VibeMathedMath problems solved by AI

Word-length spectral triples as compact quantum metric spaces

A countable discrete group with a proper length function carries a natural spectral triple on its reduced group C*-algebra. A well-studied question in non-commutative metric geometry asks whether the associated Connes pseudo-metric always recovers the weak-* topology on the state space, making it a compact quantum metric space in Rieffel's sense. It holds for groups of polynomial growth and for word-hyperbolic groups, and it was widely expected that not every word-length function works - but no explicit counterexample was known. False: for every integer d2d \ge 2 the canonical spectral triple of the Lamplighter group (Z/2Z)Fd(\mathbb{Z}/2\mathbb{Z}) \wr \mathbb{F}_d, with the word-length function of a finite symmetric generating set, fails to be a spectral metric space.

Result
Disproved(see note)
Status
Resolved
AI contribution
AI co-developed
Method
Construction
Field
Operator algebras
Posed by
Marc Rieffel
Year posed
2002
Years open
24y
Solved
2026-08-12
Model
GPT-5.6 Sol
Vendor
OpenAI
Collaborators
Mario Klisse
Verification
Unreviewed
Publication
Preprint
Significance
20 / 100
Disclosed cost
Wikipedia
No dedicated article

What was actually shown

The first explicit counterexample rather than a first suspicion: the abstract is clear that the failure was widely expected and that what was missing was a witness. It gives an infinite family, one for each d >= 2, all Lamplighter groups over free groups.

What the AI did

The acknowledgements state: "The author acknowledges the use of GPT-5.6 Sol as an exploratory tool to assist in finding the counterexample. The AI was used under the author's strict mathematical guidance. All mathematical content and arguments were rigorously reviewed, verified, and substantially revised by the author, who assumes full responsibility for the final manuscript." The model helped find the central object, which is an essential named step, but the framing is explicitly human-led and the author reports substantially revising everything - the co-developed tier rather than the discovered one.

Verification

arXiv preprint, one day old at cataloguing, not peer-reviewed and with no independent commentary yet. This site has not verified the argument; unlike a finite counterexample, this one is an infinite family and an analytic failure of a topology-recovery property, so it is not settleable by computation.

Sources

Changelog1 change
  • Rasmus Lindahladded this entry

Discussion