I’m all for trying to make the LW front page more engaging, but I am skeptical about an assumption pervading this post: that growth should be a major goal, and that faster growth is better aside from major pathologies like spammers and trolls.
For a website that exists mostly to make money (by selling things, advertising a product, gaining visibility for a person or company, etc.) this is a good assumption: all else being equal, you want more page views, lower bounce rates, higher conversion rate, etc. But LW doesn’t exist mostly to make money; it exists (in high-minded principle) to refine the art of human rationality and (more prosaically) to provide interest and entertainment for its participants, and for these purposes some visitors are much, much more valuable than others and many have negative value. (This is true even taking into account the prospective value of LW to those visitors.)
By definition, LW wants (or should want) to “grow optimally”—but that may mean “grow very rapidly” or “grow slowly” or even “not grow at all, at present”.
I concur. I have implied this in my previous comments, but I will going come out and explicitly say it here. Caeteris paribus I want LW’s growth rate to be higher than it currently is, but I am more concerned with keeping out the intellectual riff-raff than encouraging growth. Furthermore, I actively want to avoid the exponential growth scenario, unless we can create an effective user orientation or another way of reducing the time it takes for new members to acculturate.
I agree that there’s a lot of annoyance for both the new user and old users when new people join. A new user orientation is needed. That’s why I wrote an outline for a New User Orientation full of suggestions that I personally would like to see included in the new user orientation. Please go and critique it—older members are surely going to have suggestions I wouldn’t think of. I have nothing against writing one, but I can’t write a good one with no input, I’m too new.
Please distinguish “keeping out the intellectual riff-raff” from elitism. Are you talking about people who really aren’t serious about rational though? Maybe you just mean you don’t want more trolls? Or do you mean you want to outright make an IQ requirement?
I have an idea for scaring off those who are not serious about rational thought that goes like this:
The culture here is very, very honest, very confrontational when it comes to errors in reasoning. That’s one of my top five reasons for joining. But it feels a bit tentative, a bit ambiguous. People also react with hurt feelings. I think, when it comes to that, we have to choose. I know what I choose—If the truth is brutal, hurt my feelings, I want to know. I’ll be responsible for cleaning up whatever mess it makes of my emotions. I think that’s the only way rationalists can go. I would like to see a description that demands honesty—not just mentions “yeah people are more honest here” but DEMANDS honesty. I’ll show you what I mean:
I am not an employee of Amazon but a friend showed me their values page and I thought it was inspiring: It states that their employees (referred to as “leaders”) are ”...obligated to respectfully challenge decisions when they disagree, even when doing so is uncomfortable...” Amazon Leadership Principles
I want to see something just as bold, just as tough as part of the joining agreement. I’m not talking about hiding it in some website policy or rules page everyone ignores. I want to see it right by the join button:
“By pressing this join button, I agree that I am here to improve myself. I understand that my flawed reasoning will be pointed out. My feelings about that will be my own responsibility. I agree also that I will point out flawed reasoning when I see it, no matter whose it is.”
(I also wrote about how LessWrong seems to balance brutal honesty and manners and my favorite technique for that)
What are the reasons for not wanting to grow? You didn’t state any reasons I haven’t already stated.
For the record, I am aware that it’s not a good idea to assume that growth is 100% good. That’s why I put that in the OP.
I feel that LessWrong could be really important—and maybe it already is—that’s why I want to see it grow (that much is obvious.) I started writing about why, but I want to hear your reasons for not wanting it to grow.
You’re going against the grain—not a bad thing but it means you’re going to have to really lay out your reasons if you want to change the way the wind is blowing.
(I think you may have a wrong idea of what “pontificate” means. Either that or you’re being gratuitously rude, which I’m going to assume you aren’t.)
The following two propositions are different. (1) Growth is not always 100% good. (2) Growth is not always good. #1 is what you stated. #2 is what I’m saying. #2 goes much further than #1 does. The obvious inference from #1 is “make sure you grow, but take some measures to mitigate the possible downsides”; the obvious inference from #2 is “consider carefully whether growth is what you want”.
I don’t know whether I want LW to grow, or how fast, or when (which is why I said that optimal growth for LW could be any of several things, including “grow very rapidly”) so I can’t really give you “[my] reasons for not wanting it to grow”. What I can do is to give some possible reasons why growth—especially rapid growth, especially especially rapid growth by the mechanism you’re implicitly proposing—might be the Wrong Thing.
The main reason is this: It may not be possible to grow rapidly by the sort of means you describe without changing the demographics of LW in a way that would lower its quality. I don’t just mean “LW might start to attract trolls and spammers and idiots”. I mean that maybe only a small fraction of new visitors to LW are people whose presence would enhance LW, and that if that’s so then anything we do to encourage a lot more new visitors to stay will make LW worse.
