I would like to explore your belief, Louie, that increasing the rate of growth of LW is desirable.
ADDED. Clarification: interventions to increase the rate at which new people join LW are good IMHO when the current rate is low enough not to overwhelm the "mechanisms for newbie assimilation". So I guess I want to ask Louie if he is measuring the rate of new arrivals and whether he plans to discontinue the adwords if that rate gets too high.
ADDED. Last time there was a meetup at Benton house, I told Kevin that I wish he had waited with his efforts to apply "search-engine optimization" techniques (to increase participation on LW) till LW had more time to assimilate the influx from Methods of Rationality. I hope I am not misrepresenting Kevin, but I got the distinct impression from his reply in that short conversation that he recognized no good reason for LW not to grow as fast as possible. (Of course, Louie is not Kevin.)
I certainly do not object to the goal of building more rationalists (as you put it) but I worry that too fast a rate of growth might drastically decrease the usefulness of LW as a place to learn. And I worry that it would cost a lot more to reverse that decline that it would to prevent it.
Do you perceive that LW's usefulness as a place to teach and to learn has decreased since its founding?
Do you believe that as LW grows it will become a less useful resource for learning for people who are already very strong rationalists? I would not mind if that proves to be the case BTW if the growth helps larger numbers of less-strong rationalists: I just want to know what the objectives are of those making significant efforts to improve LW.
I do not think that anyone will dispute that most online conversations (forums, group blogs, boards, etc) lose their usefulness after a few years in a way that is difficult enough to reverse that one might as well just start over with new software, new rules or new owners. Paul Graham maintains that the conversation he runs (Hacker News) is constantly on the verge of failing and that only constant interventions by Graham and his helpers prevents that from happening. In particular, Graham considers growing too fast to be a potent risk to Hacker News.
To what do you attribute the uncommonly high usefulness of the LW up to now relative to other place to learn online?
Which practices (deliberate or accidental) do you believe have maintained or will maintain the usefulness of LW as a place to learn and to teach?
What signs (if any) would cause you to come to believe that LW's capacity to assimilate new voters and new contributors has been pushed to its limits and that attempts to grow LW should be suspended until the capacity has recovered?
What fraction of LW comment do you personally "vote on" (make a decision as to whether to upvote, downvote or leave alone)? ADDED. Reason I ask this is that voting on many comments was much harder work than I thought it would be before I tried it. (The hard work came from the need to avoid or recover from an overly judgmental and critical mood that would have had a deleterious effect on my non-LW life.) Consequently, unless my experience with voting was atypical, anyone contemplating an action that would add significantly to the work done by LW's voters should probably have significant experience on voting on a signficant fraction of LW's comments unless he or she has some other way to avoid overextending that particular resource.
ADDED. The tone of this comment is probably too strident. It is important to keep in mind that since it is relatively inexpensive to start a new group blog, forum or other "place to teach and learn" on the internet, the "loss" of LW is not all that expensive. Nevertheless, I am curious as to why I seem to be deep in the minority in certain perceptions:
The way LW has been since its founding is that anyone who registers can start voting, and I am aware of no plan or proposal to change that.
Perception One that I have that does not seem to be shared: it is easy to design a system better than what I just described. E.g., require new registrants to accumulate 20 karma before they can vote.
Perception Two: there are some simple numbers that could be collected by software in a straightforward way that are not being collected such that if one of the numbers took a sufficiently steep downturn, that would be a strong sign that any efforts (like sending readers of Methods of Rationality here, asking people to promote LW on StumbleUpon or buying adwords for LW) to increase the rate of growth of LW should be curtailed for the time being. One of those numbers is the rate of voting by accounts older than 2 years (those accounts being more likely to be able to vote in such a way as to keep up the standards of LW) as a fraction of total votes (or as a fraction of the votes made by those accounts during the same time frame last year). Another of those numbers is the number of comments made or number of words posted by accounts older than 2 years. In other words, it seems like a basic precaution to me to make sure that the new voters and the new writers on LW are not driving away the old ones.
Again, I am not opposed to trading off average comment quality for increased educational reach if that is the direction those who do the work to maintain LW want to take, but there's a difference between consciously making that trade-off and losing LW while neglecting to use the loss to collect any data useful for improving the robustness of future online conversations.
