Essentially, Eliezer gets negative karma for some of his comments (-13, −4, −12, −7) explaining why he thinks the new changes of karma rules are a good thing. To compare, even the obvious trolls usually don’t get −13 comment karma.
What exactly is the problem? I don’t think that for a regular commenter, having to pay 5 karma points for replying to a negatively voted comment is such a problem. Because you will do it only once in a while, right? Most of your comments will still be reactions to articles or to non-negatively voted comments, right? So what exactly is this problem, and why this overreaction? Certainly, there are situations where replying to a negatively voted comment is the right thing to do. But are they the exception, or the rule? Because the new algorithm does not prevent you from doing this; it only provides a trivial disincentive to do so.
What is happening here?
A few months ago LW needed an article to defend that some people here really have read the Sequences, and that recommending Sequences to someone is not an offense. What? How can this happen on a website which originally more or less was the Sequences? That seemed absurd to me, and so does this; as if both suggest that LW is becoming less what it was, and more a general discussion forum.
I suggest everyone to think for a moment about the fact that Eliezer somehow created this site, wrote a lot of content people consider useful, and made some decisions about the voting system, which together resulted in a website we like. So perhaps this is some Bayesian evidence that he knows what he is doing. And even in the case this would turn out to be a mistake, it would be easy to revert. Also, everyone here is completely free to create a competing x-rationalist website, if your worst nightmares about LW come true. (And then I want to see how you solve the problem of trolling there, when it suddenly becomes your responsibility.)
Recently we had also a few articles about how to make LW more popular; how to attract more readers and participants. Well, if that happens, we will need more strict moderation than we have now; otherwise we will drown in the noise. For instance, within this week we have a full screen of “Discussion” articles, some of them containing 86, 103, 191 comments. How many of those comments contain really useful information? What is your estimate, how many of that information will you remember after one week? Do you think that visiting LW once in a week is enough to deal with that amount of information? Or do you just ignore most of that? How big part of a week can you spend online reading LW, and still pretending you are being rational instead of procrastinating?
Perhaps LW needs more users, but it probably needs less text per week (certainly not more); both articles and comments. Less chatting, more thinking, better expressing ourselves. More moderation is needed. And most of you are not going to pay for human moderators, so I think you should just accept the existing rules, and their changes. Or you can always make a competing website, you know; but you won’t do it, and you also know why.
I suggest everyone to think for a moment about the fact that Eliezer somehow created this site, wrote a lot of content people consider useful, and made some decisions about the voting system, which together resulted in a website we like. So perhaps this is some Bayesian evidence that he knows what he is doing.
There’s also plenty of Bayesian evidence he’s not that great at moderation. SL4 was enough of an eventual failure to prompt the creation of OB; OB prompted the creation of LW; he failed to predict that opening up posting would lead to floods of posts like it did for LW; he signally failed to understand that his reaction to Roko’s basilisk was pretty much the worst possible reaction he could engage in, such that even now it’s still coming up in print publications about LWers; and this recent karma stuff isn’t looking much better.
I am reminded strongly of Jimbo Wales. He too helped create a successful community but seemed to do so accidentally as he later supported initiatives that directly undermined what made that community function.
Seems to me there are two important factors to distinguish:
how good is Eliezer at “herding cats” (as opposed to someone else herding cats)
how difficult is herding cats (as opposed to herding other species)
To me it seems that the problem is the inherent difficulty of herding cats; and Eliezer is the most successful example I have ever seen. I have seen initially good web communities ruined after a year or two… and then I read an article describing how exactly that happened. From outside view, LW seems to survive for surprisingly long time as a decent website.
The problem with Roko seems to me a bit similar to what is happening now—some people intentionally do things that annoy other people; moderator tries to supress that behavior; contrarians enjoy fighting him by making it more visible and rationalize their behavior as defending the freedom of speech or whatever. The Roko situation was much more insane; at least one person threatened to increase existential risk if Eliezer does not stop moderating the discussion. Today the most crazy reaction I found was upvoting an obvious trollso that others can comment on their nonsensical sequence of words without karma costs! Yay, that’s exactly the behavior you would expect to find in a super-rational community, right? Unfortunately, it is exactly the kind of behavior you will find when you make a website for wannabe smart people.
