The notion of censorship resistence is whether the individual relay operator is providing a valuable service to the social network. If censorship of CP and spammers is a valuable service, those relays will succeed. If censoring political speech, or religious speech, or bigoted speech is a service deemed valuable to the network, that relay will succeed. To globally coordinate the activities of hundreds or thousands of relay operators to censor any one kind of speech is not realistic, even if there is a shared blacklist. The Outbox model makes a coordinated censorship campaign even less effective.