Just a bit more here, after seeing the other comments come in:
My impression is that the OP is being genuine and not trying to be mean spirited. I find the style of this one post a odd and bit off putting (though admitting less so than the original crybaby post), but that’s something we’ll get when encouraging people to write on the forum.
I don’t think it is a good principle that people should trawl through blog posts from 13 years ago, on a different website, looking for something to demand a public apology for. If we start accepting posts like this then this entire forum could end up being nothing but such articles!
At the very top of this post the author says,
“As it’s many years old, I thought that everyone must assume it is no longer representative of GiveWell. However despite its age I’ve seen it quoted even quite recently.”
Unless we think the author is lying, the fact that they’ve seen it quoted (to what seems like a fair amount), seems like a valid reason to me to identify it. It doesn’t sound like the case of someone maliciously searching through their history for all bad posts.
It engages with a blog post written in 2007 in a way which will predictably lead readers to infer that as far as public evidence goes, this is still Holden’s view / style. On the contrary, GiveWell’s and OpenPhil’s styles are now wildly different from this post.
My reading is that the OP wasn’t doing this maliciously. I could easily imagine that the OP wasn’t familiar with the dramatic change in GiveWell and OpenPhil’s styles, and is rather new to the field. Or they thought that the fact that the post was “updated in 2016” (from the post) was evidence that it was still supported.
There’s some nasty subtext in the voting patterns where things are heavily upvoted and downvoted with rather little explanation. My impression is that a large crowd is very afraid of some kinds of discussions, so pounces where there’s a bit of pattern matching. These background discussions are things often the original authors are unaware of.
I remember a long while back I wrote some posts to LessWrong that were decently downvoted without much reasonable explanation and that discouraged me from writing for a long time. Here, Sanjay (from looking at other posts) seems to be the kind who is actively contributing content that’s upvoted, and I’d hate to discourage them unnecessarily.
There’s some nasty subtext in the voting patterns where things are heavily upvoted and downvoted with rather little explanation.
If I see that a post is heavily downvoted, and that several comments that criticize the post are heavily upvoted (as is the case here), I assume that people who downvoted the post generally agree with those criticisms.
In fact, I actively like the “upvoting critical comments” form of explanation if someone thinks that existing comments basically cover what they wanted to say. Otherwise, you get a Twitter-esque pile-on where a dozen people all make very similar critical comments.
(Is there some way you wish people would behave that you think avoids this scenario and the “low-explanation downvote” scenario?)
I may publish a separate comment on this post, but I thought Michelle’s critique was good and upvoted. And I downvoted this post based on my main voting criterion: “Would I want to see more content like this on the EA Forum?”
The answer is: “No, I don’t want to see more content like this on the EA Forum.” I think it generated much more heat than light, and there were many better ways to make the same point. I might not have downvoted the post had it been written by someone who was clearly new to the movement, for the reasons you outlined, but Sanjay isn’t new. On the contrary, he’s written many good posts that I upvoted because I wanted to encourage more content along those lines.
Ideally, downvotes discourage some types of posts and comments that aren’t very useful to the Forum’s goals, and upvotes encourage more posts and comments that are useful. There’s always a risk that someone whose post gets downvoted will be discouraged from writing other posts that could be better, but critical comments seem like they would create the same risk.
I think one area where we may have different feelings is in the bar for publishing to the EA Forum. My intuition was that we should have a very low bar, but have easy sorting and make things easy for readers to skip content. The feeling is something like, “We thank you for writing any content to the forum. Once it’s here, it’s easy enough to remove or hide it if really needed, so content is generally EV-positive.”
I would argue that the original Crybaby’s post may have been worse than this one, but still am fine with it being hosted publicly. If the next Holden were to write a post like that, on this Forum, I’d want them to come away learning how to do better, but not having a very negative experience.
One of the main uses of posting to the Forum is not for readership but for feedback. And some of the worst posts may be exactly those that could benefit the most from feedback.
I know a bunch of people who are reluctant to post to the Forum, and my impression is that we’re losing out a bit here.
If you see the Forum as more of a professional thing, I would hope we could eventually have some other alternative to give feedback to people on their written up thoughts and early blog posts. (Not saying that this is your responsibility, just that I would like to see someone do it).
---
Often when I see posts heavily downvoted / other comments upvoted, it’s because they seem to hit a nerve that a large part of the community deeply cares about, but the comment responses don’t make this clear (it is confusing!). For example, there have been a bunch of emotionally charged threads on transparency vs. censorship. I’m worried that people posting will touch on these issues without realizing it, then take the vote differences to be about them personally.
One of the main uses of posting to the Forum is not for readership but for feedback. And some of the worst posts may be exactly those that could benefit the most from feedback.
It seems to me like this post got a reasonable amount of feedback. The top two upvoted comments took issue with different elements of the post, and I think those commenters explained their points well.
In my experience, heavily-downvoted posts often get a lot of feedback, at least relative to the number of people who vote on them at all. I looked up a bunch of recentposts with negative karma, and they all got comments explaining why people downvoted. Even this post (with only three total votes) at least had someone asking a reasonable question about its conclusion.
