

It’s saying that copyright law doesn’t apply to AI training, because none of the data is copied. It’s more akin to a person reading an impossible amount at an impossible speed, then using what they read as inspiration for their own writing. Sure, you could ask an LLM trained on, say, Edgar Allen Poe’s works to recite the entirety of The Raven, but it can only “recall” similarly to a human, and will have just as many mistakes (probably more, really) in its recitation as a human would.
Does it matter? If you could quantify your lost potential, would such knowledge benefit you in any way? I understand the drive to understand your major life mistakes as deeply as possible, but you’ve already made great strides to learn from and correct said mistakes. You seem to be on a pretty strong upward trajectory, but if you obsess over what’s lost, it just might drag you back down.