I agree that there's a combination of higher-order structure and expressive performance in music, specifically. Spoken language audio generation is in a place r

I agree that there's a combination of higher-order structure and expressive performance in music, specifically. Spoken language audio generation is in a place roughly similar to imagery, and we cam extrapolate that to AI performers, which I think is what commercial products for audio production have focused on. Like with speech, it still has an uncanny valley to go through.