Interesting article in The New York Times about training AI. The “cut corners” in the title is too loaded, and some of it seems a bit overblown:

Their situation is urgent. Tech companies could run through the high-quality data on the internet as soon as 2026, according to Epoch, a research institute.

Fair use is murky. Reproducing long documents word for word isn’t right. But a hypothetical C-3PO should be able to read a book or watch a video. I think we’ll need to clarify what is okay in training vs. what is okay when under human command.