Concerns Surrounding CEV: A case for human friendliness first — LessWrong