I think we're actually in agreement with respect to the needs of AI in music and graphics. When I say "AI researchers" I'm talking about people in industrial and academic labs. I consider the hobbyists you describe to be the users (who obviously can also be researchers). It's the desires/needs of the latter that should drive the research agenda of the former.
I actually don't understand your position here or what you think I'm arguing for. My point is that the real musicians, artists and mathematicians (whether they're hobbyists, academics or professionals in industry) are not well served by detached AI researchers just trying to automate their work for them. They need AI researchers to understand their workflows and build tools that elevate them, i.e. bicycles for the mind (or hands?).
Again, I do recognize there may be new fully automated workflows that can come out of AI research too but I maintain that the actual artists, musicians and mathematicians today have a valuable role in guiding that development too.
I actually don't understand your position here or what you think I'm arguing for. My point is that the real musicians, artists and mathematicians (whether they're hobbyists, academics or professionals in industry) are not well served by detached AI researchers just trying to automate their work for them. They need AI researchers to understand their workflows and build tools that elevate them, i.e. bicycles for the mind (or hands?).
Again, I do recognize there may be new fully automated workflows that can come out of AI research too but I maintain that the actual artists, musicians and mathematicians today have a valuable role in guiding that development too.