This is a totally valid argument. Looking forward to see how this progresses.
News and discussions about technology
A place to share and discuss the latest news in the world of all kinds of technology. **Click here to support the development of Kbin**
I do not care that the write-anything machine got there by reading every book in the library. How else was it going to happen?
No kidding it can sound like Tolkien if you ask it to sound like Tolkien. You, the person reading this, can presumably do that as well. And if you can then it's because you read his copyrighted works. You dastardly villain.
But if you, a human person, actually do that, and then publish the result... yeah, you might be in trouble. That is the only point the NYT has, here. If some other business is ripping them off then it doesn't matter how they ripped them off. If it turns out some article they point to was written by an intern or an editor, that doesn't change the nature of their complaint against how the company is using AI. It is the application that is the problem.
Similarly, you can get an image network to render Shrek on a tricycle. The fact it can do it is fine. It'd be impossible to have a draw-anything machine that somehow draws anything except copyrighted characters. But if you stick that on a t-shirt and get sued, you can't blame the image network.
I think part of the issue is that OpenAI didn't pay for the NYT content it used to train ChatGPT.
Take a wild guess how I feel about that, given the comment you just responded to.