An Effect of “Ems”
A large population might need a far smaller proportion of scholars. Present day scholar recruitment (including the idea that the most prestigious job is that of training more scholars) might have been more appropriate to smaller populations.
Very-large populations should erase the obscurity trap, in which nobody uses a potentially-important technology because you can't find experts in it and nobody tries to learn it because you can't get a job in it. In a very large society, hobbyists can jump start nearly anything.
Increases in population size can affect moral standards. At present, in a world of a few billion humans, there are plausible-sounding calls for the currently most-powerful nation to right every wrong (e.g., “We can stop the Darfur famine.”) and plausible-sounding claims that picking and choosing between wrongs to right is hypocritical. I suspect that with a very large population, similar ideas would be regarded as obvious nonsense.
Human interactions can be classified into four types: communal sharing, equality matching, authority ranking, and market pricing. The first three have limits: Communal sharing fall apart over 150 (the Dunbar number); equality matching falls apart over a few thousand (the size of an ancient Greek city-state); and authority ranking fall apart over a few million (the size of a large hierarchical organization). Does market pricing also have a limit? If it does, we might find out in the Age of Em.