They are, but we're always going to do down this route unless relays put a cap on the amount of pubkeys reading or writing to them.

Reply to this note

Please Login to reply.

Discussion

I don't think the centralization is as much of an issue when one can create their own relay, client, et cetera. It would be one thing if the projects were closed source, but since everything is open and able to be forked one can create the experience they want permissionless.

I think there is a nuance here,

Centralization can be an issue while censorship resistance being "unhindered".

For example : twitter centralization is an issue but the internet is by and large censorship resistant if you wanna use stenography.

Hm I don’t have an answer but I hope there’s a more elegant solution that incentivizes decentralization instead of forcing relay npub caps. That’s a sign the architecture trends towards centralization which isn’t robust.

Hoping people smarter than me build "search and cache relays" that charge sats for giving you a feed. That way clients can scale and relays don't have to serve every single client.

Those "Search and cache" can be use-case specific : For example : Facebook feed, twitter feed, reddit feed etc .... but we can all just use the same protocol and have open relays all over the world.

Layering may cause some centralisation on cache servers but at the same time, you can build your own and pay for access to the "archival servers" on a post-basis.

Disclaimer :  I'm an idiot and know nothing.