I think that Scott Alexander or whatever his name is, does a much better job than say, Paul Graham, at presenting the best counterargument whenever he presents something that he believes, but it's intrinsically a game of trust.

I think what the New York Times intrinsically understands is that the art persuasion is not actually the art of being rational and critical. It is the art of convincing someone else that you have been rational and critical. I am not generally defending the New York Times, I just think that they do understand how good faith persuasive essays can be gateways to bad faith persuasive essays.

The existence and effectiveness of bad faith persuasion is nowhere clearer than in court, where you are essentially obligated by the system to act in bad faith (lawyers would not describe it this way, but presenting an argument that you know to be flawed is bad faith), and lawyers are very convincing. However without an adversarial court system, lawyers are about as good a way of getting to the truth of the matter as crystal balls.

A reasonable, but not infallible heuristic for evaluating an essay for good faith, is to just imagine how much different the essay would look if counterarguments could be inserted after reaching some threshold of support by the general population.

The aggregate problem remains, however. Persuasive essays maintain a shared credit score which is extremely problematic for the process of truth-seeking. Even essays from the same author, can operate on wildly different points on the good/bad faith spectrum.

I guess my point is that the solution is not "more rational essayists", but rather provably adversarial comment systems, in which authors are not advantaged.