x
AI alignment open problem — LessWrong