Here are several of the users I had in mind and what they’re up to now:
MichaelVassar—co-founder and CSO at MetaMed
lukeprog—now Executive Director at MIRI
Yvain—recently began medical residency
AnnaSalamon—co-founder and Executive Director at CFAR
Liron—co-founder and CTO at Quixey
Several of these users—I believe all of them, in fact—still post and comment from time to time, but less frequently than they once did. Yvain still writes his own blog, but he did that even while he was still posting on LessWrong.
This isn’t to say, of course, that the only factor preventing advanced users from posting a lot on LessWrong is that they have more important things to do. But it is a significant element, and IMO one that it’s important to be aware of.
I agree it’s an important element, and for some reason I didn’t entertain the thought before you made it available. Do you agree that some people don’t post here anymore because of the overly critical environment, and that we might have lost important contributors that way?
Do you agree that some people don’t post here anymore because of the overly critical environment, and that we might have lost important contributors that way?
It’s true that criticism deters new users much more easily than established users, but I think losing established users is orders of magnitude worse than losing new users.
That old familiar post you linked discusses karma, and I think karma has evolved to be something very different from what it was intended to be. Almost nobody has a happy trigger like that. You can’t simply dictate what kinds of signals voting is supposed to send because it will acquire new meanings by usage, and stubbornly going against the grain is going to send unintended signals to people.
There’s no shortage of internet fora which lack LessWrong’s highly critical environment. They also have much less intelligent discussion. I think there is a connection between these two.
The connection is obvious. Now that I’ve thought about this some more, maybe some good people leaving is an unavoidable side effect. This doesn’t mean you shouldn’t be tactful with the criticism, unless you want certain people to leave and not change their minds, which might sometimes be an understandable goal too.
I’ll point out that Yvain has explicitly said (in the thread I linked above) that the reason he no longer blogs much at LessWrong is because the standards at LessWrong keep it from being fun. And when he talked about those standards, the things he’s talking about are things I don’t think everyone agrees on—just things some people are vocal about (see e.g. Kaj’s reply to him).
Because of that, I seriously think that you are harming the LessWrong community.
And having said that, I’m done replying to you in this thread, because it’s far from clear to me that very many people agree with you.
it’s far from clear to me that very many people agree with you.
I agree with you that LW is sometimes too harsh, and that encouraging good content should be emphasized, and discouraging bad content perhaps de-emphasized. But I agree with katydee that this particular sequence seems inappropriate to all be in main. As written, the series seems to be about nutrition, with one (possibly two) rationality points as subtext. I get that impression primarily because each post is about a fairly short nutrition point- of the two meat posts so far, one has been about Atkins, and another about sugar.
If, instead of an n-post series arguing against Taubes on the object-level of nutrition, you had written one (potentially long!) post on those rationality points, I would be happy.
The primary rationality point seems to be “don’t choose a side in a controversy after only listening to one side.” This seems like very good advice, and Taubes seems like a good example: I haven’t read him directly, but I get the impression that both what he’s arguing against is wrong and Taubes’s proposed replacement is also probably wrong. If I were writing this post, I would write it with the halo/horns effect and meta-contrarianism in mind- rather than defending mainstream nutrition against Taubes, I would emphasize the underlying uncertainty, that low-carb diets seem to work but Taubes is probably wrong on why, and specific underhanded tricks Taubes does (like failing to point out that the FDA recommends against both sugar and fat).
But the way I would handle the first point conflicts with the second point, which is ‘trust the expert consensus.’ There was a great blog post by a physics graduate student I saw ~5 years ago, which I can’t find now, where he wrote up in a short post why he was convinced that dark matter was the best explanation for the observed data. He discussed three alternatives (like MOND), how each of them explained a portion of the data better than dark matter, but that when you considered all of the data together, dark matter was the clear winner because it did okay on three issues instead of great on one issue and terribly on two others. The link to “don’t affiliate early” is clear- a single proponent can be very convincing (and be right!) about their belief’s strong points and the weak points of other sides, but generally the experts are familiar with the weak points of all the sides, and so if they discount something it’s probably for good reason.
I, too, thought that these posts were rather short and that it would have been better as one post. But they’ve generated 100 comments per post. People like talking about nutrition. Splitting it into multiple may have better organized the discussion. Or maybe it created lots of duplication. Actually, since I think the discussion has been unproductive, actions that impeded it, like stuffing it all into one post, might have been better for everyone.
