xianshou 21 hours ago

Further coverage:

https://news.ycombinator.com/item?id=50574968

https://alpo.ge/remark.pdf

Not only has this been open for 167 years and the single most important open question in mathematics for at least a century, but the claimed proof appears even more complete (the generalized Riemann Hypothesis) than anticipated.

Besides which, what would have been the biggest news of almost any other day in the history of mathematics, unforced Navier-Stokes:

https://news.ycombinator.com/item?id=50576232

https://github.com/tristanbuckmaster/navier-stokes/

A separate claimed human proof, verdict uncertain but looking far from complete by early reports:

https://news.ycombinator.com/item?id=50579568

https://math216.wordpress.com/2026/11/20/hurried-announcemen...

Scott Aaronson on what this means:

https://news.ycombinator.com/item?id=50581084

https://scottaaronson.blog/?p=10344

Finally, Terence Tao tracking discussion and verification of all of the above:

https://terrytao.wordpress.com/2026/11/20/unforced-euler-sin...

  • sottol 21 hours ago

    How does the Alpöge/AI/Anthropic proof compare to the hurried-out in-progress version by Alessandro Masullo? I lack the understanding to judge any of it.

    • pfdietz 19 hours ago

      The latter is defective and at best unfinished.

      The former has been fully formalized, checked with Comparator and compiled with multiple Lean kernels.

      • HedonicEscal8r 19 hours ago

        While maybe true, this phrasing is rather needlessly cruel to the mathematician who spent 12 years working on the problem.

        • pfdietz 17 hours ago

          There was nothing forcing him to put out this unfinished thing right now, especially with bugs that AI review would have spotted.

  • XorNot 19 hours ago

    The interesting thing which is going to happen here is that we have AI generated proofs coming out means we may in very short order see quite a bit more work which builds on the proofs existence, where previously you could probably reliably assume years until the ramifications were fully explored.

    As a laymen and huge nerd though, if I'm being honest I see this and think "great, now task a few datacenters on how we build a practical Alcubierre drive..."

euroderf 15 hours ago

If it works out as a proof, then i'd hope to see effort devoted to understanding that crazy sort-of spiral the function makes, and the spacing of the crossing points. It all looks kinda random, but "can't be".