Power to the people.
So you want a moderator to moderate. but then you also want to have tools to see what has been moderated away and unlock those? Right? So moderate yes, but also unmoderate by the users.
Power to the people!
Mods exist and can ban/lock/block people and content but users can see everything that was banned, removed or locked, as well as the reason why; what policy did the user violate?
I think the only exception would be actually illegal content; that should be removed entirely, but maybe keep a note from the mods in its place stating "illegal content".
That way users can actually scrutinise what the mods do and one doesn't wonder whether or not the mods removed a post because they are biased or for ligit reasons, and opinions are not entirely removed, as they are still readable, but you can't respond to them
In addition crap floods? If I submit half a billion posts do you really want that handled by moderation?
Being a server operator I've seen how bad the internet actually sucks, this may be something the user base does not have to experience directly. When 99.9% of incoming attempts are dropped or ban listed you start to learn how big the problem can be.
As to moderation, why not be able to filter by several factors, like "confidence level this account is a spammer"? Or perhaps "limit tweets to X number per account", or "filter by chattiness". I have some accounts I follow (not on Twitter, I haven't used it logged in in years) that post a lot, I wish I could turn down the volume, so to speak.
What is spam... exactly? Especially when it comes to a 'generalized' forum. I mean would talking about Kanye be spam or not? It's this way with all celebrities, talking about them increases engagement and drives new business.
Are influencers advertising?
Confidence systems commonly fail across large generalized populations with focused subpopulations. Said subpopulations tend to be adversely affected by moderation because their use of communication differs from the generalized form.