x
Which types of AI alignment research are most likely to be good for all sentient beings? — LessWrong