People universally say this book is one of the most difficult (if not most difficult) of the philosophy books, and they love taking pieces out of context to show how Kant is wrong. After having listen to this masterpiece, they are misleading on both points.
First, do not listen to the overview and summary until you have listened to the whole book. Start the book at chapter 18. I made the mistake of listening to the book linearly from the beginning, and got overwhelmed by the overview and summary. I went back and re-listened to them and found them edifying. The exact opposite from how I felt when I heard them before reading the book.
Kant wants to establish absolute knowledge as real. Up to his point in time (1781), there was a dichotomy regarding knowledge, empirical v. rational (Hume v Locke). Kant does his best to bridge that gap. He'll get detailed in developing categories that we use for our conceptions (quality, quantity, relations, and modal (real v. imaginary) and he'll cross that with unity, plurality, and totality). This is an area where it got difficult to follow since he was definitely referring to tables that I had to keep recreating in my head. He's doing all this because he want's to show that concepts (ideas) can come about and will be true. Oh yeah, he's going to give us the pure category of space and time which reside within our brains. My point of saying all of this, is to just show that he is not that hard to follow.
A little more context, our perceptions give reality (i.e. the thing in itself must be constructed by our senses). Or in other words, there is the immediate v. the mediate. The thing in itself verse the filter of the brain. The thing we perceive v. reality. But, Kant is setting the reader up for his tearing down of most of philosophy. I would strongly recommend listening (or watching) the Dan Robinson 8 hour lecture series he gave at Oxford for a general audience of students and guest freely available through Itunes or on Open Culture.
After Kant lays the ground work he starts dismantling of the standard proofs for the existence of God, and the immortal soul, and the immaterial soul. He uses the standard theistic proofs: Ontological (i.e. Saint Anslem's 'since you can think of a perfect being there must be a perfect being'), Teleological (i.e. by design, he calls it 'physical theology'), and the Cosmological argument (i.e. first cause). He does finer arguments for the atheist cause than I have read in any modern atheist handbook. In the end, he 'proves' God by appealing to practical reason (contrasted with pure reason) and the certainty of man's (and woman's) morality toward well being in general.
A big part of why he wrote the book lies elsewhere. He'll say that the nature of science is to use the inductive method, to go from the particular to the general, and from the general to create a set of principals. These principals are what he calls 'apoditic' (i.e., beyond dispute). That is what gives us our necessary (and certain) truths. Truths are not contingent (and probable) but become necessary (and certain). He'll say that our understanding come about through our intuitions (both empirical and non-empirical) which determine events and lead to our concepts.
Don't be so fast to dismiss what he has to say. He's writing at the very end of the Age of Enlightenment, and Newton and his Principia are believed to be absolutely true and necessary truths. Newton says "I will feign no hypothesis'. He says that in reference to not being able to say what gravity really is, but he also believes he made no other hypotheses and statements not completely backed by data. I have many times argued with Physicist that truth is not absolute, and they will always come back "oh yeah, what about force equals mass times acceleration", and I will respond, "yes, but Einstein came up with the relativistic correction, and so that is not true", and if I haven't completely bored them I will go on to explain how F=ma is a tautology, and if they haven't yet left me due to disinterest like most people who will read this review, I will even show how in mathematics no one can define what a set is with out being circular (i.e., tautological, a word that Kant uses frequently in this book. So know that it just means the conclusion is included in the premise).
Kant will divide knowledge into synthetic and analytical. Synthetic (and the trick I used, since it begins with 'S' think senses) requires empirical knowledge gathered from the senses. Analytical, think mathematical truths. At its heart math is the study of changeless relations. Relations, are one of the four concepts that make up the twelve categories. Kant believes that mathematics is entwined with the real world. A triangle only makes sense since it can be visualized. He needs that in order to fully bridge his gap between the rational and the empirical. As for the truth regarding the nature of a triangle, your guess is as good as mine.
The reason I like this book so much I can state by paraphrasing something Kant said. He talks about Hume at length and does show him the utmost respect (I would even think that Hume would have liked this book), and says "that it's not so much that I can win the argument by reason, but that my reason I have employed is useful and the same methods can be used by others". Kant is up front by criticizing dogmatic arguments as boorish and self serving. He'll say that the loudest is not necessarily the most right, and the problem with the ignorant is they never know they are ignorant.
There are many pearls of wisdom with in this shell and it only has to be opened up and read in order to profit from it.