x
The cup-holder paradox — LessWrong