Sometimes I feel LW is very harsh for Main articles, but not harsh enough for Discussion articles. It is very difficult to write a Main article, but many kinds of trivialities get posted in Discussion.
It’s like if you measure quality from 0 to 10, then a typical Main article is 9 or 10, a typical Discussion article is 1 or 2 (you must get to 0 to get downvoted)… and the articles between 4 and 7 somehow don’t belong anywhere.
And maybe it is this category of articles—not good enough for Main, but already too good for Discussion—that people prefer to take to their own blogs.
One of the proposed splits for the site, to replace the current Main/Discussion split, is by subject matter- instrumental rationality, epistemic rationality, meetups, futurism, and so on. Everything related to instrumental rationality, from a one-sentence post with a link to a six thousand word detailed referenced article, would go in that subreddit, and there would probably be a ‘high quality’ page where you can see all of the articles that have been promoted from any of the subreddits.
It seems like that would be good at encouraging posts of medium quality- you don’t have to say “I think this is Main-quality,” you just post it where it belongs and either the editors think it’s Main-quality or they don’t, and someone sitting on a 7 post just posts it instead of agonizing over it (and eventually not posting it because of an ugh field).
If, instead of an n-post series arguing against Taubes on the object-level of nutrition, you had written one (potentially long!) post on those rationality points, I would be happy.
In retrospect, I think I may have made a mistake breaking the series up as finely as I did. However, the idea that taking a post that would be Main-suitable, and breaking it up into pieces of no less than 750 words would turn those posts into “discussion” posts strikes me as really odd.
However, the idea that taking a post that would be Main-suitable, and breaking it up into pieces of no less than 750 words would turn theses posts into “discussion” posts strikes me as really odd.
I think 750 words is pretty short, and that may be the main issue here. (I’m having a hard time teasing apart the strengths of the various reasons I think this.) My Value of Information: Four Examples was about 3k words, and the idea of splitting it up into four separate posts to Main seems odd to me. (That post is one of five (six if you count the table of contents) in a sequence which came out to something like 11k words total.) As a collection of discussion posts, sure, especially if I was posting an example whenever I came across one.
Consider Intellectual Hipsters and Meta-Contrarianism, since I linked to it a upthread. It’s about 2200 words- it seems like a good length for a post, and in particular the right length for that particular concept. It’s also got three sections- an introduction, Pretending To Be Wise, and Meta-Contrarians are Intellectual Hipsters. Splitting the post into three subposts seems like it needlessly disrupts the flow and makes the concept harder to understand and discuss- without the examples at the end, the discussion of the first section might be confused, and in the comments you’ll see various people propose other triads, which are good to have all in one place.
But even if the combination is better, what should we make of a single section? Pretending To Be Wise is an about half-length presentation of another main article, and so might make for a decent main article on its own. But the last part seems more like an “here are some examples, discuss” which is suitable for discussion because it’s missing the theory that makes it a compelling main post- which is in the first and second sections.
[Edit] I should also make clear that I don’t think it’s that odd to have a sequence that moves between Main and Discussion as appropriate- but I don’t think there are many (or possibly any) examples of that yet, and so it may be odder than I think of it as being.
I should also make clear that I don’t think it’s that odd to have a sequence that moves between Main and Discussion as appropriate- but I don’t think there are many (or possibly any) examples of that yet, and so it may be odder than I think of it as being.
I can see the point of having the main parts of the sequence in Main and appendices in Discussion, but having Part 1 in Main, Part 2 in Discussion, Part 3 in Discussion and Part 4 in Main (named that way) seems confusing to me.
I learned long ago—I think doing high-school journalism—that 250-300 words is the limit for a letter to the editor, ~750 words is a good length for a typical article or op-ed, and 3k words is a feature article. In college and later grad school, I learned 4.5k as a typical term paper length, and I made the chapters in the two books I’ve written around 6k words. Obviously, academic papers and academic book chapters can be much longer than 6k words.
I’ve been approaching this from the point of view of “blog posts should generally be like typical newspaper articles or feature articles; a term paper or a book chapter as a blog post is usually too long.” But maybe I should think of feature article / term paper as the standard?
I think that the feature article and above (though empirically the standard here seems to be about 2k, rather than 3k) as the target for Main, and the op-ed and below for discussion, is a good split. I think that shorter articles can be worth it for Main, but the quality and relevance bars are higher (and they should be standalone).
