Do you think there's a world where we could query an API for raw GPU compute for a model ran locally something like that?

I think it's not possible without trust assumptions, but all cloud has a baseline trust assumption so that's just how that is.

#asknostr @Maple AI