A 9th-grade student (Reddit user Gihyo0921) claims to have used Claude to prove that a rhombicosidodecahedron cannot pass through a copy of itself β a geometric conjecture about the 'Prince Rupert' problem for this Archimedean solid. The student reports using interval arithmetic across 12.3 million boxes over 38 CPU hours to verify singular configurations at β5. No independent verification or peer review is cited in the available material; the claim rests on a single Reddit post with computational details but no published code or formal proof artifact.
The case is a single Reddit post claiming an AI-assisted geometric proof with specific computational parameters (interval arithmetic, 12.3M boxes, 38 CPU hours) but no published code, formal proof artifact, or independent verification. It falls squarely in the 'discussed not deployed' tier of Scott's status ladder β a bare verdict without evidence packages, auditability, or mechanically different verifiers. While it topically overlaps ai-assisted-mathematics and formal-verification radar concepts, it neither challenges nor extends Scott's frameworks; it merely exemplifies the low-evidence tier those frameworks already classify. No consequential party arrives at Scott's position, and the specific claim is absent from his canon.
ip:concept.evidence-packageip:framework.discussed-is-not-deployed-status-ladderip:concept.auditabilityip:concept.deterministic-verification-before-assertionip:concept.verification-loopsip:concept.mechanically-different-verifiersradar:concept.ai-assisted-mathematicsradar:concept.formal-verificationradar:kbr-ai-formalized-math-proofradar:claude-borsuk-counterexampleradar:lean-transformer-ai-proofsradar:proofrun-local-agent-verification-receipts
queries asked of Scott's wikis
- ai-assisted-mathematics: LLM as co-author vs. verifier in formal proof workflows
- formal-verification: interval arithmetic as a trustworthy backend for LLM-generated conjectures
- reproducibility: what artifacts (code, certificates, logs) make an AI-assisted proof auditable
- geometry: Prince Rupert problems and polyhedral passage as a benchmark for AI spatial reasoning
- epistemology: when does an LLM-assisted computation constitute mathematical evidence
now 4 pts/hpeak 153 pts/hcomments 1/hpeers p88momentum: cooling2 platformsage 51h
points/hour across evidence Β· reading as of 2026-10-12 02:59:37.977291+11:00 Β· deterministic, not a model opinion