Evidence item

Speaker reports persistent rumours that early reasoning models were trained on or tested against inverting secure hash functions. HEARSAY

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 3832

Source: OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes · locator {} · excerpt hash sha256:752fec7d316b5d6e195f10aa5540a12c4dddd31a802088c5157ed9e4feec2bec

there were these very persistent rumors around the same time of Qstar and Strawberry that these early reasoning models that were purportedly alien were being trained off of or at least tested against objectives of inverting cryptographically secure hash functions like the Shaw cipher suite.