I think it is much better when the LessWrong userbase knows more about how site moderation happens, i.e. who does it, what the tools are, what actions and decisions are, who’s responsible for what, how they think about things, etc. While being careful to say that LessWrong is not a democracy and we will not care equally about the judgments of everyone on the site just because they’re an active member[1], I think transparency is valuable here for at least these overlapping reasons:
It means we can be held accountable by people either calling out decisions or policies they think are bad, or leaving because of them (vote with your feet). I really value getting feedback.
By letting ourselves be held accountable by people we wish to be accountable to, we set up good incentives and allow us to get corrective feedback.
It builds trust and confidence (assuming people like what we’re doing) that the site is somewhere worth investing your time and attention.
LessWrong team members mostly speak for themselves
I think it’s important that LessWrong team members don’t have to pretend to all agree with each other some aggregate official team belief. We each have our models which while being pretty correlated are our own, and it’s good when we just speak from them. See my post on this topic for more elaboration.
This approach is a little tricky in the context of moderation since people really need a clear sense of the policies and rules and that’s hard if different moderators say different and conflicting things. I’m not sure how to best handle this, but one idea is that moderators are always clear about what’s their judgment vs from what they think the policy we plan to endorse is. The team is currently three people who are very in-sync and work very closely, so for any bigger mod decision, we’ll have checked to clarify underlying policies. Or so I hope.
Who moderates, who’s responsible?
The LessWrong team is part of Lightcone Infrastructure. The core LessWrong team is Ruby (me, team lead), Raemon, and RobertM[2]. Habryka is head of Lightcone and also responsible for reviving LW/building LW2.0, and was for a long-time the chief moderator. He isn’t the day-to-day moderator any more, but we regularly consult him and he’ll poke us about things he thinks are an issue. Other Lightcone team members will often weigh in moderation or take some actions sometimes[3]
I (and Lightcone generally) have found generally that many collaborative situations go much better if a single person owns each decision, other people can weigh in, but ultimately that person decides and gets to be held responsible. (You can also hold people responsible for who they delegate ownership to.) This is how we aspire to operate on the LessWrong team: for any decision, we can tell you who was responsible for making it.
In that vein:
Generally, I am the final decision-maker for LessWrong matters unless I have delegated. (Habryka could fire me or in very exceptional circumstances overrule me, but that’d be surprising for that to happen.) However, there are two general large delegations that I’ve made, one very relevant for moderation. RobertM is CTO and has ownership of all technical aspects of the codebase (architecture, standards, etc). Raemon is Head of New User and <something something corrective/problem user moderation> (we haven’t figured out what exactly to call it it) moderation, i.e. Ray gets to decide which new users are welcome on the site, how they get onboarded, etc., Ray is also in charge of judgments about what we do when users seem to violating norms or makes the site worse, etc., i.e. bans, warnings, and other matters. However, decisions about overall site policies, values, norms, etc remain with me.
The Duncan and Said situation plausibly requires some kind of corrective action of one or more people’s behavior, and therefore it is Ray’s final call what happens with those users. If you want a certain decision made (e.g. disciplinary action), you should focus on addressing his cruxes, etc. To the extent there’s a broader site policy question (e.g. which behaviors are ok or not in general), that’s in my court. I care a lot of the moderation judgment of others, particular Raemon and Habryka who shape my own thinking a lot, but if you want a certain site policy, know that my cruxes are key (but if you can persuade Raemon or Habryka, good chance I’ll be convinced too.)
I (and I am pretty sure others) care a great deal about the opinions and feelings of: 1) users who we think share the core values of the site as we see them and who have good judgment about things, 2) the users who we think contribute most to LessWrong’s goals of intellectual progress, etc. It feels less important to me to appease users I’m more ambivalent about being on the site.