And having said that, I’m done replying to you in this thread, because it’s far from clear to me that very many people agree with you.
My comment saying that this post would be more appropriate in Discussion has more karma than the actual post itself. That seems like fairly clear evidence to me.
Here are several of the users I had in mind and what they’re up to now:
MichaelVassar—co-founder and CSO at MetaMed
lukeprog—now Executive Director at MIRI
Yvain—recently began medical residency
AnnaSalamon—co-founder and Executive Director at CFAR
Liron—co-founder and CTO at Quixey
Several of these users—I believe all of them, in fact—still post and comment from time to time, but less frequently than they once did. Yvain still writes his own blog, but he did that even while he was still posting on LessWrong.
This isn’t to say, of course, that the only factor preventing advanced users from posting a lot on LessWrong is that they have more important things to do. But it is a significant element, and IMO one that it’s important to be aware of.
I agree it’s an important element, and for some reason I didn’t entertain the thought before you made it available. Do you agree that some people don’t post here anymore because of the overly critical environment, and that we might have lost important contributors that way?
I agree. That said, established users generally have high status in the community, which can help mitigate this effect. I think that LW being overly critical is more of a problem for newer users, and even then it’s important to consider that one man’s “overly critical enviroment” might well be another’s “high standards of rigor.”
It’s true that criticism deters new users much more easily than established users, but I think losing established users is orders of magnitude worse than losing new users.
That old familiar post you linked discusses karma, and I think karma has evolved to be something very different from what it was intended to be. Almost nobody has a happy trigger like that. You can’t simply dictate what kinds of signals voting is supposed to send because it will acquire new meanings by usage, and stubbornly going against the grain is going to send unintended signals to people.
There’s no shortage of internet fora which lack LessWrong’s highly critical environment. They also have much less intelligent discussion. I think there is a connection between these two.
The connection is obvious. Now that I’ve thought about this some more, maybe some good people leaving is an unavoidable side effect. This doesn’t mean you shouldn’t be tactful with the criticism, unless you want certain people to leave and not change their minds, which might sometimes be an understandable goal too.
I’ll point out that Yvain has explicitly said (in the thread I linked above) that the reason he no longer blogs much at LessWrong is because the standards at LessWrong keep it from being fun. And when he talked about those standards, the things he’s talking about are things I don’t think everyone agrees on—just things some people are vocal about (see e.g. Kaj’s reply to him).
Because of that, I seriously think that you are harming the LessWrong community.
And having said that, I’m done replying to you in this thread, because it’s far from clear to me that very many people agree with you.
I agree with you that LW is sometimes too harsh, and that encouraging good content should be emphasized, and discouraging bad content perhaps de-emphasized. But I agree with katydee that this particular sequence seems inappropriate to all be in main. As written, the series seems to be about nutrition, with one (possibly two) rationality points as subtext. I get that impression primarily because each post is about a fairly short nutrition point- of the two meat posts so far, one has been about Atkins, and another about sugar.
If, instead of an n-post series arguing against Taubes on the object-level of nutrition, you had written one (potentially long!) post on those rationality points, I would be happy.
The primary rationality point seems to be “don’t choose a side in a controversy after only listening to one side.” This seems like very good advice, and Taubes seems like a good example: I haven’t read him directly, but I get the impression that both what he’s arguing against is wrong and Taubes’s proposed replacement is also probably wrong. If I were writing this post, I would write it with the halo/horns effect and meta-contrarianism in mind- rather than defending mainstream nutrition against Taubes, I would emphasize the underlying uncertainty, that low-carb diets seem to work but Taubes is probably wrong on why, and specific underhanded tricks Taubes does (like failing to point out that the FDA recommends against both sugar and fat).
But the way I would handle the first point conflicts with the second point, which is ‘trust the expert consensus.’ There was a great blog post by a physics graduate student I saw ~5 years ago, which I can’t find now, where he wrote up in a short post why he was convinced that dark matter was the best explanation for the observed data. He discussed three alternatives (like MOND), how each of them explained a portion of the data better than dark matter, but that when you considered all of the data together, dark matter was the clear winner because it did okay on three issues instead of great on one issue and terribly on two others. The link to “don’t affiliate early” is clear- a single proponent can be very convincing (and be right!) about their belief’s strong points and the weak points of other sides, but generally the experts are familiar with the weak points of all the sides, and so if they discount something it’s probably for good reason.
