Here's why I don't think this is a good idea.

First off, no rules in a centralized system can survive corrupt admins/moderators. At best, the rules can make it difficult for the admins/mods to hide their malfeisance. If we don't assume good faith from the admins, this discussion is pointless because we should just leave this instance.

Second, upvotes and downvotes already moderate discussion. The default comment sorting algorithm prioritizes upvoted comments and hides downvoted comments, and people do tend to treat downvoted comments negatively. Popularity already matters, it's just a matter to what extent each thread gets you a fresh start.

Right now, slrpnk account generation is gatekept by the mods. You have to pass a Turing test to be let in. This makes it difficult to amass a sufficient army of bots without mod assistance. It's worth looking out for, but not expected by any means.
Agreement and dislike are different things. Empirically, people can become more hardened in their opinions if they see crappy disagreement - that's why organizations like FOX NEWS show a constant cavalcade of liberals and leftists being stupid. As long as people upvote well-formulated disagreement, this could actually improve discussion because it filters out the comments that would never have convinced anyone anyway. That's a big "as long as", so it's worth seeing in practice whether or not it holds.
Lemmy instances have admins and moderators with absolute unaccountable power over bannings. It has never been decentralized or pro-free speech in the ways santabot might have destroyed in a more fundamentally anarchic social media. If you want to make use of Lemmy's decentralization, make your own instance and see who wants to let you crosspost. If you want more, make your own social media platform that is (more) fully decentralized.
Yes. Bad actors gonna act bad. Stay away from places that give them authority.
Not very well. You're leaving it up to the whims of the voting public. It would be easier to write a bot that asks ChatGPT whether a user holds certain opinions and ban them if it says yes. Or deputize more (informal) mods to ban people based on their personal opinion.

It is natural that an object can be used for bad in more ways than it can be used for good. 'Good' is a fragile concept, while 'bad' is everything else. A kitchen knife can be used for bad more easily and in more different ways than it can be used for good. So can a brick or a water bottle. The question is whether its use here pumps towards good, both now and in the future.

I understand expecting this experiment to go poorly, but I think it's excessive to say the experiment should not be run at all.

1 comments

First off, no rules in a centralized system can survive corrupt admins/moderators. At best, the rules can make it difficult for the admins/mods to hide their malfeisance. If we don't assume good faith from the admins, this discussion is pointless because we should just leave this instance.
Second, upvotes and downvotes already moderate discussion. The default comment sorting algorithm prioritizes upvoted comments and hides downvoted comments, and people do tend to treat downvoted comments negatively. Popularity already matters, it's just a matter to what extent each thread gets you a fresh start.
Right now, slrpnk account generation is gatekept by the mods. You have to pass a Turing test to be let in. This makes it difficult to amass a sufficient army of bots without mod assistance. It's worth looking out for, but not expected by any means.
Agreement and dislike are different things. Empirically, people can become more hardened in their opinions if they see crappy disagreement - that's why organizations like FOX NEWS show a constant cavalcade of liberals and leftists being stupid. As long as people upvote well-formulated disagreement, this could actually improve discussion because it filters out the comments that would never have convinced anyone anyway. That's a big "as long as", so it's worth seeing in practice whether or not it holds.
Lemmy instances have admins and moderators with absolute unaccountable power over bannings. It has never been decentralized or pro-free speech in the ways santabot might have destroyed in a more fundamentally anarchic social media. If you want to make use of Lemmy's decentralization, make your own instance and see who wants to let you crosspost. If you want more, make your own social media platform that is (more) fully decentralized.
Yes. Bad actors gonna act bad. Stay away from places that give them authority.
Not very well. You're leaving it up to the whims of the voting public. It would be easier to write a bot that asks ChatGPT whether a user holds certain opinions and ban them if it says yes. Or deputize more (informal) mods to ban people based on their personal opinion.
It is natural that an object can be used for bad in more ways than it can be used for good. 'Good' is a fragile concept, while 'bad' is everything else. A kitchen knife can be used for bad more easily and in more different ways than it can be used for good. So can a brick or a water bottle. The question is whether its use here pumps towards good, both now and in the future.
I understand expecting this experiment to go poorly, but I think it's excessive to say the experiment should not be run at all.