Andrew Wiles worked on Fermat's last theorem for six years in complete secrecy while he was at Princeton. Thus it's simply not true that isolated concentration is not possible in the American math world.
It's possible, but the culture is not conducive to it. According to Simon Singh's book on the subject, Wiles had to go as far as stockpiling several years' worth of conference papers in advance in unrelated areas so he could submit one occasionally while he was really working on FLT, just to affect the appearance of productivity. He also taught a phony graduate course in which his main collaborator enrolled and most of the other students dropped out, just so they could use the class time to thrash out some of the details.
Right, the analogy between Wiles/Fermat and Perelman/Poincaré so glaringly contradicts what she's saying ("the American model may not be able to produce a breakthrough like the proof of the Poincaré Conjecture" - huh?) that one has to wonder. Perhaps this explains it?
Masha Gessen's latest book is [...] a story of Grigory Perelman and the Poincaré Conjecture.
Isolated concentration is an issue for more than just mathematicians... Everyone needs to serve their basic survival needs (food, sleep, sex) before focus on abstract subjects can take place; isolated concentration is a theme for anyone that enjoys something abstract only for the sake of itself.
This can include computer scientists, scientists, astronomers, philosophers, mathematicians, &c... The need for psychic solitude is common amongst thinkers in general and it is difficult to achieve it within the society we are in; particularly so when what you love to do is also your primary income generator. Hence why you see many table waiting programmers or chefs that love to do math at night.
