Rauch's essential point is that we now have computer power to explore the mapping between individual charactistics and modes of person-to-person interaction on the one hand, and societal-level aggregate outcomes on the other. Sociology thus becomes, virtually, an experimental science.

Thus we can now ask--and get answers--to questions like:

--How much of a preference for living near members of one's own ethnic groups is necessary for residential patterns to stabilize in near-complete segregation?

--How much of a tendency to trust the judgment of and imitate those who you know is necessary for a stock market to become subject to recurrent irrational bubbles and crashes?

--How is it possible that a society can turn from peaceful coexistence to inter-ethnic hobbesian war overnight?

How important the computer-based virtual experimental science of sociology will become is anyone's guess. But that it is coming is very clear.