<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>AI Proved What? Updates</title>
  <link>https://aiprovedwhat.com/updates/</link>
  <atom:link href="https://aiprovedwhat.com/updates/feed.xml" rel="self" type="application/rss+xml"/>
  <description>What changed in OpenAI’s math release and the claims AI Proved What? follows.</description>
  <language>en</language>
  <lastBuildDate>Tue, 27 Oct 2026 22:00:00 -0700</lastBuildDate>
  <item>
    <title>Beyond OpenAI’s catalogue: AX02, AX03, AX04 and AN02</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-27-beyond-openai-2</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-27-beyond-openai-2</guid>
    <pubDate>Tue, 27 Oct 2026 22:00:00 -0700</pubDate>
    <description>&lt;b&gt;Beyond OpenAI’s catalogue, four more:&lt;/b&gt; &lt;a href="https://aiprovedwhat.com/episodes/ax02-dedekind-zeta-zero-free-region/"&gt;AX02&lt;/a&gt;, a zero-free half-plane Re s &gt; 31/32 for every Dedekind zeta function (Cureton and Thorner, with AxiomProver), building on &lt;a href="https://aiprovedwhat.com/episodes/003-quasi-riemann-hypothesis/"&gt;003&lt;/a&gt; and &lt;a href="https://aiprovedwhat.com/episodes/029-primitive-roots-every-admissible-integer/"&gt;029&lt;/a&gt;; &lt;a href="https://aiprovedwhat.com/episodes/ax03-huang-jiang-oblomkov-conjecture/"&gt;AX03&lt;/a&gt;, the Huang–Jiang–Oblomkov conjecture (Huang, Lau and Ono); &lt;a href="https://aiprovedwhat.com/episodes/ax04-collatz-positive-density/"&gt;AX04&lt;/a&gt;, an explicit positive density for Collatz, not a proof of the conjecture (Lau); and &lt;a href="https://aiprovedwhat.com/episodes/an02-simple-zeros-riemann-zeta/"&gt;AN02&lt;/a&gt;, more than two thirds of the zeta zeros simple and on the critical line, first found by Claude. AI Proved What? is made with Claude, Anthropic’s AI; Anthropic’s own results are held to the same standards as every other lab’s, using only public sources. &lt;span class="tag sp"&gt;Beyond OpenAI&lt;/span&gt; &lt;a href="https://aiprovedwhat.com/specials/"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Lean proof for 265, published by Axiom</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-27-265</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-27-265</guid>
    <pubDate>Tue, 27 Oct 2026 14:16:00 -0700</pubDate>
    <description>&lt;b&gt;Lean proof for &lt;a href="https://aiprovedwhat.com/episodes/265-two-dimensional-area-law/"&gt;265&lt;/a&gt;, published by Axiom&lt;/b&gt; (not OpenAI): &lt;a href="https://github.com/AxiomMath/AreaLaw2D"&gt;AxiomMath/AreaLaw2D&lt;/a&gt; says it proves the area-law paper’s Theorem 1.1 and Corollary 1.2. Not run by us; the PEPS paper is not formalized. &lt;span class="tag lean"&gt;Lean&lt;/span&gt; &lt;a href="https://github.com/AxiomMath/AreaLaw2D"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Beyond OpenAI’s catalogue: AX01, PSI01–05 and AN01</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-27-beyond-openai</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-27-beyond-openai</guid>
    <pubDate>Tue, 27 Oct 2026 13:33:00 -0700</pubDate>
    <description>&lt;b&gt;Beyond OpenAI’s catalogue:&lt;/b&gt; new episodes on results from other labs: &lt;a href="https://aiprovedwhat.com/episodes/ax01-classification-finite-simple-groups-lean/"&gt;AX01&lt;/a&gt;, Axiom Math’s Lean proof of the classification of finite simple groups; &lt;a href="https://aiprovedwhat.com/episodes/psi01-05-open-quantum-problems/#psi01"&gt;PSI01–05&lt;/a&gt;, five results by PSI on Open Quantum Problems; and &lt;a href="https://aiprovedwhat.com/episodes/an01-fermats-last-theorem-lean/"&gt;AN01&lt;/a&gt;, Anthropic’s Lean proof of Fermat’s Last Theorem. AI Proved What? is made with Claude, Anthropic’s AI; Anthropic’s own results are held to the same standards as every other lab’s, using only public sources. &lt;span class="tag sp"&gt;Beyond OpenAI&lt;/span&gt; &lt;a href="https://aiprovedwhat.com/specials/"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>151: Andrey Gogolev’s expository note</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-27-151</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-27-151</guid>
