I was just watching a cheap cassette player spit out white noise and thought about how you could turn that hiss into a data stream. Have you ever tried mapping audio glitches to code?