clay Mathematics Institute offers a $1 million bounty for each problem 6.
This suggests that AI-assisted mathematics is moving into territory previously reserved for human researchers. OpenAI's September claim that its tools solved part of the Navier-Stokes problem in 88 hours 8 points in the same direction. Analysis: the timing of Alpöge's announcement, a day after Anthropic's Nasdaq debut according to The Guardian 18, may invite scrutiny of corporate incentives, though Anthropic has not commented or involved itself in the announcement, The Verge reports 28.
Opinion
Faultline's view
The enthusiasm greeting Levent Alpöge's claim rests on a proof that, by multiple accounts, has not been independently verified. DW, TRT World and The Verge all note that verification is pending, and Kevin Buzzard is reported as saying nobody yet understands what the proof shows. The Western press has largely framed this as a triumph for Anthropic, while giving less weight to sceptical voices such as Terrance Tao, who is reported as saying verification is still some way off. The ledger also records that Anthropic itself said in August it did not expect its techniques to prove the hypothesis. Until independent mathematicians review the work, the claim remains a claim, not a confirmed result. The rush to celebrate serves corporate narratives more than mathematical rigour.
This section is Faultline’s opinion, argued only from the receipts above. It is not part of the factual record and cites nothing new.
Claims cited
- Anthropic mathematician Levent Alpöge claims he solved the Riemann hypothesis with the help of Claude. Confirmed
- Levent Alpöge said he used Claude over about six weeks to create a detailed proof of the Riemann hypothesis. Confirmed
- Levent Alpoge says the method also resolves the Generalised Riemann Hypothesis. Confirmed
- Levent Alpöge messaged Anthropic CEO Dario Amodei last Sunday night about announcing the finding. Confirmed
- The Riemann hypothesis was first proposed in 1859. Confirmed
- The Clay Mathematics Institute offers a $1 million bounty for each Millennium Prize problem solved. Confirmed
- Levent Alpoge says he used Anthropic's Claude chatbot to prove the Riemann Hypothesis. Confirmed
- OpenAI claimed in September that employees using a non-public version of its AI tools solved part of the Navier-Stokes problem in 88 hours. Confirmed
- The proof of the Riemann hypothesis still has to be verified by other mathematicians. One source
- The Riemann Hypothesis proof has not been independently verified or reviewed by outside experts. One source
- Alpöge's Riemann hypothesis solution has not been independently verified. One source
- Ravi Vakil says the General Riemann Hypothesis proof is all human, with no AI. One source
- Ravi Vakil posted a proof of the General Riemann Hypothesis written by Stanford postdoc Alessandro Masullo. One source
- Kevin Buzzard says nobody yet understands what the proof actually shows. One source
- Anthropic said in August it was working on the Riemann hypothesis with Claude but did not expect its techniques to prove it. One source
- Levent Alpöge posted a note on his personal website saying Claude helped prove the Riemann hypothesis. One source
- Levent Alpöge linked to Lean computer code that he maintained could verify the findings. One source
- Levent Alpöge's note was published one day after Anthropic went public on the Nasdaq stock market index. One source
- Alessandro Masullo released an incomplete proof of the Riemann hypothesis that he said he had worked on for almost 12 years. One source
- Kevin Buzzard says the proof is 17 million lines of Lean code. One source
- Martin Bridson calls the announcement of historic significance. One source
- Alpoge says the collaboration with AI generated more than 250,000 pages of mathematical material. One source
- Mathematician Terrance Tao said verification of the Riemann hypothesis proof is still some way off. One source
- The complete proof Alpöge put forward in Lean is over 250,000 lines long. One source
- Ravi Vakil says a postdoctoral researcher in his department independently submitted a proof of the Riemann Hypothesis. One source
- Levent Alpöge and NYU mathematician Tristan Buckmaster released another important proof related to the Navier-Stokes equations. One source
- The Poincaré conjecture is the only Millennium Prize problem officially solved in the quarter-century since the prize was established. One source
- Anthropic has not commented on or involved itself in Alpöge's announcement. One source