Doubtful. I'm interested in that as well.

If I were making it, I would listen for keywords (ex: Cash App, Onit, MyBookie.com, Manscaping, Dollar Shave Club) that could be updated like an AdBlock filter list. Combine that with obvious changes in inflection, notable differences in the waveform, etc.

I would also assume my implementation would cut the podcast down to about 4 minutes with all the false positives.

I see it as a meaningful discussion, not necessarily hate, but I see your point! I can't tell how it works, just that it is pretty cool!