> You think someone's picking up a first edition Steinbeck and destroying it so it gets into the next training corpus?

I doubt the low-paid workers stripping, scanning and shredding the books are checking what anything is. I'd imagine they are dumped with a big crate of old books and an hourly quota they have to rapidly process.

Arguably old technical manuals with limited print runs are even rarer or more valuable to historians. We have lots of copies of Steinbeck, albeit not first editions.

▲

qarl2 3 hours ago | parent [-]

> I doubt the low-paid workers stripping, scanning and shredding the books are checking what anything is.

You don't think they are, at the very least, checking to see if the book has been scanned before?

A search like that could very easily be extended to check for rare books too.

Why do you assume it's not being done?

	▲	voidhorse an hour ago \| parent [-]
		Why assume it is? We don't know, and the fact that the process is so opaque and would have been totally undisclosed if not for journalists doesn't exactly make me confident they are "doing things by the book" on this one.