An easy win for hard decisions — LessWrong