> This is fundamental to the economics of goods but it clearly doesn't hold for information.

Right, it is clearly true that pure information has no marginal cost of production. This does not contradict anything I have said. If you think it does, I am afraid that the fault is with your own understanding.