From the Department of Very-Much-Intended Consequences, you may have recently read that AI companies are purchasing massive quantities of rare books, scanning the pages to “train” their algorithms, then shredding the original books. If not, you probably will soon.
If that sounds improbable considering the almost inestimable quantities of training data available to them free via public domain access on the Web, there is indeed a sort of twisted logic behind the ‘scan-and-destroy’ strategy.

.jpg)

.jpg)




.jpg)

.jpg)


