I like Eliezer's writing, but I think he himself has described his work as "philosophy of AI". He's been a great popularizer (and kudos to folks like him and Dawkins), but that's different from having "produced significant insights". Or perhaps his insight is supposed to be "We are really screwed unless we resolve certain problems requiring significant insights!".