Wikipedia is different: it is neither a blog nor a discussion forum. And it exists at cost of hundreds of people who have no life, so they can spend a lot of time in endless edit wars. This is yet another danger for LW. Not only new users can overrule the old users, but also the old users who have no life can overrule the old users whose instrumental goals are outside of LW. Users who want to reduce their procrastination on LW will not participate in endless discussions. If there is more content per day, they will simply read less, therefore they will vote less on an average comment, and they will have less word in “community” decisions. There is a risk that the procrastinators will simply optimize the website to fit their preferences—preferences of people who don’t mind spending a lot of time online, therefore e.g. reading comments by trolls and the subsequent discussions is not a problem for them. From their point of view, strict moderation will seem too harsh and fun-reducing.
As a reminder, if someone is convinced that they (as a person, or as a group) have better skills at maintaining a rationalist website, there is always a possibility of starting a new rationalist website. It could be even interlinked with LW, similarly like OB is now. Make an experiment, bet your own money and/or time!
I don’t think any of that addresses the main point: what has Eliezer done that is evidence of good moderating skills? Who has Eliezer banned or not banned? etc.
The question isn’t: “can Eliezer spend years cranking out high quality content on the excellent Reddit codebase with a small pre-existing community and see it grow?” It is: “can Eliezer effectively moderate this growing community?” And I gave several examples of how he had not done so effectively before LW, and has not done so effectively since LW.
(And I think you badly underestimate the similarities of Wikipedia during its good phase and LW. Both tackle tough problems and aspire to accumulate high quality content, with very nerdish users, and hence, solve or fail at very similar problems.)
The problem with Roko seems to me a bit similar to what is happening now—some people intentionally do things that annoy other people
This just isn’t remotely accurate as a representation of history.
The remainder of the parent comment seems to present similarly false (or hyperbolically misrepresented) premises and reason from them to dubious conclusions.
Eliezer is not so vulnerable that he needs to be supported by bullshit.
My thoughts on the recent excitement about “trolls”, and moderation, and the new karma penalty for engaging with significantly downvoted comments:
First, the words troll and trolling are being used very indiscriminately, to refer to a wide variety of behaviors and intentions. If LW really needed to have a long-term discussion, about how to deal with the “troll problem”, it would be advisable to develop a much more precise vocabulary, and also a more objective, verifiable assessment of how much “trolling” and “troll-feeding” was happening, e.g. a list of examples.
So it seems we won’t need some specialized troll-ologists to work out all the issues. Rather than a “war on trolls” becoming a permanent element of LW political life, I’m hoping that in the long run this is just an episode in the history of LW governance. The site has transformed several times, it will undoubtedly transform again, and this is just a blip, one bump on the road.
I am somewhat interested in the larger issue of how the site might best produce intellectual progress. Viliam links to Grognor’s article, “I Stand by the Sequences” (note that Grognor has since quit LW to join the aphoristic faction on Twitter, like “muflax” and “Kate Evans”, who specialize in producing philosophical one-liners). A similar article from the same period is “Our Phyg Is Not Exclusive Enough”. These articles received some criticism as promoting dogmatism, groupthink, exclusivity, etc.; Alexander Kruel, aka XiXiDu, another LW defector, blogged about them as evidence of this.
However, I very much agree with the impulse behind those articles, even though I dissent from common LW opinion in some major ways. LW is not a site for anyone to talk about anything; it’s not even a site for anyone who considers themselves “rational” to talk about anything. The Sequences do define a philosophy and they need to remain the reference point. Perhaps elements of them will one day, by consensus, be regarded as definitively obsolete, replaced by something better, but they’re still the starting point from which any future progress begins, even if it’s progress by opposition.
A year ago, I wondered what LW would amount to, if anything. LW is protean, it has many dimensions, but I especially meant its place in the history of ideas. I’m now prepared to say that it can amount to something, that it can be a tributary feeding into the common intellectual culture of humanity, but that will require a certain amount of discipline and due diligence on the part of people who do want it to matter that way. There are many things that are working already, for example the division between Discussion and Main (and perhaps the wiki represents an even higher-level distillation). It’s good to have the rambunctious lower level where we are now, as well as the more rarefied and rigorous higher levels. It permits new possibilities to emerge.
Maybe my main message is to serious critics of LW (some of the “trolls” are just critics who aren’t doing it constructively). They can actually contribute to the overall process by being more organized in their criticisms. This is one way that intellectual progress occurs: you have a position, you have an opposite position, and both positions are refined as a result of dialogue. LW, the site and the community, does have the mechanisms and the capacity to take on alternative views and give them a fair hearing, even if they are eventually rejected. Work with that, and we can all benefit.
One more thing I want to point out. It’s often observed that hardly any sequences have been written since Eliezer. In fact, LWer palladias has written about a dozen series of posts on her blog in Sequence format. She recently achieved notoriety for converting from atheism to religion, so her sequences aren’t LW canon, but they represent an interesting example of cross-pollination between very different schools of thought.
