Doing this entirely on the client side falls down when you want to moderate something because there's too much of it (e.g. bulk spam)

The fix for that is to join your own community spam filter. It's the same way ad-block works. Community run blocklists.

I don't think servers blocking denial-of-service spam attacks is the same socially as moderators censoring speech they don't agree with. The same goes for content that is illegal for servers to host.

Right. The issue is not about client-side vs server-side, it's about the user being in control.

Yes. Exactly.

I prefer that I choose, 100% , who I talk to and what I say.

And if somebody offers me the service of filtering or editing all that, I have that option.