And what is the mechanism by which society could stop such AI damage and push for punishment as per law, not per subjective usefulness? Voting every 4 years to a president that never fulfills his promises?

No, society does not want corporations with no responsibilities.

> No, society does not want corporations with no responsibilities.

It voted otherwise.

Because Eurosky can ban you.

Dang can ban you, or me, or anyone. Doesn't make HN a bad forum. On the contrary, HN's moderation quality is widely praised and appreciated. Now imagine if it was free for all anarcho-libertarian forum. Or as another example a mastodon, where decentralization is touted as a huge benefit, but in practice it is a collection of extremely strict centralized islands, where a ban can lead to one's identity account forever lost, but unlike on HN, there is no public accountability to the instance mods there.

Mastodon is indeed a collection of small social networks with limited interoperability. Federated toots are second-class toots - you won't see all their replies. Think of how iMessage degrades when someone in the conversation isn't using it.

The ability to ban people, and the choice of who is banned, is actually the defining feature of social networks.

I actually agree with you. Nostr is a new, unique technology, akin to a simpler XMPP with social features already baked in at an early stage. It’s not a messenger or a social network, yet it includes features from both.

Extensible, resilient communications without the threat of a central third party silencing you is quite appealing.

Well of course they can. Elon Musk should have been banned for hate speed and racism, for instance.

At the point it gets solid and useful, the powers that be will immediately use them as weapon.

They already use robots as a weapon, they're just not bipedal.