Just spotted this thread. The Sequences were indeed the direct inspiration for the format of the linked series of posts I run. Though mine are on a pretty broad range of topics—most recently contrasting Sondheim’s Company with Passion and using both to talk about what the ends of marriage are.
Recently we had also a few articles about how to make LW more popular; how to attract more readers and participants. Well, if that happens, we will need more strict moderation than we have now; otherwise we will drown in the noise. For instance, within this week we have a full screen of “Discussion” articles, some of them containing 86, 103, 191 comments. How many of those comments contain really useful information? What is your estimate, how many of that information will you remember after one week? Do you think that visiting LW once in a week is enough to deal with that amount of information? Or do you just ignore most of that? How big part of a week can you spend online reading LW, and still pretending you are being rational instead of procrastinating?
Up voted for this. I can’t believe how many people don’t get it.
He got my downvotes for making terrible arguments defending a change that won’t do what it’s supposed to do, while also doing other shitty things. He was also an overconfident dick about the whole situation. The problem isn’t the rule, it’s the wrong beliefs about how the forums work and how they might be fixed.
That thread is Bayesian evidence against the new poorly thought out rule. The objections that have been raised to it have not even come close to being met. That fact that your own post is a hair breadth away from inflicting negative karma on me should be enough to give you pause.
The reaction to the new rule should not be surprising. If it was surprising, then you should update your model.
Good point about the silliness of people downvoting Eliezer to show their disagreement.
Certainly, there are situations where replying to a negatively voted comment is the right thing to do. But are they the exception, or the rule? Because the new algorithm does not prevent you from doing this; it only provides a trivial disincentive to do so.
Using the phrase ‘trivial disincentive’ looks like a deliberate reference to this article which would be an unconvincing way to argue that the change won’t cause any problems.
And in general, I don’t think that the change will have really serious side-effects but I’m in favor of changing complex systems in as small increments as possible. The only sensible, currently relevant reason for implementing the new feature (flooding of the recent comments sidebar) that was given can be solved much less invasively by not having comments from crappy threads show up in the recent comments sidebar. For additional soft paternalist goodness, you could also have replies to comments made in such threads not appear in user’s inboxes.
Do you think that visiting LW once in a week is enough to deal with that amount of information? Or do you just ignore most of that? How big part of a week can you spend online reading LW, and still pretending you are being rational instead of procrastinating?
Being able to keep up with all the conversation going on LessWrong seems incompatible with the goal of expanding the community. Reading comments and participating in conversation is a leisure activity. If I were very concerned with being “rational” about my LessWrong usage patterns I would stop reading them at all and stick to just articles (possibly only main section articles if I were really concerned).
This discussion thread is insane.
Essentially, Eliezer gets negative karma for some of his comments (-13, −4, −12, −7) explaining why he thinks the new changes of karma rules are a good thing. To compare, even the obvious trolls usually don’t get −13 comment karma.
What exactly is the problem? I don’t think that for a regular commenter, having to pay 5 karma points for replying to a negatively voted comment is such a problem. Because you will do it only once in a while, right? Most of your comments will still be reactions to articles or to non-negatively voted comments, right? So what exactly is this problem, and why this overreaction? Certainly, there are situations where replying to a negatively voted comment is the right thing to do. But are they the exception, or the rule? Because the new algorithm does not prevent you from doing this; it only provides a trivial disincentive to do so.
What is happening here?
A few months ago LW needed an article to defend that some people here really have read the Sequences, and that recommending Sequences to someone is not an offense. What? How can this happen on a website which originally more or less was the Sequences? That seemed absurd to me, and so does this; as if both suggest that LW is becoming less what it was, and more a general discussion forum.
I suggest everyone to think for a moment about the fact that Eliezer somehow created this site, wrote a lot of content people consider useful, and made some decisions about the voting system, which together resulted in a website we like. So perhaps this is some Bayesian evidence that he knows what he is doing. And even in the case this would turn out to be a mistake, it would be easy to revert. Also, everyone here is completely free to create a competing x-rationalist website, if your worst nightmares about LW come true. (And then I want to see how you solve the problem of trolling there, when it suddenly becomes your responsibility.)
Recently we had also a few articles about how to make LW more popular; how to attract more readers and participants. Well, if that happens, we will need more strict moderation than we have now; otherwise we will drown in the noise. For instance, within this week we have a full screen of “Discussion” articles, some of them containing 86, 103, 191 comments. How many of those comments contain really useful information? What is your estimate, how many of that information will you remember after one week? Do you think that visiting LW once in a week is enough to deal with that amount of information? Or do you just ignore most of that? How big part of a week can you spend online reading LW, and still pretending you are being rational instead of procrastinating?
