I think this site is dying because there’s nothing interesting to talk about anymore. Discussion is filled with META, MEETUP, SEQ RERUN, links to boring barely-relevant articles, and idea threads where the highest comment has more votes than the thread itself (i.e. a crappy idea). Main is not much better. Go to archive.org and compare (date chosen randomly, aside from being a while ago). I don’t think eternal september is the whole explanation here—you only need 1 good user to write a good article.
Discussion is filled with META, MEETUP, SEQ RERUN, links to boring barely-relevant articles
The website structure needs to be changed. “Main” and “Discussion” simply do not reflect the LW content today.
We should have a separate “Forum” (or some other name) category for all the non-article discussion threads like Open Thread, Media Thread, Group Rationality Thread, and stuff like this.
Then, the “Discussion” should be renamed to “Articles” (and possibly “Main” to “Main Articles”) to make it obvious what belongs there.
Everything else should be downvoted; optionally with a comment: “This belongs to the Open Thread”. (And if the author says they didn’t know that Open Thread exists, there is something seriously wrong… about the structure of the website.)
I feel like I wrote this to the LW discussions at least dozen times...
there’s nothing interesting to talk about anymore.
I think there are interesting things here. They are just drowned in too many less interesting things.
Let’s look at the numbers: 6 articles so far on Dec 9th; 6 articles on Dec 8th; 4 articles on Dec 7th; 11 articles on Dec 6th; 8 articles on Dec 5th; and some of the articles from Dec 4th—less than one week ago—already don’t fit on the first “Discussion” page. (The exact numbers may differ depending on your Preferences settings.) The page is scrolling insanely fast. If I stopped reading LW for one week, I would have problem to catch up with all the new stuff; I would probably just skip some of that. That’s not good if we have too much stuff, but low average quality.
We don’t downvote enough. Let me explain—if someone makes a post that is not very good, but is not completely stupid or trolling also, it will almost certainly gain more upvotes that downvotes. Because it feels wrong to punish someone only for being uninteresting. But in terms of rewarding/punishing behavior, we probably should punish them. If we try to be too friendly, the site will become boring, precisely because most of the social talk is not about giving new information.
Perhaps it could help to use some timeless deciding. If you read an article, ask yourself a question: “If the next week here would be 10 new articles like this, would it make LW better or worse?” If the answer is worse, downvote it. Because although the same author will not write 10 more articles like this during the next week, other authors will.
TL;DR—I think the Eternal September is most visible on the article level, because it is not obvious what kind of content belongs here. “Discussion” is horribly misleading—we don’t want discussion-level articles. That’s what the comments and Open Threads are for.
One issue with the LW/CFAR approach is that the focus is on getting better/more efficient at pursuing your goals, but not on deciding whether you’re applying your newfound superpowers to the right goals. (There’s a bit of this with efficient altruism, but those giving opportunities are more about moving people up Maslow’s hierarchy of needs, not on figuring out what to want when you’re not at subsistence level).
Luke’s recent post suggest that almost no one here has the prereqs to tackle metaphysics or normative ethics, but that always has seemed like the obvious next topic for rationality-minded people. I was glad when Luke was writing his Desirism sequences back at CSA, but it never got to the point where I had a decent enough model of what normative claims desirism made to be able to evaluate it.
Basically, I think these topics would let us set our sights a little higher than “Help me optimize my computer use” but I think one major hurdle is that it’s hard to tackle these topics in individual posts, and people may feel intimidated about starting sequences.
idea threads where the highest comment has more votes than the thread itself (i.e. a crappy idea)
It depends—if the higher-voted comments are expending on the original post, then I’d say the post was successful because it evoked good-quality thought, assuming that the voters have good judgement. If the higher-voted comments are refuting the original post, then it was probably a bad post.
