Rational overconfidence in the tens of billions: recent example — LessWrong