In Defence of Meta
The current mods of the Less Wrong community seem to believe that meta (I’m using meta in a broad sense to mean meta + community) discussion and community discussion are both things that need to be heavily discouraged. In part, we can see this inspired by Elizier’s post Less Meta and a general tendency for the regular inside internet communities to gradually end up spend more and more of their time in meta, instead of in the site itself. It’s generally considered rude to unilaterally end a discussion, so it’s not easy for someone to say, “Everyone, everyone, this issue really isn’t that important, let’s just move onto things that are more productive”. This is definitely a dynamic to be aware of.
However, Less Meta came at the end of entire sequence of posts about how he thought that the community should develop. Elizier wasn’t denying the value of meta discussion and, as I am arguing, neither should we. Instead, I think we should be asking two questions:
Is our meta discussion important?
Is our meta discussion productive?
1 asks whether resolving the question would significantly benefit the community. 2 asks whether further discussion is likely to push the community forward on that issue. If the community seems unable to make progress on a particular topic, it may be best wait in the hope that the community changes and resolution becomes possible in the future. If we let this guide us, then I believe we can find the middle path that ensures necessary discussion occurs without it becoming a time sink.
Three reasons why meta discussion is important at this stage:
Meaningful meta discussion is now possible again for the first time in a number of years. I seem to have been getting the impression that for a while, online meta discussion was sub-ordinate to the in person discussions happening in the bay area. Many key people within the community stopped following LW. Discussions on LW had no impact because the key players weren’t on LW. Now that LW is actually being used again, discussions which were unproductive before, may be productive now. In fact, there’s likely a bit of a backlog that has developed of important topics we need to discuss as a community.
A lot has happened in the community since Elizier wrote the Craft and the Community. It seems like we might need a new vision for our community and understanding of who we want to be and how we will get there. At the very least, at the very least, it seems that we need some kind of retrospective discussion to think about how we might update our framework given all of our new information.
We saw that the community was in a steep decline until recently, despite the fact that many rationalists wanted the community to thrive. This can only be seen as a failure of rationality. How can this inspire us to improve? I’m not just talking about patching the particular failure modes as disussed in the LW 2.0 Strategic Overview like making Vaniver BDFL, that’s making this exercise too easy on ourselves. If this was the solution, why didn’t we try this sooner? Was it that we needed someone with high political capital to post in order to make things actually happen? We were lucky enough to survive this time, but the decline still inflicted significant damage on the rationality community. We may not be lucky enough to survive next time. Indeed, we can see that many other communities like the technocracy movement and the humanist movement have lost much of their influence.
I don’t disagree that we should try to limit meta discussion in the long term given decreasing marginal utility, but in the short term I believe there are some really important discussion to be had.
Update: Anyway, I should clarify that I’m not suggesting making meta any more visible. I’m just trying to encourage people to consider the most important meta topics and start conversations on them. Not all at once though, as we don’t want to flood the site. I suspect that this kind of thinking is neglected; people tend to think, “What is frustrating me?”, not, “What is the most important discussion we should consider having?”.
Meta Meta Meta Comment: Please keep the conversation in the comments about meta discussion in general and try to avoid having the specific object level meta discussions I mentioned below. These discussions are important, but this post is not the best location for this discussion to happen.
My own thoughts (not necessarily reflective of other site admins)
tldr:
meta is okay, just not front-and-center
yes it was bad that for awhile this meant it was hard to have meta conversation at all, but hopefully the frontpage redesign makes that easier, and we can further tweak that if need be. (Community discussion on personal blogposts in particular seems fine)
most important is for meta not to be what newcomers run into, since it sets a bad signal/precedent. (in particular, the rule we were striving for Frontpage posts was not “no discussion of community” but “no discussion of community that is only useful if you’re already in the community.” I think there was some poor communication about that concept that I contributed to.)
There’s certainly a lot of good meta conversation worth having. But, after several years of participating and observing community dynamics (both in LW and in Sunday Assembly), I have a strong prior that any given “explicitly meta” conversation is a weird attractor for discussion that doesn’t go anywhere. (i.e. often a benign demon thread)
Relatedly, I think the most important ways to make progress on community problems is to just do stuff, and talk about the doing of the stuff. If you keep your eye on that, then meta conversation will come up when it’s necessary.
That all said, the priority with LW right now is not to ban all meta conversation—it’s to make it so that Meta conversation isn’t the thing that newcomers are immediately confronted with when they arrive. If I arrive at a community where much of the talk is meta-discussion, my sense is “this is a community that doesn’t have anything to offer except meandering discussion about itself.” (And there are some people who’ve given feedback that this is why they don’t participate in LW anymore)
I do think it was a problem that, for several months, the Meta and personal blog pages felt like a backwater with no discoverability. My hope is that the redesign of the front page makes it easier to participate in them. (They’re still slightly hidden away from newcomers, but only a couple clicks away, and if you choose to leave your frontpage on “Community” settings I think it should be easy to follow most discussion)
In the spirit of Phil’s suggestion on the demon thread post, I’ll probably be limiting my engagement here to 2 additional responses, since I have a strong prior that meta-meta community posts in particular don’t often go anywhere useful. (But, if you have more specific community discussions to discuss, I’d be up for discussing that in future posts)
“That all said, the priority with LW right now is not to ban all meta conversation—it’s to make it so that Meta conversation isn’t the thing that newcomers are immediately confronted with when they arrive”—that’s an important clarification, although it does seem like there is a subset of meta which newcomers should encounter relatively quickly and that is the core principles of the community and what they are trying to achieve. Of course, this is usually established more through reading write-ups, then through discussion.
