Quick Answer: No proof of the Riemann Hypothesis has been accepted by the mathematical community. A PDF claiming one, shared on Hacker News with an "Anthropic employee" attribution, should be judged on checkable evidence: a verifiable author, a public preprint record, and independent expert review. Most claims like this fail the first or second check, and the ones that survive still need years of scrutiny before anyone calls them proven.
The Riemann Hypothesis has been unsolved since Bernhard Riemann proposed it in 1859, and a proof carries a $1 million bounty from the Clay Mathematics Institute. That combination guarantees a steady stream of claimed proofs, and every few years one reaches the front page of a tech forum. The thread in question links to a PDF, and the copy I accessed is raw PDF data rather than readable text. I can't tell you what its argument says, so this piece doesn't rule on its mathematics. What I can do is show you how to evaluate a claim like this yourself, which is the more useful skill anyway.
What the Riemann Hypothesis actually asserts
The Riemann zeta function, written ζ(s), is defined for complex numbers and equals a sum over integers that converges for real parts greater than 1. Its "nontrivial" zeros, the ones that aren't simple negative even integers, all lie in the strip between real part 0 and 1. The Riemann Hypothesis says they all lie on the critical line where the real part equals exactly 1/2.
Why does this matter beyond pure theory? The position of those zeros controls how regularly prime numbers are distributed. If the hypothesis is true, the error term in the prime number theorem is about as small as it can be, which is why number theorists treat the result as a master key.
Numerical evidence is strong but doesn't count as proof. Xavier Gourdon's computations in 2004 confirmed that the first 10¹³ nontrivial zeros lie on the critical line. Checking trillions of cases tells you the conjecture is very likely true, but a counterexample could sit at height 10¹³ + 1, and no finite check rules that out.
Why the "Anthropic employee" label is a weak signal
The headline's strongest claim is about who wrote the document, and that's the easiest part to check. Affiliation is cheap to assert. A real researcher can be identified through an institutional page, a publication history, or an arXiv profile. An unsupported claim usually has none of these, or the named person has never published in number theory.
The URL here points to alpo.ge, a domain that doesn't belong to Anthropic. That doesn't prove anything about the author. Someone can host a file anywhere. But it does mean the affiliation claim comes from the poster, not from the company.
Here's the counter-intuitive part. A company affiliation doesn't make a mathematical claim more credible. Employers don't review side work in pure mathematics, and a famous name can make readers skip the checking step entirely. The credential you should care about is a track record in analytic number theory, not a job title in another field.
The track record of high-profile claims
Serious mathematicians have announced Riemann Hypothesis proofs before, and the history is instructive. In September 2018, Sir Michael Atiyah presented a short argument at the Heidelberg Laureate Forum that he said proved the hypothesis, drawing on the fine-structure constant. Coverage was wide, and the argument was not accepted by specialists. The lesson isn't that Atiyah was careless. It's that an announcement, even from a Fields Medal winner, is where verification starts, not where it ends.
The Clay Mathematics Institute's own rules reflect this caution. A proof must be published in a qualifying venue and then survive two years of community scrutiny before the prize committee considers it. That delay exists because errors in long arguments often surface slowly, after experts have tried to reproduce each step.
Most announced proofs of major conjectures fail in a predictable way: a key lemma assumes something equivalent to the conjecture, or an interchange of limits or convergence step is justified without enough care. Those are the spots where you should look first.
How mathematicians actually check a proof
Professionals don't read a 50-page argument front to back on the first pass. They look for the load-bearing claims, then test them. Here's a practical sequence:
- Read the main theorem and its stated consequences. Check whether the paper claims the full hypothesis or a weaker statement dressed up as one.
- Find the one step everything depends on. In most failed proofs, this is a single lemma, often stated without proof or justified by a citation that doesn't say what the author needs.
- Check for circularity. Ask whether the key lemma is known to be equivalent to the Riemann Hypothesis. Many equivalent formulations exist, including criteria involving the Mertens function and the Li coefficients.
