I loved the message of the book - I only found the first half (a description of the problem) to be useful and eye-opening, however, since the second half where he tries to offer solutions to overengineering or underdesigning products seems a bit haphazard and random.
Definitely worth a read though just to open people's eyes about the problem with most software engineering. I've found this issue at 2 of the 3 startups I've worked at in the last year.
I have mentioned the book before as one of my faves. Bill Buxton is a brilliant guy and all hackers should check out the book.