Why might the future be good? — LessWrong