Yes, I’ve noted the delay, and understand your point, but perhaps that would only slow the incremental rise of the feedback loop. If there is a two second delay, then what it hears from your phone will be played back two seconds later, and then again and again, each section with a two second delay. If it were instantaneous, you would get almost instant feedback, at full intensity. With a two second delay, it should rise more slowly, but the camera still hears what the app on your device outputs, and plays it that back, after 2 seconds, and so on.
Indeed, out of curiosity, I just tested this, and it take almost the full 30 seconds to reach peak amplitude. The delay is a little under 2 seconds, in my particular case.