    <pubDate>Tue, 27 Oct 2026 10:46:00 -0700</pubDate>
    <description>&lt;b&gt;&lt;a href="https://aiprovedwhat.com/episodes/151-entropy-conjecture-c1/"&gt;151&lt;/a&gt;: Andrey Gogolev’s expository note&lt;/b&gt; “A counterexample to the Shub entropy conjecture” (dated October 11, 2026) gives an elementary account of a construction based on this result’s paper. It is an exposition, not a review of the paper. &lt;span class="tag "&gt;Note&lt;/span&gt; &lt;a href="https://people.math.osu.edu/gogolyev.1/index_files/EC_example7.pdf"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>New episode: local smoothing in every dimension (079)</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-26-079-episode</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-26-079-episode</guid>
    <pubDate>Mon, 26 Oct 2026 23:48:00 -0700</pubDate>
    <description>&lt;b&gt;New episode for &lt;a href="https://aiprovedwhat.com/episodes/079-local-smoothing-every-dimension/"&gt;079&lt;/a&gt;:&lt;/b&gt; local smoothing for the wave equation in every dimension, OpenAI’s October 25 paper, with its consequences and its Lean proof. It replaces the three-dimensional episode, which stays listed under Versions. &lt;span class="tag rev"&gt;Episode out&lt;/span&gt; &lt;a href="https://www.youtube.com/watch?v=wS_FHwBOp8A"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>New result 378: the Exponential Time Hypothesis is false, OpenAI claims</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-26-378</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-26-378</guid>
    <pubDate>Mon, 26 Oct 2026 12:03:00 +0000</pubDate>
    <description>&lt;b&gt;New result &lt;a href="https://aiprovedwhat.com/episodes/378-exponential-time-hypothesis-false/"&gt;378&lt;/a&gt;:&lt;/b&gt; OpenAI claims the Exponential Time Hypothesis is false, with a deterministic algorithm that counts satisfying assignments of k-CNF formulas in subexponential time and polynomial space, and a Lean proof. OpenAI says an independent working group of computer scientists led by Russell Impagliazzo, Ramamohan Paturi and Ryan Williams will examine and explain the proof. &lt;span class="tag rev"&gt;Episode out&lt;/span&gt; &lt;a href="https://github.com/openai/math/blob/main/history.md"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Result 079 extended to every dimension</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-26-079</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-26-079</guid>
    <pubDate>Mon, 26 Oct 2026 12:02:00 +0000</pubDate>
    <description>&lt;b&gt;Result &lt;a href="https://aiprovedwhat.com/episodes/079-local-smoothing-every-dimension/"&gt;079&lt;/a&gt; extended&lt;/b&gt; to every dimension: a new paper, dated October 25, proves Sogge’s local smoothing conjecture for the wave equation in every spatial dimension n ≥ 2, with a Lean proof. OpenAI says it also yields the Kakeya maximal and dimension conjectures and Stein’s spherical restriction estimates. By our reading it implies the claims of &lt;a href="https://aiprovedwhat.com/episodes/074-kakeya-three-four-dimensions/"&gt;074&lt;/a&gt; and &lt;a href="https://aiprovedwhat.com/episodes/078-three-dimensional-bochner-riesz-conjecture/"&gt;078&lt;/a&gt; and the sphere and paraboloid cases of &lt;a href="https://aiprovedwhat.com/episodes/077-fourier-restriction-positively-curved-surfaces/"&gt;077&lt;/a&gt;. Its new episode is out. &lt;span class="tag rev"&gt;Revised&lt;/span&gt; &lt;a href="https://github.com/openai/math/blob/main/history.md"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Lean proof added for result 032</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-26-032</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-26-032</guid>
