Astra's proof sidesteps three deep Mathlib results by inventing a new 'capacity' quantity

IgorCarron · x · 2026-11-19

The quoted post notes that a published proof leans on three deep results not in Mathlib — and Astra used none of them. Instead it defined a new quantity it calls "capacity" and built the rest of the proof on it, doing genuinely novel mathematics along the way. The commenter's "And they fall..." suggests traditional library-dependent proof strategies are being upended by AI finding independent proof paths — a notable datapoint for frontier AI in formal mathematics beyond search-and-assemble over existing libraries.

Original post →

More from Models

Models channel →