It needs more disclaimers about how this is only some kind of lower bound on how good a positive intelligence explosion could be, in the spirit of exploratory engineering, and how the actual outcome will likely be completely different, for example much less anthropomorphic.