Riemann Hypothesis proof passes independent Lean kernel checks: 17M lines, zero rejections
flngr · x · 2026-11-21
- MIT's Andrew Sutherland and others independently compiled the 17-million-line, 175,000-file Lean proof of the Riemann Hypothesis produced by Levent Alpöge with Claude. The Lean kernel accepted it using only the three standard axioms, with Mathlib's own RiemannHypothesis definition, no sorry and no custom axioms.
- Independent kernel checks then passed: nanoda (a Rust checker) accepted the proof in 28 minutes; con-leche accepted all 1.47M declarations and confirmed the five theorems (including RH) match Mathlib-only statements; lean4lean passed all 175,187 modules; Lean's own kernel replay had zero rejections.
- With every independent checker agreeing, a Lean bug is essentially ruled out. Two slower checks are still running with nothing failing so far.
- What remains is the human part: no human-readable proof exists yet, and the Clay Millennium Prize requires publication plus two years of review. As the author puts it: RH is proven, but not yet understood.
More from AGI Musings
- "RH is not even a sliver of what is coming": investor says AI math breakthroughs validate his forecasts — burny_tech · 2026-11-21
- If AI proves the Riemann Hypothesis, it also proves verifiable tasks aren't general intelligence — bendee983 · 2026-11-21
- OpenAI's littmath: low p(doom), but p(gloom) higher than she'd like — mjdramstead · 2026-11-21
- Poster predicts Bitter Lesson will dominate everyone's minds in 2027 — akbirthko · 2026-11-21
- Pearl weighs in: JEPA world models can answer counterfactuals with two predictor calls — yudapearl · 2026-11-21
- Truth-seeking vs community-seeking: why AI doomers and boomers can't understand each other — gleech · 2026-11-21