Regarding the content of this interview:
>If you compiled an enormous dataset of everything Borges read, and combined it with an exquisitely sensitive record of every sensory experience he ever had, could you create a Borges LLM?
This is my Kantian way of thinking about epistemology, but I don't think that LLMs can create synthetic a priori knowledge. Such knowledge would be necessary to create Borges out of a world without Borges.
In this interview, Simon's view feels much more like the way Hume viewed people as mechanical "bundles of sensations" rather than possessing a transcendent "self". This led to his philosophical skepticism, which was (and still is I guess) a philosophical dead end for a lot of people. I think such epistemological skepticism is accurate when applied to machines, at least until some way of creating synthetic a priori knowledge is established (Kant did so with categories for humans, what would the LLM version of this be?)
Yes, his writings are short, but man they are dense!
To anyone who cares, do this exercise: read short story by Borges, probably the shorter the better. Then go ahead say, next day, and try to write it down again in your own words. I tried a couple of times, and I ended with at least twice the number of pages. Amazing.
Do you think that a LLM has the ability to identify a new a priori knowledge?
It seems like it would be a lower threshold to meet but If you combine that with a stochastic process then it seems inevitable that it would be able to ruminate until it came up with new a priori knowledge.
Sadly, almost no one talks about it. Ditch the form and embrace the substance. ← It also nods to the mystery behind The Navidson Record.
I wish I had known this when I first read it.
The Navidson Record is on youtube if you dare =P
Thanks!
Hmm, what if you could recreate, word-for-word, the great works of an author like Borges (or, say, Cervantes) by so thoroughly understanding their life that the words themselves came out of you, not memorized and recapitulated, but naturally and unbidden? What an interesting idea for a story, maybe an LLM will be able to write that one day.
With AI tools, though, I can "read" Borges in his native language: with my phone + OCR + translate I have an English language companion. Or, using the voice interface I can try narrating the Spanish text and ask clarifying questions whenever I'm confused.
An author like Borges makes it well worth the extra effort. And, his puzzles often involve language, so the extra layer of mental translation can mirror the work itself, e.g. in his poem La luna [1]. (though, I envy your native Spanish)
(I agree!)
I wrote about the connection between Borges, AI, Wikipedia, Kafka (the messaging system, not the author), GPUs, and cryptography in the small print on page 7 of this:
My read is that most likely, it was recorded on an old school reel-to-reel tape recorder. It's entirely possible that the tapes are still sitting on a shelf somewhere in Argentina, though the chances of actually tracking them down are pretty low. I worked with some reel-to-reel tapes that Alan Ginsberg made (now held at Stanford) in the mid-60s (including one where he is talking to Bob Dylan!) and they held up pretty well. Had to use audio editing software to remove tape hiss, but they were not as badly preserved as I expected.
Anyway, I uploaded the Simon book chapter at https://gwern.net/doc/borges/1996-simon-2.pdf
https://borgestodoelanio.blogspot.com/2017/05/jorge-luis-bor...
An LLM trained on Sartre would be amazing because the logical extensions of many of his positions and postulations would be uncomfortable in polite society. Even as a human being he quite frequently espoused concepts counter the grain of civility or notions of what ethics are or should be. An unrestrained, uncensored LLM in this vein could be scary and gut wrenching and yet a good reminder of our less-than-ideal state of refinement of thought and behavior as a species.