Do any counterexamples come to mind? Posts or comments with a lot of downvotes and little-to-no feedback in the form of critical replies?
Often when I see posts heavily downvoted / other comments upvoted, it’s because they seem to hit a nerve that a large part of the community deeply cares about, but the comment responses don’t make this clear (it is confusing!). For example, there have been a bunch of emotionally charged threads on transparency vs. censorship.
Again, I’d be interested to see examples. I’ve written at least twoposts that touch on issues of transparency and/or censorship, and they both got plenty of critical attention that (to me) made it clear what people were concerned about. Other posts on controversial topics also seem to fit this description (when they get more than a couple of votes overall).
If you see the Forum as more of a professional thing, I would hope we could eventually have some other alternative to give feedback to people on their written up thoughts and early blog posts.
I also think that feedback on the Forum tends to be more helpful (on average) than you’d get on almost any other free online platform. My main criticism of the Forum’s commentariat is that they don’t write enough comments (I’d love to see people get more feedback), but I don’t know what alternative platform would be better in that regard.
*****
A question: Do you think the Forum would be a better site, overall, if it had only upvotes and comments, but no downvotes? This would reduce the chance of people getting discouraged by downvotes, but it would also lead to an atmosphere where posts were (by default) ranked by how much attention they received, rather than by how good people thought they were. That seems worse to me.
It seems to me like this post got a reasonable amount of feedback
That was kind of my point. I interpreted you before as saying you wanted to see fewer posts like this, and was pointing out that if we did, those posts wouldn’t get feedback like this. (That said, I think that the feedback on this could have been better, I’m responsible for this too)
Do any counterexamples come to mind?
I think going through this would take a lot of time, especially if you really don’t share this intuition. Happy to discuss in a call. I would note that “lots of critical comments” doesn’t exactly mean that “what’s going on” is obvious. There are many mediocre posts that have bad ideas that get no comments. I would expect that if one had a post that got a ton of bad comments, they would assume it could be because their post was much worse than those that got few comments, or that people dislike them personally.
I also think that feedback on the Forum tends to be more helpful (on average) than you’d get on almost any other free online platform. My main criticism of the Forum’s commentariat is that they don’t write enough comments (I’d love to see people get more feedback), but I don’t know what alternative platform would be better in that regard.
I wouldn’t really disagree, but expect that we can still aim for much better. Many of the forums I encounter (especially ones that border on discussions of Morality) are really really bad.
A question: Do you think the Forum would be a better site, overall, if it had only upvotes and comments, but no downvotes?
I’m not excited about this idea. The upvote/downvote system is crude enough as it is; if anything, I’d like to see more specificity come to it (like, I upvote the reasoning of this post, but not its conclusions). I am excited about things like important community members (myself included) learning how to better handle emotionally sensitive communication, though there are many useless ways of attempting this. Eventually, it could be interesting to have ML bots or similar to help here.
I think I’d also like to recommend having the EA Community Health team or similar jump in on situations like this, and hopefully have a call with Sanjay and the top critics. Social disagreements on sensitive issues are really tough to have in person, let alone on a public forum with everyone in the community judging you.
Just a bit more here, after seeing the other comments come in:
My impression is that the OP is being genuine and not trying to be mean spirited. I find the style of this one post a odd and bit off putting (though admitting less so than the original crybaby post), but that’s something we’ll get when encouraging people to write on the forum.
At the very top of this post the author says,
Unless we think the author is lying, the fact that they’ve seen it quoted (to what seems like a fair amount), seems like a valid reason to me to identify it. It doesn’t sound like the case of someone maliciously searching through their history for all bad posts.
My reading is that the OP wasn’t doing this maliciously. I could easily imagine that the OP wasn’t familiar with the dramatic change in GiveWell and OpenPhil’s styles, and is rather new to the field. Or they thought that the fact that the post was “updated in 2016” (from the post) was evidence that it was still supported.
There’s some nasty subtext in the voting patterns where things are heavily upvoted and downvoted with rather little explanation. My impression is that a large crowd is very afraid of some kinds of discussions, so pounces where there’s a bit of pattern matching. These background discussions are things often the original authors are unaware of.
I remember a long while back I wrote some posts to LessWrong that were decently downvoted without much reasonable explanation and that discouraged me from writing for a long time. Here, Sanjay (from looking at other posts) seems to be the kind who is actively contributing content that’s upvoted, and I’d hate to discourage them unnecessarily.
If I see that a post is heavily downvoted, and that several comments that criticize the post are heavily upvoted (as is the case here), I assume that people who downvoted the post generally agree with those criticisms.
In fact, I actively like the “upvoting critical comments” form of explanation if someone thinks that existing comments basically cover what they wanted to say. Otherwise, you get a Twitter-esque pile-on where a dozen people all make very similar critical comments.
(Is there some way you wish people would behave that you think avoids this scenario and the “low-explanation downvote” scenario?)
