[cavoirom]

One thing I like about the ATProto is the stable identity, I could change the hosting as the author did recently and the other users see no difference.

The one down side of the system is the cost. It's cheap to host a PDS but expensive for other components. Users could not relies on "someone" for running those components for free forever.

[Zambyte]

The cost of ATProto is only high if your AppView has a goal of "every user should be able to see and access every post and interaction indefinitely". If one limits the scope of what users see on their AppView to, say, a similar scope as what popular ActivityPub applications do (only posts and interactions from users that use that instance, and people that they interact with / follow are available / retained), it becomes not that expensive at all.

The difference is that it's easy to scale ATProto AppViews up beyond the reach of their users. It can scale down though. It is not easy to scale ActivityPub instances up beyond the scope of the people who use it, and it would probably be way more expensive if you tried.

[danabramov]

I wouldn't say other components are expensive. Common ones are:

- Relay as an optimization. That's cheap-ish ($30/mo) or free-ish if you pool with others.

- Your own app server. That's on par with normal web apps as long as you can keep a socket open. What's expensive is if the app you're making is a fully capable copy of Bluesky itself with gigabytes of existing posts — but is that the app you're making? The economics here are identical to normal web app stuff.

[cavoirom]

> What's expensive is if the app you're making is a fully capable copy of Bluesky itself with gigabytes of existing posts

This is the implicit idea in my comment when I said it's expensive. If Bluesky banned me and I could not find another AppView with comparable reach and audience, I lose, the ban is effective.

Another problem is ATProto users don't usually associate then with an AppView the same way Mastodon users associate with the instance, it's hard to raise fund to sustain the infrastructure cost.

[danabramov]

>This is the implicit idea in my comment when I said it's expensive. If Bluesky banned me and I could not find another AppView with comparable reach and audience, I lose, the ban is effective.

Okay, but that's always the case when you get banned from a service. The difference here is that it's possible to build a competing service with different moderation decisions being a lens over the same data.