For me, a comment like this one: http://news.ycombinator.com/item?id=4034170 exemplifies some of the deterioration. The commenter asks for substantiation of a claim that refers back to the original article that was submitted. If he'd read the article, he'd know that.
Now, it's easy to argue that this is a one-time example, and alone that's true. But I'm noticing these kinds of comments more and more—ones in which people comment without knowing anything, or without reading the original article, or even without closely reading the person who they're replying to.
No one has found a good solution to this problem. As an open community grows it will inevitably experience Eternal September and content quality will decrease. It happened with Usenet, Slashdot, Reddit, and Hacker News is following their footsteps. The upper echelon will migrate, others will follow.
Perhaps closed communities with a curated member list is the answer[1][2]. Closed communities are not without its own drawbacks, but the trade off may be acceptable to some.
Open communities allows for anyone's input, but you also get drama like arguing about Javascript semicolons or GitHub's commit messages. However, closed communities are susceptible to group think. How do you judge a minority opinion's validity?
HN probably slower than the others for the same reason that PostgreSQL seems to be perpetually getting better....
I think the decay you note is specific to communities who are not coming in with any sort of common self-interest at stake. In essence they are what I might call "intellectual entertainment" communities. People read and argue on Slashdot primarily because it's fun. People argue on PostgreSQL email lists because they all want PostgreSQL to be as good as it can be. In the former, a troll can stir up a lot of argument, but in the latter, the most that can be hoped for is a collegial discussion on the merits of a specific approach.
HN leans towards the former. I am betting most commenters are mostly interested in fun conversations. However, the fact is that it is also an integral part of a VC ecosystem and so presents an aspect of common economic interest as well.
I think you're hitting on an important point by discussing different types of conversations. My theory (based on many years of Usenet) is that there are three basic types of online participants: "cocktail party", "scientific conference", and "debate team". In "cocktail party", the participants are having an entertaining conversation and sharing anecdotes. In "scientific conference", the participants are trying to increase knowledge and solve problems. In "debate team", the participants are trying to prove their point is right.
HN was originally largely in the "scientific conference" mode, with very smart people discussing areas in which they were experts. Now HN has much more "cocktail party" flavor, with smart people chatting about random things they often know little about. And certain subjects (e.g. economics, Apple, sexism, piracy) bring out the "debate team" commenters.
Any of the three types can carry on happily by themself. However, much of the problem comes when the types of conversation mix. The "cocktail party" conversations will annoy the "scientific conference" readers, since half of what they say is wrong. Conversely, the "scientific conference" commenters come across as pedantic when they interrupt a fun conversation with facts or "citation needed". A conversation between "debate team" and one of the other groups obviously goes nowhere.
I think the comment karma in HN encourages "cocktail party" conversation since you're as likely to get upvoted for trivial chitchat as a carefully-reasoned expert statement. (See http://news.ycombinator.com/bestcomments) Also, as I just found out, "expired link" on HN encourages quick comments rather than slowly written ones.
(And yes, I realize the irony that this is a "cocktail party" style comment of with my random opinion. I've actually considered ways of making this categorization quantitative, but it would take way too much time.)
For me, a comment like this one: http://news.ycombinator.com/item?id=4034170 exemplifies some of the deterioration. The commenter asks for substantiation of a claim that refers back to the original article that was submitted. If he'd read the article, he'd know that.
Now, it's easy to argue that this is a one-time example, and alone that's true. But I'm noticing these kinds of comments more and more—ones in which people comment without knowing anything, or without reading the original article, or even without closely reading the person who they're replying to.
Unfortunately, I don't have a solution.