As an old-school video editor who learned editing on analog VTRs (tape machines) and progressed to digital non-linear editing, I think the time saving potential of AI-assisted audio editing is substantial.
I still think it will be a while before AI is ready to create more than a first draft edit but if it handles 95% of the easy cases that's still a huge time-saving - especially on long-form projects. I spent so, so many hours of soul-sucking, rote drudgery doing things like manual audio ducking back in the day.
Hopefully something like this catches on w/ smaller podcasters. Listening to shit-quality compressed internet audio in a podcast interviewing someone really interesting is just such a shame!
I edit one and clean all that crap out; usually 2/3 of the material ends up being cut, and I've had to write up explainers for journalists on how giving an interview is a different kind of journalism from writing an article. But doing all those micro edits, volume balancing etc. etc. is draining.
A 60-90 minute episode can take 10 hours of intense editing by hand, longer if there are quality issues on material recorded in the field. My stuff sounds good but I've been making arrangements to hand off a lot of it to a junior editor and incorporate machine transcription/partial editing to take care of the 'first pass', and I can't tell you how relieved I am.
You have my sympathy. In doing my own interview recording I've found that starting with good environmental control and properly placed high-quality mics independently recorded at high bandwidth makes a profound difference and saves 70-80% of post labor. I think these AI editing assists could yield another 10% netting an order of magnitude cost reduction for high-quality interview production.
I still think it will be a while before AI is ready to create more than a first draft edit but if it handles 95% of the easy cases that's still a huge time-saving - especially on long-form projects. I spent so, so many hours of soul-sucking, rote drudgery doing things like manual audio ducking back in the day.