VibeMathedMath problems solved with AI

The Hovey-Strickland conjecture: the thick tensor ideals of dualizable K(n)-local spectra

Fix a prime pp and a height n≥1n\ge1, and let D(n,p)\mathcal D_{(n,p)} be the homotopy category of dualizable K(n)K(n)-local spectra, with tensor product LK(n)(X∧Y)L_{K(n)}(X\wedge Y). A thick tensor ideal is a full subcategory closed under retracts, cofiber sequences, suspensions and tensoring with every object. For 0≤k≤n0\le k\le n the localization of a finite type-kk spectrum generates a thick tensor ideal Dk\mathcal D_k, giving a chain D0⊋D1⊋⋯⊋Dn⊋0\mathcal D_0\supsetneq\mathcal D_1\supsetneq\cdots\supsetneq\mathcal D_n\supsetneq0. By analogy with the Hopkins-Smith thick subcategory theorem for finite spectra, Hovey and Strickland (1999) conjectured that this chain is everything. Barthel, Heard and Naumann proved it at height two and showed it follows from Chai's invariant-ideal conjecture. For every prime and every height, are the thick tensor ideals of D(n,p)\mathcal D_{(n,p)} exactly the n+2n+2 ideals D0,…,Dn,0\mathcal D_0,\ldots,\mathcal D_n,0, so that the Balmer spectrum is a chain of n+1n+1 points?

Result
Proved(see note)
Status
Candidate (review pending)
AI contribution
AI-discovered
Method
Argument
Field
Chromatic homotopy theory; tensor-triangular geometry
Posed by
Mark Hovey and Neil Strickland (Morava K-theories and localisation, Mem. AMS 1999, Section 12); restated by Barthel, Heard and Naumann (2022)
Year posed
1999
Years open
27y
Solved
2026-09-24
Model
Unreleased internal OpenAI model
Vendor
OpenAI
Collaborators
—
Verification
Unreviewed
Publication
Announced
Collection
OpenAI math release (October 2026), version adc7f12
Significance
30 / 100
Disclosed cost
—
Wikipedia
No dedicated article

What was actually shown

Theorem 1.2: for every prime pp and n≥1n\ge1 the thick tensor ideals of the dualizable K(n)K(n)-local category are exactly D0⊋⋯⊋Dn⊋Dn+1=0\mathcal D_0\supsetneq\cdots\supsetneq\mathcal D_n\supsetneq\mathcal D_{n+1}=0, so there are n+2n+2 of them and each Dk\mathcal D_k does not depend on the chosen type-kk spectrum; a corollary identifies the Balmer spectrum with the n+1n+1 proper members of the chain, with its specialization order. It is deduced from Theorem 1.1 (Chai's conjecture, a separate entry) by the known Barthel-Heard-Naumann implication. It concerns dualizable objects only and does not classify localizing or other subcategories of the whole K(n)K(n)-local category. Height two was already known.

What the AI did

The release README says all results in the release were produced by an unreleased internal OpenAI model with one fixed procedure, on average about three hours of ChatGPT Pro thinking compute per result. This family is not among the README's exceptions (the Hodge conjecture for CM abelian varieties, and the Re(s) > 11/12 zero-free region whose write-up was human-edited). The manuscripts are authored 'OpenAI' and name no human author. One manuscript proves Chai's invariant-ideal conjecture as its main step and deduces the Hovey-Strickland conjecture from it through the implication already established by Barthel, Heard and Naumann; the same manuscript underlies the separate catalog entry for Chai's conjecture.

Verification

No independent mathematician has checked this yet. Checked here: the introduction, Theorem 1.1 (stabilizer-invariant ideals) and Theorem 1.2 (Hovey-Strickland conjecture) were read against the conjecture as Hovey-Strickland and Barthel-Heard-Naumann state it. The release has no Lean formalization for this family (no lean/docs page at the pinned commit). The spectral classification is not proved from scratch: it uses the Barthel-Heard-Naumann implication from invariant ideals to thick tensor ideals (their Theorems 4.9 and 4.15), resting on Mathew's descent and Balmer's nilpotence criterion. The new input is the classification of invariant primes in characteristic p via perfectoid and Fargues-Fontaine curve methods; characteristic-zero primes use the Gross-Hopkins period map. The proof was not refereed here.

Sources

Changelog1 change

Discussion