Intelligence Amplification Open Thread — LessWrong