I am strongly opposed to this measure, because I believe it both treats symptoms rather than causes and is negative expected value for users.
[EDIT]: There was an html error in the original version of this comment, and so several paragraphs disappeared.
Beware Trivial Inconveniences seems relevant. I, for one, neither have a Reddit account nor visit that site, and would like to keep things that way. Even were I to make an account, to remain on top of meta discussions I'm involved in, I would have to regularly visit that thread, rather than just seeing it in my LW inbox. A predictable consequence of it being annoying is that less conversations will happen, and it's not clear to me that's a step in the right direction.
I think it's also worth considering the boundaries between meta conversations and 'object level' conversations. Anybody want to join a math club? seems, strictly speaking, like a meta conversation- but to move it offsite would totally kill it.
These two sentences stuck out to me:
Moreover, moderators can't easily ignore those conversations when they encounter them, because they're usually attacks on the moderators themselves
bitterly meta in a thread that's not originally about a LW site meta issue
If you look at a thread like Who owns LessWrong?, it's a meta conversation that shows up in Recent Comments just like the Recent Unpleasantness with V_V, but it's in its own thread. Does that make it the sort of thing that should be in Discussion, or is it the sort of thing you would want moved off site?
This is a perfect recipe for a flamewar taking over Top Comments even when most people don't care that much.
Would you characterize the Recent Unpleasantness as a flamewar, or are you thinking more of something like this?
I'm also curious about whether or not LW has a long tail. Are most comments on LW things that most users don't care about? Every now and then, I realize that I'm not voting as much as I could, and am thus slacking in being informative to other users. But when I go through Recent Comments to try and upvote or downvote, I find I often only care enough about ~2 comments out of 10 to upvote or downvote them, which implies that the majority of comments on LW are ones I don't have even a mild opinion about.
[...] I find I often only care enough about ~2 comments out of 10 to upvote or downvote them, which implies that the majority of comments on LW are ones I don't have even a mild opinion about.
I'm confused. I thought the voting system ostensibly represents quality of reason, not concurrence of values or opinions.
I'm confused. I thought the voting system ostensibly represents quality of reason, not concurrence of values or opinions.
It does; quality of reason is one of the larger things I care about. Even so, most comments don't have noticeable quality, such that I would upvote or downvote them on those grounds.
OK. I agree this is a legitimate issue with the proposal, but it's the best idea I know of that can be implemented without site changes (and we all know how easily site changes happen).
Do you agree, though, that the sporadic meta flame wars are a problem? (If not, it's worth noting that other people do find it a significant ugh field- beware trivial inconveniences to their use of the site too.)
I apologize, there was a html error in my comment and so most of it was accidentally eaten; I recommend rereading it. In particular, I'm curious if you think the Recent Unpleasantness with V_V was a flamewar or not.
Do you agree, though, that the sporadic meta flame wars are a problem?
I disagree with the words "flame" and "a": I think that the meta discussions that generate the most heat are discussions about the use of moderator power, and I think those discussions are often gone about in a suboptimal manner. I think that there are significant cheap improvements to the way those discussions occur, and significant cheap improvements that make those discussions less frequent.
I don't think that meta discussions as a whole should be avoided, because there are many meta topics that are useful. If people get the sense that there is too much meta discussion going on, I suspect that's generally a disguised complaint that there's not enough object discussion going on, and it is better cured by subsidizing / generating object discussion than penalizing meta discussion.
Edited to add: I didn't elaborate on my disagreement with "a" enough. I think there are several related problems that meta discussions bring up, and I think that targeting those problems individually is superior to a blanket ban / penalization.
Yes, the recent unpleasantness was the reason I made this post. And I don't think there's much evidence to the effect that meta arguments happen more when there's less other content on the site; I think it flares up at pretty random intervals.
I don't find meta discussions annoying, and I would rather have them here, because any discussion about site policy seems important enough that everyone should have an opportunity to see and comment on it. (Indeed, I've sometimes posted meta discussions in Main rather than Discussion, because I was worried that there might be people who wouldn't see it if I only put it in Discussion.)
