safety
opinion
bearish
Lack of value generalisation is a fundamental reason for the hardness of the AI alignment problem.
I believe that lack of value generalisation is a fundamental reason for the hardness of the problem
AI Alignment Forum01 Sept 2026