Err... It's just a POD from scans of books.
http://news.cnet.com/8301-19882_3-10379830-250.html
I expect it will only be useful for researchers who can't access that library, and prefer a paper edition to an ebook. I don't really the problem, I expect Google will do very similar things (or who knows, maybe use HP's system) once the Google Books Settlement gets ironed out.
I assume they are currently just doing the public domain books. I'm a little surprised Amazon is involved, since last I heard they were strong-arming people to use their own POD services.