There’ve been a couple of surveys of LW participants. They suggest that the userbase of LW is very unusual. So, for that matter, does a casual look at the discussions here. LW participants are unusually interested in rationality (of course, since that’s the central topic of the site), something like three standard deviations above average in measured intelligence (according to self-reports, so take that with a pinch of salt, but I don’t find it hard to believe that something of the kind is true), and willing to read and in many cases write long, well researched, sometimes highly technical material. This combination of characteristics is rare. It’s surely a large part of what gives LW the distinctive character it has. I am not optimistic about the prospects of keeping it if LW grows rapidly by keeping a much larger fraction of its new visitors.
(I am aware that the foregoing paragraph sounds as if I’m saying “look at us, we’re superior to everyone else”. That is not, even slightly, the point. LW regulars, as a population, are unusually good at some things, unusually bad at others, and very average at others, and treating one group of people as “superior” to others is a road to ruin for all sorts of reasons. All I’m saying is that a community with the particular strengths LW has is a very unusual thing, and I think a lot of LW’s distinctive merits derive from those strengths.)
Now, I know that you said things like “We’d better choose what audience to target”. But I think you underestimate what an … odd … audience it is that LW might need to target, and the general shape of your proposals—make it so people are immediately grabbed without having to read all that text; base the design on general “web marketing” principles—seems like it’s implicitly aimed at exactly the people who don’t have the peculiar characteristics that make LW what it currently is.
I think this lies behind your perception that people here “seem to be totally unaware of the field of web marketing”. Probably many are, but please consider the possibility that in some cases the issue isn’t others’ obtuseness but that they simply aren’t interested in the “web marketing” goals you have in mind, and for a possibly-good reason.
A large increase in LW’s audience could have big benefits; I’m not denying that. They might outweigh the likely loss in quality of discussion that (I think) would go along with it. Or maybe there’s a way to increase the audience a lot without a big growth in the population of active participants. Or maybe “basic web marketing principles” can in fact be applied in a way that pulls in a huge new population of people with the extremely unusual characteristics LW’s current regulars have. So I’m not (to repeat myself) saying that LW should not aim to grow a lot. I don’t know whether it should. It just makes me really uneasy to see growth being treated as axiomatically good and important (which, repeating myself again, is not the same as saying that it has no downsides, and which I know you weren’t saying).
Oops sorry. I removed pontificate. Thank you for not assuming I was being rude.
“I mean that maybe only a small fraction of new visitors to LW are people whose presence would enhance LW, and that if that’s so then anything we do to encourage a lot more new visitors to stay will make LW worse.”
Good problem. However, that can happen whether it grows quickly or slowly.
“I am not optimistic about the prospects of keeping it if LW grows rapidly by keeping a much larger fraction of its new visitors.”
Yes! I did warn that this could happen, and that it could happen whether we like it or not.
I like that things can be voted up and down, but I think it would be a heck of a lot more effective if specific feedback was provided. As a new user, I will adjust a lot faster if I know WHAT I am doing wrong/right not just THAT I am doing something wrong/right. Also, I wouldn’t write off the votes to “bad attitude” and “trolls” when I don’t understand them. That’s what I’m doing now, lol, and I don’t like that, but I don’t know what else to think of them, lol.
I think if the votes required a reason, which could be as quick as the word “Trolling” or the name of a logical fallacy, that would make a huge difference. That would ALSO force us to get conscious of our reasons for voting, which would provoke conscious review of the reason, which would probably result in better decisions.
And to ensure the culture doesn’t change too quickly, what if it took, say, three months and X number of posts before you’re allowed to vote? That way, the oldest users get to influence the culture, and users who just aren’t compatible with LessWrong will wander away before they’ve exerted an influence.
This is basically the same as how you have to be a resident of a country before you can vote.
From a developer’s standpoint, that would be a very easy change, but more importantly, it would protect the site no matter whether it grows faster or at the current pace.
“please consider the possibility that in some cases the issue isn’t others’ obtuseness but that they simply aren’t interested in the “web marketing” goals you have in mind, and for a possibly-good reason.”
Oh, I did! I did consider that. (: That was the entire reason I created this thread. I figured you would all let me know what you wanted and if growth wasn’t it, you’d shoot me down. (:
“They might outweigh the likely loss in quality of discussion that (I think) would go along with it. ”
Has anyone thought much about what would improve the quality of discussion? Because that could be improved at the same time. Or before making marketing changes—to be sure that the safeguard is in place first.
“Or maybe “basic web marketing principles” can in fact be applied in a way that pulls in a huge new population of people with the extremely unusual characteristics LW’s current regulars have. ”
Yes! That’s what I was hoping for. (: That’s why I asked who your target demographic is. I might think I know, but you guys are the ones that have been talking to them for years.
However, that can happen whether it grows quickly or slowly.