Perception Three: Since a person who has figured out how to use one web site will have very little trouble figuring out how to use another one, the higher the quality of a conversation on the web which anyone can join, the more difficult it will be to prevent that quality from regressing to the mean quality of conversations on the web that anyone can join -- and a conversation has to be far above the mean quality for it to start to compete with a competently-chosen college or textbooks as a way for young aspiring rationalists to learn.
Perception Four: Even though LW is vastly better as a place to teach and to learn and for people to cooperation on projects to improve the world than the average place on the internet, it is possible to imagine places on the internet that are much better than LW. Effort put into understanding how to preserve or maintain the quality of the conversation on LW can probably be applied to making it even better. I.e., the optimistic outcome of work on "online conversation quality" has very high utility. I.e., this is not just about avoiding the need to find another place to teach and to learn because LW has become useless for that purpose.
Specifically, there are probably tricks we can learn to improve online group rationality without needing to improve the rationality of the group's members. By "improving group rationality" I am imagining mostly (1) making teaching and learning online less irritating, less addictive and more time-efficient and (2) making world-improving projects more effective without increasing the individual effectiveness or rationality of the teachers, learners or project members. What people like Paul Graham, Clay Shirky and Teresa Nielsen Hayden know about maintaining online conversations is quite relevant here IMHO.
Thanks for thinking about these things. Very useful comments and questions.
Having a karma requirement for moderating comments seems like a really, really good idea. I'm gonna sleep on that and see if I can think of any downsides. At the very least, we should probably have a requirement that you must have positive karma (which I don't think is currently required?). I wonder how much moderation is done by people with less than a certain level of karma? It's possible significant amounts of moderation is being done by people with zero or negative karma which is probably not good and it might improve things if only we change that. I'll have to look at things before I implement this change to see where the best cutoff might be. You guestimate of 20 sounds like about the right ballpark but the data will be more informative.
I was actually a little sad, thinking about how maybe this campaign could dilute LW... until I remembered something subtle about LW: 96% of LW participants are lurkers -- and amazingly, this is not a made-up statistic! Even though there are 5000 regular, daily readers, we only have 200 participants who actively comment in any given month. I've only attracted 2-3 new participants to LW with my current AdWords campaign, so you you'll have to wait until I do 10x as much promotion before you'd expect to see a new user register and make a comment. The growth rate from this campaign is under a person a day right now and probably won't attract more than 5 new commenters / yr unless I hit on a really good landing page or start pouring obscene amounts of money into it. My goal is to get to 1 new participant / day but I'm still only about 20% there.
A totally different way to look at things is that there are roughly 500,000 English speakers worldwide who are IQ 145+. So our community is 100x smaller than the theoretical limit of highly qualified new readers who would be able to join without lowering the intellectual quality of our community. This is just multiplying English speakers by 0.15%... I've been trying to think of more rigorous ways to estimate theoretical upper-bounds or lower-bounds on the potential community size of LW.
I'm not saying we could support 100x as many people as a community if we grew too fast, but there are lots of communities with 100x as many members. Obviously the participants don't think they've imploded. For instance, we have less posts than all 2246 message boards tracked by Big Boards and less users than almost all of them as well. We don't even have 15% of the content of their lowest ranked forum #2246... which happens to be about Dance Dance Revolution. So unfortunately, rationality is in no eminent danger of growing out of control unless we re-brand the site to be about "refining the art of rationality, motor-sports, video games, and hot chicks".
Generic method is to search google for "site:lesswrong" to see what already has high page rank and is therefore considered useful by people linking to the site from outside. I found that a lot of tags come up that way so I tried this search: "site:lesswrong.com -tagged".
Going down a page or two and then looking around I found these that might be cool:
I like the idea of an ad saying "A human's guide to words", thought I don't know what article it should link to.
I just realised something: could SIAI/FHI demand you take down those ads, if they didn't like them? If not, what's to stop malicious people from advertising another's website in a way that makes it look bad, or that has unwanted consequences (say, "4Chan - for all your sparkly vampire needs")?
