Yeah also presumably most of the AI happens on servers somewhere else with only low-power functions being done on-board. How much network bandwidth is used? How many servers per robot? If they're cloud-based, how much is the subscription?
rimu
Check this out for general background discussion https://piefed.social/post/205362. The idea to differentiate by trusted instances was mine and not discussed there. Pretty sure there was some discussion about it in the Matrix channel which is lost to time.
During the recent roadmap planning one of the potential units of work was to sort all this out https://piefed.social/post/411591 but it didn't garner significant interest and didn't make it through to the final version of the roadmap.
I hear ya. There was quite a bit of back-and-forth about it and we ended up with a compromise. It would be good to have more configurability of this to suit different preferences.
There's a niche out there for a max-privacy instance. No server logs, no email verification, automatic deletion of old content. And if it was running PieFed, no trusted instances set.
Not a niche I want to pursue but someone could.
Bummer.
It depends what your threat model is. Admins being dickheads about who downvoted what was the main issue at the time so I made it about choosing which admins to trust.
If future Lemmy versions show votes to mods (not just admins) then Pandora's box would be well and truly open so we'd need to rethink this.
MailCow is similar except uses docker. I expect that will mean easier maintenance as it is less tightly bound to the underlying OS.
Is the transmogrification series peak Calvin & Hobbes? Maybe so!
One time, in 2008. It was an envelope, tho.
Last chance to strengthen their bargaining position
How do we feel about the idea of reply controls on Lemmy threads? Limit replies to members of the community, same instance, etc...