Oh god, I'm not insane. This is definitely not a new surveillance vector and there are clear techniques to overcome it without meaningfully interfering with sig

Oh god, I'm not insane. This is definitely not a new surveillance vector and there are clear techniques to overcome it without meaningfully interfering with signal or performance. It also reminds me of some of the older copy-protection technologies which relied on a comparable idea of checking for expected variation at a level above the actual data stream.