Perhaps LW needs more users, but it probably needs less text per week (certainly not more); both articles and comments. Less chatting, more thinking, better expressing ourselves. More moderation is needed. And most of you are not going to pay for human moderators, so I think you should just accept the existing rules, and their changes. Or you can always make a competing website, you know; but you won’t do it, and you also know why.
There’s also plenty of Bayesian evidence he’s not that great at moderation. SL4 was enough of an eventual failure to prompt the creation of OB; OB prompted the creation of LW; he failed to predict that opening up posting would lead to floods of posts like it did for LW; he signally failed to understand that his reaction to Roko’s basilisk was pretty much the worst possible reaction he could engage in, such that even now it’s still coming up in print publications about LWers; and this recent karma stuff isn’t looking much better.
I am reminded strongly of Jimbo Wales. He too helped create a successful community but seemed to do so accidentally as he later supported initiatives that directly undermined what made that community function.
Seems to me there are two important factors to distinguish:
how good is Eliezer at “herding cats” (as opposed to someone else herding cats)
how difficult is herding cats (as opposed to herding other species)
To me it seems that the problem is the inherent difficulty of herding cats; and Eliezer is the most successful example I have ever seen. I have seen initially good web communities ruined after a year or two… and then I read an article describing how exactly that happened. From outside view, LW seems to survive for surprisingly long time as a decent website.
The problem with Roko seems to me a bit similar to what is happening now—some people intentionally do things that annoy other people; moderator tries to supress that behavior; contrarians enjoy fighting him by making it more visible and rationalize their behavior as defending the freedom of speech or whatever. The Roko situation was much more insane; at least one person threatened to increase existential risk if Eliezer does not stop moderating the discussion. Today the most crazy reaction I found was upvoting an obvious troll so that others can comment on their nonsensical sequence of words without karma costs! Yay, that’s exactly the behavior you would expect to find in a super-rational community, right? Unfortunately, it is exactly the kind of behavior you will find when you make a website for wannabe smart people.
Wikipedia is different: it is neither a blog nor a discussion forum. And it exists at cost of hundreds of people who have no life, so they can spend a lot of time in endless edit wars. This is yet another danger for LW. Not only new users can overrule the old users, but also the old users who have no life can overrule the old users whose instrumental goals are outside of LW. Users who want to reduce their procrastination on LW will not participate in endless discussions. If there is more content per day, they will simply read less, therefore they will vote less on an average comment, and they will have less word in “community” decisions. There is a risk that the procrastinators will simply optimize the website to fit their preferences—preferences of people who don’t mind spending a lot of time online, therefore e.g. reading comments by trolls and the subsequent discussions is not a problem for them. From their point of view, strict moderation will seem too harsh and fun-reducing.
As a reminder, if someone is convinced that they (as a person, or as a group) have better skills at maintaining a rationalist website, there is always a possibility of starting a new rationalist website. It could be even interlinked with LW, similarly like OB is now. Make an experiment, bet your own money and/or time!
I don’t think any of that addresses the main point: what has Eliezer done that is evidence of good moderating skills? Who has Eliezer banned or not banned? etc.
The question isn’t: “can Eliezer spend years cranking out high quality content on the excellent Reddit codebase with a small pre-existing community and see it grow?” It is: “can Eliezer effectively moderate this growing community?” And I gave several examples of how he had not done so effectively before LW, and has not done so effectively since LW.
(And I think you badly underestimate the similarities of Wikipedia during its good phase and LW. Both tackle tough problems and aspire to accumulate high quality content, with very nerdish users, and hence, solve or fail at very similar problems.)
This just isn’t remotely accurate as a representation of history.
The remainder of the parent comment seems to present similarly false (or hyperbolically misrepresented) premises and reason from them to dubious conclusions.
Eliezer is not so vulnerable that he needs to be supported by bullshit.
My thoughts on the recent excitement about “trolls”, and moderation, and the new karma penalty for engaging with significantly downvoted comments:
First, the words troll and trolling are being used very indiscriminately, to refer to a wide variety of behaviors and intentions. If LW really needed to have a long-term discussion, about how to deal with the “troll problem”, it would be advisable to develop a much more precise vocabulary, and also a more objective, verifiable assessment of how much “trolling” and “troll-feeding” was happening, e.g. a list of examples.
