I want to talk about the elephant in the room. AI in music.
I'm sick of real artists getting overshadowed by a generated idea in a computer that can be recreated by anyone who takes the uniqueness out of the original work that it was trained on replicating. that said, I love the concept of generative synthesis as a whole, because it's how notes are fit together that really makes a song and a mood fit together.
AI makes mistakes and it doesn't correct the mistakes by learning in the same way we do. If it forgets what it's doing in the middle of generating a song, it can't go back and change those parts. You have to direct it to recreate the song, and the parts that it got wrong and hope it splices the finished product back together properly. Which, at best is a tool in its infancy and I believe, just makes the entire thing sound a little worse than it already did. Especially when AI fuses instruments into each other or forgets overtones of one thing like voices and synth pads for example, or guitars and harps.
You really begin to notice this in musical styles that focus on replicating fast passages of some content while leaving others untouched such as rap, and heavy metal. Orchestral pieces it also struggles with a lot, especially because its watermark of synthesis generation adds high frequency static to everything it creates, making nuance difficult to listen to for long periods of time.
That said, I'm totally in favor of generative synthesis for music. this has been done through sampling of actual instruments at their most in tune peak for digital replication, and works flawlessly. Electronic music was born entirely from wave synthesis in repeatable patterns the artist directly controls, as opposed to the software, or its internal memory.
When AI first came out, I thought it was going to be able to focus on how these different types of waves could remember existing loops of patterns and turn music into the progressive rock style of creativity we saw from the latter half of the 20th century. But what got built instead, was a total demolishment of musical artistry through prompts that centered around replication of instrumental waves digitally reverse engineered from samples already recorded and isolated, as opposed to using already recorded or prerecorded samples of isolated instruments to extend their melodies, patterns or styles. Which would have completely preserved original artists and their work.
This tells me one thing. AI is a piracy haven for people's copyright complaints in its current generation in the musical, and media space overall. If I put a lot of hard work into an idea, I don't care who takes it. I care who preserves it, rather than replaces it with something else that ultimately forgets to keep who I was in mind. That is most apparent to me as a musical artist myself and I wish we lived in a world, where media wasn't a business, it was just a creation of personal interest and that's it.
That's the world of music I care to remember.