“And there are some people who’ve given feedback that this is why they don’t participate in LW anymore”—in the past, there were serious problems that needed to be addressed. So someone would post about it and then nothing would happen and then later someone else would post on the same thing and nothing would happen again. So I think that part of meta being so aversive was that a) it didn’t achieve anything in the past b) we kept repeating the same conversations.
Anyway, I should clarify that I’m not suggesting making meta any more visible (though I do think major site changes should be visible on the front page). I’m just trying to encourage people to consider the most important meta topics to discuss and start conversations on them, but not all at once as we don’t want to flood the site. I suspect that this kind of thinking is neglected; people tend to think, “What is frustrating me?”, not, “What is the most important discussion we could be having?”.
Yeah, previously a large part of meta was essentially “LW sucks”, which is a bad kind of meta. There are also other kinds of bad meta.
There are also good kinds of meta. Arguably, rationality itself may belong to this category. But I meant the kind that inspires people to do stuff.
I guess I tried to say that we need more good meta and less bad meta, but it’s a bit difficult to explain what exactly makes the difference. (People sometimes come up with rules like “it must include a proposal for how to do things, not just complaints”, but it is still possible to write a bad meta that formally fulfills the requirement.)
Why do you believe this? Or by “community” do you mean “the LessWrong website”?
From my vantage point, it looks like the overall, in-person+online community has been growing slowly ever since I joined it ~6 years ago.
I meant the community of people on LW was in decline, I wasn’t talking about the broader rationalist community.
I suppose one advantage of the decline of LW is that it led to further community formation elsewhere. But I can imagine alternative timelines in which LW never recovered. We got lucky.
I see. In that case
seems very non-obvious to me. Though of course the decline of LW1 was very bad for people not near any in-person community or involved with any of the LW diaspora online, I am not sure that it had a bad effect on the community as a whole.
But then, I’m not sure how to define “bad effect on the community as a whole,” either, short of the entire thing dissolving.
“Was very bad for people not near any in-person community or involved with any of the LW diaspora online”—I feel that there were sufficient people with an interest in improving LW that it was a failure that we didn’t find a way to achieve this/that it took so long. I don’t dispute that there are many people who would have had sufficient in person community that trying to fix LW might not have been relevant to their goals.
It seems to me to have been a mix of nobody feeling responsible and tragedy of the commons.
I will note that the-thing-that-worked was approximately 2 people making this their full-time job for a year. This is a level of commitment that I think goes beyond “if we’d just tried a bit harder and/or if there had been a clear source of responsibility, we could have fixed this.”
(I think a less-ambitious version of the revitalize plan was possible, but still would have been something like 8-person-months-worth-of-work, and by the time you’re willing to do that, you might as well get more ambitious)
Yeah, I don’t think this attempt “succeeded” because of nobody feeling responsible, and more that everyone who did feel responsible was extremely busy, and nobody else who had the skills and resources was available.
Resources and people are very scarce and generally taking the initative on projects like this means giving up a large amount of job security and salary and related things, and I nobody else comes to mind who would have been happy to run the LessWrong project who could have done it much earlier (the primary failure of rationality I see is me working about three quarters of a year too long at CEA and wish I instead had started working on Arbital and LessWrong three quarters of a year earlier). And there is also no simple community institution that I know of that we could have built that would have significantly reduced the cost to whoever wanted to make this project happen, without detracting significantly from other comparably impactful projects (i.e. CFAR or MIRI).
Also, most people greatly underestimated how much work would it take.
There were at least five situations when someone said “okay, let me look at this, we only need to make a few small changes in the Reddit codebase, how difficult could that be?” and a few weeks later they admitted that they are running out of time and energy and they barely scratched the surface. And a few weeks later a new person came and said “okay, now let me look at this, because changing a few little things in the Reddit codebase cannot be that difficult...”
In other words, it was difficult to coordinate on doing a large project, because it was difficult to even believe that a large project was actually necessary.
(It went against common knowledge: “Don’t reinvent the wheel.” Reddit is a large modern debate site; how likely it is that their software is difficult to maintain, and cannot resist a single dedicated attacker?)
Less Wrong, as a community, has an unusually high percentage of programmers. My intuition is that sufficient status incentives could have solved this problem and/or a kickstarter could have funded development work. We could have offered a student an internship over the summer.
I think it should have been solvable in 2 or 3 months worth of development. The difficulty was that you simply needed to pick the right features to add. Challenging, but not impossible.
Not just alternative timelines, but also all the people living in places with not enough rationalists to organize a meetup. (Another lucky thing is that we also have Slate Star Codex. I wonder how in a world without LW and SSC I would even explain to other people what I imagine by saying “a rationalist community”.)
EDIT: Oh, I see you already made this point. How do I delete my own comment?