Skip to content
Hot eventDeveloping

Mathematician Rebuilds Master's Thesis Proof on Riemann Hypothesis Using Multi-Agent AI

1 reports1 reporting sourcesUpdated 1 h ago

Start here

AI digest

A mathematician used a self-built harness of Claude Code and Codex agents, running on $20 subscription plans, to rebuild the proof from his 2011 master's thesis on squarefree numbers in arithmetic progressions with polynomial moduli. In the neutral run, the system reached the main-term asymptotic for x^(7/9+δ) ≤ f(y) ≤ x^(1−δ); Claude Sonnet 5.5 wrote nine proposition proofs, while Codex handled planning, numerics, refereeing and integration. The AI readings also flagged bookkeeping gaps in the printed thesis and a factor-of-two error in its leading constant. The experiments corrected that constant to about 0.138083 for f(k)=k²+1.

Generated from reports · updated 1 h ago

Report timeline

Follow the reports to see the event from different sides.

November 18
  1. Hacker News · AI
    Rebuilding a Master's Thesis Proof on the Riemann Hypothesis With a Multi-Agent AI Harness

    A mathematician rebuilt the proof from his 2011 master's thesis on squarefree numbers in arithmetic progressions with polynomial moduli using a self-built harness of Claude Code and Codex agents on $20 subscription plans. The neutral run reached the main-term asymptotic for x^(7/9+δ) ≤ f(y) ≤ x^(1−δ), with nine proposition proofs written by Claude Sonnet 5.5 and Codex handling planning, numerics, refereeing and integration. The AI readings also flagged bookkeeping gaps in the printed thesis and a factor-of-two error in its leading constant, which the experiments corrected to about 0.138083 for f(k)=k²+1.

Heat for this event

Not enough continuous observations yet to draw a trend.