On knowing the Future — LessWrong