Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

And here you can e.g. find Ficino's correspondence translated into English, with commentary, https://archive.org/details/lettersofmarsili0000fici

If you make a cursory search you can also find other translations of his works, various biographies, and a wide range of commentary and criticism by later authors.

Many of Ficino's originals are also in the corpus of scanned and OCRed or recently republished texts. I'm sure there are archives here or there with additional materials which have not been digitized, but it seems questionable whether those would make any significant difference to a process as indiscriminate and automatic as LLM training.



Yes, but many of his books are not translated or ocr’d. For instance, La pestilenzia or de mysteriis.

And he is one of the most central figures of the renaissance. Less than 20% of neolatin has been digitized, let alone translated.

It is fine to question whether including neolatin, Arabic or Sanskrit in AI training will make AI better.

But for me, it is a core set of humanism that would be a shame to neglect.


Consiglio contro la pestilenza was apparently written in the Florentine language. You can find a nice scan at https://archive.org/details/ita-bnc-in1-00000486-001/ and the corresponding mediocre OCR at https://archive.org/stream/ita-bnc-in1-00000486-001/ita-bnc-... (using better OCR software could give a version with few errors; I don't think anyone has produced a carefully checked digital text). There's some discussion at https://www.jstor.org/stable/40606241 as well as plenty of other commentary around. If you are trying to figure out specific details about this book, you should just check the book. I wouldn't expect adding a carefully produced copy to an LLM training corpus would make that much difference unless you have niche questions about it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: