What? Your theorem doesn't even match with the basic prisoner's dilemma. True, in the classic version of this game there's no communication... but the point is the same - just because an outcome is Pareto efficient doesn't mean that it's a stable equilibrium - people's rational self interest can result in worse outcomes. I don't know what kind of economics you're talking about. Most of the absurd simplifying assumptions really have to do with the idea that people have perfect foresight and their statistical guesses are unbiased - and that we perfectly understand variance/risk and all that jazz. It doesn't have as much to do with how we communicate with one another, rather that we are super awesome flawless statistical machines when in reality we all suck at statistics (have all kinds of biases and heuristics to rely on) in every day life.

And just because information is cheaper doesn't mean it's free. At the very least it costs us time - at least in the current form of the internet it's not like there's simultaneous, effortless, instant communication between every human being on Earth. Now that might change things.

But I don't think you fully made the case that the internet is fundamentally different than a "better" or more effective printing press. My point, I guess, is that maybe we approach that zero cost world, but will never reach it. And that there's a fundamental distinction between being at zero cost and being at any positive cost... so that the internet is more like a printing press rather than a zero cost communication mechanism.