I think if clients did a better job on relay discovery, we already have enough compute to serve tens of millions of users
Like I doubt Twitter has hundreds of servers but we already have hundreds of relays.
One relay operator can serve thousands of users for just like $10+$20/month
I think that could change depending upon what the users are doing. In my experience, most Bitcoiners are mostly using text. A lot of normies are expecting Instagram, TikTok, YouTube, etc and that gets really expensive quickly. Number of relays isn't necessarily the important metric. It's the resources required to keep all of this up. You say 'just like' as if someone is just going to pay that for millions of people for nothing. We just end up right back with centralization, spying, ads and the old world bullshit if people don't understand the principles that got us here. At which point we just end up with a segment of Nostr that's probably smaller than we are now and a segment running relays owned by Facebook. Is that better? Maybe, but we need a better financial model and a lot of kinks worked out before trying to dump millions of normies onto Nostr. Most people aren't even using password managers. Lol Understanding, protecting and using private keys is just going to rek and frustrate a lot of people. And there are other kinks and things people just won't understand. Nostr was never going to replace everything in a couple years.
> A lot of normies are expecting Instagram, TikTok, YouTube, etc and that gets really expensive quickly.
No, it doesn't get expensive because most of these users are failing to decentralize the storage of images and videos anyways, and merely linking to them.
Links are text, no more expensive than any other kind of text.
There is still an advantage in sharing links this way because, although the linked assets are not resilient, one's audience is and if some hosts start censoring stuff one won't have to rebuild their following, one can just use a different host for new images.
