that's absolutely a valid point. the audience is tricky.

I'll use twitter as the example of an owned network, and analogize to nostr terms. they have all the relays, and no one can stand up their own relays. they custody all private keys, and can magically prevent random people from creating their own key pairs.

the nth node adds n-1 (potential) new connections to the network. this is the source of the exponential growth, what we mean by "network effect". in the twitter example, they own the network, and as such, they can monetize or not all the value that arises from that growth.

under the nostr protocol, it is not possible to own the network itself. the nth pubkey still provides n-1 possible new connections to the network, but no single entity can monopolize the returns from that growth.

in the example of a large relay with a lot of audience, there is certainly an incentive to use them, but they can't block you from taking your network connections to another relay. since switching costs are so low here (from the perspective of the reader), just add another relay to read from, it's hard to argue that any one relay will be able to exert much force in keeping their audience captive.

but absolutely if the pay-to-relay phenomenon takes off and spam is so bad that prices go way up, the audiences become stickier. so it's worth thinking through all these scenarios.

would you agree that an innovation of nostr is to break centralized ownership of the network?