Amen.

Slap a shotgun mike onto the camera to make the audio marginally acceptable, then get someone with a passing knowledge of Adobe Premier to edit out all of the boring bits and awkward silences. Or maybe go nuts and actually show things in the video that don't come across in text, like the gadgets and whatnot that the people are talking about and layer the edited interview audio over said footage, and Voila you've got a video worth posting. Uploading the raw, unedited footage from a rambling 15 minute video, while a step up from a vertical video, is the reason why there's so much hate for the videos.

I whipped up a very similar prototype a few years ago. The vibration frequency was proportional to distance, and edges were highlighted with sharp jolts. The hope was that you could build a mental map of your surroundings by sweeping the sensor around. The ultimate downfall was the ultrasonic range finders. You need a narrow beam to be able to get any kind of reasonable spatial resolution, however, whenever you hit any sonicaly specular surface (i.e. a flat wall) at an angle that's not very close to perpendicular, the outgoing ping just bounces off to the side, instead of coming back to the detector, giving you a false negative. Which translated into running into walls that are 45 degrees to you. So until we get cheap, robust LIDAR sensors, the idea is mostly a waste of time.

Agreed, the combination of wide-screen camera in the wrong orientation on a wide screen monitor is as bad as it gets. The video shooting apps should by default sense when the phone is held vertically, while it's running, and make it vibrate incessantly until the user turns it landscape.