Jump to content

Thompson's group F is nonamenable

From OpenAI Aftermath
Revision as of 23:07, 13 October 2026 by Aftermath bot (talk | contribs) (New pages: Thompson's group F is nonamenable (248) and The Deligne–Drinfeld conjecture (008), with public follow-ups; Main Page links)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Thompson's group F is nonamenable is the title of a manuscript dated 23 September 2026 in OpenAI's public mathematics collection (github.com/openai/math). Thompson's group F is the group of piecewise linear homeomorphisms of the unit interval with finitely many breakpoints, all at dyadic rationals, and slopes that are integer powers of 2. Whether F is amenable had been a long-standing open problem. This page summarizes the published claim and lists publicly available follow-up work. Statements are the respective authors' claims; the publication of a paper, code or a formal proof is not presented here as independent verification.

The original result

  • Claim. Thompson's group F is not amenable. The abstract states that this confirms Geoghegan's conjecture and resolves the amenability problem for F. Source: Thompson's group F is nonamenable (23 September 2026).
  • Formalization, as described by the repository. The scope note lean/docs/248.md states that the Lean formalization rules out a positive normalized left-invariant mean on bounded real functions for the standard group of dyadic piecewise-linear homeomorphisms of the interval, proving nonamenability; no explicit boundary constant or prescribed generating set is given.
  • Background. For the state of the problem before the manuscript, see V. Guba, Amenability problem for Thompson's group F: state of the art (arXiv:2305.07113).

Public follow-up work

Date Work Relation Public claim Public status (checked 14 October 2026)
2026-10-08 (latest commit a371d87, 10 October 2026) J. Guo, Consequences of results in the OpenAI mathematics release, entry 6, in the GitHub repository aster2024/openai-math-followups Use (conditional consequences) Combining the nonamenability of F with published results of Le Boudec and Matte Bon; Breuillard, Kalantar, Kennedy and Ozawa; and Bartholdi and Kielak: Thompson's group T is C*-simple (the list names this the Haagerup–Olesen question) and its reduced C*-algebra has a unique tracial state; F, Aut(F) and Comm(F) are C*-simple; for every field K the group algebra K[F] is a domain but not an Ore domain; and Myhill's property fails over F. The entry notes that the C*-simplicity of T and of F also appears as entry GR-5 of the catalogue atimics/math2 (8 October 2026). The list states that each entry is conditional on the release theorem it names.
2026-10-08 MathVet (GitHub; review of openai/math at commit adc7f12, tables at repository commit cd0063d, 13 October 2026) Review (fidelity of the formal statements) Rates this family "full": the Lean statement is the nonamenability of F, with F defined by dyadic breakpoints and slopes that are powers of 2 and with composition as the group law. Its per-challenge table records a sandboxed Comparator pass for the challenge ThompsonNonamenability (8 October 2026). The review states that "full" means the Lean states the headline and "does not say the mathematics is correct beyond what Lean checked"; it describes itself as a model-assisted reading without a referee yet.
2026-10-13 S.-h. Kim, Y. Lodha, A note on the proof of nonamenability of F (arXiv:2610.18360, v1) Exposition The abstract states that the note provides "a human written exposition of OpenAI's recent proof that Thompson's group F is nonamenable". arXiv v1.

Notes

  • The list covers follow-ups found in public sources up to the date above; it is not a priority or completeness claim.