Alan Jacobs makes a point here that I’ve been trying to make for a year or so, but have been unable to phrase anywhere near so succinctly:
If, as celebrants from Vico to Northrop Frye and Joseph Campbell have said, myths and archetypes are deeply and pervasively embedded in all our cultural productions — and pause for a moment to reflect on the enormous significance of this — then, per necessitatem, they are also deeply embedded in our large language models. Which means, first, that GAI endeavors will be thoroughly shaped by those myths and archetypes; and second, that if human beings are able to create artificial general intelligence, if the Singularity really does happen, then it will be foundationally constituted by those very myths and archetypes.
Perhaps it’s because I’m a reader and writer of genre literature—one of the places where, as Jacobs observes, these mythic underpinnings of culture are permitted to show, and to be played with—that this is so obvious to me?
Though there’s another step in the chain of reason, here, at least as I see it: this is to observe that the very prospect of “general artificial intelligence” is itself a mythic artefact, fossicked out of the myth-midden of science fiction by people whose lack of humanities education renders them incapable of seeing it for what it really is.
Which in turn means that “the Singularity” will never happen, in much the same way that the Augean stables will never be cleaned. “The Singularity” was only ever a metaphor; the foundational “if” of Vinge’s original pitch for it—namely “whether we can create human equivalence in a machine”, a question that begs a whole cascade of questions, all of which the scientific episteme is simply unequipped to even begin to ask, let alone answer—is a wonderful basis for science fictional extrapolation, but a really poor basis for a secular religion that mistakes itself for an industrial-economic policy.
(That said, we appear to have hit a stage with Singularitarianism that’s analagous to the shift from simple Creationism to “intelligent design”; witness this man with an AI-flavoured bridge to sell you, as he expends thousands of words returning to the essential “number go up, but exponentially!” thesis of the faith, while reframing it in Pinker-esque “everything’s improving, inevitably” liberal self-congratulation1. This may presage the fusion of a sort of Transhumanism Lite with legacy neoliberalism, which is not at all the philosophical paradigm we need, but—when seen from a cosmic perspective—might well be the one we deserve.)
So in one sense, Jacobs is totally correct: a post-Singularity reality would be entirely constituted of the mythologies that preceded it. The truth of that can be observed in the already-extant stories in which various writers have already built that (fictional, metaphorical) world.
I also think it worth noting that most writers of sf, even those most known for their work in this particular conceptual space, have long since moved on to other questions. The Singularity is a dead end. Hell, if you read Vinge’s essay, linked above, against its rather techno-triumphalist grain, he’s telling us it’s a dead end! It’s a vision of human obsolescence; there’s no viable sequel to that story.
A work of science fiction is never about the time in which it is supposedly set; it’s about the time in which it is written. Which means that the question we should be asking ourselves is why this particular myth—the myth of an exponential escalator of extractive and exploitative magic technology which will carry us to a heaven of disembodied and optimal perfection—is the one that’s bubbling through right now.
Vinge’s essay was written in 1993, at what might have been the peak of GBN-wave internet utopianism; back then, this was a new idea, or at least a new remix of a very ancient mytheme. So why is the Singularity back in the discourse now, three decades later?
Because the kids who grew up reading that stuff stopped reading fiction, and started reading management-fad literature instead.
An S-curve always looks like an exponential when you’re stood in the middle of it. That’s true of technological development, but it’s also true of a human life. The appeal of the Singularitarian exponential narrative has always been its implicit promise to a certain sort of person that their life shouldn’t have to slow down and tail off like the lives of their customers or employees.
I recall a piece by the transhumanoid and climate change denier Ronald Bailey, in which he wrote that “egalitarian opponents of enhancement want to make sure that the rich and the poor remain equally diseased, disabled and dead”.
To which my response was then, is now, and will always be, “yes, exactly“.
- This link came via Patrick Tanguay’s indispensable Sentiers newsletter. Poor Patrick had to endure yours truly being very snarky about the linked essay in the comments of a LinkedIn post, and I would like to apologise for that invasion of space—though not at all for the sentiment expressed, nor the manner of its expression. Screw these grifters and their secular salvation story. ↩︎
Leave a Reply