Dan Luu

Also, what happens if you try to optimize for "people I know are good readers understanding" instead of "minimize Internet static"? I'd worry that the latter is too noisy (hah) for most improvements to be noticeable

I think that this may result in fewer of what you call good readers because of the way so much traffic is driven through news aggregators & social media. I think (?) a lot of people decide what to read based on reading the comments and/or the title and

from having spent too much time commenting, I've noticed that a seeming refutation will get upvoted about the same regardless of whether or not it's right, and if someone replies with a seeming counter-refutation, the upvote rate will slow down, again regardless of correcntess

People don't seem to actually care about the information in the comment and judge things by how right-sounding the comment is. I think (?) an incorrect refutation that sits at the top of a comment section will cause people to not read the article.

If I have some anti-nitpick clause about the thing someone is complaining about, on HN, someone will respond that this was addressed in the article, which will become the top response and knock out the comment

If it's on reddit, the 2nd level response won't get any traction unless it's humorous and there's no effect, and on slashdot the 2nd level response usually won't happen, and if it does, it will end up +2 vs. a +5, insightful or something

Maybe this doesn't actually matter? It's hard to tell. But I think that potential readers sometimes don't read because the top comment is something like "this guy obviously doesn't know anything about CPUs since this idea was first seen in 1973".

This also maybe doesn't matter as much nowadays (if it ever did) because most of my traffic no longer comes from news aggregators now that my blog has been around for a long time.