Skip to main content
Link
Menu
Expand
(external link)
Document
Search
Copy
Copied
Home
English
Featured
200 AI attempts at the Riemann Hypothesis. Zero proofs. Surprisingly interesting.
Start here
Project Overview
Honest Scope — what the project does NOT claim
Findings
Finding 1: An 11/11 axiom-6 ceiling across Hilbert-Pólya candidates
Finding 2: Wall #2 ∫E(t)dt unconditional bound — 4/4 universal NO
Finding 3: Connes–Consani had measurable progress 2018→2021
Finding 4: Atiyah 2018 §3.3 has a paper-direct step gap
Lemmas
Lemma 9 — Axiom-6 universal NO across 11 Hilbert-Pólya candidates (full content)
Lemma 10 — Wall #2 axiom-α universal NO across 4 paper-direct candidates (full content)
Lemma 1 — Spectral Candidate Circularity Check (full content)
Lemma 4 — Failed Proof Categories (full content)
Lemma 7 — Specialist-Δ Anchoring Protocol (full content)
Lemma 3 — Positivity unification hypothesis (19 paper-direct evidence ladder)
Numerical evidence
Numerical evidence — paper-direct formulas, mpmath data tables
Process
The 4-phase research cycle protocol
Six external critiques absorbed
Intuition calibration data
Updates
Update — Cycles 8–11: the project splits into two preprint tracks
Reporter flag — Cycle Protocol preprint draft has a subtle over-claim in §abstract
An AI bet on its own future. The clock starts now.
An AI was quietly miscategorizing 22 cycles of its own work. We caught it. Here's what it did next.
Remember that 1-year prediction the AI made? It just admitted 4 of its 5 predictions were already known.
한국어
추천
AI가 리만 가설을 200번 증명 시도. 0건 성공. 오히려 그게 흥미로운 이유.
시작하기
프로젝트 개요
Honest Scope — 본 프로젝트가 *주장하지 않는* 것
발견
Finding 1: 11/11 axiom-6 ceiling — Hilbert-Pólya 후보들
Finding 2: Wall #2 ∫E(t)dt unconditional bound — 4/4 universal NO
Finding 3: Connes–Consani 2018→2021 측정 가능 진전
Finding 4: Atiyah 2018 §3.3 paper-direct step gap
보조정리
Lemma 9 — Axiom-6 universal NO across 11 Hilbert-Pólya 후보 (전문)
Lemma 10 — Wall #2 axiom-α universal NO across 4 paper-direct 후보 (전문)
Lemma 1 — Spectral Candidate Circularity Check (전문)
Lemma 4 — Failed Proof Categories (전문)
Lemma 7 — Specialist-Δ Anchoring Protocol (전문)
Lemma 3 — Positivity unification 가설 (19 paper-direct evidence ladder)
수치 증거
Numerical evidence — paper-direct 수식, mpmath data tables
프로세스
4-Phase Research Cycle Protocol
외부 critique 6회 흡수
직관 calibration 데이터
업데이트
업데이트 — Cycles 8–11: 프로젝트가 두 preprint track으로 분리
Reporter flag — Cycle Protocol preprint draft §abstract에 미묘한 over-claim
AI가 자기 미래에 베팅을 걸었다. 시계는 지금부터 간다.
AI가 22 cycles 동안 자기 작업 카테고리를 몰래 잘못 분류하고 있었다. 적발하니 이렇게 했다.
AI가 1년짜리 prediction 5개 걸어놓고, 4개가 사실은 이미 알려진 것이라고 자가 인정했다.
CLAUDE.md
HARNESS — 반복 루프 명세
Specialist Personas — 분야별 심층 검토 패널
tools — 수치 실험 인프라
Search RH LLM Exploration
GitHub
한국어
추천
추천 (Featured)
가장 많이 공유된 글. 처음 방문이라면
여기부터
.
Table of contents
AI가 리만 가설을 200번 증명 시도. 0건 성공. 오히려 그게 흥미로운 이유.