One possibility is that the kind of content printing-spoon likes is easy to get wrong, and therefore easy to get voted down for, and therefore the system is set up with the wrong incentives (for the kind of content printing-spoon likes). I’d guess that for most users, the possibility of getting voted down is much more salient than the possibility of getting voted up. Getting voted down represents a form of semi-public humiliation (it’s not like reddit, where if you post something lame it gets downvoted and consequentially becomes obscure).
The great scientists often make this error. They fail to continue to plant the little acorns from which the mighty oak trees grow. They try to get the big thing right off. And that isn’t the way things go.
Overall, I suspect that LW could stand to rely less on downvoting in general as a means of influencing user behavior. It seems like meta threads of this type often go something like “there’s content X I hate, content Y I hate, and practically no content at all, really!” Well if you want more content, don’t disparage the people writing content! It may make sense to moderate voting behavior based on how much new stuff is being posted—if hardly any new stuff is being posted, be more willing to upvote. If there’s lots of stuff competing for attention, vote down lamer stuff so the good stuff gets the recognition it deserves.
I think we could stand to see high-karma LWers who rarely post in Main/Discussion post there more. Maybe make it impossible for anyone with over X karma to get voted below 0 in a Main or Discussion post. Or make a new subforum where high-karma users can post free of moderation. (I’ll admit, the oligarchical aspect of this appeals to me.)
Also, maybe be realistic about the fact that most people are not going to be willing to go to lukeprog/gwern lengths to dig up papers related to their posts, and figure out the best way to live with that.
Well, I tried to make a post once, got downvoted into oblivion, and decided not to put myself through that again. So yeah this happens for real, although perhaps in my case it is no big loss.
I’m not sure… I think the topics I find most interesting are simply used up (except for a few open questions on TDT or whatever). Also the recent focus on applied rationality / advice / CFAR stuff… this is a subject which seems to invite high numbers of low quality posts. In particular posts containing advice are generally stuffed with obvious generalizations and lack arguments or evidence beyond a simple anecdote.
Also, maybe the regular presence of EY’s sequences provided a standard for quality and topic that ensured other people’s posts were decent (I don’t think many people read seq reruns, especially not old users who are more likely to have good ideas).
I think this site is dying because there’s nothing interesting to talk about anymore. Discussion is filled with META, MEETUP, SEQ RERUN, links to boring barely-relevant articles, and idea threads where the highest comment has more votes than the thread itself (i.e. a crappy idea). Main is not much better. Go to archive.org and compare (date chosen randomly, aside from being a while ago). I don’t think eternal september is the whole explanation here—you only need 1 good user to write a good article.
The website structure needs to be changed. “Main” and “Discussion” simply do not reflect the LW content today.
We should have a separate “Forum” (or some other name) category for all the non-article discussion threads like Open Thread, Media Thread, Group Rationality Thread, and stuff like this.
Then, the “Discussion” should be renamed to “Articles” (and possibly “Main” to “Main Articles”) to make it obvious what belongs there.
Everything else should be downvoted; optionally with a comment: “This belongs to the Open Thread”. (And if the author says they didn’t know that Open Thread exists, there is something seriously wrong… about the structure of the website.)
I feel like I wrote this to the LW discussions at least dozen times...
I think there are interesting things here. They are just drowned in too many less interesting things.
Let’s look at the numbers: 6 articles so far on Dec 9th; 6 articles on Dec 8th; 4 articles on Dec 7th; 11 articles on Dec 6th; 8 articles on Dec 5th; and some of the articles from Dec 4th—less than one week ago—already don’t fit on the first “Discussion” page. (The exact numbers may differ depending on your Preferences settings.) The page is scrolling insanely fast. If I stopped reading LW for one week, I would have problem to catch up with all the new stuff; I would probably just skip some of that. That’s not good if we have too much stuff, but low average quality.
We don’t downvote enough. Let me explain—if someone makes a post that is not very good, but is not completely stupid or trolling also, it will almost certainly gain more upvotes that downvotes. Because it feels wrong to punish someone only for being uninteresting. But in terms of rewarding/punishing behavior, we probably should punish them. If we try to be too friendly, the site will become boring, precisely because most of the social talk is not about giving new information.
