I think that what is being suggested is that there be multiple, user-selectable, opt-in providers of crowd-sourced moderation services that are orthogonal to the unfiltered communication platform.
Something like the 3rd-party spam filtering services for blog comments or email.
I imagine, though, that a 3rd-party "view" of the Twitterverse would also be a Twitter client, have it's own web presence, and basically be a white-label version of Twitter. Twitter itself would be largely relegated to being a backend pipe that the different sites have in common for interoperability.
I can't imagine how these federated Twitter clones would themselves be able to provide a curated experience without running into the same liability problems.
The alternative is that these optional collaborative moderation services are somehow integrated into Twitter itself, which, while not impossible, poses some serious issues of combinatorial complexity, scalability, and sustainability.
Yes, we have federated services now, but the largest and most popular services are non-federated walled gardens.
If your position is "never remove anything ever", you circumvent this. But when your position becomes, "actually, you can delete only spam, but nothing else" then you're back in the exact same situation we are currently in - some things are removed by judgement call and not everyone is going to agree on every judgement. It completely removes the very core of the argument - that judgement calls should be unnecessary or not allowed.
If you're suggesting crowd sourcing moderation in a democratic fashion, I would contend this both isn't effective and also that by providing systems that facilitate this, companies would be implicitly affecting the "censorship" of their platform. After all, someone has to decide what the voting algorithm is, what the chance of being a spammer before you are blocked is, etc.