
How do you feel about generative sounds and AI?
Alongside the rise of AI-generated music there's also some serious pushback, very visible around Reddit.
I'm curious about people's opinions around here.
I've only had a brief play with systems like Suno, where the AI models create mostly from their memories of what they were trained in. It may be better with a pro account, but I found it both amazing and disappointing. Amazing because you get quasi-finished tracks out of thin air, disappointing because there's always a familiarity, heard it before.
Ok, that's an extreme case of machine-assist.
I've recently been working a lot recently with generative plugins. None (as yet) actually incorporates AI at run time, but they are algorithmic. I have a bass line generator that operates according to certain rules (genre, key, scale etc), another that does counterpoint against it, another that adds chords, another that sequences drums.
I also now have an AI vocalist. A friend sent me a speaking toy AI bot which is hooked up to a language model in China. So I've strapped it to a microphone. I can tell it a prompt and it delivers its response into the mic. Great fun.
So, thoughts?
A tool is just a tool, you should use anything you like, still legit (my personal opinion)? Or do you draw the line somewhere, some systems are somehow *wrong*?
Here's a video I recorded yesterday making a lot of use of the generative plugins, together with the AI vocals : https://youtu.be/CCkw3s7abps?is=mZqG7rjYsmWFYeqx
The generative plugins are around here : https://github.com/danja/downspout
I just released an album that uses some of the plugins here and there : https://github.com/danja/attone