I have yet to read Steven Pinker's Better Angels of Our Nature, though I've read one or two of his articles expounding it's central thesis: that over time humans have become less violent.
I'd tend to think that the basic thesis is probably true, at least to the extent that we consider this as personal, physical violence. Humanity has become more affluent over the centuries, and in general more affluent people are engage in personal violence less. Further (and this Pinker does not seem to take into account) European/Western culture has become increasingly dominant the world over, and that brings with it a bit of vestigial Christianity and Christianity's opposition to revenge and needless violence.
However, while Pinker's central point seems to have some amount of validity, he seems to have engaged in a fair amount of "too good to check" acceptance of very shoddy figures in support of his thesis. Humphrey Clarke of Quodlibeta has been doing a very good series of posts digging into the bogus statistics, odd assumptions, and misrepresentations that crept into Pinker's work. His most recent highlights some particularly egregious claims. For instance:
The three founders of Protestantism, Luther, Calvin, and Henry VIII, had thousands of heretics were burned at the stake, as they and their followers took Jesus literally when he said, “If a man abide not in me, he is cast forth as a branch, and is withered; and men gather them, and cast them into the fire, and they are burned.”
Protestants certainly killed their share of heretics over the years, but as Humphrey points out, the three individuals that Pinker accuses can't really be accused of killing "thousands". More like "about a hundred". That distortion, however, is nothing compared to the next:
Christian conquistadors massacred and enslaved native Americans in vast numbers, and perhaps twenty million were killed in all (not counting unintentional epidemics) by the European settlement of the Americas.
However, not only is it hard to come up with any rational way of arriving at this number after you exclude diseases, but Pinker apparently arrived at this number simply by averaging some estimates provided on the Necrometrics.com website -- estimates which the author of the page states to be unreliable guesses on the part of the authors of the studies involved.
This may not undercut Pinker's basic thesis, but it does certainly call into question whether he had any business writing a book (rather than a brief essay or two) if his research was going to be so sketchy in places.