I consider Three Worlds Collide a very strong candidate for the best piece of writing on LW (and, for that matter, miles ahead of HP:MoR, which I still enjoy immensely).
As a bonus, it also seems one of the easiest to market through a two-line AdSense. The most effective option might be going for the "WTF?!" angle; say, Aliens eat babies and want to sex us. A tale about rationality. I think curiosity would not be an uncommon reaction in the face of that.
People who go specifically looking for information on how to think ought to be directed here, so perhaps some keywords related to thought in general: "thinking", "confusion". And some likely terms people would make up for thinking-about-thinking: "metathinking", "metathought", "wrongness".
Surprisingly, there's not much advertising if you search for "politics". I feel like that ought to redirect to Politics is the Mind Killer.
All of the advertisements, and the top result for "philosophy" are, strangely, about cosmetics. I don't know what's up with that, but people searching for that would be better off here.
I'm buying ads for Policits is the Mind Killer already. That's one of the other few I'm experimenting with. I don't think people search for "politics" in general. Again, people search for very specific things and already have their minds made up.. so it's more likely that someone would search for "proof that Obama is a Muslim" or "why do liberals hate America?" than something as neutral and high-level as "politics". So coming up with decent keywords for this one has been harder than expected.
Call me crazy, but I actually think we'll do better among the people searching for "politics" -- fewer in number though they may be -- than people searching for "proof that Obama is a Muslim" and the like.
Funny thing about that... Philosophy is actually a major cosmetics brand these days. Who do you think owns philosophy.com? I've actually been to their stores and it always occurs to me that most of the store's patrons probably don't know the other meaning of the word.
People who go specifically looking for information on how to think ought to be directed here, so perhaps some keywords related to thought in general: "thinking", "confusion".
I second this, but I'm not sure how many people will think to look for information on how to think. We might also look for people who are heading in this direction (like people interested in philosophy of science or AI).
Hey Bongo! I like that page but it seems like a pretty hard landing page to dump someone on given how it immediately refers to other resources it assumes you've just read. I can see how it would work to send to smart friends, but I worry that it would not work for people who were casually clicking an ad.
The first paragraph seems like a hard landing, but the rest of the article seems not to require other background knowledge.
The primary causal reason I'm here is because Eliezer's prose is entertaining and engaging. I'd link to the most entertaining posts, the ones that are fun to read, like Disputing Definitions or maybe Stranger than History. I'd also avoid posts that look like they're advocating a particular viewpoint, since humans tend strongly to dismiss such things as "oh, he's trying to persuade me of something".
I agree that posts like these would be the best sort to promote for exactly the reasons you like them. I guess the problem with those two is that they don't answer a question someone might have or solve a specific problem. Hmm...
Can you think of keywords or ads that might work for either of those pages?
I really like That Alien Message and it has done well (more than once) on Hacker News and reddit.
http://lesswrong.com/lw/qk/that_alien_message/
Not sure what keywords to bid for -- probably science fiction related keywords.
For many terms that are relevant (Bayesanism, rationality), LW is already on the top set of Google hits. So I don't know how much return this sort of thing would get.
People don't search for terms like Bayesian or rationality. What they search for is answers to specific questions -- which we don't actually ever come up in searches for.
I've looked at the analytics and I know. Most people coming to Less Wrong through searches are clearly regular users trying to find specific articles or just using the search bar as a URL bar (so they don't have to remember if it's less wrong dot com or dot org or whatever).
If karma points count for anything, Generalizing From One Example should be in the running as an attractive article. I'm not sure what good key words for it would be-- maybe "psychology".
I recommend doing this as a top-level post.
Yudkowsky.net is actually well optimized for that, although not #1 these days. From what I heard, his "TMOL" page, which he has since marked "obsolete", used to be the #1 result for "What is the meaning of life?" on AskJeeves several years ago.
Apparently many people who ended up on SL4 or later involved with SIAI first wandered into our community through this search.
[a web page by Eliezer on the meaning of life] used to be the #1 result for "What is the meaning of life?" on AskJeeves.
It was (if Eliezer of the year 2000 can be believed) almost exactly 10 years ago. (I remember because I remember when I first saw the page.)
