I've said nasty things about the videos I've made; this is one of the more recent ones. My comments usually boil down to: "they just sit there," and don't supply the visual interest viewers expect.
I think of them as video poems -- where the subtitles are the star -- and maybe that's where I go wrong: the whole idea of having to pay close attention to text over a video background is silly. Even the (vaguely similar in technique) "Deep Thoughts" bits on Saturday Night Live had more of the standard values -- like a human voice reading along, some camera movement and scene transitions -- that suggest an active view.
I make these things with a program I wrote that plays sound and video files, and drives software character agents according to a screenplay I feed it. The subtitles here are the word balloons of an invisible narrator.
The video clips used above are from a fast food commercial, an old movie, a crude webcam video I took of gas storage tanks being imploded in Brooklyn, and another one of the Union Square subway station.