Sutherland and others compile Alpöge’s proposed Riemann proof

Further checks matched five theorem statements and found only standard starting assumptions, but mathematical review of the argument continues.
In this article
MIT mathematician Andrew Sutherland and other participants in a Lean community discussion have independently compiled Levent Alpöge’s proposed solution to the Riemann hypothesis, TechCrunch reported. The checks follow Alpöge’s release of an argument developed with Claude, Anthropic’s superintelligence (SI) model, for the longstanding problem and an extension of it.
Lean is a programming language that expresses mathematical arguments as code so their logical steps can be checked. Compilation alone does not show that the code states the intended result without extra assumptions, or settle the mathematical review. Alpöge, a mathematician at Anthropic, has committed to pursuing an explanation other mathematicians can understand and discuss.
Earlier: Anthropic mathematician Levent Alpöge claims Riemann proofs with Claude
What the further checks found
Sutherland’s verification page said that, as of 20 November, an independent comparison had matched five theorem statements against separately written targets. They included the original Riemann hypothesis and an analogous statement for Dirichlet L-functions, a family of related mathematical functions. An audit found that the five theorems depended only on three standard axioms, or starting assumptions: propext, Classical.choice and Quot.sound.
Sutherland’s page said every module had been replayed through Lean’s kernel, the core that checks proofs. Two independently written kernels had accepted the full exported proof, and another checker, lean4lean, had accepted every compiled module. A complete lake comparator run and a Lean-kernel replay of the export were still in progress; the page said it would be updated as they finished.
The page cautioned that passing checks rely on the kernels and on the definitions in Mathlib, a library used by Lean, being what they are intended to be. The build also ran without a sandbox, an isolated computing environment, and its full output had not been compared with a sandboxed build.
What the hypothesis says
Bernhard Riemann proposed the hypothesis in 1859. It says that every nontrivial zero of the Riemann zeta function has a real part of one-half. A zero is an input for which a function’s output is zero; the real part is one component of a complex number. The generalized hypothesis extends the assertion to Dirichlet L-functions.
The zeta function is closely linked to the distribution of prime numbers. An accepted proof would sharpen bounds on estimates of how those numbers are distributed. The Clay Mathematics Institute lists the hypothesis as an unsolved Millennium Prize Problem.
Separate proposed proofs
Stanford professor Ravi Vakil also released a separate proposed proof of the generalized hypothesis by Stanford postdoctoral researcher Alessandro Masullo. Vakil said Brian Conrad had examined the draft for months and given extensive feedback, and asked others to check and formalize it and fill any small gaps.
Separately, Alpöge and NYU mathematician Tristan Buckmaster released a paper and Lean formalization concerning finite-time blow-up for the Navier–Stokes equations. Their repository says a certificate and the computation checking it were not included, and the patched Lean package had not been rebuilt when the repository was assembled.
What comes next
Before the Clay Mathematics Institute will consider a proposed Millennium Prize solution, the work must be published in a qualifying outlet, at least two years must pass after publication, and the global mathematics community must generally accept it. The institute does not accept direct submissions.
Follow AveNews
Physics, Maths & DiscoveriesAnthropic mathematician Levent Alpöge claims Riemann proofs with Claude
The Anthropic mathematician’s claimed proofs of the original and generalized hypotheses have not been independently verified or publicly accepted by the Clay Mathematics Institute. A separate code check is underway.
3 min
Sources
Show all 9 sourcesShow fewer
TechCrunch AIAI claimed a solution to the 167-year-old Riemann hypothesis — and a human did, too (opens in a new tab)
- github.comRH/remark.md at main · TomasOrtega/RH · GitHub (opens in a new tab)
- math.mit.eduRH Lean proof: verification files (opens in a new tab)
- grokipedia.comRiemann hypothesis (opens in a new tab)
- grokipedia.comGeneralized Riemann hypothesis (opens in a new tab)
- claymath.orgRiemann Hypothesis - Clay Mathematics Institute (opens in a new tab)
- math216.wordpress.comHurried Announcement: Alessandro Masullo GRH | Math 216: Foundations of Algebraic Geometry (opens in a new tab)
- github.comGitHub - tristanbuckmaster/navier-stokes · GitHub (opens in a new tab)
- claymath.orgRules for the Millennium Prize Problems - Clay Mathematics Institute (opens in a new tab)
