It’s not great, but it’s the only really effective model we have.
Yeah, agreed. It’s a much harder problem than it seems like on the surface because it’s by definition happening behind the scenes – when everything’s working, it looks like moderation is not needed.
AI just isn’t there, and opens up a whole new can of worms about the programming of said AI even if it were.
Not sure this is true. I’ve been messing around with it seems surprisingly plausible to get it to work. Of course… how it works in practice is very very different from how it works playing around with it.
I think the issue is that the mods on .world are allowing their biases to affect their willingness to remove or ban something. There’s a lot more leniency for typically far-left viewpoints.
I don’t think this is true. Contrast it with lemmy.ml - the mods there are clearly just removing viewpoints they don’t like. I haven’t seen that on lemmy.world. You sent me examples, but as far as I can tell they are very clearly rule violations being removed, with the viewpoint of the comment not really being relevant.
Like I’ve been talking about “yay Biden” for weeks now, and not a word about removing anything until I started copy-pasting stuff.
Edit: Another example… just out of curiosity I just ran your user through my little moderation-tester, and my instant reaction looking at the first thing that came out was “man… I don’t want to submit that to OpenAI; I’m gonna get in trouble with the content filter.”
You sent me examples, but as far as I can tell they are very clearly rule violations being removed
Of course, but the rules are extremely broad. It allows for selective enforcement. While on .ml the mods ignore any pretext of rules and just blatantly censor, on .world if your comment breaks a rule, it is more likely to be removed if it is pro-Biden than if it is anti-Biden. That’s why I say the .world mods are more subtle about it.
Per rule 3, if you call someone an idiot, your post is in violation. If you call someone a tankie, your post is in violation. But guess which insult is almost always removed, and which insult is comparatively rarely removed?
Yeah, agreed. It’s a much harder problem than it seems like on the surface because it’s by definition happening behind the scenes – when everything’s working, it looks like moderation is not needed.
Not sure this is true. I’ve been messing around with it seems surprisingly plausible to get it to work. Of course… how it works in practice is very very different from how it works playing around with it.
I don’t think this is true. Contrast it with lemmy.ml - the mods there are clearly just removing viewpoints they don’t like. I haven’t seen that on lemmy.world. You sent me examples, but as far as I can tell they are very clearly rule violations being removed, with the viewpoint of the comment not really being relevant.
Like I’ve been talking about “yay Biden” for weeks now, and not a word about removing anything until I started copy-pasting stuff.
Edit: Another example… just out of curiosity I just ran your user through my little moderation-tester, and my instant reaction looking at the first thing that came out was “man… I don’t want to submit that to OpenAI; I’m gonna get in trouble with the content filter.”
Of course, but the rules are extremely broad. It allows for selective enforcement. While on .ml the mods ignore any pretext of rules and just blatantly censor, on .world if your comment breaks a rule, it is more likely to be removed if it is pro-Biden than if it is anti-Biden. That’s why I say the .world mods are more subtle about it.
Per rule 3, if you call someone an idiot, your post is in violation. If you call someone a tankie, your post is in violation. But guess which insult is almost always removed, and which insult is comparatively rarely removed?