A centralized platform with marginally cleverer cryptography can technically allow posts and comments to belong to a user, but not be traceable back to the user! When asked by the government who made a particular post or comment, it should not be possible for the platform to readily identify who made it. Of course the IP address should not be tracked either, certainly not beyond 1h. The logged in user would still be able to view and manage all of their posts and comments because they would have the cryptography key to do so. So what about spam and guardrail control -- one solution is to let AI classify it. Much more can be possible with cleverer uses of cryptography. In short, it is not formally necessary to switch to a decentralized or federated solution to address the anonymity issue.