x
Towards A Unified Theory Of Alignment — LessWrong