it's an interesting approach, since you can personalize the moderation, that I see used in some communities but is not perfect. The implementation that I see requires good tagging (where you mark what tag are you not interested in) or has poor discoverability and high manageability cost (where you have to find the correct blocklist, and this blocklist has to be constantly updated).

More importantly, it doesn't solve the other problem: I don't want to use a website that supports a certain ideology or promotes dangerous content. So the website operator needs to have some sort of serverside moderation otherwise part of the community will not comfortable using it.

Some imageboards use both system, a client-side taglist that hide undesirable comment, and some "server-side" rules about what is generally allowed. I don't think client-side only moderation is sufficent