Yes. It’s an issue regardless. But if pursuing a strategy of rapid growth guarantees an influx of negative-net-value-to-LW new users, it’s probably a bad thing. (We will all die eventually, but I still prefer there to be fewer mass murders.)
I wonder whether you think this idea would work
I don’t have a good enough mental model of the typical first-time visitor to LW to have a strong opinion. What I do think likely is that either it wouldn’t work or it would work at the cost of getting rid of that rapid growth you were hoping for. Because I think it’s likely that rapid growth by increasing visitor retention implies the sort of change in LW’s demographics that I described.
I think it would be a heck of a lot more effective if specific feedback was provided
I have proposed something similar myself. The UI would be tricky to get right. It might well make LW a better place (or it might not; these things have a way of producing unexpected consequences). But I don’t see how it would do much to solve the problem I described, if (as I suspect but don’t claim to know) it really is a problem.
what if it took, say, three months and X number of posts before you’re allowed to vote?
Might work. Has anyone reading this got experience of such systems? (The obvious concern is that it would drive away “good” users as effectively as it would drive away “bad” ones, so that it would slow growth without actually making the overall pattern of growth any more favourable.)
I figured you would all let me know what you wanted and if growth wasn’t it, you’d shoot me down.
According to my quick count, 28 different people have commented here [EDITED to add: not counting “metatroll”], of whom 7 seemed definitely in favour of growth (perhaps having thought it through, perhaps just because it’s a sort of default goal), one seemed definitely against growth, and 6 seemed definitely skeptical (i.e., saying “growth might not be what we want”, basically my position). The other 14 expressed no opinion on the matter that I could detect. Make of that what you will.
Has anyone thought much about what would improve the quality of discussion?
I think a lot of people have, but so far as I know no one (here or elsewhere) has a silver bullet that ensures that a community of unusual people will retain its distinctively valuable characteristics as it grows. (Or for that matter as it doesn’t.)
what if it took, say, three months and X number of posts before you’re allowed to vote?
Might work. Has anyone reading this got experience of such systems? (The obvious concern is that it would drive away “good” users as effectively as it would drive away “bad” ones, so that it would slow growth without actually making the overall pattern of growth any more favourable.)
I’m not sure if I like the basic idea, but tying it to the new users’ karma would favor good users over bad (for certain values of good).
Hacker News uses this sort of system: there are thresholds for things like being able to vote, being able to downvote, etc., and they are all based on your karma score. The same is true on Stack Overflow, though that’s a very different kind of site.
Both of them see frequent complaints that they’re going downhill, but it’s hard to be sure whether that’s anything more than standard “the world was better in my young days” thinking (which I think results from a general tendency to remember good things better than bad things).
However, that can happen whether it grows quickly or slowly.
Yes, but one of the things that effects wether someone is a good user or bad user is how well they acculturate. From what I understand, sites with a high old timer to newb ratio usually have an easier time acculturating the newbs.
That’s not what I meant… Imagine a curve that gets more and more steep as it progresses. In the first quarter of a year, say a website gets 1000 new users. In the second quarter of the year, it gets 1500 new users. In the third quarter of the year, 2250 new users.
In the second quarter, 2⁄3 of the members are old members, 1⁄3 of the members are new. In the third quarter, the ratio changes because there are so many new members—nearly half the members are new. In the fourth quarter, an even larger number of new members joins, and suddenly there are more new members than old ones. If this keeps up for any length of time, the culture will be totally destroyed.
Whether you start with 20 members and increase it by an increment of 1 and 1⁄4 the amount of users every quarter, or start with 1000 and increase it by an increment if 1 and 1⁄2 the number of users every quarter, if the number of users keeps increasing this way, eventually it will get to a point where the ratio flips and there are more new users than old ones.
It doesn’t matter whether that type of growth happens quickly or slowly. If the growth is exponential there’s a good risk of that eventually happening. Slow growth might mean it happens next year, fast growth might mean it happens next month. Either way, it’s important to protect the community from that problem.
I think you’re missing my point. Does your model have any allowance for the idea that new members can eventually become old members, and that this process is speed up by new members interacting with old members? Because the main point that I was trying to make is that a large portion (though nowhere near all) of a user’s desirability is how well they absorb the culture, and that culture can survive (more or less) intact through multiple “generations” if at any given point in time a large enough proportion of people have already acculturated. Consider a toy model were users can be discreetly divided into “old-hands” and “newbs”. In this toy model newbs turn into old-hands in x weeks iff at least 50% of the other users are old-hands (who can set a proper example), otherwise the culture collapses. In this case the culture can survive indefinitely iff the doubling time is longer than x.
Of course it could still be a good idea to rewrite the home page, since maybe the sort of people who would be attracted to a site with LW’s current home page have on the whole negative value for LW.