Some meta notes about moderation process
Preamble: I like transparency
I think it is much better when the LessWrong userbase knows more about how site moderation happens, i.e. who does it, what the tools are, what actions and decisions are, who’s responsible for what, how they think about things, etc. While being careful to say that LessWrong is not a democracy and we will not care equally about the judgments of everyone on the site just because they’re an active member[1], I think transparency is valuable here for at least these overlapping reasons:
It means we can be held accountable by people either calling out decisions or policies they think are bad, or leaving because of them (vote with your feet). I really value getting feedback.
By letting ourselves be held accountable by people we wish to be accountable to, we set up good incentives and allow us to get corrective feedback.
It builds trust and confidence (assuming people like what we’re doing) that the site is somewhere worth investing your time and attention.
LessWrong team members mostly speak for themselves
I think it’s important that LessWrong team members don’t have to pretend to all agree with each other some aggregate official team belief. We each have our models which while being pretty correlated are our own, and it’s good when we just speak from them. See my post on this topic for more elaboration.
This approach is a little tricky in the context of moderation since people really need a clear sense of the policies and rules and that’s hard if different moderators say different and conflicting things. I’m not sure how to best handle this, but one idea is that moderators are always clear about what’s their judgment vs from what they think the policy we plan to endorse is. The team is currently three people who are very in-sync and work very closely, so for any bigger mod decision, we’ll have checked to clarify underlying policies. Or so I hope.
Who moderates, who’s responsible?
The LessWrong team is part of Lightcone Infrastructure. The core LessWrong team is Ruby (me, team lead), Raemon, and RobertM[2]. Habryka is head of Lightcone and also responsible for reviving LW/building LW2.0, and was for a long-time the chief moderator. He isn’t the day-to-day moderator any more, but we regularly consult him and he’ll poke us about things he thinks are an issue. Other Lightcone team members will often weigh in moderation or take some actions sometimes[3]
I (and Lightcone generally) have found generally that many collaborative situations go much better if a single person owns each decision, other people can weigh in, but ultimately that person decides and gets to be held responsible. (You can also hold people responsible for who they delegate ownership to.) This is how we aspire to operate on the LessWrong team: for any decision, we can tell you who was responsible for making it.
In that vein:
Generally, I am the final decision-maker for LessWrong matters unless I have delegated. (Habryka could fire me or in very exceptional circumstances overrule me, but that’d be surprising for that to happen.) However, there are two general large delegations that I’ve made, one very relevant for moderation. RobertM is CTO and has ownership of all technical aspects of the codebase (architecture, standards, etc). Raemon is Head of New User and <something something corrective/problem user moderation> (we haven’t figured out what exactly to call it it) moderation, i.e. Ray gets to decide which new users are welcome on the site, how they get onboarded, etc., Ray is also in charge of judgments about what we do when users seem to violating norms or makes the site worse, etc., i.e. bans, warnings, and other matters. However, decisions about overall site policies, values, norms, etc remain with me.
The Duncan and Said situation plausibly requires some kind of corrective action of one or more people’s behavior, and therefore it is Ray’s final call what happens with those users. If you want a certain decision made (e.g. disciplinary action), you should focus on addressing his cruxes, etc. To the extent there’s a broader site policy question (e.g. which behaviors are ok or not in general), that’s in my court. I care a lot of the moderation judgment of others, particular Raemon and Habryka who shape my own thinking a lot, but if you want a certain site policy, know that my cruxes are key (but if you can persuade Raemon or Habryka, good chance I’ll be convinced too.)
I (and I am pretty sure others) care a great deal about the opinions and feelings of: 1) users who we think share the core values of the site as we see them and who have good judgment about things, 2) the users who we think contribute most to LessWrong’s goals of intellectual progress, etc. It feels less important to me to appease users I’m more ambivalent about being on the site.
jimrandomh also works heavily on the LessWrong site, that not as part of the core team.
Less so at the moment as policies are unclear since LW Team is adjusting moderation policy
Very informative (and thanks for your efforts)! Who owns the domain name lesswrong.com?