A good way for me to maintain motivation when slogging through information-dense nonfiction books (like textbooks, programming language manuals) is to read through actively (writing notes/questions in margins or post-its) and then to quickly "generate" something from what I just learned. For instance, if I am reading a math textbook, I will play with the theorems, adding and subtracting conditions and generating counterexamples, and then do all the problems and TeX them up. If I am reading a philosophy book, I will summarize the accounts of the section and then argue my own perspective, essentially writing a brief philosophy paper. If I am reading something that I tend to mostly disagree with, like GEB, I will even try and write a "devil's advocate" paper, where I try suspend by disbelief and figure out why the particular perspective has appeal by arguing against myself.

In the short run, this may mean that you take a lot more time to finish books, but it has drastically improved my ability to recall what I have learned and my ability to not abandon the process after I hit a wall. This not only because you are immediately reinforcing what you have just learned, but I've found that if I see the fruit of my labors, even if it's just a steadily growing PDF on my desktop or a short, elegant program in my home directory, it provides just enough to motivation to finish the next chapter, and the next chapter, until you're finally done.