Multi-device support in an encrypted messenger is an authorization problem. It is not a sync problem, not a UX problem, not a storage problem. The question that must be answered before anything else can proceed is simple: how does a second device prove it belongs to the same person, in a way that every group member can independently verify, without any central authority vouching for the claim? Marmot PR #44 introduces MIP-06 to answer exactly this question for White Noise.

I wanted to write you, but can't find you on White Noise, did you use a different npub?

Did you try with the latest release? It's less broken

That's because many relays don't support auth, especially not in combination with protected events. There is a big push to get that implemented and deployed to more relays in the last weeks. In anycase, the security depends on the encryption to a large extend anyhow.

Dev/Hack/Day + Cypherpunk Meetup = If you're coming to , don't miss it. Get your early bird tickets at freedomtechsummit.com

No, latest is actually 1.0.1, you should update, maybe uninstall and download again from zapstore.dev

Yes exactly. We just need to spec it out and build the libraries so that it works smoothly across clients. Not impossible, just takes time.

Well marmot is a really sophisticated standard that solves a lot of issues and keeps evolving. For example we just now picked up work on the anonymous mobile native push notifications scheme. We're hopefully soon at a point where there ate enough reliability and features to make the migration a nobrainer.

Is zapstore up to date?

When you install , make sure you use a different npub than in , multi client support with the same npub is not yet speced out.

seems like Wisp doesn't work when not giving DM decrypt permissions to Amber.

Basically yes, any consolidation decreases privacy. It doesnt matter if they have the same anon score, that's just coincidence.

Love it.

Even then you get a popup for each npub that doesnt have a LN address.

Oh yeah, I had a brainfart conflating ringsig with multisig... Of course you're right.

Yes, seems like that would work too. But not with publishing key material. If Alice uploads her private key, and Bob does not, That's not full deniability since Alice couldn't sign again if she doesn't know Bobs, neither could a third party. MACs are fully symmetric secrets, so any one uploading it is enough.