Blackburn is one of the main figures in contemporary metaethics, and this is his big book on the issues. He tries to lay out the general details of his approach to ethics here, so it's a book that's essential reading for anyone interesting in metaethics. And it's a book that's slowly grown on me as I've studied it.
That said, it's evasive and frustrating at places, and it's longer than it needs to be. Blackburn spends too much time on extraneous stuff, and provides too little detail about the absolutely crucial material in the first, third, and ninth chapters. These are the chapters in which Blackburn lays out the fundamentals of his favored form of noncognitivism, explains the nature of the quasi-realist project, and attempts to answer the objection that his views lead to subjectivism or relativism. Furthermore, there's just not enough engagement with the literature in large parts of this book--the obvious exception being chapter 4, which may be the book's best chapter. (For those who'd like a better introduction to Blackburn's views in ethics, I'd recommend chapters 5 and 6 in his earlier Spreading the Word; and the essays on metaethics in Essays in Quasi-Realism are probably provide a better account for the expert.)
Here's a recap of what goes on in the main chapters here. Not much of signal importance happens in chapter 1, but Blackburn very briefly explains why he thinks ethics is essentially practical and tries to say a bit about the sort of emotional and attitudinal states that are relevant to ethics. Chapter 3, which is the heart of the book, gives us the basics of Blackburn's expressivism, introduces quasi-realism, and takes up some challenges to it (e.g. unasserted contexts, motivational externalism, moral truth and moral facts, etc.). Chapter 4 is Blackburn's attack on all forms of cognitivism and moral realism; he argues against reductivist realisms, Cornell realism, and McDowell's and Wiggins's views. Chapters 7 and 8 are concerned with the respective places of reason and sentiment in ethics; Blackburn's on the side of sentiment. Chapter 9 is his attempt to answer challenges to expressivism alleging that it leads to subjectivism or relativism. And there is a very helpful appendix that clarifies what Blackburn takes his quasi-realist project to be and why he thinks expressivism is preferable to other metaethical views.
The remaining chapters are interesting, but inessential. The discussion of issues in normative ethics in chapter 2 is underdeveloped and largely unnecessary. Blackburn comes out in favor of consequentialism on the grounds that virtue theories and deontological theories need to appeal to consequentialist considerations in order to make sense of virtues and duties. The material about egoism and game theory in chapters 5 and 6 is true and important--though none of it is terribly original and it's hard to see why it plays a crucial role here.
With all that out of the way, I'll try to put some philosophical meat on the bone by outlining what I take to be Blackburn's central metaethical views. First, Blackburn's expressivism. Blackburn's expressivism is a noncognitivist account of moral language; it claims that moral language is (primarily) used to express attitudes. If theory is correct, our moral practice is guided by the aim of expressing our own attitudes about parts of the natural world and coordinating our attitudes with those of other people. Consequently, we do not need to posit moral facts, nor do we need to posit any special faculty for arriving at moral knowledge.
The aim of Blackburn's project is pretty straightforward: Blackburn's is a project of naturalizing ethics. He wants to understand ethical thought and language as part of a naturalistic conception of human nature. The most obvious way to naturalize ethics would be to attempt a reduction of the moral to the natural. But this isn't the route that Blackburn takes. Indeed, his expressivism is inconsistent with taking this approach to reconciling moralizing with his naturalism. Blackburn thinks we ought to "synthesize" moral propositions in order to understand moralizing naturalistically. And synthesizing the moral proposition is not a matter of reducing it to anything else; it is a matter of understanding its place within a broader naturalistic account of human beings and of moralizing as a human activity.
For Blackburn, then, the emphasis is on explanation rather than reduction. In this explanation the expressivist starts with the activity of moralizing. Why do we have an activity like this? What naturalistic explanation do we have for the practice of moralizing and for the existence of moral language and thought? Starting with answers to these questions amounts to synthesizing the moral proposition rather than analyzing it. We don't begin with ordinary moral claims and try to find some natural facts that make them true or false. Rather, we begin with a naturalistic account of the world and our place within it, and we try to explain why we think morally and why we use moral language in the way we do.
Now, it's not that those who aren't expressivists cannot offer any explanation of moralizing. The problem for cognitivists is that there's a central and essential element of morality they simply cannot explain. According to Blackburn, the cognitivist's explanation cannot account for the practical dimension of morality. Moralizing is a practical activity: that is, it's an activity that leads to and coordinates action in a group of people. There's an essential tie between moralizing and acting, and the cognitivist's explanation appears to leave this out of the picture. Why is it left out? It's open to a person to simply not care about the moral facts. Some people might care about these moral facts but that turns out to be a contingent fact about human psychology.
The fundamental virtue of expressivism, Blackburn thinks, is that it alone succeeds in explaining moralizing in a way that is consistent with naturalism, that it alone makes sense of why we moralize in a way that is consistent with the best account of the world and of human beings that is provided by the natural sciences. Expressivism accounts for the essential practicality of moralizing. For moralizing, if the expressivist is right, is primarily a matter of expressing one's attitudes, and attitudes possess a necessary connection to action.