Well since the front page is so generic, it might attract people who react positively to the mere idea of “rationality comminity” who are more likely to be people like stereotypical Objectivists and entry-level atheism fans, as opposed to the kinds of critical and savvy people that LW would need to prove itself to by eg. showing them some awesome posts from the archives and not expecting them to be excited merely about the idea of a site about rationality.
Yes, yes! I mentioned the same thing in a different comment and I’m glad to finally see that someone else gets it. (: Other people probably see this also, but it seems most are not really interested in analyzing how new people are reacting to the front page and about page and how it affects who does and does not join. Thanks, Yli. I’ll definitely be interested in hearing any other theories you may have.
I’m happy you see where I’m coming from. Another opinion I have about issues like this is that to fix them, it wouldn’t be enough to read complaints in comments in threads like this one and try to fix things you find people complaining about. You would actually need to find a person (or persons) with a good eye for and intuitions about these things, who has good taste and who knows what they’re doing, and just let them take control of the whole design and let them change it at their discretion. I think there must be a few people on LW who would be both capable and willing to do it, though of course the site as it currently is would repel them.
I feel capable of doing the intuition part because I’m a psychology enthusiast and have done a lot of free coaching and emotional support, so I’m good at intuiting things about people’s feelings. It’s a matter of knowing what these guys care about the most. Without a feel for their culture, I won’t be able to target people who already have the same culture—this is true—but since this site is creating it’s own culture, that’s not the objective. Finding people compatible to the culture, since it’s a rationalist culture, isn’t quite the right problem to solve either. Firstly, these guys seem to want NEW information, not just hear the same old stuff—they’re not looking for people perfectly following their conventions, conforming to a stereotype, the way that most cultures do—that would be boring. Secondly, it’s a culture of rationalists. A true rationalist wants to be proven wrong and to change in the event that their ideas weren’t as good as somebody else’s. So what we’re looking at here is not the challenge of “Find more people with the exact same ideas.” but the challenge of “Find people who can contribute to this culture.”
Of course there is a fine line to draw when making the distinction between those who contribute to the culture and those who do not. The most important way to make the distinction seems to be “don’t subject us to tedium”. I theorize that boredom is the brain’s way of punishing you for not constantly learning, which gives one a survival advantage since we live in a world where even one small spot of ignorance can cause a problem, yet we can’t know everything. The only solution is: learn constantly. So, it’s no wonder humans experience boredom whenever they’re idle or doing something far too easy.
Depending on one’s intelligence level, one might be fully immersed by the human world at all times, because it’s designed to put people at that level into a state of flow, or they may be totally left out and experience flow very rarely. I think a lot of people here are the type that have to seek out flow experiences and are subjected to too much tedium elsewhere. The last thing they want is to be stuck reading the same stuff over and over, or explaining old ideas to new people. Their brains punish them for that, and they get too much of that type of punishment as it is.
So, because they’re rationalists who know better than to assume that just because somebody disagrees, they’re automatically wrong, and because they’re hungry for new information, and won’t be able to tolerate more of the same stuff all the time, I think the best way to go with this is not to attempt to attract people that match the details of the culture, but to attempt to attract people that match the root of the culture.
The root is caring about using good reasoning as one of your top priorities.
Do you think it might be possible to make Less Wrong more interactive? There are a lot of simple rationality tests which most people get wrong; like the classic tests used for Confirmation Bias or The Conjunction Fallacy.
Could we make fun versions of such tests for people to try out, perhaps with animation, etc? With a little work, we could even come up with little games that test Bayesian reasoning skills.
This would select for active, interested people who like to try things out, and if we quickly explained mistakes as soon as people made them, the idea that “I, too, am irrational” would be much more salient.
Rationality is a skill we have to practice, in the end, not just a thing we read about.
(I apologize if this would be too hard to implement to be worth it; I have very little programming experience.)
I’ve been thinking about this for a little while. This is a really, really good idea. It’s not too hard to implement. I feel there are some other things to do first (prevent endless September for instance) but other than that I think it’s a great idea and it’s just a matter of constructing these with the right questions. Questions have to be worded very carefully, and tests constructed cautiously, in order to have the scientific properties that give test results their accuracy. On the other hand, fun internet quizzes can bring in users and do not necessarily need to be scientifically sound (though for this site in particular, I’d figure that wording even a fun quiz as scientifically as possible would be the way to go, as that would attract more like minded people and gain more respect). It would actually take a lot more time to think of all the questions, consider how scientific the series of tests was, and word all the questions correctly than to set up a script that allows you to take the tests.
At least to begin with, we don’t have to come up with things on our own. There’s a whole literature of psychological studies we can comb through to recreate. The Sequences cite a whole lot of iconic studies, and a sufficiently motivated person could dig up more obscure follow-ups, too.
Converting them into a format that would work on the Internet is a bit trickier, but a lot can be done with Java applets.
