Mikey Shulman, the CEO of AI music generator company Suno AI, argued that most people don’t enjoy the “majority of time” making music.
Mikey Shulman, the CEO of AI music generator company Suno AI, argued that most people don’t enjoy the “majority of time” making music.
As someone who don’t know anything about music, AI could enable me to make some music. But I doubt the next one-hit-wonder will be done purely with AI, and especially by some tone-deaf novice.
But would it be you making the music? If I go to a restaurant and ask the waiter for a burger, hold the onions and add a fried egg, did I make that burger, or did I ask for it?
Good question, if I start digging into this, and let’s ignore lyrics to begin with.
In your example the prompt could be, “Make me some instrumental 80’s rock music”, and here you just get a fusion of what the LLM think is 80’s music, the prompt could be refined and tuned to get some specific sound, but it can’t go past what the LLM knows.
Alternatively someone who knows more about music could train their own LLM on what they need and make different tracks that they mix together. In this example an artist uses the LLM as an instrument.
One thought is “Art equals effort” so the more effort/experience an artist puts into a thing the more it becomes art (is there a term for this?).
One thing is sure: It’s complicated.
If we can resolve the ethical complications of AI, I agree, AI could be a net-positive, beneficial tool for learning and accessibility. Suno isn’t really that, though. It’s more like a vending machine.