A proposed complex structure on the six-sphere
An explicit compact complex threefold, built as a family of 2-tori over the (3,4,infinity) orbifold, is argued to be diffeomorphic to the six-sphere, which if it holds settles a question Hopf raised in 1948.
- Lab
- Anthropic
- Model
- Claude, version not stated
- Field
- Mathematics
- Date
- 2026-08-23
- Human collaborators
- Levent Alpöge
- Problem posed
- 1948 · open 78 yrs
Sources
Original work
Media coverage
What was found
Whether the six-sphere admits a complex structure is the surviving case of a question going back to Hopf: the two-sphere and the six-sphere are the only spheres carrying an almost complex structure, and only the six-dimensional case is open. The manuscript builds a threefold X fibred over the projective line by complex 2-tori, using the period functions of the (3,4,infinity) triangle group, degenerating to a degree-six del Pezzo fibre at one special point and to bielliptic multiple fibres of multiplicities 3 and 4 at the other two. It argues that X is simply connected with the integral homology of the six-sphere, hence diffeomorphic to it, and that its algebraic dimension is exactly 1. That last number is where the paper collides with the literature, and it says so itself rather than passing over it.
Novelty check
The problem is named, ancient and heavily attempted: Hopf posed it in 1948, and the modern literature runs through Chern, Bryant, LeBrun and the Campana-Demailly-Peternell papers the manuscript cites and disputes. Its 60-item bibliography is real and checkable. The specific construction, a family of 2-tori over the (3,4,infinity) orbifold completed at its three special points, does not appear in the prior literature, and no accepted complex structure on the six-sphere exists. Atiyah's 2016 attempt is the best-known recent failure and is not this argument. The construction is new; whether it is correct is a separate question this entry does not settle.
Caveats and known objections
Enters at claimed, and there are three separate reasons not to grade it higher. First, the AI attribution rests on a social-media post and nothing else. The 108-page manuscript was read here in full: it has no acknowledgements section, no methods note and no mention of Claude, Anthropic or any paraphrase anywhere in its text. The author's public post names the model and credits it substantively, but says nothing about which parts of the work it did, and no stable direct link to that post could be confirmed. Autonomy is graded ai-assisted because that is all the disclosure can carry, not because the contribution is known to be small. The model version is not stated. Second, the result contradicts a published theorem. The paper carries a section headed on why the argument of the Campana-Demailly-Peternell corrigendum does not apply, and a remark stating that its theorem contradicts that paper's Corollary 2.3 as published, locating the divergence in a non-vanishing higher direct image and in a lemma there that assumes trivial monodromy where here the monodromy is non-trivial. Worth reading both ways: the 1998 paper's Lemma 1.5 was itself wrong and the 2020 paper is the corrigendum, so that literature is known to be fragile, which makes the collision survivable without making the claim likelier. Third, the problem has a long history of false proofs. Not peer-reviewed, not formalized, and no independent check is recorded.
Also recorded at
vibemathedmodular-family-of-2-tori-as-a-complex-structure-on-s6
Nothing mechanically. It is a parallel listing of the same result, carrying its own verification label rather than an independent check.
Nobody outside the lab has checked this yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
Entry history (1 event)
- AddedEntered the registry graded Claimed and AI-assisted.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Graded claimed for verification and ai-assisted for autonomy. What these mean.
Cite this entry
whataifound.org. (2026). A proposed complex structure on the six-sphere. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-08-23-s6-complex-structureBibTeX
@misc{whataifound-anthropic-2026-structure,
title = {A proposed complex structure on the six-sphere},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Anthropic. Verification: Claimed. Autonomy: AI-assisted.},
url = {https://whataifound.org/finding/2026-08-23-s6-complex-structure}
}