... I'm going to tell people to start a thread on the LW Uncensored Reddit Thread instead.
This is intended to mean STFU, right?
Firstly, there is no convenient way to find new comments in Reddit comment threads, so telling people to use an old already long reddit comment thread always reads pretty much like shut up.
Secondly, the point of meta discussions is to have mods or site admins see them, so if a certain meta discussion is something that the mods should be shielded from then STFU is the appropriate thing to say.
Non-hypothetical meta test case: I have a book under review at a publisher. When it comes out, I'd like to post some of it here for criticism. Hooray for free content or boo hiss for commercial posting? Because yes indeed I hope LW readers are inspired to buy my book after I give part of it to them for free.
What is the subject of the book? Is it to do with the art of human rationality?
We're happy in general for commercial posting if relevant to rationality (CFAR, donation drives, job adverts) and unhappy with free content unconnected to rationality.
Aside from opinions on the particular case, I think this is an example of a question you could post on an Open Thread. If for some reason people start getting bitter over it, then I'd politely ask everyone to go to Reddit.
Thoughts?
I'm indifferent.
Also, I'm not quite sure what the point of this is.
Our stated downvote policy is "downvote what you want less of."
If we don't want to see "meta" discussions on LW, it follows that we should downvote them.
If enough people follow this policy and share this judgment (net), meta discussions will quickly be downvoted to invisibility. If not, then they won't.
Since meta discussions aren't quickly being downvoted to invisibility, I infer that not many people share this judgment and/or that not many of the people who do share it follow existing stated policy. Either way, it would surprise me if stating a new policy changed much of anything, though I don't object to you trying the experiment.
That said, committing to following existing downvote policy (either instead of or in addition to) might do more good.
Along the same lines, encouraging downvoting... for example, by periodically praising in public users who have used a lot of downvotes, if there's some way of looking that up.. might be a good idea.
I also suspect that this whole concern with whether a discussion is "meta" or not is serving as an indirect measure of discussion quality. I distrust optimizing for indirect measures generally, as it tends to decouple the correlation more than it tends to optimize for what I actually care about.
That said, I admit that "tends to" != "always" and this might be one of the exceptions. Again, I don't object to you trying the experiment, though I expect the results to be indeterminate.
Doesn't solve the problem of flooding Recent Comments without a site update, which ain't happening.
There are at this point at least four or five serious site issues that need work. Suggestion: Someone should start a Kickstarter.
What issues are you seeing? The only one I've noticed is search returning subtly defective links for discussion articles.
All meetups spamming discussion.
User profile page creation spamming the "recent wiki edits" feed.
Sequence_reruns devouring the tag space in discussion.
Among other problems that have been mentioned before.
Edited to add: Also, there's this old list of issues and requests that could be run through and hopefully add a few of the accepted-but-never-implemented features.
DaFranker gave a pretty good summary of a lot of the issues. There are also minor interface issues (e.g. there's been prior discussion about how being able to preview comments would be nice).
Meta discussions can sometimes be very productive, like when they result in some beneficial change in site policy or an insight that makes subsequent discussions go a bit better (see this post for an example). I'm curious why orthonormal (or others) find meta discussions so annoying. Is it just that they have a tendency to be low quality and unproductive, or is it something inherent in being meta?
There are a lot of terrible discussions on this site, any time you can them lump together into a category that category becomes salient and acquires horns for people. "Meta" "Pua" "Gender issues" etc.
Why not a periodic (every 2 months) meta-thread? Similar to the media recommendations thread etcetera.
Also, while it's within your capabilities, please don't unilaterally engage in policy-regulated systematic downvoting.
thread on the (less moderated) LW Subreddit
Actually, the Official LW uncensored thread on reddit is this one. That's the place we should link people to, I would think.
And by "less moderated" you mean "moderated by reddit's usual moderators, whoever they are, not moderated by LW or MIRI."
