But that's where I disagree with you. I think Bostrom et al's take on things is correct (at least potentially), and has nothing at all to do with the Kurzweilian norm. Bostrom's argument requires far fewer background assumptions.