Turns out less is more ...
Private local AI on the go is now practical with LMStudio, including secure device links via Tailscale and fast model ...