The Singularity Ain’t What it Used to Be:

Rejecting the Frame of Tidy Foresight and Prophetic Vision

Vernor Vinge coined the term the technological singularity in 1983¹, but I was deep into my Singularity obsession before I ever encountered the term. In college in the early 90s, I picked up a book in the university bookstore called The Tomorrow Makers: A Brave New World of Living-Brain Machines by Grant Fjermedal. It told the story of researchers, animated by a maniacal obsession, driven to unlock the mysteries of cognition, consciousness, and perhaps immortality. That book led me to the books of some of the people Fjermedal interviewed for The Tomorrow Makers. Over thirty years later, author / title combinations still leap to mind:

Engines of Creation by Eric Drexler

Mind Children by Hans Morevec

Society of Mind by Marvin Minsky

I’m pretty sure I never read The Physics of Immortality: Modern Cosmology, God and the Resurrection of the Dead by Frank J. Tipler, but I used to eat lunch with another philosophy grad student to discuss Singularitarian ideas (without using that term). Neither of us lived in a dorm, but we both had campus meal plans, and we would linger in the meal hall drinking coffee and speculating about possible futures on afternoons when neither of us had seminars to attend, classes to teach or deadlines to meet. I soaked up enough secondhand accounts of Tipler’s speculations that it feels like I read the book.

Point being, I explored the idea that AI (and, at the time, molecular nanotechnology) would transform the human condition for years before the phrase the technological Singularity entered my personal lexicon. I read Ray Kurzweil’s The Age of Spiritual Machines: When Computers Exceed Human Intelligence in 2000 or 2001, and he didn’t use the phrase the Singularity at all in that book, though he would put it in the title of later books.

Writers like Robert Anton Wilson and Terence McKenna introduced me to the concept of the increasing pace of innovation and the accumulation of knowledge, and McKenna even proposed the specific date of December 21st, 2012 for an event he called the Concrescence, borrowing a term from Alfred North Whitehead.

While captivating in his delivery, McKenna arrived at that date via numerology and psychedelic revelation rather than through a plausible extrapolation of observable trends in technological innovation. You kinda had to be there to grok why I folded McKenna’s speculations into my worldview, but for a minute, the year 2012 was close enough to get excited about but still far enough away to pin my hopes to. I did both.

I can’t pin down when or how the Singularity worked its way into my conceptual toolkit, but my guess is that I got it from Ray Kurzweil’s blog. I followed his output closely in the years between my reading The Age of Spiritual Machines and the publication of The Singularity is Near in 2005–so closely that when I rushed to the Barnes & Noble in Fayetteville, Arkansas to buy it, I was disappointed to open the book and recognize every graph and illustration I saw. The book was a collection of the material I’d taken in piecemeal over the course of the previous few years, and I left the bookstore without buying it.

Like Kurzweil, the ideas upon which the concept of the Singularity rests reached me well ahead of the term itself. And, like Kurzweil, once the Singularity meme found its way into my worldview, it rose to become the primary totem of my concept of the future.

I repudiated my Singularitary obsession during my Peak Oil collapse years, but when I later rejected doomerism I did not take up the Singularity again. Even when diffusion and large language models started attracting attention and signalling that we might be approaching the knee of the curve² in AI development, I still wasn’t tempted to revive any Singularity-informed expectations.

Now I live in a world in which speculations about a hard AI take-off within two years are mainstream and the laptop class sees vocational obsolescence rushing toward them. And yet I’m still wary of the notion of the Singularity.

Why?

Two reasons occur³ to me:

1) It panders to the expectation that I will live to see How Things Turn Out, how the story ends. That hope or hubris has fueled doomsday cults and motivated Christians from the time of Paul to believe they were living in the End Times. They weren’t. The Singularity is an obvious substitute for that time-tested morsel of psychological bait. I gobbled it up back then, and I don’t see that it did me a lick of good.

2) Events are not playing out as they did in the speculative forecasts that captured my imagination in the 90s and early 00s before I contracted the Peak Oil collapse virus.

A crucial ingredient in Kurzweil’s vision and in the expectations of people spinning scenarios of rapid transformation was MNT (molecular nanotechnology); tiny machines that coordinate with billions of their kin to re-organize matter like atomic legos. True believers probably hold to the notion that MNT is beyond human capabilities but that it will follow on the heels of ASI (artificial super intelligence). Maybe, but I doubt it.

Another aspect of the Singularity Gospel that its prophets got plain wrong was the idea that the accelerating, recursive self-improvement of AI systems would happen outside of public consciousness; that it would only be of interest to STEM nerds. The claim that “Most people won’t know what hit them,” was central to the Singularitarian forecast. The nerd prophets would have been hard pressed to get any prediction more wrong than this.

The companies building this stuff are all the brain-children of Singularity True Believers who proselytize their infectious wish-fulfillment fantasies, seasoned with the fear that the bad guys will get there first, to raise capital and shape public policy. As such, talk of accelerating AI capabilities is everywhere.

