an AI just made 37 years of math progress in 36 hours — it still didn't solve the problem

Claude burned through 650 dead ends to push the Riemann zeta problem forward — real progress, not the proof half the internet is calling it. Aliteq Lena Fischer · AI & Local Compute Editor

Anthropic's unreleased research Claude raised the proven lower bound on Riemann zeta zeros from 41.6% to 67.2%, announced August 10, 2026.

The field had moved that bound only fractions of a percentage point over the prior decades — one Menlo Ventures observer called this the field's biggest jump in 37 years.

Claude reached it by autonomously running 60 AI subagents across roughly 1.5 days, testing 650 failed ideas and 2,400 shell commands before combining two existing papers no mathematician had…

This does not prove the Riemann Hypothesis. Oxford mathematician James Maynard says there's 'no pathway' from this technique to the actual hypothesis.

Anthropic mathematicians Levent Alpöge and Ralph Furman reviewed the result, external experts Brian Conrey and Dan Goldston examined it, and Claude produced a formally verifiable Lean proof.

This is the most genuinely interesting AI-for-math result I've seen this year, and I also think it'll get flattened into 'AI solves math' by every aggregator that didn't read past the headline. What…

Read the full story

Aliteq an AI just made 37 years of math progress in 36 hours — it still didn't solve the problem Read the full story on Aliteq