x
An AI-2027-like analysis of humans' goals and ethics with conservative results — LessWrong