5 comments

[ 6.3 ms ] story [ 173 ms ] thread
Presumably it’s floating around torrents.

It’s a strange world though where meta can use it without legal consequences yet take downs succeed.

Distributing copyrighted books is different from learning from copyrighted books.

That said, most of the major AI companies have pending copyright lawsuits. The one against OpenAI specifically mentions this dataset. Nobody knows how it's going to play out in court.

> Distributing copyrighted books is different from learning from copyrighted books.

I don’t think that has been established

Maybe people want to read copyrighted books to learn. Then we’re down to saying it’s ok for machines but not humans?

It is part of the pile. So, ya don't know why this matters. You can easily get it.
Another chunk of data gone from the-eye.eu, unfortunate. That's what you get for hosting things publicly. Nothing torrents can't solve.