I don’t recall this being discussed by the community at all. It seems like a bad idea. Valuable conversations can extend from comments that are already negative. −3 is also not that negative. This also discourages people from actually explaining why someone is wrong if there are a lot of people who downvote the comment. This will both make it harder for that person to become less wrong and make it more likely that bystanders who are reading the conversation will not see any explanation for why the comment is downvoted. Overall this is at best a mixed idea that should have been discussed with the community before implementing.
I don’t recall this being discussed by the community at all.
This isn’t terribly relevant. Moderators that discuss every decision with the community and only act when they’ve built consensus fail prey to vocal minorities, e.g., Wikipedia. Then they tend to stagnate.
Instead of trying to build a consensus, Eliezer could have asked the community “Here are the consequences I intend/foresee with this change. Are there any important ones I may have overlooked?”, which has no obvious downsides that I can see, other than the opportunity cost of writing out the question.
Instead of trying to build a consensus, Eliezer could have asked the community “Here are the consequences I intend/foresee with this change. Are there any important ones I may have overlooked?”, which has no obvious downsides that I can see, other than the opportunity cost of writing out the question.
The obvious downside in such cases is that asking for and being told what the downsides are and then ignoring them is often perceived as even worse than not asking at all. If Eliezer anticipated that he would go ahead with his change regardless of what downsides are pointed out then it could be detrimental to ask.
(Note: That is the downside of asking, not a claim that asking would be a net negative.)
The obvious downside in such cases is that asking for and being told what the downsides are and then ignoring them is often perceived as even worse than not asking at all.
Do you think this is true even if one made it clear that one is not seeking a consensus but reserving the right to make the final cost/benefit judgement? If so, it’s contrary to my expectations (i.e., I don’t see why that would be perceived as being worse than not asking at all), and I would appreciate any further explanations you might have.
Moderators that discuss every decision with the community and only act when they’ve built consensus fail prey to vocal minorities, e.g., Wikipedia. Then they tend to stagnate.
Yes, discussing every decision with the community is probably a bad idea. But that doesn’t mean that specific, large scale changes shouldn’t be discussed.
Because the community has additional experience and may have thoughts about a proposal. The impression one gets when moderating something can be very different from the impression one gets in the general case. Discussing such issues in advance helps prevent severe unintended consequences from occurring.
Valuable conversations can extend from comments that are already negative.
If a reply to a downvoted comment is not downvoted, replies to that reply are not punished, so good subthreads are unaffected.
-3 is also not that negative.
5 Karma points is not that much as well, so if it’s really worth replying to, it’s possible to continue the conversation. It’s usually not worth replying though, and when not feeling any cost people would ignore that consideration, giving fuel to bad conversations as a result. The motivation I agree with is to stop bad conversations, not necessarily replies to individual bad comments, which is just the means.
I don’t recall this being discussed by the community at all. It seems like a bad idea. Valuable conversations can extend from comments that are already negative. −3 is also not that negative. This also discourages people from actually explaining why someone is wrong if there are a lot of people who downvote the comment. This will both make it harder for that person to become less wrong and make it more likely that bystanders who are reading the conversation will not see any explanation for why the comment is downvoted. Overall this is at best a mixed idea that should have been discussed with the community before implementing.
This isn’t terribly relevant. Moderators that discuss every decision with the community and only act when they’ve built consensus fail prey to vocal minorities, e.g., Wikipedia. Then they tend to stagnate.
Instead of trying to build a consensus, Eliezer could have asked the community “Here are the consequences I intend/foresee with this change. Are there any important ones I may have overlooked?”, which has no obvious downsides that I can see, other than the opportunity cost of writing out the question.
The obvious downside in such cases is that asking for and being told what the downsides are and then ignoring them is often perceived as even worse than not asking at all. If Eliezer anticipated that he would go ahead with his change regardless of what downsides are pointed out then it could be detrimental to ask.
(Note: That is the downside of asking, not a claim that asking would be a net negative.)
Do you think this is true even if one made it clear that one is not seeking a consensus but reserving the right to make the final cost/benefit judgement? If so, it’s contrary to my expectations (i.e., I don’t see why that would be perceived as being worse than not asking at all), and I would appreciate any further explanations you might have.
Yes, discussing every decision with the community is probably a bad idea. But that doesn’t mean that specific, large scale changes shouldn’t be discussed.
Very well, then: why should specific, large scale changes be discussed?
I’m intentionally ignoring the implication that this specific change was a “large scale” one.
Because a community is made up of its users, and if people find the changes negative enough, they will stop using the site.
Because the community has additional experience and may have thoughts about a proposal. The impression one gets when moderating something can be very different from the impression one gets in the general case. Discussing such issues in advance helps prevent severe unintended consequences from occurring.
In short, you’re hoping for the positive part of WWIC, while hoping the negative half doesn’t happen.
See references therein for applications to social websites.
If a reply to a downvoted comment is not downvoted, replies to that reply are not punished, so good subthreads are unaffected.
5 Karma points is not that much as well, so if it’s really worth replying to, it’s possible to continue the conversation. It’s usually not worth replying though, and when not feeling any cost people would ignore that consideration, giving fuel to bad conversations as a result. The motivation I agree with is to stop bad conversations, not necessarily replies to individual bad comments, which is just the means.