Claim Collision Engine
Where papers disagree, the platform writes its own testable hypotheses.
How to read this page. The engine takes falsifiable claims from different papers and checks, pair by pair, whether they genuinely disagree. Every judgment here is AI reasoning over the claims (provenance tier inferred) — not a new measurement. A Resolved verdict means the tension was reconciled from what the papers state, not that anything was proven. An Unresolved collision becomes a brand-new hypothesis the platform proposes — a suggestion to test, never a finding. The engine is instructed never to manufacture a conflict where none exists.
Loading collisions…