Some Claims Are Just Too Extraordinary — LessWrong