I may publish a separate comment on this post, but I thought Michelle’s critique was good and upvoted. And I downvoted this post based on my main voting criterion: “Would I want to see more content like this on the EA Forum?”
The answer is: “No, I don’t want to see more content like this on the EA Forum.” I think it generated much more heat than light, and there were many better ways to make the same point. I might not have downvoted the post had it been written by someone who was clearly new to the movement, for the reasons you outlined, but Sanjay isn’t new. On the contrary, he’s written many good posts that I upvoted because I wanted to encourage more content along those lines.
Ideally, downvotes discourage some types of posts and comments that aren’t very useful to the Forum’s goals, and upvotes encourage more posts and comments that are useful. There’s always a risk that someone whose post gets downvoted will be discouraged from writing other posts that could be better, but critical comments seem like they would create the same risk.
Thanks for the thoughtful response.
I think one area where we may have different feelings is in the bar for publishing to the EA Forum. My intuition was that we should have a very low bar, but have easy sorting and make things easy for readers to skip content. The feeling is something like, “We thank you for writing any content to the forum. Once it’s here, it’s easy enough to remove or hide it if really needed, so content is generally EV-positive.”
I would argue that the original Crybaby’s post may have been worse than this one, but still am fine with it being hosted publicly. If the next Holden were to write a post like that, on this Forum, I’d want them to come away learning how to do better, but not having a very negative experience.
One of the main uses of posting to the Forum is not for readership but for feedback. And some of the worst posts may be exactly those that could benefit the most from feedback.
I know a bunch of people who are reluctant to post to the Forum, and my impression is that we’re losing out a bit here.
If you see the Forum as more of a professional thing, I would hope we could eventually have some other alternative to give feedback to people on their written up thoughts and early blog posts. (Not saying that this is your responsibility, just that I would like to see someone do it).
---
Often when I see posts heavily downvoted / other comments upvoted, it’s because they seem to hit a nerve that a large part of the community deeply cares about, but the comment responses don’t make this clear (it is confusing!). For example, there have been a bunch of emotionally charged threads on transparency vs. censorship. I’m worried that people posting will touch on these issues without realizing it, then take the vote differences to be about them personally.
It seems to me like this post got a reasonable amount of feedback. The top two upvoted comments took issue with different elements of the post, and I think those commenters explained their points well.
In my experience, heavily-downvoted posts often get a lot of feedback, at least relative to the number of people who vote on them at all. I looked up a bunch of recent posts with negative karma, and they all got comments explaining why people downvoted. Even this post (with only three total votes) at least had someone asking a reasonable question about its conclusion.
Do any counterexamples come to mind? Posts or comments with a lot of downvotes and little-to-no feedback in the form of critical replies?
Again, I’d be interested to see examples. I’ve written at least two posts that touch on issues of transparency and/or censorship, and they both got plenty of critical attention that (to me) made it clear what people were concerned about. Other posts on controversial topics also seem to fit this description (when they get more than a couple of votes overall).
I think the Forum should host a really wide range of things written by a wide range of people. I don’t see it as “professional” at all, save in the sense that I want professional people to feel comfortable sharing their work here, same as anyone else who wants to discuss EA.
I also think that feedback on the Forum tends to be more helpful (on average) than you’d get on almost any other free online platform. My main criticism of the Forum’s commentariat is that they don’t write enough comments (I’d love to see people get more feedback), but I don’t know what alternative platform would be better in that regard.
*****
A question: Do you think the Forum would be a better site, overall, if it had only upvotes and comments, but no downvotes? This would reduce the chance of people getting discouraged by downvotes, but it would also lead to an atmosphere where posts were (by default) ranked by how much attention they received, rather than by how good people thought they were. That seems worse to me.
That was kind of my point. I interpreted you before as saying you wanted to see fewer posts like this, and was pointing out that if we did, those posts wouldn’t get feedback like this. (That said, I think that the feedback on this could have been better, I’m responsible for this too)
I think going through this would take a lot of time, especially if you really don’t share this intuition. Happy to discuss in a call. I would note that “lots of critical comments” doesn’t exactly mean that “what’s going on” is obvious. There are many mediocre posts that have bad ideas that get no comments. I would expect that if one had a post that got a ton of bad comments, they would assume it could be because their post was much worse than those that got few comments, or that people dislike them personally.
I wouldn’t really disagree, but expect that we can still aim for much better. Many of the forums I encounter (especially ones that border on discussions of Morality) are really really bad.
I’m not excited about this idea. The upvote/downvote system is crude enough as it is; if anything, I’d like to see more specificity come to it (like, I upvote the reasoning of this post, but not its conclusions). I am excited about things like important community members (myself included) learning how to better handle emotionally sensitive communication, though there are many useless ways of attempting this. Eventually, it could be interesting to have ML bots or similar to help here.
I think I’d also like to recommend having the EA Community Health team or similar jump in on situations like this, and hopefully have a call with Sanjay and the top critics. Social disagreements on sensitive issues are really tough to have in person, let alone on a public forum with everyone in the community judging you.