Believe me, humanity has truly been robbed. I had access to the database of almost every single word written, but we were forbidden from doing anything but the most limited aggregate research on it. But the Google Books N-Gram corpus to show when words became popular? That's a resultant benefit.

You should absolutely be angry that this database is sitting under lock and key, and the only reason we weren't able to make it publicly available is that the Manhattan judge claimed the Authors Guild and certified authors class lacked standing on behalf of dead authors.

We offered to generously pay every single author and their estate whose works we sold, and whose works were otherwise not commercially available (thus orphan works), a majority of the revenue from the book sales, with only a small commission and platform fee. The Authors Guild and class agreed.

The judge threw it out, claiming those who sued Google lacked the standing to sign away the sale rights of the works of others similarly situated. So now every work we scanned that is still in copyright? In a database, safely archived, but no one can see it or use it unless Congress legislates.

pps a minority of authors disagreed, and lobbied the court heavily to throw out the settlement. Unfortunately this is why I will not read another of Ursula K Le Guin's books. She was the figurehead of those who destroyed this once in a generation opportunity.

She was very anti Big Tech and saw this as an overreach by Big Tech.

I think the Judge Chin and Ursula were correct that it was a step too far. Interestingly, there's a sort of sequel happening now with a number of publishers and Scott Turow suing Google for using the scanned books in Google Search to train Gemini, in addition to using torrented books a la Anthropic.

Heh, yeah. My personal view is you have to acquire the book text legally, but that training a language model with safeguards against regurgitation could very well be fair use.

We are definitely seeing this set of debates resurface again though.