This is exactly what you want to avoid Goodhart's law. Mesaoptimization isn't a bug to be solved, it is literally the mechanism by which humans can avoid Goodharting and be moral beings in a way not…
Posters: jd_pressman. jd_pressman
2 likes recorded by the source corpus.
Topics: Consciousness, AI safety
View Post: https://x.com/jd_pressman/status/1812203751698207088