FULL STORY

The S^6 Complex Structure Saga: AI-Assisted Claim Shakes Math

Anthropic mathematicians claimed a Claude-assisted proof that S^6 has a complex structure; subsequent talk of AI refutation sparked debate over prestige and the loss of elegant problems.

2026-08-24 ~ 2026-08-24 · 2 episodes · 8 posts

Episode 1 · Anthropic Mathematician Claims Claude Helped Prove S^6 Has a Complex Structure (2026-08-24, 6 posts)

TL;DR: Posts relayed by @Singularitarian, @soumitrashukla9, @scaling01 and others report that an Anthropic mathematician, working with Claude, claims to have proven that the six-dimensional sphere (S^6) admits a complex structure—a Hopf problem open since 1947/1948. The original poster's initial check found no errors and the breakthrough reportedly took about 3 days, but details await peer review.

Confirmed

  • Multiple relays agree on the core event: Claude contributed to a positive resolution of whether S^6 has a complex structure.
  • @soumitrashukla9 and @scaling01 note the problem has been open since 1947 and that Atiyah gave a flawed proof in 2016.
  • @scaling01 and @eldonredwards credit mathematician Alpoge with sharing the result; Ethan Mollick reshared it with amazement.
  • @Singularitarian relays that no errors were found in a preliminary check and the result may have taken only 3 days.

Unconfirmed

  • Final correctness of the proof: only a preliminary check has been done; relays stress details need verification.
  • The exact model version involved: relays variously mention Claude 3 and "Claude" generically.

Why it matters

  • If correct, this resolves a nearly 80-year-old problem notorious for attracting failed proofs by experts, so rigorous review is essential.
  • It is seen as a milestone for LLM-assisted research in highly abstract mathematics, with Claude credited with helping discover a long-sought geometric object.

Episode 2 · Debate Over AI Proving Complex S^6 (2026-08-24, 2 posts)

Mathematicians are debating the implications of AI disproving the Complex S^6 problem. While some lament the loss of an elegant problem, others argue that although prestige is scarce, interesting problems are not, countering pessimistic views on AI depleting mathematical resources.