I'd create a separate thread. Some people may not want to see basilisk discussion, but do want to talk meta.
You decide it's time for less meta, and more beta.
Counter-proposal: Everyone, let's just agree to downvote any discussion of moderation, moderators, downvoting, c?ns?rsh?p, and so on. Here's why:
Prolonged discussions about moderation attract the attention of three sorts of people:
However, any prolonging of the discussion just makes it worse. Shoot it. That is what the downvote button is for.
(Feel free to apply the above policy to this comment.)
I completely agree with your three points.
Now it seems to me the problem with meta comments is that they are more expensive than normal comments.
As an example, if I write "tomorrow a robot army from China will destroy the world", people can agree or disagree, upvote or downvote, but almost nobody thinks it is the topic which needs to be discussed, and there is almost zero risk that it will spawn half dozen meta articles tomorrow.
On the other hand, if I write "Eliezer is censoring this website and I am so going to write an article on RationalWiki about that", suddenly most people want to express their agreement or disagreement, then I can go more meta and say "everyone who downvoted me only proves that LW hive mind is insane", then a few confused contrarians will upvote my previous comments, to keep the controversy alive, then Eliezer joins, then a new moderation rule is proposed, half dozen meta articles are written, etc. Am I right?
Then it seems to me the best solution is to make writing meta comments more expensive, too. To make sure that people don't start a nuclear war over triviality. If it is not worth for anyone to spend a few minutes preparing a solid argument, then it probably is not so important.
I propose this rule: Meta discussions should be forbidden and censored in comments on threads with a different topic. (How would you feel if you spent an afternoon writing a nice article for LessWrong, only to see that the discussion goes off topic and finishes in a flamewar about something completely unrelated to your article.) However, meta discussions and other forms of criticism are OK in form of well-written articles. Within some reasonable limits, of course, to prevent dozen meta articles on the same topic, or by the same person.
If it is not worth making a well-written article about, then it probably is not so important.
Also I would recommend a policy that meta comments and articles should be downvoted unless you agree with the argument and the way it was presented. (In other words, don't be "useful idiots" and upvote trolls only to protect the diversity and fight censorship.) No, it does not mean that all constructive ideas would be supressed. Let me remind you that ideas like "make a separate subreddit for meetups" are typically upvoted heavily. Even the discussions about what is taboo and what is not, or the current minisequence connotationally accusing LessWrong of sexism, are relatively well accepted. The criticism of SIAI by Holder was upvoted and thoroughly discussed. So if something is not accepted well, let that be an evidence that there is something seriously wrong about that, and that it deserves to be downvoted and/or removed.
Every time you feed a troll, remember about the costs you have caused to everyone who has to deal with that. (I should more often remind myself about that, too.)
Meta discussions should be forbidden and censored in comments on threads with a different topic. (How would you feel if you spent an afternoon writing a nice article for LessWrong, only to see that the discussion goes off topic and finishes in a flamewar about something completely unrelated to your article.)
The problem with this rule is that on the same topic is not well defined. Arguably, the appearance of rituals to outsiders was on topic in a thread about rituals performed by some folks in this community.
If you think it's not, downvote. If you're not sure enough or think it is, don't. The community's opinion will become obvious.
Or were you suggesting a Great Definition Decree from the Benevolent Dictator is necessary? I don't see why a significant amount of people would follow that definition rather than their own thoughts.
I'm not asking for a decree - I am concerned that Villiam_Bur underestimated the difficulty of creating such guidance. Specifically, I'm doubtful there is any consensus on what is or is not on topic from a particular discussion post.
Would writing "I find LW rituals creepy" be off topic on the Schelling Day post? Or rude for some other reason? Or just fine? I suspect there is no consensus at all, so suggestions otherwise should be met with a bunch of skepticism.
Now it seems to me the problem with meta comments is that they are more expensive than normal comments. [...] Then it seems to me the best solution is to make writing meta comments more expensive, too.
Upvoted for this, which is an admirably plausible formulation.
