The Deligne-Drinfeld conjecture

Follow-up works

• Thomas Willwacher, The Deligne–Drinfeld conjecture after OpenAI (arXiv:2610.15592, v1 11 Oct 2026, v2 16 Oct 2026). Presents a streamlined, simplified version of this paper's proof built around a reduction to characteristic 2, and states that the author "claims no originality regarding the proof strategy"; its AI usage statement says AI tools helped restructure the arguments.
• Minoru Hirose, Confluence relations, Hoffman basis over $\mathbb{Z}_{(2)}$, and the Deligne–Drinfeld conjecture (arXiv:2610.23719, 16 Oct 2026). Gives an alternative proof via a $\mathbb{Z}_{(2)}$-integral Hoffman-basis result for multiple zeta values, combined with Brown's theorem and Furusho's confluence characterization of GRT; cites this paper as the earlier proof, notes both arguments use reduction modulo 2, and discloses AI assistance with author verification.
• Minoru Hirose (@hirose_minoru), post on X (21 Oct 2026). The author states (in Japanese) that combining his result with results of Brown and Furusho and with motivic results also gives an alternative proof of the Deligne–Drinfeld conjecture announced by OpenAI.

Comments and discussion

• Leila Schneps, The OpenAI proof of the Deligne-Drinfeld conjecture (HMI workshop talk abstract, October 2026). Announces an overview of the paper's proof strategy and main ideas, distinguishing straightforward steps from technical parts.
• Thomas Willwacher / AHGT, On the Deligne–Drinfeld conjecture (Taming the Flow seminar, scheduled for 22 October 2026). Announces a seminar on this manuscript, situating its claim relative to earlier work of Drinfeld, Brown and Willwacher.
• Christian Haesemeyer (@CHaesemeyer), post on X (20 Oct 2026). Two-post thread saying the now three-page proof "works by writing down a representation that comes out of nowhere" and then checking well-definedness, and arguing that where it comes from "is where the mathematics is."
• Bjørn Kjos-Hanssen (@kjoshanssen), post on X (11 Oct 2026). Asks whether the argument in question is about the localization ℤ₍₂₎ rather than the 2-adic integers ℤ₂.

Formalization

• OpenAI, The Deligne–Drinfeld conjecture: Lean scope note (lean/docs/008.md) (OpenAI Math Release, 6 October 2026). States that the main statement (freeness on odd generators for the solution space of antisymmetry, three-term and pentagon equations with the Ihara bracket, plus the completed isomorphism) is formalized, and that the regularized-transport proposition and graph-complex consequences are not. [family-level]

Related prior literature mentioned by others

• Hidekazu Furusho, The pentagon equation and the confluence relations (arXiv:1809.00789; American Journal of Mathematics 144, 2022). Hirose's follow-up uses its equivalence between the pentagon equation and confluence relations to turn Hoffman generation into an alternative proof of the conjecture treated by this paper.

Possibly relevant literature (found by the agent, not checked)

• Hidekazu Furusho, Pentagon and hexagon equations (arXiv:math/0702128; Annals of Math. 171 (2010)). Shows the pentagon implies the hexagon equations, relevant to the paper's definition of W through antisymmetry, the three-term relation and the pentagon only.
• Maria Podkopaeva, On the Jacobson element and generators of the Lie algebra grt in nonzero characteristic (arXiv:0812.0772, Dec 2008). Studies grt and its generators in characteristic p, which bears on the paper's main step, a degeneration of the pentagon in characteristic two.

Digests and expositions

• AI Proved What?, One per odd weight: the Deligne–Drinfeld conjecture (Episode 008, October 2026). Animated introduction and transcript outlining the statement, prior results, characteristic-two proof strategy and stated Lean scope, with AI-assisted production disclosed by the site. [AI-generated per source] [family-level]
• Platzi, The Grothendieck–Teichmüller Lie algebra and the Deligne–Drinfeld conjecture (lesson 14 of the "OpenAI Math Preprints" course) (online course, undated, no author stated). Short video and written summary outlining the paper's characteristic-two dimension bound and the free-Lie-algebra lower bound.

News and blogs

• Trinity College Dublin News, Leading mathematicians gather in Trinity to examine new AI-generated mathematical proof (15 Oct 2026). Reports that a Hamilton Mathematics Institute workshop examined this proof, that Leila Schneps presented her understanding of it, and that Willwacher reduced the roughly 40-page argument to about eight pages; it quotes Anton Alekseev and Florian Naef.