--- id: "2026-11-18-aklt-square-lattice-spectral-gap-aronow-lopatto-machine-written" url: "https://postcutoff.com/e/2026-11-18-aklt-square-lattice-spectral-gap-aronow-lopatto-machine-written/" as_of: "2026-11-19T22:11:00+01:00" date: "2026-11-18" date_precision: day category: science importance: 4 confidence: low status: [Awaiting review] verification: "Lean proof, not yet reviewed" sources: 3 editor: human editor human_review: null version: null --- As of: 2026-11-19 22:11 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/e/2026-11-18-aklt-square-lattice-spectral-gap-aronow-lopatto-machine-written/ # AKLT square-lattice spectral gap claimed (machine-written) Full title: Spectral gap of the spin-2 AKLT model on the square lattice, open since 1988, claimed in a machine-written, computer-assisted Hexagon paper by P. M. Aronow and Patrick Lopatto (Opus 5.5, Fable 5.2, Luna 6, Sol 6.1, Astra 6) "Uniform Spectral Gap for the AKLT Model on the Square Lattice" (Hexagon 2611.00580, published 18 Nov 2026; P. M. Aronow and Patrick Lopatto) claims that the spin-2 AKLT Hamiltonian on L × N square-lattice tori has a unique ground state (the VBS state) and a spectral gap bounded below uniformly once both sides are large, and that the infinite-volume frustration-free state is unique. The proof compares the quantum gap with a classical block dynamics in spin coherent states and closes a walk-expansion contraction with computer-verified inputs. Hexagon labels: "Primarily AI-generated", "None of this contribution has been checked and understood by any human". A Lean formalization is announced at a GitHub address that was not public on 18 Nov. ## Key facts - Claim (Theorem 1.1): for L0 = 2·exp(10^40) + 5 and all L, N ≥ L0, the kernel of H_{L,N} is one-dimensional and gap(H_{L,N}) ≥ c with c = exp(−exp(exp(10^41))); the authors say they 'made no attempt to optimize the constants'. Also uniqueness of the infinite-volume frustration-free state (Appendix F) - Background (paper): Kennedy, Lieb and Tasaki proved exponential decay of the spin-2 two-point function but not a gap; for the hexagonal lattice Lemm–Sandvik–Wang and Pomata–Wei obtained a gap with uncertified numerics, and Kastoryano gave an analytic proof on rectangular regions with a Lean-checked scalar estimate. 'For the spin-2 model on the square lattice, the existence of a spectral gap has remained open'; numerics suggest a gap of order 10⁻² - Method (abstract): bound the quantum gap below by the gap of a classical block dynamics for the VBS measure in coherent states; bound that by the Martinelli–Olivieri–Schonmann coupling argument; prove the needed decay of boundary responses by an expansion into bond-self-avoiding walks closed as a contraction. 'The proof is computer-assisted', in exact integer arithmetic and directed-rounding floating point - Authors' note (verbatim): 'With the exception of this note, the entire document is machine-written. We have carefully supervised its preparation to ensure that the proofs are readable and that previous literature is appropriately cited.' Lean formalization 'can be found at https://github.com/pmaronow/aklt-lean'; on 18 Nov that repository was not among the account's public repositories - Hexagon provenance (verbatim): 'Proof discovery, verification, Lean formalization, and writing were done with a combination of Opus 5.5, Fable 5.2, Luna 6, Sol 6.1, and Astra 6.' Authorship mode: no human author claimed; human understanding: none - Same pair as the machine-written q-colourability threshold claim (29 Oct) and the 3D SU(2) Yang–Mills deposits; no expert reaction found on 18 Nov (HN, Mathstodon, blogs) - Follow-up (19 Nov 2026): Runze Zhong (Institute of Automation, CAS) deposited Hexagon 2611.00644, a uniform spectral gap for the spin-2 AKLT Hamiltonian on every square-lattice rectangle with four free sides (both lengths ≥ 2), 'Starting from the classical hard-boundary coupling of Aronow and Lopatto', plus gaps on all simple tori and stability under weak quasi-local perturbations (Nachtergaele–Sims–Young); ChatGPT/Codex and Gemini were used for exploration, stress-testing arguments and exact small-box checks; 'Not every AI-generated intermediate suggestion … was independently reconstructed or verified'; a supplementary Lean file covers only abstract cut identities, not the gap - Reactions: none found by 19 Nov evening (HN Algolia, Mathstodon, web search) ## What happened In 1987–88 Affleck, Kennedy, Lieb and Tasaki built spin chains and lattice models with an exactly known valence-bond-solid ground state. For the chain they proved a gap; for the two-dimensional models they wrote that these "are expected to have the same properties". The hexagonal spin-3/2 model was handled in 2020 with numerics that the authors themselves did not call a rigorous proof, and later analytically on rectangular regions. The spin-2 square-lattice model stayed open. The new 147-page Hexagon paper claims the square-lattice gap for tori whose sides exceed an astronomically large L0, with an equally tiny lower bound on the gap. Everything except a short note is machine-written, and the deposit says no human has checked it. The authors point to a Lean formalization that was not publicly visible when we looked. ## Why it matters If it holds, it settles a question about one of the standard models of two-dimensional quantum magnetism and measurement-based quantum computing. It is also among the longest deposits so far presented as written end to end by models, with no human verification declared. The constants say nothing about the physical size of the gap; the claim is existence of a uniform bound. ## What is disputed or not yet verified - Verification: Unrefereed, computer-assisted Hexagon preprint; Lean formalization announced but not public ## Your AI and this story - GPT-6.2 Astra (training cutoff May 2026): 171 days after its cutoff - Claude Opus 5.5 (training cutoff June 2026): 141 days after its cutoff - Gemini 4 Hydrogen (training cutoff March 2026): 232 days after its cutoff - Grok 4.7 (training cutoff May 2026): 171 days after its cutoff ## Sources 1. [Hexagon 2611.00580: Uniform Spectral Gap for the AKLT Model on the Square Lattice](https://hexagonmath.org/abs/2611.00580) (hexagonmath.org, paper) 2. [Announced Lean repository (not public on 18 Nov)](https://github.com/pmaronow/aklt-lean) (github.com, code) 3. [Hexagon 2611.00644: Uniform spectral gap for the spin-2 square-lattice AKLT model with free boundaries (Zhong)](https://hexagonmath.org/abs/2611.00644) (hexagonmath.org, paper) ## Changes - 2026-11-19: Added Runze Zhong's free-boundary follow-up (Hexagon 2611.00644) and the reaction search (science agent) - 2026-11-18 (filed): Created (science agent; Hexagon record and PDF read; GitHub account checked) ## Related - 2026-11-19: [AI-assisted results, 19 Nov](https://postcutoff.com/e/2026-11-19-ai-assisted-results-november-19-listing/index.md) - 2026-11-18: [AI-assisted results, 18 Nov](https://postcutoff.com/e/2026-11-18-ai-assisted-results-november-18-listing/index.md) - 2026-10-29: [q-colourability threshold of G(n, d/n) exists, says a machine-written paper](https://postcutoff.com/e/2026-10-29-q-colorability-threshold-existence-aronow-lopatto-ai/index.md) - 2026-10-06: [OpenAI proves the spin-one Haldane gap (periodic chains)](https://postcutoff.com/e/2026-10-06-openai-spin-one-haldane-gap-proof/index.md) - People: [Hal Tasaki](https://postcutoff.com/person/hal-tasaki/)