The first page I linked to shows a Confirmation Bias test designed by a LWer, based on the classic experiment. The Conjunction Fallacy has a simple written multiple choice test.
Anchoring and adjustment could be tested for by providing people with high or low random numbers and asking them to answer a bunch of estimation questions, like the “How many countries in Africa?” test. It would work out best if we already had people take this test and had gathered data to show people upon completion.
and for these purposes some visitors are much, much more valuable than others and many have negative value
That’s what downvotes are for, isn’t it? OTOH, if a potential valuable contributor sees the front page, doesn’t fully realize what the site is about, and walks away, that’s too bad.
I’m all for trying to make the LW front page more engaging, but I am skeptical about an assumption pervading this post: that growth should be a major goal, and that faster growth is better aside from major pathologies like spammers and trolls.
For a website that exists mostly to make money (by selling things, advertising a product, gaining visibility for a person or company, etc.) this is a good assumption: all else being equal, you want more page views, lower bounce rates, higher conversion rate, etc. But LW doesn’t exist mostly to make money; it exists (in high-minded principle) to refine the art of human rationality and (more prosaically) to provide interest and entertainment for its participants, and for these purposes some visitors are much, much more valuable than others and many have negative value. (This is true even taking into account the prospective value of LW to those visitors.)
By definition, LW wants (or should want) to “grow optimally”—but that may mean “grow very rapidly” or “grow slowly” or even “not grow at all, at present”.
I concur. I have implied this in my previous comments, but I will going come out and explicitly say it here. Caeteris paribus I want LW’s growth rate to be higher than it currently is, but I am more concerned with keeping out the intellectual riff-raff than encouraging growth. Furthermore, I actively want to avoid the exponential growth scenario, unless we can create an effective user orientation or another way of reducing the time it takes for new members to acculturate.
I agree that there’s a lot of annoyance for both the new user and old users when new people join. A new user orientation is needed. That’s why I wrote an outline for a New User Orientation full of suggestions that I personally would like to see included in the new user orientation. Please go and critique it—older members are surely going to have suggestions I wouldn’t think of. I have nothing against writing one, but I can’t write a good one with no input, I’m too new.
Please distinguish “keeping out the intellectual riff-raff” from elitism. Are you talking about people who really aren’t serious about rational though? Maybe you just mean you don’t want more trolls? Or do you mean you want to outright make an IQ requirement?
I have an idea for scaring off those who are not serious about rational thought that goes like this:
From my post New User Orientation
Do you think this would work?
What are the reasons for not wanting to grow? You didn’t state any reasons I haven’t already stated.
For the record, I am aware that it’s not a good idea to assume that growth is 100% good. That’s why I put that in the OP.
I feel that LessWrong could be really important—and maybe it already is—that’s why I want to see it grow (that much is obvious.) I started writing about why, but I want to hear your reasons for not wanting it to grow.
You’re going against the grain—not a bad thing but it means you’re going to have to really lay out your reasons if you want to change the way the wind is blowing.
Elaborate, please.
(I think you may have a wrong idea of what “pontificate” means. Either that or you’re being gratuitously rude, which I’m going to assume you aren’t.)
The following two propositions are different. (1) Growth is not always 100% good. (2) Growth is not always good. #1 is what you stated. #2 is what I’m saying. #2 goes much further than #1 does. The obvious inference from #1 is “make sure you grow, but take some measures to mitigate the possible downsides”; the obvious inference from #2 is “consider carefully whether growth is what you want”.
I don’t know whether I want LW to grow, or how fast, or when (which is why I said that optimal growth for LW could be any of several things, including “grow very rapidly”) so I can’t really give you “[my] reasons for not wanting it to grow”. What I can do is to give some possible reasons why growth—especially rapid growth, especially especially rapid growth by the mechanism you’re implicitly proposing—might be the Wrong Thing.
The main reason is this: It may not be possible to grow rapidly by the sort of means you describe without changing the demographics of LW in a way that would lower its quality. I don’t just mean “LW might start to attract trolls and spammers and idiots”. I mean that maybe only a small fraction of new visitors to LW are people whose presence would enhance LW, and that if that’s so then anything we do to encourage a lot more new visitors to stay will make LW worse.
There’ve been a couple of surveys of LW participants. They suggest that the userbase of LW is very unusual. So, for that matter, does a casual look at the discussions here. LW participants are unusually interested in rationality (of course, since that’s the central topic of the site), something like three standard deviations above average in measured intelligence (according to self-reports, so take that with a pinch of salt, but I don’t find it hard to believe that something of the kind is true), and willing to read and in many cases write long, well researched, sometimes highly technical material. This combination of characteristics is rare. It’s surely a large part of what gives LW the distinctive character it has. I am not optimistic about the prospects of keeping it if LW grows rapidly by keeping a much larger fraction of its new visitors.
