x
Beliefs about formal methods and AI safety — LessWrong