x
Learning can be deciding — LessWrong