«

Podcasts about the Topic

“The Topic” being, of course, the Mathocalypse, except sometimes with AI alignment and even quantum computing and complexity theory and other stuff thrown in.

With Ryan Peterman:

Notably, near the end of the above podcast, you can watch my live reaction to the news that an Anthropic model found algorithms for the 3SUM and the All-Pairs-Shortest-Path problems beating records that stood for generations—news that hadn’t yet reached me when I woke up and logged in. (If you don’t feel like watching, my reaction is basically: “uh, wow … if I hadn’t been so overstuffed with delicacies right now from history’s biggest math breakthrough buffet, I’d surely be giving this one the individual amazement that it deserves…”)

And here’s with “Modern CTO”, on a similar set of topics:

Enjoy, and thanks to the hosts of both podcasts!

Email, RSS Follow

This entry was posted on Tuesday, October 13th, 2026 at 3:26 pm and is filed under Announcements, Complexity, Quantum, The Fate of Humanity. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

2 Responses to “Podcasts about the Topic”

  1. Odd Anon Says:
    Comment #1 October 13th, 2026 at 4:19 pm

    So interestingly, Wikipedia’s article on the topic is full of doubts about the validity of the proofs, and the editors have been reverting any updates to relevant math articles, pending “publication in reliable peer-reviewed journals”. Am I in a media bubble, and OpenAI’s results are actually very uncertain, or is this just Wikipedia being weird as usual?

  2. Scott Says:
    Comment #2 October 13th, 2026 at 4:47 pm

    Odd Anon #1: I think the smart prediction is that most (not all) of these proofs will ultimately be found to be valid—given that we know (from vetted results) that AI is now capable of solving problems at this level, and given also that many of the results come with Lean certificates.

    Those who pretend otherwise strike me as still being in the Denial stage of the Five Stages of Grief.

    On the other hand, the Wikipedia editors are correct that the solutions haven’t yet been accepted by academic peer review, or in the vast majority of cases, even understood yet by any human. And even the Lean certificates don’t provide absolute certainty, when we factor in bugs in the Lean kernel, the possibility that Lean’s formal system is unsound (!), and the possibility that the Lean formalizations fail to match the informal statements.

    So, it seems to me like the right way for Wikipedia to handle this would simply be to inform readers “Smith proved X. Jones improved that to Y. In its October 2026 release, OpenAI claimed to prove Z, a claim still being evaluated by experts.”

Leave a Reply

You can use rich HTML in comments! You can also use basic TeX, by enclosing it within $$ $$ for displayed equations or \( \) for inline equations.

Comment Policies:

After two decades of mostly-open comments, in July 2024 Shtetl-Optimized transitioned to the following policy:

All comments are treated, by default, as personal missives to me, Scott Aaronson---with no expectation either that they'll appear on the blog or that I'll reply to them.

At my leisure and discretion, and in consultation with the Shtetl-Optimized Committee of Guardians, I'll put on the blog a curated selection of comments that I judge to be particularly interesting or to move the topic forward, and I'll do my best to answer those. But it will be more like Letters to the Editor. Anyone who feels unjustly censored is welcome to the rest of the Internet.

To the many who've asked me for this over the years, you're welcome!