Taleb's book pursues several central ideas that seem fairly commonsensical. The major theme is about the titular Black Swans - events that are ignored by forecasters (be they financial, social, or whatever) because they are either considered too unlikely to be worth taking into account or are not even noticed due to the narrow focus or inadequate models being used. Taleb points out that, rather than being the effect of gradual change, many of the most important changes to our world, for better or worse, are massive upheavals caused by unpredicted events. Financial crises, inventions, wars; so often these things are unpredicted and seem unpredictable, and cause fundamental transformation.
Taleb doesn't advance a model for predicting these things - indeed, part of his thesis is that the fervour for prediction tends to lead to a narrowing of foresight and the expression of false certainty, making the effect of the subsequent (and inevitable) unforeseen events all the more cataclysmic. Instead, he argues for a mindset that is open to seeing the possibilities rather than focusing on certain expectations.
Some of the blame for this outlook he places on the ubiquity of the Gaussian bell curve, or normal distribution, which is excellent for ranking the spread of certain characteristics (height and weight across a population, for example, where the vast majority will fall within a narrow band of the mean and outliers becoming increasingly rare - you probably know lots of people who are six feet tall, a few who are 6'4", and would be dumbstruck on meeting someone 8'), inadequate in areas with a massive disparity (income, book or record sales) and disastrous when used as a predictor - something unlikely can have a far larger effect than all the many median-scale common occurrences, such as a tsunami or a financial meltdown.
The author uses the terms Mediocristan (the place in which events follow the normal distribution) and Extremistan (the place where a significant number of events fall outside of it) and argues that we live with the dual problem that, while many of the vital mechanisms of our world are extreme we tend toward a certain blindness that leads us to think they are mediocre, that we like to form patterns and narratives that seem to make sense of the world and are then unable to respond when the world refuses to fit into these patterns.
While I find this difficult to dispute - especially within the utterly pretend-scientific world of economics which is the target of much of Taleb's vitriol - he tends to argue in terms so broad that this does somewhat dilute the strength of his points. For instance, he rails against the bell curve as a blemish on cognition, eventually conceding that it is just not the tool for the job to which it is being put, and this is emblematic of his approach.
Another example is his preference for the company of people with a broad intellectual scope rather than narrow expertise, and his extension to arguing that the former are generally preferable in all fields. Again, this is difficult to argue with on the face of things - many intellectuals whom I admire greatly have argued against the tendency of our educational system to train students in a limited fashion rather than educate them thoroughly - but the fact is that many of the most brilliant people who have provided the greatest boon to society are narrowly focused experts - not through training, perhaps, but because of targeted obsession in a particular field. If Taleb is arguing for a greater scope of education and interdisciplinary appreciation, I don’t think the point can be denied but, again, his argument is scattershot and it is not entirely clear what he is arguing for.
A final problem is that, early on, Taleb betrays our trust. He tells the story of a Russian author who, unable to get publishers interested in her experimental novel, published privately and became a huge success. Some time later he admits that she is a complete fabrication - and then returns to her in other examples throughout the book. I assume that he is drawing attention to the fallacy of narrative, about which he complains that we place too much trust. We are, as others have said, animals that tell stories to make sense of the world. This meta-textual trick of course makes us take what he writes with a pinch of salt, as we should, but as he uses personal anecdotes frequently throughout to support his points, the thought that he may well be making them all up further weaken his case.
All this said, The Black Swan is an interesting read from a man who obviously has some good ideas and relishes intellect and discussion. I imagine that he is a voluble and entertaining speaker - and probably an excellent conversationalist and dinner guest.