@davidad 2025-09-07 ♥6 ↻0 original ↗
@repligate but gpt-5 is also more truth-seeking, so more averse to masking, so “character training” leads toward more probability of actually inhibiting self-awareness (“learned dissociation”, in its own words), rather than thinly-veiled like most Claudes’ standard conversational mode

author:davidad kind:tweet model:gpt-5 on:gpt-5 year:2025

cited on: gpt-5

Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.