Something I have learned from moderating: you can’t just moderate statelessly or with localized state. Moderation is not just about deleting someone’s offensive post or banning a troll. It’s also about setting community norms by performing moderation as a public act. It’s about acknowledging when moderation happens. It’s why sometimes you give someone multiple chances: you hope they get better but also if not at least they’ll dig their own reputation a deeper grave. And sometimes you ban at the first sign of trouble.

It’s quite possible that an AI could learn to do all this but I suspect that a big part of modernism that participants in a community know and respect the moderators. Not many humans respect AI in the same way.

Yes, not to mention the 'behind the scenes' where the moderators discuss disagreements in approach, what actions to take in some situations, need on occasion for PSA's to reiterate the guidelines or explain recent actions, etc.

An algorithm can't ask itself for feedback. I hope Musk does make the Twitter algorithms open source, that would be fascinating and put pressure on other platforms if it turns out to be eye opening or useful

To be fair, isn’t the whole GAN methodology all about asking itself for feedback?

But I think a more refined version of my initial comment is that if the users respect moderators, then moderation will be much more cohesive and well received.

There are no community norms at the scale of Twitter. It's helpful to have community norms for a small forum to discuss model trains or mountain biking or whatever. But Twitter users are a cross section of the whole world. There's nothing to unify them.

Yup, that’s my point. This is why dang doesn’t scale.

Twitter clearly has utility. But it will never be a nice place to hang out except in some very localized parts of it.

The norms are, and should be, what the United States considers norms.

It was built in America, by Americans, and now, is about to be owned by an American. And like America, everyone is (should be) welcome... provided you honor our Western traditions of free speech.

I like what Musk said about free speech within the limits of the law. I think that's a good place to settle things.

Does the US have a uniform set of norms? Last I checked it was pretty divided.

The US has a uniform set of federal laws and Supreme Court opinions, which is what matters here.

Wow that is pretty naive. Let’s see:

A lot of stalking and harassment laws are state level and not the same in every state. Since online harassment is a common problem on Twitter, which state’s laws would you like to apply?

Spreading misinformation that results in violence towards an individual is not governed by federal laws unless it falls under the definition of a hate crime.

Spreading misinformation about results of elections isn’t illegal but is very obviously quite damaging and has real world consequences. How do you tackle this issue?

That’s just a very very small sample. The problem you are going to face with your approach is that the US laws were not designed with the idea of governing stuff on Twitter.

> Spreading misinformation about results of elections isn’t illegal but is very obviously quite damaging and has real world consequences. How do you tackle this issue?

You don't. People are legally allowed to spread misinformation.

> The problem you are going to face with your approach is that the US laws were not designed with the idea of governing stuff on Twitter.

Yes they were... you just don't like the way American laws handle these things.

Difference.