From the article:

Imagine a square matrix in which our n commodities are rows and columns. How much real information is there in this matrix? Obviously, the oil-oil exchange rate is always 1, and the oil-wheat and wheat-oil rates are the same thing. So half our boxes, plus n, become blank. But this is by no means enough blanking. Because we note an interesting fact - if this entire marketplace clears, it cannot be possible to construct a cycle of exchanges through which one ends up with more wheat, or oil, or silver, or anything, than one started with. This is obviously a desirable trade, and yet in a cleared market there are no desirable trades.

This statement is misleading economics - markets work because individuals preferences do not match the general consensus. Because of comparative advantage, there's no particular reason to expect individual valuations to ever exactly match the market rates.

Later, Moldbug notes people want accumulate money, but rejects the explanation that it is valued because it is the medium of exchange. It is confusing when the monetary object is actually of some value (i.e. gold or silver), but no one uses the gold standard any more. There's really no reason to treat fiat money as a good. And all of the discussion assumes that treating money as a good makes sense.

Finally, Moldbug's discussion of the transition from gold-standard to fiat money without discussing private currency at all seems like more selective history. All selective history is worthless - it's like throwing out all the data from an experiment that does not agree with your hypothesis.