2010-ish: copying is not theft, stealing a thing leaves one less left, copying it makes one thing more, that's what copying's for... 🎵🎶🎵🏴‍☠️🏴‍☠️🏴‍☠️

2025:

I'm myself not overly enthusiastic about AI, not at all. I think LLMs are mostly bullshit generators with very little practical use. But from a public policy point of view, the stance that training AI should require consent of or payment to the copyright holders is IMHO wrong; if we accept it, what is next? Humans summarizing and paraphrasing sources (e.g. for a wiki article) should require consent/payment? Humans learning from sources should require consent/payment? I think wanting a free and open society means taking the side of AI operators, not copyright holders, in this matter.