I wish this argument would die. It's so comically false, and is just used to allow people to pave over their cognitive dissonance with the real misfortunes of a small minority.
I am a millennial and rode the wave of piracy as much as the next 2006 computer nerd. It was never, ever, about not being able to afford these things, and always about how much you could get for free. For every one person who genuinely couldn't afford a movie, there were at least 1000 who just wanted it free.
I happily pay for my media when there's a way to do so, without simultaneously supporting the emplacement of telescreens everywhere you look.
You have this backwards. There are way more poor people who can't afford things than there are people who can afford whatever they want
Genuinely cannot afford means you don't have the $15 to buy the movie after paying for necessities.
Cannot afford tends to mean "I bought a 72" OLED last week so no way I'm spending another $1400 on a movie collection".
If you have to use credit to "afford" such things, then you can't actually afford them
Stricter IP laws won't slow down closed-source models with armies of lawyers. They'll just kill open-source alternatives.
a) The model and the data
b) Why are we meeting in the middle?
For example, copyright makes it illegal to take an entire book and republish it with minor tweaks. But for something short like an HN comment this doesn’t apply; copyright always permits you to copy someone’s ideas, even when that requires using many of the same words.
I think most people think that AI training means copying vast troves of data onto ChatGPT hard drives for the model to actively reference.