Generally, this book is a good layman's introduction into embodied AI. However, there are some flaws:
Embodied AI has actually been around for a long time, so most of the conclusions, advantages and disadvantages
are already known to an expert - but still nice to hear and compare with one's own experience, and of course new to non-experts.
There were some new ideas which were quite refreshing. But still, one cannot fail to get the impression that the authors have
done the easy parts of this new approach to AI, and expect others to do the hard parts, which may or may not be
infeasible. That is normally not a good sign for an upcoming research field, and contradicts their conclusion that
we may make more progress in this direction. Only time will tell this, but for now I would bet against it.
Also, two authors and using I makes for cumbersome constructs such as I (Josh) vs. I (Rolf), which could easily have been avoided.
A good layman's introduction to new AI, although some chapters seem very contrived (e.g. applying new AI to business models)