Very up-to-date, high-quality summaries of all the current methods in practical, adversarial AI. A must-read for anyone doing work in the field or students pursuing a career in AI.As a professional AI researcher and developer, I originally planned to just read a few of the articles touching on techniques I use or am interested in, but ended up reading it cover to cover. The quality of the articles and applicability to current research is phenomenal. Kudos to Steven Rabin for really excellent selection and editing. All scholarly mumbo jumbo has been fully excluded. The descriptions of the techniques are clear and easy-to-understand.This second volume is a big improvement over volume one, which tended to have a lot of idiosyncratic articles you would not care about unless you were interested in the exact, squirrely case for which the author was decribing the solution. Volume two has much better coverage of key methods like A*, MCTS, minimax, constraint-based programming and other techniques which have general applicability. As an example of this, volume one had an article describing the architecture used for implementing the kill bots in Killzone 3 which was very interesting and informative, but not anything you would be likely to use directly, even if you were solving that same problem. Volume two has articles more like the one in Chapter 3 on dual-utility reasoning by Kevin Dill. Dill gives a nice two-page summary of dual-utility and describes how he used it in Zoo Tycoon 2. For the AI developer this is really useful and something you might end up actually implementing or leveraging in some way. Also, volume 2 has better literature references and better grounding in the scholarly state of the art.Game AI Pro 1 was excellent, but this new edition ups the ante and is even better, providing even more practical and apropos articles.