I recently bought the book after having read the raving reviews here on Amazon. Before I turn to my critique, it has to be said that Naughton makes a fine effort in bringing together the whole context (or at least some sort of plausible context) of "the internet". There are some not uninteresting bits of information on various side issues such as ham radio.
Anyway, as for the core topic -- the internet -- it turns out the book is little more than a mix of pieces taken very much in sequence from the awesome and much underrated Hafner and Lyon book, some actually very funny manual type of sections on things like how to use a browser to click on hyperlinks, and towards the end a little bit of Raymond and Lessig inspired musings about how much open source is better than proprietary software, and how the internet is threatened by corporate giants.
For a serious researcher this book is almost totally useless as an original source of information. Also, there are some strange asides such as on page 147 "'Real e-mail dates from 1970" with a footnote stating that "For some reason, Hafner and Lyon ... date it as 'one day in 1972', but this must be wrong because the RFC archive shows a flurry of discussions of a mail protocol in the summer and autumn of 1971." This explanation makes no sense to me, for there have been all sorts of dead end RFCs, especially in the very early days. I could elaborate the discussion on what qualifies as the "first email ever" much further, but the crucial point is that Naughton offers very little authoritative information and introduces quite a bit of subjectivity on the sources he builds on.
As an aside, don't even waste your time with the Abbate book, just get the Hafner and Lyon book and get to the original sources of the BBN guys, the NWG, Pouzin, Cerf, and the more recent Dave Clark papers on design principles and the internet.