Again, intent matters. If you ban a site because the admins intentionally platform hatespeech, that’s within federation policy. The same for users. You also own the hardware. You can do whatever you want with it.
But suppose you organized a ban across the entire federation of users who happened to be on a problematic instance but who themselves did not engage in abusive conduct. I mean, you made a list and every instance admin banned them.
That’s what these automod autoban scripts do. They ban by association. Even when the association doesn’t exist.
There is a difference between banning because of promoting violence or engaging in abusive and hateful conduct versus banning because someone espoused a generic political view. And that’s happening on Reddit. For PR.
EDIT: I think your argument is build upon a foundation of false equivalency. That is, you argue your bans are the same as how bans are used on Reddit. And I’m countering, no I think not. There are genuine differences in the weeds of why. Or, similar outcomes (ban) based on different circumstances (reasons why) are not the same. Context matters.
That’s not my point. Reddit intentionally creates social media bubbles for audience accumulation, and in the process mods ban any and every account that speaks against a mod (or admin) driven narrative.
Hell, they’ll autoban you just for commenting in opposing subs, even when you’re just countering a nonsense statement there. They also coordinate voting off site (Discord/telegram), to fake comment and post scores.
It’s gamed in every way imaginable to manipulate your biases and gaslight you.
Here’s what Lemmy does for you. There are no corporate and political mods banning you or controlling the feed for their own self aggrandizement or PR. There is no faked voting algorithm. There is no gaslighting, which is central to the reddit business model.
Lemmy doesn’t have the audience (yet). But in every way reddit is infuriating, Lemmy is not.
There’s a book out of MIT Press which details the history of CyberSyn of the Allende regime. I think the approach taken here will not work. It’s computationally impossible with classical computing. It may be possible with quantum computers, but we don’t have enough accelerated algorithms yet to test, much less there also being no hardware yet either. That’s gonna take decades of pure computer science research.
Flatpak was a problem. This is a good move.