Also, I think the bitlaw interpretation is incorrect. “Sweat of the brow” doesn’t magically produce copyright protection, and they don’t mention that.

Taking their example, if you had a collections from quotes from presidents, and I got a bunch of similar collections, then made my own ultimate definitive collection based partially on your list, then there’s very little chance I’d be liable for violating your copyright. If I copied the list and typesetting verbatim, you’d have a better case.

Also, modern rulings about LLM training (the topic of this thread) certainly mean copyrights on compilations of facts don’t survive training + inference cycles.