I'd definitely push back on the way you characterized _Applied_ and _Practical_ (now called _Cryptography Engineering_). The two books are very different and are the product of different authors; Schneier wrote _Applied_ but co-authored _Practical_.
_Applied_ is a broad survey of cryptographic techniques that was good for its time but has aged terribly and probably done more harm than good for the industry†.
_Practical_ is an engineering book; unlike _Applied_, which is a sightseeing tour of the field, _Practical_ is a book about actually building system with cryptography. Also unlike _Applied_, _Practical_ is diligent about recognizing the limitations of mass-market technical books; it explains things just as often to convince you not to implement things as it does to motivate you to implement them. _Practical_ does not have a lot of "rah-rah" in it. It's not an evangelistic book. The tone of its prose, a mix of casual, clear, and precise, is all Schneier, but the content is wildly different from _Applied_. It's a narrower book, not a "bigger" one.
You might be right about the progression of Schneier's other books. I'm not sold on the idea of a career in information security as a springboard to public policy research on security. Real-world security is not very much like information security at all.
On the specific issue of the book he's promoting right now, I think it's worth pointing out that Schneier has been wrong as often as he's been right about macro- Internet security. Example: Schneier was an early opponent of vulnerability research, for instance; he used his Crypto-Gram newsletter to single out eEye for irresponsibly disclosing Windows vulnerabilities. eEye was a pioneer of what we'd now call "Responsible Disclosure" (an Orwellian term whose basic function is to marginalize Metasploit) and employed people Riley Hassell, Derek Soeder, and Barnaby Jack, all of whom are now vulnerability research authorities.
† (though you can't necessarily say that for the field of cryptography, which he did more to popularize than anyone; perhaps there are lot of great postdoc crypto people today who got into the field because of _Applied_)