I hope people start using these models to try to learn more about the system. Not even to overload them in a financial DDoS-style attack where a bunch of people start trying to overload the model's usage to make it so it costs tons of money (which is actually technically legal). But having crowd-sourced projects to collect the data from these models.

And if these companies complain about their intellectual property, they are wrong for these two reasons. First, they "stole" other people's content. (I don't think it's really stealing in my opinion but still.) Second, the responses to the user are now or at least should be the property of the user, and the copyright and patent rights of these corporations should not override the users when they are doing their own thing

And btw, the AI companies DID steal the world's intellectual property for training data. However much they try to excuse that theft now, there's plenty of evidence they knew what they were doing was unethical and illegal.

The AI companies were using data illegally and that is not okay. What I was saying is that if they are allowed to, other people should be too (no double standards for the rich).

The thing about hallucinations is very interesting because they actually imagine things when they can't see what the dta is.