Disagree with the scale limit, in design considerations. If someone is willing to pay for it, especially, with work or satoshis, the system should scale to handle it, like any economy does to a population. As you said, we're not doing money/finance, so we can give up some things. Scale is one we want back in a global chat/speech system.
#Nostr isn't making guarantees about censorship, no, but it does make tools available to verify. Perfection is the enemy of the good here. We can't give up scale so we shouldn't try to get a guarantee. We can still verify. If you go to another connection and act like another user, you can check your visibility. It would be nice to augment the protocol, I think, to make it better at detecting and prompting a verification check.
If you're being censored or moderated, make a change in relays. That power to the users is the economic means of enforcement in #Nostr, where we lack a more rigid protocol level verification and enforcement of censorship resistance.
There are always other means available to reach users, even within #Nostr itself, like you've mentioned in another post a DM with ack requested would be hard to censor. It is sufficiently functional.