With Riemann Done, DeepMind's Irving Proposes Harder Prime-Generation Conjecture for AI

geoffreyirving · x · 2026-11-21

Following AI's formalization of Riemann-related theorems, DeepMind's Geoffrey Irving suggests moving to harder targets, proposing a 2020 conjecture: there is no O(n²) expected-time randomized algorithm to generate ≥n-bit primes — a next proving ground for AI mathematical reasoning.

Related event: DeepMind researcher asks what comes after Riemann — number theorists have no answer(8 posts)→

Original post →

More from Research

Research channel →