In a nutshell, everyone in the games industry should read this book! It doesn't matter if you're a designer, programmer, artist or producer, a beginner or a veteran - if you don't find something in the book that justifies the asking price and time invested then either a) you're too stubborn to learn or b) why haven't you written your own book yet?
OK, that's the good bit, and it's broad, so for the rest of the review I'll concentrate on the weaknesses that make it not quite perfect.
Firstly, the name of the book is rather misleading. Whilst the book does contain some good advice on game architecture and game design, it is actually not what the majority of the book is about. A quick glance at the table of contents shows that there is a major section on project management, and actually that's what the majority of the architecture and design sections focus on to. Whilst there is some specific advice on techniques, algorithms or whatnot to add to your game, the main focus throughout is on developing an effective *process*. It's basically a manifesto for making better games. That's what makes the book so strong... the games industry is full of people with great ideas for games, and great programming skill etc, but as a rule we have the management skill of a dead slug. This book seeks to address that problem. Even if you're not at management level, the advice and ideas in the book will be very useful to you in your self-management, and hopefully will help you to streamline your team.
That said, sometimes the ideas will be useful by giving you better ideas when you disagree with them. The book preaches a method called the "software factory", which does have many merits and probably is a very efficient way to make solid games - but I doubt it could ever lead to an exceptional one.
Although it occasionally denies the fact, the book does seem to promote a notion that is very common to people outside the industry, and occasionally present and harmful within it - that of a solitary game designer heroically bringing his unique vision and genius to life, either with the aid of a team of artists and programmers eager to lap up his sage advice or (more commonly, one expects) struggling against the team that fails to appreciate his special skills. The book may give the wannabe game designer the idea that he will be responsible for creating the story, characters, game mechanics, level features, art direction and even code architecture for his project, leaving the rest of the team to fill in the details. Well, maybe there's a dozen people in the world capable of filling all those roles on a modern game project - and at most that many development teams willing to work with them.
The book states at one point that "the future belongs to the visionary, the dreamer". Not in the games industry it doesn't - it belongs to those designers with the technical skill and knowledge to turn a vision or a dream into a solid, consistent and satisfying set of game mechanics and level designs. Contrary to popular belief, good ideas are extremely common, and can come from any member of a development team. The skills to take those ideas and turn them into a real fully fleshed out system that results in player satisfaction are much rarer, and much more direly needed. BUT, this book does give any game developer a better chance of developing those skills, I think. However, I think the creation of a modern AAA title needs to be a much more collaborative process, drawing on the creativity and passion of the whole team, not just a few individuals on it. If nothing else, the team will produce much better work if they're involved on the creative level.
Back to the problems with the book - organisation is another. The book is grouped into 3 sections - design, production and architecture. Logically, the order should be design, architecture, production, and the book would read a lot better for it (and a bunch of production-related ideas that end up in the architecture section could be put in the appropriate place).
Lastly, although the book claims it is "not a programming book", it spends an inordinate amount of time on issues like bracing styles and commenting conventions. This information is very specific to programmers, unlike the rest of the book, and does not really belong here will it will doubtless make non-programmers stop reading. Furthermore, the issues and information are covered better in a myriad of other books, so the authors would have been better to simply refer the reader to one of those.
Another comment, though this book suffers less from it than most books written on games - the authors often seem to think of pc games as the only games, and the "real time strategy" types of games are the most frequently used as an example. It seems to be that programmers on these types of projects are the most likely to write about it, as a huge amount of the literature on games seems to have this bias that is unfortunately not reflective of the real distribution of the games on the store shelves. It hurts this book less than some others though.
OK, so there are some flaws with the structure and some of the advice needs to be taken with a pinch of salt, but hopefully those will be fixed in the next edition - and even with the warts, the book is a very readable source of invaluable ideas, written by people that clearly have battle-won experience in development. As such, unless you've just shipped your 10th million seller and have nothing left to learn, I guarantee that it is worth your time to read this book. Your next game will be better for it.