Yep, but better than nostr devs going to jail on CSAM charges. Hopefully it will be temporary.
Discussion
The issue is DoA until major relay softwares have moderation tools.
Storing this data in a DB is also not a great plan.
I don’t understand how offsetting the liability on someone else is a good plan either.
Personally, I think making it easier for people to spin up their own stuff with Nostr integration is a better plan. With a competitive market of bigger fish for those who don’t want to deal with it.
My picsur instance for example. Picsur is super easy to spin up. If we had ShareX api capabilities in clients, you could even have uploads directly in the client to your private image hosting instance.
Does picsur have moderation tools? I wouldn't mind using an existing product or protocol as long as relays can recommend a good service provider
I wouldn’t recommend picsur for scale. It also uses DB backend, and lacks some moderation tools. I used it because it provided a lot of really useful features, such as URL based image conversion. It would be perfect for a self hoster though.
If some of the proposals end up working the way you described in your other post, and have a standard API, then that’s great. It’s almost not “relay hosting the image” in that case though, it’s more like “relay+hosting” configuration.
Yeah, agreed, it should almost be a different thing, except I think the network topology already in place with relays is valuable
Relays don't have to store in a db, be limited to one AZ, or serve over websockets. It would also be easy to support imgproxy syntax as part of the spec. A relay is just a server, anything a server can do a relay can too. The benefit is access to >1 image server with the same API.
That said, you're right that moderation (or at least operator willingness to take responsibility for what they're hosting) is a hard requirement. This is a much bigger ask than for regular relays, it's ok if we have big operators here IMO.