so im not a fan of AI in music. but, i am a fan of Giorgio Sancristoforo and his compositions on the SFTMC Buchla 100, through which i discovered his AI/ML (SampleRNN, TensorFlow) based album Mellophonics. and, i have to admit, i think its quite good.
Mellophonics transports glitch music into the 21st Century; this time it is not CD players and hard drives that fail, but artificial intelligences located hundreds or thousands of miles away from the programmer; Mellophonics establishes a contemporary form of sampling through machine learning, exalting the defects and the hallucinations of the machine, putting a spotlight on the near future of music and of the non-human intelligences that will dominate the field.
How will we deal with it?
I was listening to Nils Frahm’s Mussel Memory from Music for Animals when I launched above Sancristoforo’s album… I have to concede both really work well together.
This notebook was authored by leonsarmiento@gmail.com. It may request access to your data stored with Google, or read data and credentials from other sessions. Please review the source code before executing this notebook. Please contact the creator of this notebook at leonsarmiento@gmail.com with any additional questions.
I think a lot of it comes down to people not understanding how the “AI” works… that said most of these AIs were trained on data sets that used copywritten content, AI isn’t “intelligent” and really is just a clever ways of parsing huge data sets, also is debatable if you can make something with AI that is truly yours using any of the current AI models, even doing custom trainings doesn’t eliminate what built the AI in the first place and the more something diverges from what work you trained it with the more it is other peoples work defining what it is producing.
I think people prefer to think of AI as magic though, the more you understand it the more boring it becomes imo. Could be/can be a useful tool but we have really yet to see where it will all shake out legally, anyone using it should probably truly learn how it works (I would say my understanding is still pretty low) and then make a personal ethical choice on if they want to use it or not.