Interesting read: "Can we make a moral, high integrity AI?" is a better question than 'alignment' & P(doom)