everyone has heard of archive[dot]org which contains the largest database of books in the world in electronic format. It is a colossal project, much bigger than Scribd or Project Gutenberg
And ChatGPT should also have a database with all the books on the Internet Archive. For now ChatGPT is limited, if I ask him about any book on archive[dot]org he doesn’t know anything about it.
A special crawl should be made to index in the ChatGPT database all books, in all languages, from archive[dot]org
I read a lot of books in Romanian, Russian, and Hungarian, not just English or French. So, in the future, take this into account.
At the same time, on Internet Archive you can find very good specialized books, from various fields. Mathematics, physics, accounting, statistics, mechanics, chemistry, etc… But the OCR used by the Internet Archive is not very good in terms of rendering the formulas in these books, but it could have used a high-performance OCR such as Mathpix Snipping Tool which is a much clearer reader and knows how to read formulas and copy them into text files, using a special notation