I applaud the author for the effort, but it's not good to mix philosophy with AI/mathematics. Or you end up with wrong conclusions similar to Zeno's.
The AIs in this are more human than machines. The pseudo-moral dilemmas don't help either. It's still fairly good if one is to disregard these though.