As for the proposed rule and policy, I continue to think that increasing compliance with "downvote what you want less of" would subsume most of the benefits of the various new policies being proposed, and I continue to be skeptical of the benefits of stating new policies where compliance with existing stated policies is the issue.
Then the right meta thing we need to do is: remind people that their (not) downvoting (or even upvoting of things they consider wrong, but feel it is fair to give them visibility because of some abstract principle) has consequences. Most obviously, the consequence of getting more of what they don't downvote, or even upvote.
Next thing would be to emphasise that there are different forms of criticism, and that well-written criticism is typically well accepted (see Holden). Therefore we do not have to pay attention to trolling-style criticism.
Then, if someone sees a comment saying "okay, I know this is stupid and offensive, and was already said hundred times, but I upvoted it anyway because I think LessWrong needs dissenting voices", it can be responded by a link to that article.
Sure, I endorse all of this. (Well, at least if "the right thing we need to do" is replaced by "a useful thing we can do".)
You decide it's time for less meta, and more beta.
That doesn't make sense.
Counter-proposal: Everyone, let's just agree to downvote any discussion of moderation, moderators, downvoting, c?ns?rsh?p, and so on. Here's why:
That's a complementary proposal, not a counter-proposal. Unless you are advocating everyone getting reddit accounts and going there to systematically downvote.
Meta nerds (including rules lawyers), who should go play Nomic, Calvinball, Mao, or something.
Or, you know, they could go have meta discussions on reddit where it is none of your business unless you force it to be.
However, any prolonging of the discussion just makes it worse.
Getting such conversation off lesswrong has already achieved the practical objectives. Opposing outright exile of conversation in favour of the "counter-proposal" of keeping it present but everyone cooperating to actively disapprove of it via downvotes amounts to several steps in the wrong direction.
Your solution is optimised for moral indignation signalling, not practical consequences.
(Feel free to apply the above policy to this comment.)
Done.
Moreover, moderators can't easily ignore those conversations when they encounter them, because they're usually attacks on the moderators themselves;
I'm not sure how to parse this statement in such a way that it doesn't sound like a litany of Tarski violation.
s/Tarski/Gendlin, right?
I was confused for a bit, until I realized that's what you probably meant, so hopefully you meant it.
I'd imagine if attacks on me consisted of information only accessible to me, they'd be easier to ignore than attacks consisting of public attempts to reduce my credibility.
Officially banning all meta conversations, however, is a bad precedent, and I don't want LW to do that.
Why?
Otherwise, this seems a good idea. Seconded.
Meta-threads discussing the value of meta-threads...
I support this. Trolls test the willpower of the reader, a finite resource better spent elsewhere.
After a recent comment thread degenerated into an argument about trolling, moderation, and meta discussions, I came to the following conclusions:
Ideally, Less Wrong would implement a separate "META" area (so that people can read the regular area for all the object-level discussions, and then sally into the meta area only when they're ready). After talking to Luke (who also wants this), though, it seems clear that nobody is able to implement it very soon. So as a stopgap measure, I'm personally going to start doing the following, and I hope you join me:
Whenever a conversation starts getting bitterly meta in a thread that's not originally about a LW site meta issue, I'm going to tell people to start a thread on the LW Uncensored Reddit Thread instead. Then I'm going to downvote anyone who continues the meta war on the original thread.
I know it's annoying to send people somewhere that has a different login system, but it's as far as I can tell the best fix we currently have. Since some meta conversations are important, I'm not going to punish people for linking to meta thread discussions that they think are significant, and the relevant place for those links is usually the Open Thread. I don't want LessWrong to be a community devoted to arguing about the mechanics of LessWrong, so that's my suggestion.
Thoughts? (And yes, this thread is obviously open to meta discussion. I'm hopefully doing something constructive about the problem, instead of just complaining about it, though.)
EDIT: Changed the link to the uncensored thread more specifically, at Luke's request; originally I linked to the general LW subreddit, which is more heavily moderated.