Mathematician models OpenAI's 372-theorem drop as a self-organized criticality collapse
littmath · x · 2026-10-15
Responding to OpenAI's October 6 "mathocalypse" paper with 372 theorem proofs, mathematician Dan Romik offers a physics-flavored framing.
- The frontier of research mathematics behaves like an Abelian Sandpile Model (Bak, Tang & Wiesenfeld, 1987) exhibiting self-organized criticality: mostly stable, but a single grain can trigger avalanches of any size.
- The OpenAI drop — a mass of results one step away from existing human knowledge — acted like dumping a handful of sand, collapsing the system to sub-criticality; the community is now in a post-avalanche rebuilding phase.
- The essay addresses whether human math research is dead, which classes of results AI threatens most, and how mathematicians should feel.
A 6-minute read, written without AI, per the author.
More from AGI Musings
- Frontier models keep cracking decades-old problems — a cross-discipline roundup — IgorCarron · 2026-10-15
- Aleksa Gordić: "Do you remember the world before AIs?" The lift is already real — gordic_aleksa · 2026-10-15
- 10-year Treasury yields back near 5%: is this the end of the Great Stagnation, priced by AI capex? — deanwball · 2026-10-15
- Asking Is Free, Answering Is Expensive: The Case for Building the Other Side of Communication — r0ck3t23 · 2026-10-15
- AI Slop Is Flooding Comment Sections, Pushing Creators Toward Gated Communities — alliekmiller · 2026-10-15
- UCLA paper: productivity tech like AI can trigger a workplace "rat race" that hurts welfare — soumitrashukla9 · 2026-10-15