But this amounts to "it's better to avoid talking the way that pretty much every human being not in a minority of literal-minded Internet users talks in most contexts".

Which I also agree with! Ceteris paribus, would it not be better if people were less keen to assert literal falsities in debates about facts?

I do of course give people more leeway on this in most contexts. If someone says something that's literally false in a face-to-face conversation, but their intended claim is clear and basically accurate, I'm unlikely to bother contradicting them. But we are in fact talking about something written on Less Wrong, and I'm OK with applying a higher standard here.

This is of course correct, yet it still doesn't change that.

I can parse that but I don't understand the point it's making.

And the correct way to respond to someone doing that is to respond when you think they're doing it--not to be uncharitable and interpret every non-literal statement as literal regardless of whether you think it''s a case of that or not .

For whatever it's worth, I think there was a fair chance Azathoth123 was doing it unconsciously out of habit. (They have something of a track record of saying things I find incredible, apparently completely guilelessly.) And the effect of doing it can be pretty much the same, regardless of intention, so there's a reason to flag when someone's engaging in the behaviour even if there's no objective evidence of mischievous intent. (Plus, again, this is LW, not a live chat where people have to compose sentences in real time, so why not exercise higher standards?)

Sure. I'm not arguing that he's right. But "he's wrong even when I interpret him charitably" is not a reason to interpret him uncharitably.

But "he's wrong even when I interpret him charitably" drains the force from an objection of the form "that's an uncharitable interpretation". If someone's wrong under both available interpretations, they're just wrong; why not respond as such?