Draft proof of GRH by Alessandro Masullo released, then problems reported
A blog post announces a proposed proof of GRH (the Generalized Riemann Hypothesis) by Alessandro Maria Masullo. The post's author writes in a hurry: they are at the American Mathematical Society Executive Committee and Board of Trustees meeting and about to chair an hours-long meeting, but call the matter urgent and want to get the file out quickly.
According to the author, Brian Conrad has been looking closely at the proof for months and has given extensive feedback, which is being incorporated. It is so far looking good to him, the author says. The author also states that the proof is all human, with no AI.
The original plan was to have Conrad work with Masullo for a year to polish the writing, and then announce the proof after running it by enough people. The author now thinks it best to put out the current draft. The post hopes people will quickly check it, formalize it and fill any small gaps if they occur. The author says they really hope it works and have great faith in Conrad's judgement, adding that Conrad has put a lot of thought into checking what Masullo has done. There is a long story to tell, the author says, but for now they want to get the file out, and the post ends that paragraph with "Here it is."
An update added about 8 hours later says the author is out of the meetings and catching up. Several people have contacted the author to report problems in the proposed proof, identified with the assistance of different AI models. The author thanks several mathematicians in particular who were first to get in touch. As usual, the update says, all issues need to be resolved to the satisfaction of the community before the proof would be accepted as true. The author adds in parentheses that they make no claim as to the correctness of the proof strategy.
Key facts
- The post announces a proposed proof of GRH by Alessandro Maria Masullo and releases the current draft file.
- Brian Conrad has reviewed the proof for months and gave extensive feedback; per the post's author it is so far looking good to him.
- The author says the proof is entirely human-written, with no AI.
- The plan had been a year of polishing with Conrad before any announcement; the author decided to release the draft instead.
- About 8 hours later an update reported that several people had found problems in the proof with the help of different AI models; the author makes no claim as to the correctness of the proof strategy.
Why it matters
GRH is a famous open problem, and a claimed proof would be a major event if it held up. The story is as much about process as result: the author skipped the planned year of polishing and released a draft early, then within about 8 hours had to add that problems had been found. It is also a small example of AI models being used by outside readers to probe a human-written proof.
Who it affects
Mathematicians who may be asked to check or formalize the draft, and the author's community, which the post says must be satisfied before the proof is accepted. Masullo and Conrad are the named people directly involved. For general readers, it is a story to watch rather than a result.
How to use it
Treat this as an announcement of a draft awaiting scrutiny, not as a settled proof. The post asks people to check the draft quickly, formalize it and fill any small gaps. Anyone following the story should wait for the issues raised to be resolved.
How solid is it
Low. The post is a hurried note, and the author says they make no claim as to the correctness of the proof strategy. Conrad's view is relayed second-hand: the proof is so far looking good to him, per the author. The update says several people reported problems found with the help of AI models; the post does not say what the problems are, who reported them beyond thanks to several mathematicians, or whether they can be fixed.
Risks and caveats
Early releases of claimed proofs of major problems often fail, and this one already has reported problems. The post gives no details of the proof's method or of the issues found, and no reaction from Masullo or Conrad to the reports. The author's statement that no AI was involved in writing the proof is the author's own claim. Wait for the community's verdict before treating the result as true.
“It is all human, with no AI.”
— The post's author, on the proposed proof of GRH