I, too, thought that these posts were rather short and that it would have been better as one post. But they’ve generated 100 comments per post. People like talking about nutrition. Splitting it into multiple may have better organized the discussion. Or maybe it created lots of duplication. Actually, since I think the discussion has been unproductive, actions that impeded it, like stuffing it all into one post, might have been better for everyone.
Sometimes I feel LW is very harsh for Main articles, but not harsh enough for Discussion articles. It is very difficult to write a Main article, but many kinds of trivialities get posted in Discussion.
It’s like if you measure quality from 0 to 10, then a typical Main article is 9 or 10, a typical Discussion article is 1 or 2 (you must get to 0 to get downvoted)… and the articles between 4 and 7 somehow don’t belong anywhere.
And maybe it is this category of articles—not good enough for Main, but already too good for Discussion—that people prefer to take to their own blogs.
One of the proposed splits for the site, to replace the current Main/Discussion split, is by subject matter- instrumental rationality, epistemic rationality, meetups, futurism, and so on. Everything related to instrumental rationality, from a one-sentence post with a link to a six thousand word detailed referenced article, would go in that subreddit, and there would probably be a ‘high quality’ page where you can see all of the articles that have been promoted from any of the subreddits.
It seems like that would be good at encouraging posts of medium quality- you don’t have to say “I think this is Main-quality,” you just post it where it belongs and either the editors think it’s Main-quality or they don’t, and someone sitting on a 7 post just posts it instead of agonizing over it (and eventually not posting it because of an ugh field).
In retrospect, I think I may have made a mistake breaking the series up as finely as I did. However, the idea that taking a post that would be Main-suitable, and breaking it up into pieces of no less than 750 words would turn those posts into “discussion” posts strikes me as really odd.
I think 750 words is pretty short, and that may be the main issue here. (I’m having a hard time teasing apart the strengths of the various reasons I think this.) My Value of Information: Four Examples was about 3k words, and the idea of splitting it up into four separate posts to Main seems odd to me. (That post is one of five (six if you count the table of contents) in a sequence which came out to something like 11k words total.) As a collection of discussion posts, sure, especially if I was posting an example whenever I came across one.
Consider Intellectual Hipsters and Meta-Contrarianism, since I linked to it a upthread. It’s about 2200 words- it seems like a good length for a post, and in particular the right length for that particular concept. It’s also got three sections- an introduction, Pretending To Be Wise, and Meta-Contrarians are Intellectual Hipsters. Splitting the post into three subposts seems like it needlessly disrupts the flow and makes the concept harder to understand and discuss- without the examples at the end, the discussion of the first section might be confused, and in the comments you’ll see various people propose other triads, which are good to have all in one place.
But even if the combination is better, what should we make of a single section? Pretending To Be Wise is an about half-length presentation of another main article, and so might make for a decent main article on its own. But the last part seems more like an “here are some examples, discuss” which is suitable for discussion because it’s missing the theory that makes it a compelling main post- which is in the first and second sections.
[Edit] I should also make clear that I don’t think it’s that odd to have a sequence that moves between Main and Discussion as appropriate- but I don’t think there are many (or possibly any) examples of that yet, and so it may be odder than I think of it as being.
I can see the point of having the main parts of the sequence in Main and appendices in Discussion, but having Part 1 in Main, Part 2 in Discussion, Part 3 in Discussion and Part 4 in Main (named that way) seems confusing to me.
This is useful.
I learned long ago—I think doing high-school journalism—that 250-300 words is the limit for a letter to the editor, ~750 words is a good length for a typical article or op-ed, and 3k words is a feature article. In college and later grad school, I learned 4.5k as a typical term paper length, and I made the chapters in the two books I’ve written around 6k words. Obviously, academic papers and academic book chapters can be much longer than 6k words.
I’ve been approaching this from the point of view of “blog posts should generally be like typical newspaper articles or feature articles; a term paper or a book chapter as a blog post is usually too long.” But maybe I should think of feature article / term paper as the standard?
I think that the feature article and above (though empirically the standard here seems to be about 2k, rather than 3k) as the target for Main, and the op-ed and below for discussion, is a good split. I think that shorter articles can be worth it for Main, but the quality and relevance bars are higher (and they should be standalone).
Thanks. I will follow this rule in the future.
My comment saying that this post would be more appropriate in Discussion has more karma than the actual post itself. That seems like fairly clear evidence to me.