Bayeswatch 9: Zombies — LessWrong