Thank you. That article gave me some insights.
- AI firms fear AI generated content and train their models with books from before. The fact that AI waters down the source data quality, because AI generated data becomes part of the source data. It's worrying, let's see where it leads and how the AI effects can be mitigated.
- Lawsuits seem to be an issue for anthropic and a driver for the scanning program too. Because illegal copies are problematic.
- The books are destructively scanned and disposed afterwards.
- In rare cases a 18th century book might be lost. But the ordered books seem to include anything and nothing. The term rare book seems to refer to the fact that a small amount of copies have been printed and only a few copies are available for sale (and the larger part is stored in libraries).
- Booksellers are both worried and happy at the same time.
Jetzt bin ich aber doch überrascht.
Next they find out that meta also didn't care about all the other users, when they rolled out the addictive mechanisms. That they needed to please the gods of revenue... I mean... shareholders.