"We live on an island surrounded by a sea of ignorance. As our island of knowledge grows, so does the shore of our ignorance." - John Wheeler
Training in highly specialized knowledge is a lot like building and growing an island. You learn more and more, but learn that you actually know very little. That is, as you learn more you learn more about what you don't know.
Building islands is still useful, however, as the bigger your island of knowledge gets, the more complex and interesting the edifices you can build on that island become. So what that you go a little stir crazy? So what that you never leave your island? Careers are made out of lush, ever-growing, ever more sophisticated islands.
"A philosopher is a person who knows less and less about more and more, until he knows nothing about everything. A scientist is a person who knows more and more about less and less, until he knows everything about nothing." - John Ziman
The alternative is little better, and less-well paid. Instead of a single massive island, the generalist builds an archipelago. Generalized knowledge forms atolls, barely subsisting above the waves, perhaps big enough to be substantial, but hardly big enough to exploit.
Instead, the generalist is most productive not in building on individual islands, but in traversing the space between them. But once there are too many you start to know less and less about more and more, and you spend all of your time travelling between islands accomplishing nothing.
I've characterized both of these in the negative, but they have their virtues. The specialist certainly can construct impressive edifices of knowledge, and the generalist can become quite expert in communicating across boundaries. The problem is that, with few exceptions, specialization is vastly better paid and more respected than generalization, even though both are essential to a functioning knowledge economy.