(I am aware that the foregoing paragraph sounds as if I’m saying “look at us, we’re superior to everyone else”. That is not, even slightly, the point. LW regulars, as a population, are unusually good at some things, unusually bad at others, and very average at others, and treating one group of people as “superior” to others is a road to ruin for all sorts of reasons. All I’m saying is that a community with the particular strengths LW has is a very unusual thing, and I think a lot of LW’s distinctive merits derive from those strengths.)
Now, I know that you said things like “We’d better choose what audience to target”. But I think you underestimate what an … odd … audience it is that LW might need to target, and the general shape of your proposals—make it so people are immediately grabbed without having to read all that text; base the design on general “web marketing” principles—seems like it’s implicitly aimed at exactly the people who don’t have the peculiar characteristics that make LW what it currently is.
I think this lies behind your perception that people here “seem to be totally unaware of the field of web marketing”. Probably many are, but please consider the possibility that in some cases the issue isn’t others’ obtuseness but that they simply aren’t interested in the “web marketing” goals you have in mind, and for a possibly-good reason.
A large increase in LW’s audience could have big benefits; I’m not denying that. They might outweigh the likely loss in quality of discussion that (I think) would go along with it. Or maybe there’s a way to increase the audience a lot without a big growth in the population of active participants. Or maybe “basic web marketing principles” can in fact be applied in a way that pulls in a huge new population of people with the extremely unusual characteristics LW’s current regulars have. So I’m not (to repeat myself) saying that LW should not aim to grow a lot. I don’t know whether it should. It just makes me really uneasy to see growth being treated as axiomatically good and important (which, repeating myself again, is not the same as saying that it has no downsides, and which I know you weren’t saying).
You had a good point in your suggestion so I changed my “100% good” statement.
I also responded (different comment).
Oops sorry. I removed pontificate. Thank you for not assuming I was being rude.
“I mean that maybe only a small fraction of new visitors to LW are people whose presence would enhance LW, and that if that’s so then anything we do to encourage a lot more new visitors to stay will make LW worse.”
Good problem. However, that can happen whether it grows quickly or slowly.
“I am not optimistic about the prospects of keeping it if LW grows rapidly by keeping a much larger fraction of its new visitors.”
Yes! I did warn that this could happen, and that it could happen whether we like it or not.
I wonder whether you think this idea would work:
“http://lesswrong.com/r/discussion/lw/e5r/lesswrong_could_grow_a_lot_but_were_doing_it_wrong/78r2″
I also had another idea:
I like that things can be voted up and down, but I think it would be a heck of a lot more effective if specific feedback was provided. As a new user, I will adjust a lot faster if I know WHAT I am doing wrong/right not just THAT I am doing something wrong/right. Also, I wouldn’t write off the votes to “bad attitude” and “trolls” when I don’t understand them. That’s what I’m doing now, lol, and I don’t like that, but I don’t know what else to think of them, lol.
I think if the votes required a reason, which could be as quick as the word “Trolling” or the name of a logical fallacy, that would make a huge difference. That would ALSO force us to get conscious of our reasons for voting, which would provoke conscious review of the reason, which would probably result in better decisions.
And to ensure the culture doesn’t change too quickly, what if it took, say, three months and X number of posts before you’re allowed to vote? That way, the oldest users get to influence the culture, and users who just aren’t compatible with LessWrong will wander away before they’ve exerted an influence.
This is basically the same as how you have to be a resident of a country before you can vote.
From a developer’s standpoint, that would be a very easy change, but more importantly, it would protect the site no matter whether it grows faster or at the current pace.
“please consider the possibility that in some cases the issue isn’t others’ obtuseness but that they simply aren’t interested in the “web marketing” goals you have in mind, and for a possibly-good reason.”
Oh, I did! I did consider that. (: That was the entire reason I created this thread. I figured you would all let me know what you wanted and if growth wasn’t it, you’d shoot me down. (:
“They might outweigh the likely loss in quality of discussion that (I think) would go along with it. ”
Has anyone thought much about what would improve the quality of discussion? Because that could be improved at the same time. Or before making marketing changes—to be sure that the safeguard is in place first.
“Or maybe “basic web marketing principles” can in fact be applied in a way that pulls in a huge new population of people with the extremely unusual characteristics LW’s current regulars have. ”
Yes! That’s what I was hoping for. (: That’s why I asked who your target demographic is. I might think I know, but you guys are the ones that have been talking to them for years.
Yes. It’s an issue regardless. But if pursuing a strategy of rapid growth guarantees an influx of negative-net-value-to-LW new users, it’s probably a bad thing. (We will all die eventually, but I still prefer there to be fewer mass murders.)
I don’t have a good enough mental model of the typical first-time visitor to LW to have a strong opinion. What I do think likely is that either it wouldn’t work or it would work at the cost of getting rid of that rapid growth you were hoping for. Because I think it’s likely that rapid growth by increasing visitor retention implies the sort of change in LW’s demographics that I described.
