In thinking more about this question of who should run a “public option” AI, it seems obvious that the state is likely not going to be the best actor (depending on the state), though they would surely have some role in it. What’s the model then to follow? Trust corporations to benevolently something something? Yeah, good luck with that.

Given their experience in public administration of commitment to storage and universal access to knowledge though, what about public libraries?

The first dial-up internet I had as a teenager came from our local library. Why can’t we do the same thing with AI?

There’s no default that says this all has to be run by corporations and national security interests. That might be the arc we’re stuck in at the moment, but I don’t believe that’s aligned with the long arc of time, which is probably the best scale for us to measure AI against in the end…

It might be a hare-brained idea, but it might just be that we need the outsider angles to get through the coming impasse…