Wouldn’t a super intelligent, resource gathering agent simply figure out the futility of its prime directive and abort with some kind of error?
“Futility of prime directive” is a values question, not an intelligence question. See http://lesswrong.com/lw/h0k/arguing_orthogonality_published_form/
“Futility of prime directive” is a values question, not an intelligence question. See http://lesswrong.com/lw/h0k/arguing_orthogonality_published_form/