I think Nostr was created to be censorship-resistant. That is to say, no matter how much a relay or client may choose to censor content, users are not limited by those censorship choices. Even if all relays and/or all clients chose to censor particular content, anyone is free to spin up their own relay and/or client instances to circumvent those censorship choices of others. i.e. I think Nostr was created to render the value judgement of censorship moot.
Yes, anyone CAN spin up their own relay, but do we really want to make that our freedom of speech proposition? "Hey folks, Nostr is great--but you have to spin up your own relay to make it work if you don't want to be censored" One of our greatest (current) problems is easy of use...I'd say going down the "spin up your own relay" isn't really our best answer...
Don't conflate "CAN spin up their own relay" with "HAVE TO spin up your own relay". Again, only "if all relays" censored you would you be required to spin up your own. That means it would only take one like-minded individual as yourself to spin up their own censorship-free relay so that you (and all other like-minded individuals) wouldn't have to. At worst, this would-be issue presents opportunity (either to you or to anyone else) to cater to you and like-minded individuals.
Yes, I do get it...yet I still believe (STRONGLY) that we need to develop tools that let USERS control and filter their own content--and to not rely on someone else (e.g., relay operators) to do it. Frankly, the functionality to "focus" a user's feed is really missing from Nostr currently...developing such a framework would help to solve both problems. Decentralization is the primary core tenet of Nostr--and any "filtering" should be decentralized as well.
"we need to develop tools that let USERS control and filter their own content" This is where we agree. Ultimately, users ought to be able choose a stream of content to their liking (i.e. choose a set of relays) AND have the tools available to further curate that stream to suit their preferences with reasonable ease. I see no reason to assume that work isn't being done to eventually achieve this.