(Once again, ignoring the language itself to focus on the book).
Lisp is one hell of a language. Lots of things that are getting traction today already exist in Lisp for a long time. And, still, Lisp is not a common language. And this book shows why.
Just before reaching 1/4 of the book, there was a little piece of code that had nothing less than six "close parenthesis" in it. Just to have an idea of how bad it is, one of my friends that know Lisp inside and out rewrote the same piece of code using two new functions and hell, it was a lot easier to read -- simple because Lisp must be read from right to left, which most written languages use the other way around and reducing the closing parenthesis by moving pieces of code to other functions really does help.
The book also fails in a lot of places: There is the constant abuse of global variables, which breaks the idea of functional programming, one of the pillars of the language; the whole code focus on the REPL and nowhere it talks about using an external editor and running the code, although in the very end a magical "load" appears, out of the blue; there is some serious lack of contextualization about concepts: things are simply thrown out and then explained (there is a piece about minimax which the author simply explains the code after showing it and, in the next chapter, he finally explains the minimax concept, only to tell that the code could be "fixed" to follow it in only 23 lines -- no shit, Sherlock, you wrote half of it without ever explaining why you were writing it).
In the end, it's not a "oh so terrible" book, but it could really use some editing help.