Happy to be thinking about such topics, but I don’t think this model of incentives and capabilities of various kinds of posters is all that useful. Real money is a HUGE barrier to many good posters, and a surprisingly small barrier to many bad ones. It’s a disconnected filter—orthogonal to what you care about. It DOES have the advantage of revenue, so you can hire moderators, but I don’t think that comes close to balancing out the downsides.
Speaking as someone who’s used this pseudonym (or variations thereof, as it’s gotten popular and often already taken) since before there was an internet, and has done a lot of work on account and identity systems, I don’t think your focus on real name vs pseudonym is quite as cut-and-dried as you seem to imply here. There are always plenty of legitimate reasons to have multiple separate accounts on a site, and the cost/benefit of any given level of verification and enforcement almost NEVER justifies being very strict.
I don’t know how big a problem it is to have new accounts making personal blogposts or comments (they can’t get onto frontpage without a mod promoting the post, right? Obvious spam gets downvoted to oblivion pretty quickly, low-value content less quickly but still doesn’t get upvoted). Isn’t voting sufficient here? Maybe adjust the default threshold for readers, so unless someone has opted in, they only see +3 and above. This distributes the moderation work among everyone, and if pure spam and puppet-voter-accounts becomes a problem, then further adjustments (like “ignore votes from accounts with less than 50 karma” or whatever) can be made.
Basically, even though I hate how much voting (especially strong votes) gets into the collective consciousness, I do recognize that there’s information there, and it’s way better information than real-money willingness/ability to pay.
Real money is a HUGE barrier to many good posters, and a surprisingly small barrier to many bad ones. It’s a disconnected filter—orthogonal to what you care about. It DOES have the advantage of revenue, so you can hire moderators, but I don’t think that comes close to balancing out the downsides.
Upvoted because this is an interesting point about the possible asymmetrical relationship.
That may be true, but the ‘many good posters’ could just use their real name instead. I imagine the set of folks who both can’t do that, for whatever reason, and are unwilling to spend real money, is a lot smaller.
Basically, even though I hate how much voting (especially strong votes) gets into the collective consciousness, I do recognize that there’s information there, and it’s way better information than real-money willingness/ability to pay.
There’s also the issue that folks will naturally be more skeptical of any voting system, and the results, when they know creating multiple accounts, to vote, is almost effortless.
I personally treat everything within 50 votes of zero as giving approximately the same credibility for nearly everyone, with the exception of online celebrities like Eliezer or gwern which would probably be closer to within 100-200 votes of zero (though I haven’t thought about what the correct proportion should be).
If I knew it was actually onerous to create multiple accounts my confidence would be increased resulting in a more fine differentiation. I imagine many readers would have similar sentiments.
That may be true, but the ‘many good posters’ could just use their real name instead.
We should probably separate that into a different subthread (which I guess this now is). I hadn’t responded to it because I thought your concern was mostly about new user filtering, not use of chosen handle to determine initial standing.
Are you saying “real name”, including human verification into government ID, or are you saying “real-sounding name”, which mods just guess as to validity? Or just self-assed checkbox “this is my real name”? None of them are particularly worthwhile as requirements, IMO.
“real name” with some amount of proof for moderators to review. It could be a link to a blog with a credible amount of archived history/posts, github profile with real code commits, linkedin, etc.. it just has to be accessible and reasonably difficult to duplicate or fake.
The point is that if the moderators do decide to restrict or ban their account, they can feel confident that the user is actually gone (or paid real money to compensate for their time spent, if that’s implemented).
As I say, I have zero idea whether smurfing (multiple accounts or creating new accounts to bypass a ban) is a real problem for LW, nor if it is just how much friction and pain we want to impose on people to make it slightly harder.
Practically, this is a LOT harder than you think. First, simple impersonation—it’s not difficult to imagine creating as many accounts as I like claiming to be any not-already-on-LW social media or github account. Unless some sort of auth scheme or posting protocol (put this code on your GH profile so LW knows it’s you) is in place, there’s no actual verification. Second, many names are common enough that they will conflict, and it’d be really upsetting to find you were banned due to a DIFFERENT John Doe misbehaving.