I have proposed something similar myself. The UI would be tricky to get right. It might well make LW a better place (or it might not; these things have a way of producing unexpected consequences). But I don’t see how it would do much to solve the problem I described, if (as I suspect but don’t claim to know) it really is a problem.
Might work. Has anyone reading this got experience of such systems? (The obvious concern is that it would drive away “good” users as effectively as it would drive away “bad” ones, so that it would slow growth without actually making the overall pattern of growth any more favourable.)
According to my quick count, 28 different people have commented here [EDITED to add: not counting “metatroll”], of whom 7 seemed definitely in favour of growth (perhaps having thought it through, perhaps just because it’s a sort of default goal), one seemed definitely against growth, and 6 seemed definitely skeptical (i.e., saying “growth might not be what we want”, basically my position). The other 14 expressed no opinion on the matter that I could detect. Make of that what you will.
I think a lot of people have, but so far as I know no one (here or elsewhere) has a silver bullet that ensures that a community of unusual people will retain its distinctively valuable characteristics as it grows. (Or for that matter as it doesn’t.)
I’m not sure if I like the basic idea, but tying it to the new users’ karma would favor good users over bad (for certain values of good).
(Your quoting is slightly broken.)
Hacker News uses this sort of system: there are thresholds for things like being able to vote, being able to downvote, etc., and they are all based on your karma score. The same is true on Stack Overflow, though that’s a very different kind of site.
Both of them see frequent complaints that they’re going downhill, but it’s hard to be sure whether that’s anything more than standard “the world was better in my young days” thinking (which I think results from a general tendency to remember good things better than bad things).
Frustratingly, the help doesn’t say how to do nested quotes.
Hmm, let’s experiment.
This line is normal.
This line is normal.
All those lines are separated by blank lines. (Not doing so produces bad results.)
Ok, I think I have it working.
Yes, but one of the things that effects wether someone is a good user or bad user is how well they acculturate. From what I understand, sites with a high old timer to newb ratio usually have an easier time acculturating the newbs.
That’s not what I meant… Imagine a curve that gets more and more steep as it progresses. In the first quarter of a year, say a website gets 1000 new users. In the second quarter of the year, it gets 1500 new users. In the third quarter of the year, 2250 new users.
In the second quarter, 2⁄3 of the members are old members, 1⁄3 of the members are new. In the third quarter, the ratio changes because there are so many new members—nearly half the members are new. In the fourth quarter, an even larger number of new members joins, and suddenly there are more new members than old ones. If this keeps up for any length of time, the culture will be totally destroyed.
Whether you start with 20 members and increase it by an increment of 1 and 1⁄4 the amount of users every quarter, or start with 1000 and increase it by an increment if 1 and 1⁄2 the number of users every quarter, if the number of users keeps increasing this way, eventually it will get to a point where the ratio flips and there are more new users than old ones.
It doesn’t matter whether that type of growth happens quickly or slowly. If the growth is exponential there’s a good risk of that eventually happening. Slow growth might mean it happens next year, fast growth might mean it happens next month. Either way, it’s important to protect the community from that problem.
I think you’re missing my point. Does your model have any allowance for the idea that new members can eventually become old members, and that this process is speed up by new members interacting with old members? Because the main point that I was trying to make is that a large portion (though nowhere near all) of a user’s desirability is how well they absorb the culture, and that culture can survive (more or less) intact through multiple “generations” if at any given point in time a large enough proportion of people have already acculturated. Consider a toy model were users can be discreetly divided into “old-hands” and “newbs”. In this toy model newbs turn into old-hands in x weeks iff at least 50% of the other users are old-hands (who can set a proper example), otherwise the culture collapses. In this case the culture can survive indefinitely iff the doubling time is longer than x.
Of course it could still be a good idea to rewrite the home page, since maybe the sort of people who would be attracted to a site with LW’s current home page have on the whole negative value for LW.
Why is it attracting the wrong ones, yli? I have a vague sense that you might be right, so I’m very curious about your reasoning.
Well since the front page is so generic, it might attract people who react positively to the mere idea of “rationality comminity” who are more likely to be people like stereotypical Objectivists and entry-level atheism fans, as opposed to the kinds of critical and savvy people that LW would need to prove itself to by eg. showing them some awesome posts from the archives and not expecting them to be excited merely about the idea of a site about rationality.
Yes, yes! I mentioned the same thing in a different comment and I’m glad to finally see that someone else gets it. (: Other people probably see this also, but it seems most are not really interested in analyzing how new people are reacting to the front page and about page and how it affects who does and does not join. Thanks, Yli. I’ll definitely be interested in hearing any other theories you may have.
