You could argue that eventually the finance industry becomes so large that it reaches equilibrium within the markets it operates (Mervyn king has suggested the same about the City of London), thus becoming zero-sum, but this nuance is completely absent.

Or perhaps you should consider that in a zero-sum game individual agents can produce great value, while the system as a whole does not.

This is a macroeconomics article, after all.

The staggering assertion that "facilitat[ing] the exchange of already-existing assets" is an activity "that cannot possibly increase human welfare" suggests that you're right about this, however much you may have thought you were being sarcastic.

One group digs the holes and the other fills them, and it can appear that productivity for both is good while the system doesn't particularly make any headway.

I don't care for peanuts (in fact, I'm allergic to them). The subjective value that I assign to them is zero (or negative)

Same thing, you don't care for shellfish (religious reasons).

Suppose we move around existing assets -- trade peanuts for shellfish.

All of a sudden both of us have something we can use, of non-zero subjective value (the only kind of value that exists, btw).

To say that no value has been created in the transaction is quite simply to ignore the basic universal motivation for trade.

Every voluntary transaction happens only because both parties believe that they're getting the better part of the deal (I'd rather have shellfish than peanuts).

This is completely false. As I've pointed out elsewhere, the possibility of trading existing assets means that one entity (say, a farm) can produce more of a good (say, rice) than it can directly consume. Without exchanging goods, this would make no sense, specialization of labor would be lost, and the production of all types of goods would crater.