Also what’s the mechanism to prevent AT from turning into a firehouse of usenet bot spam?
The mechanism for spam filtering is the same as any other network. You can drop posts from low-trust PDSes, or use Bluesky's existing labeling system to apply advisory "spam likely" labels to accounts or posts that some heuristic decides are spam. This can be done in a composable way with their algorithmic feed aggregator and labels.
I feel like the XKCD perfectly reflects the free speech arguments I see online.
If a big chunk of your society make a habit of censuring people for disliked speech, organizations that are in the business of relaying speech make a habit of dropping customers over disliked speech, etc? Then your society is not practicing the principle of freedom of speech.
The stupidest thing I've read today