Aren't we talking about blocking and muting known user handles? This is easy, the spam problem would be identifying which handles to block.
> What happens if B sends enough to DOS A's link? Or sends CSAM and now A is committing a crime just by downloading the content?
What happens if B does the same before A had a chance to block them? And why would A's client download the content after reading that B published it from the metadata?
> Which increases the friction on B if they want to harass A
True, but I'm afraid we already know from Twitter that B will do it anyway, aided by either fast user switching or a Tweetdeck-like interface.
> Aren't we talking about blocking and muting known user handles?
This is the same as blocking known spam domains and IP blocks in email. They just switch to a new one. Which is best dealt with at the remote end since the server has a better chance of being able to identify new accounts which are related to blocked ones - Instagram even suggests this for you when you block someone: "...and any future accounts they may create". Having to do this in A's client is much more difficult because it lacks the information that would be useful.
> What happens if B does the same before A had a chance to block them?
Aren't we talking about blocking and muting known user handles? This is easy, the spam problem would be identifying which handles to block.
> What happens if B sends enough to DOS A's link? Or sends CSAM and now A is committing a crime just by downloading the content?
What happens if B does the same before A had a chance to block them? And why would A's client download the content after reading that B published it from the metadata?
> Which increases the friction on B if they want to harass A
True, but I'm afraid we already know from Twitter that B will do it anyway, aided by either fast user switching or a Tweetdeck-like interface.