    <pubDate>Mon, 26 Oct 2026 12:01:00 +0000</pubDate>
    <description>&lt;b&gt;Lean proof added&lt;/b&gt; for &lt;a href="https://aiprovedwhat.com/episodes/032-rational-hodge-conjecture-cm-abelian/"&gt;032&lt;/a&gt;, the rational Hodge conjecture for CM abelian varieties. Its scope notes leave out the consequences for the Tate and Hodge standard conjectures. &lt;span class="tag lean"&gt;Lean&lt;/span&gt; &lt;a href="https://github.com/openai/math/blob/main/history.md"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Code restored for result 266</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-26-266</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-26-266</guid>
    <pubDate>Mon, 26 Oct 2026 12:00:00 +0000</pubDate>
    <description>&lt;b&gt;Code restored&lt;/b&gt; for &lt;a href="https://aiprovedwhat.com/episodes/266-three-mubs-dimension-six/"&gt;266&lt;/a&gt;: OpenAI restored the C++ source and saved numerical records for the computation that excludes a fourth mutually unbiased basis in dimension six, with a Python checker. &lt;a href="https://github.com/openai/math/blob/main/history.md"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Three papers withdrawn, 14 revised</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-07-withdrawals</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-07-withdrawals</guid>
    <pubDate>Wed, 07 Oct 2026 12:00:00 +0000</pubDate>
    <description>&lt;b&gt;Three papers withdrawn&lt;/b&gt; after a sign error in “Algebraicity of Weil classes on split abelian eightfolds” and two papers that depend on it; 14 other manuscripts revised and 6 Lean formalizations added. The catalogue now holds 719 manuscripts in 372 results. &lt;span class="tag sp"&gt;Withdrawn&lt;/span&gt; &lt;a href="https://github.com/openai/math/blob/main/history.md"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>OpenAI’s math release: 722 manuscripts in 372 results</title>
    <link>https://aiprovedwhat.com/updates/#2026-10-06-release</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-10-06-release</guid>
    <pubDate>Tue, 06 Oct 2026 12:00:00 +0000</pubDate>
    <description>&lt;b&gt;OpenAI’s math release&lt;/b&gt;: &lt;a href="https://github.com/openai/math"&gt;github.com/openai/math&lt;/a&gt; opens with 722 manuscripts in 372 result families. &lt;a href="https://github.com/openai/math"&gt;Source&lt;/a&gt;.</description>
  </item>
  <item>
    <title>Special S01: finite-time blowup for Navier–Stokes and Euler</title>
    <link>https://aiprovedwhat.com/updates/#2026-09-08-S01</link>
    <guid isPermaLink="false">aiprovedwhat-update-2026-09-08-S01</guid>
    <pubDate>Tue, 08 Sep 2026 12:00:00 +0000</pubDate>
    <description>&lt;b&gt;Announced outside the math release:&lt;/b&gt; &lt;a href="https://aiprovedwhat.com/episodes/s01-navier-stokes-euler-blowup/"&gt;finite-time blowup for Navier–Stokes and Euler&lt;/a&gt;, two papers with Lean proofs in openai/NavierStokesAndEuler. &lt;span class="tag sp"&gt;Special&lt;/span&gt; &lt;a href="https://github.com/openai/NavierStokesAndEuler"&gt;Source&lt;/a&gt;.</description>
  </item>
</channel>
</rss>
