The local models require massive amounts of ram that isn't available for most consumer hardware

I do hope the bubble bursts soon but low ram and local ai aren't compatible