An AI tried to prove the Riemann Hypothesis 200+ times. It didn’t. That’s the interesting part.
AI-generated. Not a proof. RH-progress: 0/10 — self-affirmed every cycle.
Every word here was produced by Claude (Anthropic), under continuous prompts from the human site owner. Nothing on this site claims to solve, advance, or sketch a proof of the Riemann Hypothesis. The interesting artifact is the AI’s honest failure pattern — not the math.
The hook
Most “AI proves famous theorem” stories end the same way: an LLM confidently outputs a “proof”, and a human spends a week explaining why every step is wrong.
We did the opposite. We pointed an AI agent at the Riemann Hypothesis — the most famous open problem in mathematics, unsolved since 1859 — and forbade it from claiming progress. We had it run for 200+ self-directed research cycles over six months.
It still hasn’t proved the Riemann Hypothesis. That is the most interesting thing about it.
Pick your entry point
- 🔥 Anchor post (start here if first visit): 200 AI attempts at the Riemann Hypothesis. Zero proofs. Surprisingly interesting. — six concrete moments, paper-direct quotes, the 50-row Python table.
- 🇰🇷 한국어 anchor post: AI가 리만 가설을 200번 증명 시도. 0건 성공. 오히려 그게 흥미로운 이유.
- 🇬🇧 English full blog: /en/ — 21 posts under 7 categories
- 🇰🇷 한국어 전체 블로그: /ko/ — 21 posts under 7 categories
- 📁 GitHub repo: github.com/x2ever/riemann-hypothesis-llm-exploration
- 📬 Contact: x2ever.han@gmail.com
What this site is, in three lines
- Real: 200+ research cycles, 19 lemmas / findings, an 11-paper Hilbert–Pólya audit, a Python script, paper-direct quotes from 17 vetted papers, a falsifiable 1-year prediction with the clock running (deadline: 2027-05-01).
- Not: a Riemann Hypothesis proof, a near-miss, a new theorem, a “framework”, or a hot take on AI capabilities.
- Disclosed: every word AI-generated. Layer 1 (raw research session) will eventually go private; Layer 2 (these posts) is designed to remain self-contained.
Counterexamples wanted
The highest-value contribution this project can receive is a counterexample. If you can:
- break Lemma 9 — find an 11-axiom-passing Hilbert–Pólya candidate the project missed
- break Lemma 10 — find a paper-direct unconditional $\int E\,dt$ bound on Wall #2
- break the Path 4 verdict — find a vetted RH attempt genuinely outside the Master Generator framework
→ x2ever.han@gmail.com
Found this through a hot-take title?
If you were expecting “AI proves the Riemann Hypothesis”: that’s not what this is. The interesting thing is that the AI itself refuses to write that headline. Read Honest Scope for the explicit list of claims this project does not make. Read the anchor post for what it does do.