Meaningful research into the ethics of AI requires both a deep understanding of the hard science of AI, and the hard philosophy of ethics, yet much of the research into AI ethics treats it as a "soft" field, i.e. as license for populist activism and naive sociology.
I respect Eliezer very much, and his arguments are never shoddy, but I don't see how and to which degree he hopes to preserve human esthetics once he treats minds as mutable. Presently, AI ethics are more about governance and social regulation, and he seems less interested there.