Adam

You’ve not factored in egress costs. Which on Amazon can add up quite quickly.

“For example, if every time I post a new update on BlueSky, if I had to send my post to every single one of my followers’ repositories, that would be extremely inefficent”

Somewhat ironic to have this posted on and activitypub driven fediverse.

In nostr’s case, that’s tech bros and freedom of speech “absolutists”,

Don’t forget the crypto bros.

But, the theory goes, you’re not supposed to be reliant on third parties as you should be in control of your own domain (or within a few degrees of the person who is).

Large instances are what are antithetical to decentralisation.

Of course, the reality of it is that, it just hasn’t worked out like that.

Activitypub makes it next to impossible to “move” an instance to a new domain.

Every post/comment/and user is uniquely identified using the domain. In the eyes of ActivityPub changing the domain just makes each of those things a completely new thing.

You can set up a new service at your new domain and potentially get most all your users to migrate but they’ll be leaving behind their entire histories and as a “new” fediverse user they’ll only be discoverable via the historical posts for as long as the original server is reachable.

a lemmy instance can act as a censor and push the biases of their admins,

This is a strength of the federated model. In an ideal world instances are small and a user’s values align with those of their instance’s admins.

The problem here is that a single instance has grown so large that a decision like this has had such an impact.

There’s a lesson to be learned here: don’t push a massive update close to the holidays because, ya know, you might break federation and go on vacation for a week.

I mean, no one has forced any admins to deploy a new version. But yeah, this is an annoying one.