It would be great if there was a visual tool that could take the camera stream and overlay scripts, such as face recognition, or specific person recognition. The ability to overlay sound from another camera, crop the image size. Change image parameters. Add a clock to the camera image. Save the camera stream to your hard drive. Save the stream to disk only when motion is detected. Also overlay the stream of two or more cameras picture-in-picture. Those, develop visual camera editor without yaml.
And the next step is to make HA a video set-top box, with HDMI input and output. So that, for example, when watching a movie, a picture could be superimposed as someone enters the house. It will be nice if HA can output video to multiple TVs. During the broadcast, could display a part of the Lovelace interface that would accept voice commands. So that a certain event would be shown on the TV or on the monitor and displayed in a voice - you have such an event that you want to take.