Jun 15, 2016
ohn_Maxwell_IV argued that revitalising Less Wrong is a lost purpose. I'm also very skeptical about Less Wrong 2.0 - but I wouldn't agree with it being a lost purpose. It is just that we are currently not on a track to anywhere. The #LW_code_renovation channel resulted in a couple of minor code changes, but there hasn't been any discussion for at least a month. All that this means, however, is that if we want a better less wrong that we have to do something other than what we have been doing so far. Here are some suggestions.
Systematic changes, not content production
The key problem currently is the lack of content, so the most immediate solution is to produce more content. However, not many people are an Elizier or a Scott. Think about what percentage of blog are actually successful - now throw on the extra limitation of having to be on topic on Less Wrong. Note that many of Scott's most popular posts would be too political to be posted on Less Wrong. Trying to get a group of people together to post content on Less Wrong wouldn't work. Let's say 10 people agreed to join such a group. 5 would end up doing nothing, 3 would do 2-3 posts and it'd fall on the last 2 to drive the site. The odds would be strongly against them. Most people can't consistently pump out high quality content.
The plan to get people to return to Less Wrong and post here won't work either unless combined with changes. Presumably, people have moved to their own blogs for a reason. Why would they come back to posting on Less Wrong, unless something was changed? We might be able to convince some people to make a few posts here, but we aren't going to return the community to its glory days without consistent content.
Why not try to change how the system is set up instead to encourage more content?
Decide on a direction
We now have a huge list of potential changes, but we don't have a direction. Some of those changes would help bring in more content and solve the key issue, while other changes wouldn't. The problem is that there is currently no consensus on what needs to be done. This makes it so much less likely that anything will actually get done, particularly given that it isn't clear whether a particular change would be approved or not if someone did actually do it. At the moment, what we have is people coming on to the site suggesting features and there is discussion, but there isn't anyone or any group in charge to say if you implement this that we would use it. So people will often never start these projects.
Before we can even tackle the problem of getting things done, we need to tackle the problem of what needs to be done. The current system of people simply making posts in discussion in broken - we never even get to the consensus stage, let alone implementation. I'm still thinking about the best way to resolve this, I think I'll post more about this in a future post. Regardless, practically *any* system, would be better than what we have now where there is *no* decision that is ever made.
Below I'll suggest what I think our direction should be:
Less Wrong is the website for global movement and has a high number of programmers, yet some societies in my university are more capable of getting things done than we are. Part of the reason is that university societies have positions - people decide to run for a position and this grants them status, but also creates responsibilities. At the moment, we have *no-one* working on adding features the website. We'd actually be better off if we held an election for the position of webmaster and *only* had that person working on the website. I'm not saying we should restrict a single person to being able to contribute code for our website, I'm just saying that *right now* implementing this stupid policy would actually improve things. I imagine that there would be at least *one* decent programmer for whom the status would be worth the work given that half the people here seem to be programmers.
If we want more content, then an easy way would be to have a links section, because posting a link is about 1% of the effort of trying to write a Less Wrong post. In order to avoid diluting discussion, these links would have to be posted in their own section. Given that this system is based upon Reddit, this should be super easy.
The other easy way to generate more content would be to change the rules about what content is on or off topic. This comes with risks - many people like the discussion section how it is. However, if a separate section was created, then people would be able to have these additional discussions without impacting how discussion works at the moment. Many people have argued for a tag system, but whether we simply create additional categories or use tags would be mostly irrelevant. If we have someone who is willing to build this system, then we can do it, if not, then we should just use another category. Given that there is already Main and Discussion I can't imagine that it would be that hard to add in another category of posts. There have been many, many suggestions of what categories we could have. If we just want to get something done, then the simplest thing is to add a single new category, Open, which has the same rules as the Open Threads that we are already running.
John_Maxwell_IV points out that too many posts are getting downvotes and critical comments. We could try to change the culture of Less Wrong, perhaps ask a high status individual like Scott or Elizier to request people to be less critical. And that might even work for even a week or a month, before people forget about it. Or we could just halve downvotes. While not completely trivial, this change would be about as simple as they come. We might want to only halve downvotes on articles, not comments, because we seem to get enough comments already, just not enough content. I don't think it'll lower the quality of content too much - quite often there are more people who would downvote a post, but they don't bother because the content is already below zero. I think this might be worth a go - I see a high potential upside, but not much in the way of downside.
If we could determine that a particular set of features would have a reasonable chance of improving LessWrong, then we could crowd-source putting a bounty on someone implementing these features. I suspect that there are many people who'd be happy to donate some money and if we chose simple, well defined features, then it actually wouldn't be that expensive.