Keith Devlin is one of those rare research mathematicians who is able to make recent advances in mathematics understandable and interesting to those whose mathematical education is obsolete or incomplete. I'm in the former category, having done my graduate work in pure math 50 years ago; although I've tried to keep up, constraints of time and other obligations have made it difficult.
Most modern texts on set theory put the reader to sleep, either because they avoid the important parts ("Set Theory for Those who Don't Want to Know It") or because they employ a degree of formalism that is quite difficult to grasp ("Set Theory Derived by Pure Propositional Logic, Step by Step"). Devlin's book avoids both traps. He presents modern advanced material that illuminates the subject admirably, but is careful not to submerge the reader in overwhelming finicky details. His discussions of constructive set theory, of independence proofs in set theory, and of non-well-founded set theory, are the first ones I've seen that get me excited enough to put the book aside and start exploring some of the implications on my own.
If I search for anything about the book to criticize, I find only one very minor thing. The sequence of proofs that show "Zorn's Lemma", the Axiom of Choice, the well-ordering principle, "Tukey's Lemma", etc to be equivalent to one another as an addition to the traditional Zermolo-Frankel axioms would be clearer if prefaced by an intuitive discssion of why the various steps in the chain of reasoning "ought" to work as they do; such a discussion helped me a lot many years ago to internalize what's going on. But that comment is just a nit.
On the other extreme, having once, 30+ years ago, being forced by the exigencies of a real-world problem to blunder through the creation of my own version of fragments of non-well-founded set theory, it gives me much joy to see it exounded as a coherent mthematical topic.
I read and reread this book, and drag it off the shelf when it occurs to me to ponder on some aspect that I don't fully recall. There are a number of other books on topics in pure mathematics about which I feel the same way, but they are a tiny minority among the deluge of texts that will never be read by anyone who doesn't have to. It's obviously an excellent text for advanced undergraduates and beginning graduate students, but beyond that, I recommend it to anyone with a working knowledge of pure math whose knowledge of set theory is somewhat behind current knowledge.
In short, buy a copy!