However, it seems that people are already moving on. For future reference, here are all the articles in Discussion which arose directly from the appearance of the new penalty and the ensuing debate: “Karma for last 30 days?”, “Dealing with trolling”, “Dealing with meta-disussion”, “Karma vote checklist?”, “Preventing endless September”, “Protection against cultural collapse”, and hopefully that’s the end of it.
So it seems we won’t need some specialized troll-ologists to work out all the issues. Rather than a “war on trolls” becoming a permanent element of LW political life, I’m hoping that in the long run this is just an episode in the history of LW governance. The site has transformed several times, it will undoubtedly transform again, and this is just a blip, one bump on the road.
I am somewhat interested in the larger issue of how the site might best produce intellectual progress. Viliam links to Grognor’s article, “I Stand by the Sequences” (note that Grognor has since quit LW to join the aphoristic faction on Twitter, like “muflax” and “Kate Evans”, who specialize in producing philosophical one-liners). A similar article from the same period is “Our Phyg Is Not Exclusive Enough”. These articles received some criticism as promoting dogmatism, groupthink, exclusivity, etc.; Alexander Kruel, aka XiXiDu, another LW defector, blogged about them as evidence of this.
However, I very much agree with the impulse behind those articles, even though I dissent from common LW opinion in some major ways. LW is not a site for anyone to talk about anything; it’s not even a site for anyone who considers themselves “rational” to talk about anything. The Sequences do define a philosophy and they need to remain the reference point. Perhaps elements of them will one day, by consensus, be regarded as definitively obsolete, replaced by something better, but they’re still the starting point from which any future progress begins, even if it’s progress by opposition.
A year ago, I wondered what LW would amount to, if anything. LW is protean, it has many dimensions, but I especially meant its place in the history of ideas. I’m now prepared to say that it can amount to something, that it can be a tributary feeding into the common intellectual culture of humanity, but that will require a certain amount of discipline and due diligence on the part of people who do want it to matter that way. There are many things that are working already, for example the division between Discussion and Main (and perhaps the wiki represents an even higher-level distillation). It’s good to have the rambunctious lower level where we are now, as well as the more rarefied and rigorous higher levels. It permits new possibilities to emerge.
Maybe my main message is to serious critics of LW (some of the “trolls” are just critics who aren’t doing it constructively). They can actually contribute to the overall process by being more organized in their criticisms. This is one way that intellectual progress occurs: you have a position, you have an opposite position, and both positions are refined as a result of dialogue. LW, the site and the community, does have the mechanisms and the capacity to take on alternative views and give them a fair hearing, even if they are eventually rejected. Work with that, and we can all benefit.
One more thing I want to point out. It’s often observed that hardly any sequences have been written since Eliezer. In fact, LWer palladias has written about a dozen series of posts on her blog in Sequence format. She recently achieved notoriety for converting from atheism to religion, so her sequences aren’t LW canon, but they represent an interesting example of cross-pollination between very different schools of thought.
Just spotted this thread. The Sequences were indeed the direct inspiration for the format of the linked series of posts I run. Though mine are on a pretty broad range of topics—most recently contrasting Sondheim’s Company with Passion and using both to talk about what the ends of marriage are.
Up voted for this. I can’t believe how many people don’t get it.
He got my downvotes for making terrible arguments defending a change that won’t do what it’s supposed to do, while also doing other shitty things. He was also an overconfident dick about the whole situation. The problem isn’t the rule, it’s the wrong beliefs about how the forums work and how they might be fixed.
That thread is Bayesian evidence against the new poorly thought out rule. The objections that have been raised to it have not even come close to being met. That fact that your own post is a hair breadth away from inflicting negative karma on me should be enough to give you pause.
The reaction to the new rule should not be surprising. If it was surprising, then you should update your model.
Good point about the silliness of people downvoting Eliezer to show their disagreement.
Using the phrase ‘trivial disincentive’ looks like a deliberate reference to this article which would be an unconvincing way to argue that the change won’t cause any problems.
And in general, I don’t think that the change will have really serious side-effects but I’m in favor of changing complex systems in as small increments as possible. The only sensible, currently relevant reason for implementing the new feature (flooding of the recent comments sidebar) that was given can be solved much less invasively by not having comments from crappy threads show up in the recent comments sidebar. For additional soft paternalist goodness, you could also have replies to comments made in such threads not appear in user’s inboxes.
Being able to keep up with all the conversation going on LessWrong seems incompatible with the goal of expanding the community. Reading comments and participating in conversation is a leisure activity. If I were very concerned with being “rational” about my LessWrong usage patterns I would stop reading them at all and stick to just articles (possibly only main section articles if I were really concerned).