Mitt Romney's $10,000 bet — LessWrong