Perhaps impersonation or name formatting will be a problem in the future if this site becomes even more popular, but they’re not a present day concern.
Whereas many present day concerns could be addressed by some combination of the aforementioned or other proposals.
Happy to be thinking about such topics, but I don’t think this model of incentives and capabilities of various kinds of posters is all that useful. Real money is a HUGE barrier to many good posters, and a surprisingly small barrier to many bad ones. It’s a disconnected filter—orthogonal to what you care about. It DOES have the advantage of revenue, so you can hire moderators, but I don’t think that comes close to balancing out the downsides.
Speaking as someone who’s used this pseudonym (or variations thereof, as it’s gotten popular and often already taken) since before there was an internet, and has done a lot of work on account and identity systems, I don’t think your focus on real name vs pseudonym is quite as cut-and-dried as you seem to imply here. There are always plenty of legitimate reasons to have multiple separate accounts on a site, and the cost/benefit of any given level of verification and enforcement almost NEVER justifies being very strict.
I don’t know how big a problem it is to have new accounts making personal blogposts or comments (they can’t get onto frontpage without a mod promoting the post, right? Obvious spam gets downvoted to oblivion pretty quickly, low-value content less quickly but still doesn’t get upvoted). Isn’t voting sufficient here? Maybe adjust the default threshold for readers, so unless someone has opted in, they only see +3 and above. This distributes the moderation work among everyone, and if pure spam and puppet-voter-accounts becomes a problem, then further adjustments (like “ignore votes from accounts with less than 50 karma” or whatever) can be made.
Basically, even though I hate how much voting (especially strong votes) gets into the collective consciousness, I do recognize that there’s information there, and it’s way better information than real-money willingness/ability to pay.
Upvoted because this is an interesting point about the possible asymmetrical relationship.
That may be true, but the ‘many good posters’ could just use their real name instead. I imagine the set of folks who both can’t do that, for whatever reason, and are unwilling to spend real money, is a lot smaller.
There’s also the issue that folks will naturally be more skeptical of any voting system, and the results, when they know creating multiple accounts, to vote, is almost effortless.
I personally treat everything within 50 votes of zero as giving approximately the same credibility for nearly everyone, with the exception of online celebrities like Eliezer or gwern which would probably be closer to within 100-200 votes of zero (though I haven’t thought about what the correct proportion should be).
If I knew it was actually onerous to create multiple accounts my confidence would be increased resulting in a more fine differentiation. I imagine many readers would have similar sentiments.
We should probably separate that into a different subthread (which I guess this now is). I hadn’t responded to it because I thought your concern was mostly about new user filtering, not use of chosen handle to determine initial standing.
Are you saying “real name”, including human verification into government ID, or are you saying “real-sounding name”, which mods just guess as to validity? Or just self-assed checkbox “this is my real name”? None of them are particularly worthwhile as requirements, IMO.
“real name” with some amount of proof for moderators to review. It could be a link to a blog with a credible amount of archived history/posts, github profile with real code commits, linkedin, etc.. it just has to be accessible and reasonably difficult to duplicate or fake.
The point is that if the moderators do decide to restrict or ban their account, they can feel confident that the user is actually gone (or paid real money to compensate for their time spent, if that’s implemented).
As I say, I have zero idea whether smurfing (multiple accounts or creating new accounts to bypass a ban) is a real problem for LW, nor if it is just how much friction and pain we want to impose on people to make it slightly harder.
Practically, this is a LOT harder than you think. First, simple impersonation—it’s not difficult to imagine creating as many accounts as I like claiming to be any not-already-on-LW social media or github account. Unless some sort of auth scheme or posting protocol (put this code on your GH profile so LW knows it’s you) is in place, there’s no actual verification. Second, many names are common enough that they will conflict, and it’d be really upsetting to find you were banned due to a DIFFERENT John Doe misbehaving.
Not to mention all of https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/ .
Perhaps impersonation or name formatting will be a problem in the future if this site becomes even more popular, but they’re not a present day concern.
Whereas many present day concerns could be addressed by some combination of the aforementioned or other proposals.