One potential problem with the scenario outlined here by Krugman is the assumption that an economy organised principally around information will precipitate the devaluing of that knowledge. What's true for oil and steel is not necessarily true for information (a generic categorisation if there ever was one), and Krugman draws this analogy far too glibly. If anything, possessing information makes it easier to acquire more and more valuable information; the technological exponentiation we are living through today makes that point clearly. If information does not obey the basic rule of scarcity that physical goods do, it seems unreasonable to assume automatically that the effect of "information abundance" on the economy will be analogous to a glut of oil or grain.
A lot is also made of the outsourcing of "knowledge workers"; hardly a new phenomenon, even in 1996. And it will almost certainly still exist in 2096. But the inequities in living standards and expectations that makes outsourcing practical currently is steadily being eradicated by the gradual improvement in economic conditions which outsourcing itself helps to make possible. In some sense, the condition is probably self-throttling. The presumptive end of all this is a high-entropy state, where there is no strong "potential difference" between economies to make outsourcing worthwhile: it would eventually become pointless for everyone. I don't imagine we'll reach that point by 2096, though :-)
"But the inequities in living standards and expectations that makes outsourcing practical currently is steadily being eradicated by the gradual improvement in economic conditions which outsourcing itself helps to make possible."
In effect, it is an arbitrage opportunity where you have a large number of well educated people with a low standard of living, combined with cost of living differences compared to western countries.
Certainly, over the next 100 years such arbitrage opportunities should abate, as people around the world with valued skills have similar standards and costs of living, and thus demand similar wages. We seem to be seeing the beginnings of that process already.
A lot is also made of the outsourcing of "knowledge workers"; hardly a new phenomenon, even in 1996. And it will almost certainly still exist in 2096. But the inequities in living standards and expectations that makes outsourcing practical currently is steadily being eradicated by the gradual improvement in economic conditions which outsourcing itself helps to make possible. In some sense, the condition is probably self-throttling. The presumptive end of all this is a high-entropy state, where there is no strong "potential difference" between economies to make outsourcing worthwhile: it would eventually become pointless for everyone. I don't imagine we'll reach that point by 2096, though :-)