Perhaps it could help to use some timeless deciding. If you read an article, ask yourself a question: “If the next week here would be 10 new articles like this, would it make LW better or worse?” If the answer is worse, downvote it. Because although the same author will not write 10 more articles like this during the next week, other authors will.
TL;DR—I think the Eternal September is most visible on the article level, because it is not obvious what kind of content belongs here. “Discussion” is horribly misleading—we don’t want discussion-level articles. That’s what the comments and Open Threads are for.
One issue with the LW/CFAR approach is that the focus is on getting better/more efficient at pursuing your goals, but not on deciding whether you’re applying your newfound superpowers to the right goals. (There’s a bit of this with efficient altruism, but those giving opportunities are more about moving people up Maslow’s hierarchy of needs, not on figuring out what to want when you’re not at subsistence level).
Luke’s recent post suggest that almost no one here has the prereqs to tackle metaphysics or normative ethics, but that always has seemed like the obvious next topic for rationality-minded people. I was glad when Luke was writing his Desirism sequences back at CSA, but it never got to the point where I had a decent enough model of what normative claims desirism made to be able to evaluate it.
Basically, I think these topics would let us set our sights a little higher than “Help me optimize my computer use” but I think one major hurdle is that it’s hard to tackle these topics in individual posts, and people may feel intimidated about starting sequences.
The problem is that there is an unfortunate tendency here, going all the way up to EY to dismiss philosophy and metaphysics.
It depends—if the higher-voted comments are expending on the original post, then I’d say the post was successful because it evoked good-quality thought, assuming that the voters have good judgement. If the higher-voted comments are refuting the original post, then it was probably a bad post.
Do you have a theory as to why there aren’t enough good users, or why they are not writing good articles?
One possibility is that the kind of content printing-spoon likes is easy to get wrong, and therefore easy to get voted down for, and therefore the system is set up with the wrong incentives (for the kind of content printing-spoon likes). I’d guess that for most users, the possibility of getting voted down is much more salient than the possibility of getting voted up. Getting voted down represents a form of semi-public humiliation (it’s not like reddit, where if you post something lame it gets downvoted and consequentially becomes obscure).
You and Your Research
See this thread for more: http://lesswrong.com/lw/5pf/what_were_losing/
Overall, I suspect that LW could stand to rely less on downvoting in general as a means of influencing user behavior. It seems like meta threads of this type often go something like “there’s content X I hate, content Y I hate, and practically no content at all, really!” Well if you want more content, don’t disparage the people writing content! It may make sense to moderate voting behavior based on how much new stuff is being posted—if hardly any new stuff is being posted, be more willing to upvote. If there’s lots of stuff competing for attention, vote down lamer stuff so the good stuff gets the recognition it deserves.
I think we could stand to see high-karma LWers who rarely post in Main/Discussion post there more. Maybe make it impossible for anyone with over X karma to get voted below 0 in a Main or Discussion post. Or make a new subforum where high-karma users can post free of moderation. (I’ll admit, the oligarchical aspect of this appeals to me.)
Also, maybe be realistic about the fact that most people are not going to be willing to go to lukeprog/gwern lengths to dig up papers related to their posts, and figure out the best way to live with that.
Well, I tried to make a post once, got downvoted into oblivion, and decided not to put myself through that again. So yeah this happens for real, although perhaps in my case it is no big loss.
See my comment for one possibility.
I’m not sure… I think the topics I find most interesting are simply used up (except for a few open questions on TDT or whatever). Also the recent focus on applied rationality / advice / CFAR stuff… this is a subject which seems to invite high numbers of low quality posts. In particular posts containing advice are generally stuffed with obvious generalizations and lack arguments or evidence beyond a simple anecdote.
Also, maybe the regular presence of EY’s sequences provided a standard for quality and topic that ensured other people’s posts were decent (I don’t think many people read seq reruns, especially not old users who are more likely to have good ideas).