randomize
winning no relay dev support from this quarter at least
following mike's path is not gonna make you popular with the people whose work your work depends on, just don't do it
We got some questions about the gossip/outbox model, so I’ll clarify Primal’s position.
We like the outbox model. We think it will help nostr scale and become more censorship resistant. Primal will roll out support for the outbox model in several steps. We recently upgraded all Primal clients to manage relays per NIP-65. We are currently in the process of upgrading our clients to leave relay hints on all published events to make sure that outbox clients work well with the content published by our clients. Next we will randomize the default list of relays for new users to help spread content out. The final step will be to implement reading content directly from relays, which will be used in conjunction with our caching service.
Imho, the winning approach involves using the outbox model together with caching nodes, giving us the best of both worlds. I described this approach a year ago when we launched the preview of Primal:
https://habla.news/u/miljan@primal.net/Lightning-fast%20UI%20for%20Nostr
randomize
winning no relay dev support from this quarter at least
following mike's path is not gonna make you popular with the people whose work your work depends on, just don't do it
How do you suggest we should set the initial set of relays for new users?
i've discussed this with nostr:npub1jlrs53pkdfjnts29kveljul2sm0actt6n8dxrrzqcersttvcuv3qdjynqn - use a referral scheme and pull in the referrer's follow list, what else makes sense?
the trending global feed is totally the wrong answer