And yes, participants (prob on SL4) reported that their participation began when they googled for meaning of life and hit that page.
Generic method is to search google for "site:lesswrong" to see what already has high page rank and is therefore considered useful by people linking to the site from outside. I found that a lot of tags come up that way so I tried this search: "site:lesswrong.com -tagged".
Going down a page or two and then looking around I found these that might be cool:
Thanks for thinking about these things. Very useful comments and questions.
Having a karma requirement for moderating comments seems like a really, really good idea. I'm gonna sleep on that and see if I can think of any downsides. At the very least, we should probably have a requirement that you must have positive karma (which I don't think is currently required?). I wonder how much moderation is done by people with less than a certain level of karma? It's possible significant amounts of moderation is being done by people with zero or negative karma which is probably not good and it might improve things if only we change that. I'll have to look at things before I implement this change to see where the best cutoff might be. You guestimate of 20 sounds like about the right ballpark but the data will be more informative.
I was actually a little sad, thinking about how maybe this campaign could dilute LW... until I remembered something subtle about LW: 96% of LW participants are lurkers -- and amazingly, this is not a made-up statistic! Even though there are 5000 regular, daily readers, we only have 200 participants who actively comment in any given month. I've only attracted 2-3 new participants to LW with my current AdWords campaign, so you you'll have to wait until I do 10x as much promotion before you'd expect to see a new user register and make a comment. The growth rate from this campaign is under a person a day right now and probably won't attract more than 5 new commenters / yr unless I hit on a really good landing page or start pouring obscene amounts of money into it. My goal is to get to 1 new participant / day but I'm still only about 20% there.
A totally different way to look at things is that there are roughly 500,000 English speakers worldwide who are IQ 145+. So our community is 100x smaller than the theoretical limit of highly qualified new readers who would be able to join without lowering the intellectual quality of our community. This is just multiplying English speakers by 0.15%... I've been trying to think of more rigorous ways to estimate theoretical upper-bounds or lower-bounds on the potential community size of LW.
I'm not saying we could support 100x as many people as a community if we grew too fast, but there are lots of communities with 100x as many members. Obviously the participants don't think they've imploded. For instance, we have less posts than all 2246 message boards tracked by Big Boards and less users than almost all of them as well. We don't even have 15% of the content of their lowest ranked forum #2246... which happens to be about Dance Dance Revolution. So unfortunately, rationality is in no eminent danger of growing out of control unless we re-brand the site to be about "refining the art of rationality, motor-sports, video games, and hot chicks".
So I'm trying to build more rationalists. To do this, I've invested a few hundred dollars of my own money to promote Less Wrong by buying low-cost AdWords on Google for different LW pages. I want to reach smart people with a really good article from Less Wrong that answers their question and draws them into our community so that the site's content can help improve their rationality. Based on buying AdWords before, I'd estimate that only 0.5-1% of people who click through to Less Wrong will actually get involved after reading an article, but since clicks only cost ~$0.04, that means it only costs me ~$6 to build a new rationalist and drastically improve someone's life. Seems like an excellent return on investment.
But to get a strong 1% conversion rate and really make an impact, I need to identify REALLY EXCELLENT Less Wrong content. Right now I'm experimenting by buying a lot of keywords related to quantum mechanics and sending people to http://lesswrong.com/lw/r8/and_the_winner_is_manyworlds/
My hope is that this page is useful and memorable enough that some small % of readers stick around and click through to other pages. My guess is that this isn't the ideal page to do this with but it's aiming in the right direction.
What page would you would want a new Less Wrong reader to find first? What answers a specific question they might have in such an impressive way that they would want to learn more about our community (perhaps many different pages for many different questions)?? Which articles are most memorable? Just looking at "Top" didn't yield any obvious choices... I felt like most of those articles were too META-META-META ... you'd need too much back knowledge for many of them. An ideally article would be more or less "stand-alone" so that any relatively intelligent person who doesn't have the whole LW corpus in their head already could just jump in and understand it immediately... and then branch out and explore LW from there.
So what do you think? Give me links to any landing pages you think would be worth promoting this way. You can write rough mini-ads as suggestions too if you'd like to be even more helpful. I'm looking forward to hearing your suggestions!