I also ran into the hitlery porn shit, that is rude man. No wonder people say what they do. I still like you though, even though our bacon number is 2. You know Tim May, Anton, Lizard, Jamey Donnald? WoT!!!

The problem with the "economic agents know they don't know" statement is that it requires possibly one of the most difficult mental skills for the human race. For some reason this is just... hard!

We are not discussing the same thing here. In my model the car has an objective value attached to it, which would be its “true value”. In your model the car has no objective value attached to it, and instead it is represented by context-values. Thus it ends up having one value in the buyer’s context (presumably some sort of utility-value) and another in the seller’s (presumably some sort of “what can it be sold for” value). Thus in my perspective the market operates to discover the “true value” of the car, within arbitrary precision, whereas in your perspective the market operates to negotiate translation between contexts.

While I agree that when discussing a more complex problem, this difference would be very important, in this case objective-value approach can be translated to one presented by me quite easily. The value for buyers should be exactly the same as value for sellers. Then, obviously, the market will be broken. Still, I believe that my question remains valid: not every market is the one for goods with objective value (marginal value for marginal buyers seems much more frequent in reality) and thus to determine if asymmetric information is harmful one must inspect that particular market and not invoke the theory.

As to the alleged petitio principii, it might be a result of inadequate clarification from my part. My reasoning was as follows:
The theory points that upon asymmetric information, it will be either relatively irrelevant (as in my case A) or destructive (case B). If market exists, then it is probably case A. Unless participants are deluded and do not know that they do not possess perfect knowledge, or equilibrium is not yet reached in case B.
I of course agree that market participants' delusion wreaks havoc on both the market and participants themselves, I noted it explicitly:

For an exchange to be mutually beneficial, it is only necessary that the economic actors know that they do not have perfect information, which is much weaker assumption than symmetric information.

For the Indians, the utility-value of cars is normally distributed in the [1, 100] interval. For the Dutch, the utility-value of cars is normally distributed in the [200, 400] interval. (This was actually the case historically, with various spices, tobacco and whatnot. Pomodori are ethimologically reflective of the situation.)

In this highly contrived nonsensical context, should the buyers cartel (as they normally would in similar, real circumstances) and offer 50 per car, they will acquire all the cars in the [1, 49] interval, or roughly half the cars. This again is patent nonsense, as the Indian who sees his friend sell his 13 worth car for a 300% profit is not likely to sell his 49 car for a 2% profit, and this is also how the Dutch find out about how to value cars in the Indian system, but hey.

The average profit of the Indian in this theoretical example would be 25 per sold car. The average profit of the Dutchman would be 250 per bought car. Together they realise a profit of 275, which is unevenly split.

Admitting no further cars are produced as a result of this, and the demand on the Dutch side is not satisfyied, the next round will result in as many cars being sold, for a price of 100 each, yielding the Indians the same 25 per sold car in profit, but the Dutch merely 200, for a total of 225, which is somewhat more evenly split this time.

At this point it's game over and the Dutch have all the fishies (which is not exactly unlike the result of "free market" interactions between the actual Indians and the actual Europeans in the continental US, as we can find today in the field). The Dutch continue to happily trade their cars among themselves for something in the [200, 400] range and that's that. The price has increased towards average values and the profit distribution has pushed towards balance, slightly, somewhat, for as long as the Indians still had any cars at least. After which it was a 100% Dutch affair and good luck Daisy.