If anything, my take home message is that the reach of copyright law is too long and needs to be taken down a peg.

Exactly! Copyright law is terrible. We need to hold AI companies to the same standard that everyone else is held. Then we might actually get big corporations lobbying to improve copyright law for once. Giving them a free pass right now would be a terrible waste of an opportunity in addition to being an injustice.

I think the photocopying thing models fairly well with user licenses for software. Without commenting on whether that’s right in the grand scheme of things, I can see that as analogous. Most folks accept that they need individual user licenses for software right? I get that photocopying can’t be controlled the same way software can but the case was in the 90s? I mean these things aren’t about whether the provider of the article/software faces increased marginal cost for additional copies/users but that the user/company is getting more use than they paid for. License agreements. Seems like a problem with the terms of licenses and laws rather than how they were judged as following them or not. Their use didn’t seem to be transformative and the for profit nature of their use sort of overruled the “research” fair use.

I also think the mp3.com thing sucks, but again, the way the law is, that’s a reasonable/logical outcome. Same thing that will kill someone offering ebooks to people who show a proof of purchase.

I don’t know the solution to the situation with NYT/open AI. It’s a pretty bad look to be able to spit out an article nearly verbatim. We do need copyright reform, but I think that’s at the feet of the legislators, not judges. I only need to see the recent Alabama IVF court ruling to be reminded of the danger of more… interpretative rulings.