He concerns himself a lot with the ideas of scraping, which I think is fine, but this point specifically is salient and one that I really think people miss:<p>> The first thing to consider is the possible collateral damage of such a law.<p>The appeal to copyright law as some giant hammer of legislation to bring down on ai companies feels misguided at best and cataclysmic at worst. The implications of the type of law that would need to hone in on the technical aspect of how models work would need to, by necessity, be almost so vague as to suggest that, once any pixels of anything are accessed at all, the owner of the accessed work can claim copyright infringement on anything you do subsequently after that. He brings up a similarly glib idea, but I don't think it's implausible that, if given the opportunity, large companies will very much abuse any permutation of these laws:<p>> Unless you think search engines and the Internet Archive shouldn't exist, then you should support scraping at scale.