I’m happy you see where I’m coming from. Another opinion I have about issues like this is that to fix them, it wouldn’t be enough to read complaints in comments in threads like this one and try to fix things you find people complaining about. You would actually need to find a person (or persons) with a good eye for and intuitions about these things, who has good taste and who knows what they’re doing, and just let them take control of the whole design and let them change it at their discretion. I think there must be a few people on LW who would be both capable and willing to do it, though of course the site as it currently is would repel them.
I feel capable of doing the intuition part because I’m a psychology enthusiast and have done a lot of free coaching and emotional support, so I’m good at intuiting things about people’s feelings. It’s a matter of knowing what these guys care about the most. Without a feel for their culture, I won’t be able to target people who already have the same culture—this is true—but since this site is creating it’s own culture, that’s not the objective. Finding people compatible to the culture, since it’s a rationalist culture, isn’t quite the right problem to solve either. Firstly, these guys seem to want NEW information, not just hear the same old stuff—they’re not looking for people perfectly following their conventions, conforming to a stereotype, the way that most cultures do—that would be boring. Secondly, it’s a culture of rationalists. A true rationalist wants to be proven wrong and to change in the event that their ideas weren’t as good as somebody else’s. So what we’re looking at here is not the challenge of “Find more people with the exact same ideas.” but the challenge of “Find people who can contribute to this culture.”
Of course there is a fine line to draw when making the distinction between those who contribute to the culture and those who do not. The most important way to make the distinction seems to be “don’t subject us to tedium”. I theorize that boredom is the brain’s way of punishing you for not constantly learning, which gives one a survival advantage since we live in a world where even one small spot of ignorance can cause a problem, yet we can’t know everything. The only solution is: learn constantly. So, it’s no wonder humans experience boredom whenever they’re idle or doing something far too easy.
Depending on one’s intelligence level, one might be fully immersed by the human world at all times, because it’s designed to put people at that level into a state of flow, or they may be totally left out and experience flow very rarely. I think a lot of people here are the type that have to seek out flow experiences and are subjected to too much tedium elsewhere. The last thing they want is to be stuck reading the same stuff over and over, or explaining old ideas to new people. Their brains punish them for that, and they get too much of that type of punishment as it is.
So, because they’re rationalists who know better than to assume that just because somebody disagrees, they’re automatically wrong, and because they’re hungry for new information, and won’t be able to tolerate more of the same stuff all the time, I think the best way to go with this is not to attempt to attract people that match the details of the culture, but to attempt to attract people that match the root of the culture.
The root is caring about using good reasoning as one of your top priorities.
What do you think about that?
Do you think it might be possible to make Less Wrong more interactive? There are a lot of simple rationality tests which most people get wrong; like the classic tests used for Confirmation Bias or The Conjunction Fallacy.
Could we make fun versions of such tests for people to try out, perhaps with animation, etc? With a little work, we could even come up with little games that test Bayesian reasoning skills.
This would select for active, interested people who like to try things out, and if we quickly explained mistakes as soon as people made them, the idea that “I, too, am irrational” would be much more salient.
Rationality is a skill we have to practice, in the end, not just a thing we read about.
(I apologize if this would be too hard to implement to be worth it; I have very little programming experience.)
I’ve been thinking about this for a little while. This is a really, really good idea. It’s not too hard to implement. I feel there are some other things to do first (prevent endless September for instance) but other than that I think it’s a great idea and it’s just a matter of constructing these with the right questions. Questions have to be worded very carefully, and tests constructed cautiously, in order to have the scientific properties that give test results their accuracy. On the other hand, fun internet quizzes can bring in users and do not necessarily need to be scientifically sound (though for this site in particular, I’d figure that wording even a fun quiz as scientifically as possible would be the way to go, as that would attract more like minded people and gain more respect). It would actually take a lot more time to think of all the questions, consider how scientific the series of tests was, and word all the questions correctly than to set up a script that allows you to take the tests.
At least to begin with, we don’t have to come up with things on our own. There’s a whole literature of psychological studies we can comb through to recreate. The Sequences cite a whole lot of iconic studies, and a sufficiently motivated person could dig up more obscure follow-ups, too.
Converting them into a format that would work on the Internet is a bit trickier, but a lot can be done with Java applets.
Ooh… But what about copyrights? And if they are copyright expired, would they be any good? Maybe. Hmm.
Do you have suggestions for specific materials we could start with?
The first page I linked to shows a Confirmation Bias test designed by a LWer, based on the classic experiment. The Conjunction Fallacy has a simple written multiple choice test.
Anchoring and adjustment could be tested for by providing people with high or low random numbers and asking them to answer a bunch of estimation questions, like the “How many countries in Africa?” test. It would work out best if we already had people take this test and had gathered data to show people upon completion.
That’s what downvotes are for, isn’t it? OTOH, if a potential valuable contributor sees the front page, doesn’t fully realize what the site is about, and walks away, that’s too bad.