So making libgen is illegal, but using it to train LLMs is legal? I know there's a whole issue of transitive liability (maybe you couldn't know you were getting an illegal thing from the thief, so it doesn't always make sense for you to to be liable too), but this kind of thing seems to power way too much of my industry for me to be comfortable.
There's the concept of inducing copyright infringement (a la MGM v. Grokster), so much depends on whether those who train LLMs were inducing libgen's operations in some way, for example if payment or resources were being contributed to libgen.
Welcome to the future! Companies will make illegal or very expensive to access original information, like scientific papers. However, guess what, your friendly AI LLM, trained by your friendly tech monopoly on stollen data, will allow you to access all this research that was paid with your taxes, through monthly payments. But don't ask the AI where it got this information from, because it can get really upset with you...
It's actually amusingly easy to have ChatGPT criticize some OpenAI practice or another. Tell it to do a search for some controversial story, then to "analyze it from an ethical standpoint".
Rule of Capitalism #1: If it is found to be good, but one can't make money off it, it must be made illegal to produce run, then replaced with something that sucks horribly, but can potentially be profitted off of.
I'm not sure I can agree with that, because historically speaking would include the time we had nobility. And in that time period, having money would not provide you with power, as nobles were beyond the law and could simply cease it for themselves.