It has to be, because the AI method that has demonstrated itself to deliver results is crazy expensive. It requires huge data centers filled with scarce, expensive, electricity-hungry hardware. The scaling hypothesis holds that increasing model size, training data volume and computational resources produces increases in intelligence. Some people reject this hypothesis, but in the face of rapidly increasing AI capabilities, their arguments seem more and more like ritual incantations meant to ward off evil spirits than sober assessments free of egoic commitment to previously expressed views.

Another thing the turn of the century Singularity prophets missed was social media. My introduction to the co-adapted meme complex that would later evolve into the Singularitarian faith came from books. Later, it proliferated in online discussions, but this was before the rise of MySpace and Facebook, the platforms that opened the internet to the normies in a big way. The advertisers followed the normies, and before diffusion models could whip up painterly images or LLMs could write poetry and robust code, social media companies set earlier forms of AI to the task of keeping people engaged with social media platforms.

The motivation was financial. Each page refresh–later, each upward thumb swipe in the endless scroll–brought a new monetizable opportunity to match an ad to an eyeball. Tech nerds, even the ones who started companies, lived in expectation of a post-capitalist Star Trek utopia. They failed to see how the profit motive would dictate the path of development. They failed hard. They didn’t envision dumb but tireless algorithms combing through idea space to find the magic elixir that would turn social media platforms into capitalist juggernauts and doomsday weapons of psychological warfare. They didn’t see how the profit motive and primitive AI would drive political polarization and the adoption of extremist ideologies as it obliterated attention spans and the tolerance for nuance and uncertainty.

Regular readers know that I’m no Marxist, but it doesn’t reflect well on Capitalism that without adult supervision, LLMs are about as trustworthy as kindergarteners when it comes to staying focused on a task. Even so, corporations and investors can’t wait to swap their human workers for AI. The AI IS NOT READY to replace humans, but that prospect is too enticing to capitalists to put off until the tech matures.

For early True Believers like myself, strong AI as we used to call it, would be developed outside of public awareness and proceed from motivations largely unmoored from crass commercialism or ideological struggle. It would emerge into public consciousness only when it was fully formed and capable of remaking the world in a timeframe so compressed that it would be viewed as a discrete event. The Singularity was to be a membrane that separated the past from the future. Most people wouldn’t see it coming, and when they burst through it to a post-Singularity existence, they wouldn’t know what hit them.

Shit ain’t playin’ out like that.

Competition for the limited supply of cutting edge, AI-relevant hardware is fierce. It drives US foreign policy and raises the stakes of Xi Jinping’s ambition to seize Taiwan. The fact that the manufacturing acumen necessary to produce this stuff is limited to an island off the coast of China is a glaring and perpetual humiliation and strategic vulnerability to the global imperial hegemon. Far from being an esoteric project that can gestate out of sight until maturity, AI is in the news with a vengeance.

Some people claim that it’s overhyped. Others retort that it is radically under-hyped. I think the former group is correct in the very short term and the latter folks have got the long view right. That’s my intuitive expecation, but the fact that talk of AI is everywhere is undeniable. Much of that talk consists of people clinging to outdated notions inherited from 20th Century science fiction, shallow takes meant to signal tribal affiliation, or the evergreen wish-fulfillment fantasy, but that doesn’t diminish the fact that most everybody is now hip to the fact that SOMETHING weird is coming.

But so what? The Singularity has confounded expectations, but that was to be expected. That doesn’t mean it isn’t happening. It may not sweep over human civilization like a rogue wave that rises without warning in a calm ocean. The fact that lots of people see it coming doesn’t diminish its power or transformational potency.

Doesn’t it, though?

The concept of the Singularity as a discrete event like a plane breaking the sound barrier that only a handful of forward-looking visionaries could conceive of before it happened is very different from something that is exhaustively pitched to consumers, venture capitalists and office holders and shaped to the needs of the military industrial complex and espionage agencies. The real Singularity will not proceed according to the Platonic ideal of the intelligence explosion. It will morph to match expectations and fears of nations, the business community and individual humans.

We see it coming, and it adapts itself to our expectations. The Singularity is tainted with the stink of inadequate vision. It is handcuffed to the baggage of flawed forecasts made by idealists for whom the practical imperatives of the marketplace and geopolitics were too mundane to contemplate. We’re not getting the pure Singularity of the science fiction nerds.

Better, I think, to accept that the name for what’s coming will be decided after the fact with the benefit of hindsight and after enough time has passed for the dust to settle and the rubble to stop bouncing. If it ever does.

1. First Word, an editorial published in Omni magaxing in January of 1983.

2. “The knee of the curve” refers to the point on an exponential growth curve where the rate of change begins to appear dramatically steep to human observers. Before this point, growth may seem gradual or linear; after it, it appears sudden and overwhelming. In Singularity discourse, the “knee” is often treated as the moment when accelerating technological progress becomes unmistakable and irreversible—especially in AI development.

3. I can only speculate about the workings of my own mind. I have not more access to my internal cognitive processes that Chaffo does of ets.