Yesterday I snagged a library book on Generative AI (GenAI), which is what ChatGPT is along with other GenAI we're hearing about, like Suno, as I understand it anyway.

And hey, I've barely scratched the surface, but something the author pointed out early on is that GenAI operates on very advanced 'predictive' software involving Large Language Models, but the output is an illusion. ChatGPT doesn't understand the prompt. It doesn't understand its own response. It doesn't know 'true' or 'false' or fact from fiction. It generates new things from parts of old things in its database.


My stuff:
https://soundcloud.com/scott-h-olson
https://www.youtube.com/@ScottHOlson

Sonar Platinum, Studio One Pro 4, Windows 10, Hybrid Home Studio