Borges, Averroes, ChatGPT, Myself
A short reflection on humans' lack of introspection
From my privileged position as a human being, it’s easy to glimpse the impossibility that grips ChatGPT in its attempt to understand—and perhaps model—the human mind and the human world. Despite having read everything that’s ever been transcribed into ones and zeros and uploaded to the infinite internet, it’s still trivial to make it stumble, lacking even the most mundane human abilities.
Its absence of common sense, its inability to avoid mistakes that betray a lack of consistent mental structure—all this stems from the fact that ChatGPT doesn’t exist in this reality. It exists in another one, which it commands with godlike skill, but in this physical realm of oxygen, love affairs, and the ceaseless annoyance of traffic jams, ChatGPT has yet to set foot.
And then I realize, as easily as I spot the limits of any artificial intelligence, that I’m just as incapable of recognizing my own and that ChatGPT, were it to have free will and a body to extend a condemning finger, could point at me with the same accusation: Just as it cannot comprehend me, I cannot comprehend it, and therefore I cannot model its mind from mine, as superior as I might consider myself to be.
Like Borges at the end of that tale, where he tried to recount Averroes’ search for Aristotle’s notions of tragedy and comedy, I have to accept defeat. The Argentine writer, clever as he was, realized that just as that Arab physician had been separated by fourteen centuries from the Greek philosopher, making his quest difficult, Borges himself was separated from Averroes by another eight or nine—both fated to chase after an understanding they’d never attain.
And even though I am a contemporary of ChatGPT in space-time, our alien natures—one made of bits, the other of atoms—create between us an unbridgeable abyss.
So, recognizing the humility I lack when pointing out the flaws of something I do not truly understand, while failing to acknowledge my own—which exist in equal measure—I offer an apology for my strong arrogance and weak introspection: it is far easier to attack what one does not fully grasp than to recognize—and accept—the impossibility of ever doing so.
If anything, perhaps this little piece brings me closer to that goal than the very jokes we overuse, at the expense of this young chatbot; jokes that could be turned right back on us with equal sharpness and effect, if only we dared truly look in the mirror.
REMINDER: The current Three-Year Birthday Offer—20% off forever—runs from May 30th to July 1st. Lock in your annual subscription now for $80/year.
Starting July 1st, The Algorithmic Bridge will move to $120/year. Existing paid subs, including those of you who redeem this offer, will retain their rates indefinitely.
If you’ve been thinking about upgrading, now is the time.
Increasingly, I think that embodiment is everything.
One basic assumption leading us astray is ascribing individuality to ChatGPT (and other LLMs). It is not a he, or a she, or even an ‘it’. Interacting with AI is interacting with very loosely connected iterations of a set of algorithms. Like visiting different shops at a mall: the meaning is in the shops, not the mall