We present Remot-IO, a system for mobile collaboration and remote assistance around Internet connected devices. The system uses two Head Mounted Displays, cameras and depth sensors to enable a remote expert to be immersed in a local user's point of view and control devices in that user's environment. The remote expert can provide guidance through the use of hand gestures that appear in real-time in the local user's field of view as superimposed 3D hands. In addition, the remote expert is able to operate devices in the novice's environment and bring about physical changes by using the same hand gestures the novice would use. We describe a smart radio where the knobs of the radio can be controlled by local and remote user alike.