Half of the occult books are talking about magic and irrelevant stuff. The other half is philosophy and spirituality hidden behind materialistic concepts (think Freemasons for example).
All those books would most likely be useless or detrimental for LLMs I guess.
Most of the books are the outcomes of the Renaissance. The relationship between “science” and spirituality was much closer then than now.
Further, most books published in Europe between 1300-1700 were written in Neo-Latin. Most of these books, therefore, have not been digitized and translated.
Now, to me, it seems like a real shame if this humanist core of European thought is deemed too dangerous for consumption. But it wouldn’t be the first time. The library behind these works, the Biblioteca Philosophica Hermetica, specializes in books banned by various church authorities.
I personally believe that these materials should definitely be part of large model training. The renaissance, esoteric though it may be, deserves to be part of the diversity of thought used to train LLMs.
We can easily imagine an AI apocalypse - maybe these books might even help us imagine an AI renaissance…
> I personally believe that these materials should definitely be part of large model training.
Already done: https://news.ycombinator.com/item?id=37752272 It turns out that the real safety risk with AI is not Mecha-Hitler, it's just that it might end up reading the wrong sorts of books and accidentally conjure a horde of demons.
Probably absolutely no. I was studying about the corresponding names between tarot card and Shem Hamphorash and gave me incorrect names, it gave me a correct angel name but not the correct one of several cards.
Half of the occult books are talking about magic and irrelevant stuff. The other half is philosophy and spirituality hidden behind materialistic concepts (think Freemasons for example).
All those books would most likely be useless or detrimental for LLMs I guess.
Most of the books are the outcomes of the Renaissance. The relationship between “science” and spirituality was much closer then than now.
Further, most books published in Europe between 1300-1700 were written in Neo-Latin. Most of these books, therefore, have not been digitized and translated.
Now, to me, it seems like a real shame if this humanist core of European thought is deemed too dangerous for consumption. But it wouldn’t be the first time. The library behind these works, the Biblioteca Philosophica Hermetica, specializes in books banned by various church authorities.
I personally believe that these materials should definitely be part of large model training. The renaissance, esoteric though it may be, deserves to be part of the diversity of thought used to train LLMs.
We can easily imagine an AI apocalypse - maybe these books might even help us imagine an AI renaissance…
> I personally believe that these materials should definitely be part of large model training.
Already done: https://news.ycombinator.com/item?id=37752272 It turns out that the real safety risk with AI is not Mecha-Hitler, it's just that it might end up reading the wrong sorts of books and accidentally conjure a horde of demons.
2 replies →
> books banned by various church authorities
A book that "explains why" the church is still against it, is the Anti-christian conspiracy, Msgr. Henri Delassus, 1911 (I guess)
1 reply →
More than useful for running a d&d campaign
Thinking of a RAG with the entire Ritman library collection as a GM.
Probably absolutely no. I was studying about the corresponding names between tarot card and Shem Hamphorash and gave me incorrect names, it gave me a correct angel name but not the correct one of several cards.
So for studying? Nope, for practicing neither.