- Look at the publication path. A preprint on arXiv gets public feedback. A journal submission gets referee reports. Neither guarantees correctness, but a claim that avoids both deserves more doubt.
- Wait for specialists. Experts in analytic number theory typically respond to serious claims within weeks. Silence from the field is itself information.
Here's where most readers go wrong: they treat the length and polish of a document as evidence. A clean LaTeX PDF with numbered theorems looks exactly like a real paper, and that's precisely why it's not informative. Crank proofs and real proofs share the same format.
| Check | What a credible claim looks like | Common red flag |
|---|---|---|
| Author identity | Named researcher with a verifiable publication record | Affiliation only, no prior work in the field |
| Venue | arXiv preprint plus journal submission | Posted directly to a file host or forum |
| Core argument | Explicit reduction to a clearly stated lemma | Key step asserted as "well known" without citation |
| Scope | Claims the hypothesis as stated, with proofs of each case | Claims a stronger or altered statement |
| Expert response | Specialist commentary, blog posts, or errata | Silence, or only non-experts praising it |
Where formal verification fits in
Proof assistants offer a different path. Lean, together with its Mathlib library, lets you encode a theorem and check each inference mechanically. This approach has caught real errors in published work. Its limitation is the one people skip: the formal statement must match the informal claim. A misformalized theorem can verify perfectly and still prove the wrong thing. So when someone says a proof is "formally verified," ask which statement was formalized.
Machine assistance is also becoming part of how mathematicians draft and check arguments. That makes the evaluation problem harder, not easier, because fluent, well-formatted text can now be produced cheaply. The verification steps above don't change. They just matter more.
What to do with a thread like this
If you want to evaluate this kind of claim yourself, start with the abstract and the stated main theorem. Then find the lemma the whole argument hinges on and test whether it's already known to be equivalent to the hypothesis. Search for the author's publication history and check whether any number theorist has commented publicly. If you can't complete those steps, the honest conclusion is that you don't yet know whether the proof works, and that's a perfectly reasonable place to stay.
Frequently Asked Questions
Has the Riemann Hypothesis been proven?
No. As far as the published mathematical record shows, the Riemann Hypothesis remains an open problem. Many claimed proofs have been announced, including a 2018 claim by Michael Atiyah, and none has been accepted by specialists. It is one of the Clay Millennium Prize problems, which carries a $1 million award.
How do mathematicians verify a proof of the Riemann Hypothesis?
They locate the load-bearing lemma, check whether it's equivalent to the hypothesis itself, and look for public scrutiny through arXiv or journal refereeing. Experts usually respond within weeks if a claim is serious. Formal verification in Lean can check each inference mechanically, but only if the formal statement matches the claim.
Can a company employee's proof of the Riemann Hypothesis be trusted?
Not on the strength of the job title alone. Affiliation is easy to claim and is unrelated to competence in analytic number theory. Judge the argument itself and the author's publication history in the field. A real proof will hold up under independent checking regardless of who wrote it.
Why do so many Riemann Hypothesis proofs turn out to be wrong?
The hypothesis connects to many deep results, so a proof must be correct at every step, and errors often hide in convergence arguments or in lemmas assumed without proof. Some claims also reduce to a statement already known to be equivalent to the hypothesis. Errors of this kind are hard for the author to spot, which is why outside review matters.
What does the Clay Millennium Prize require before awarding the Riemann Hypothesis prize?
Clay's rules require the proof to be published in a recognized journal, and then to survive two years of community acceptance before the scientific advisory board considers an award. That waiting period lets errors surface and gives specialists time to examine the argument.
Before you share a claimed proof, run it through the checks above: verify the author, find the key lemma, and see whether specialists have responded. If you're reading this thread now, pick one step from the PDF and test it yourself, then tell someone who's wrestling with the same question which step you checked and what you found.