@mattsheffield aside: subjective outputs coming out of genAI has a significantly different context to (supposedly) objective outputs coming out of gen AI.
When there is no need for genAI outputs to be objectively correct (e.g. 2 + 2 = 4 according to defined axioms), the picture looks different.
I don't have an easy way of mentally reconciling the following personally:
1. a human person learning music and organically creating music which necessarily has been influenced by how they learnt music and their lived experiences.
2. a gen-AI tool processing musical inputs and generating music which is necessarily informed by inputs and randomness inherent in the gen-AI tool (for sake of argument lets say authors of the inputs were suitably financially compensated)
3. when some people (I'll put me in that category) listen to music, I don't typically have an appreciation of the background of the artist. I'll also disclose that I haven't been looking for AI music.
Is it as reductive as - a gen-AI tool is less creative because it can only blend inputs? And then its a matter of incremental improvement & expanding inputs until creative equilibrium is reached.
Conversely looking at it from the outside from a purely utilitarian viewpoint, why spend all the effort, resources and externalities into making music when we already have musicians?