Antidote to Pascal's Wager — LessWrong