My current take on logical uncertainty — LessWrong