why do trivial inconveniences matter? — LessWrong