The dystopia of abundance
As much as I want to avoid being another guy talking about AI in this AI-slathered moment, I’m writing this because I can’t help but feel like there’s something important missing from the discussion. Of course, there’s no shortage of doomsday scenarios floating around, and to be clear, I lean fairly pessimist on them, so if I’m skipping over them it’s not because I don’t think they’re likely, but more just because I don’t think there’s much we can do about it: Are we on the cusp of a Skynet situation? An AI-induced nuclear apocalypse? An all-encompassing, totalitarian surveillance state? Honestly, probably. But then what?
But there’s the other view on AI—one so many oddball silicon valley types get very excited about—which is that if properly managed, AI will bring about utopia. When asked to describe such a utopia they all describe the same thing: a world of pure decadence (or as they call it, abundance) where everything anyone could ever want is available the instant it pops into their mind—where thrift and prudence become legacy virtues to be written off as side effects of a more primitive time in our species’s history—where suffering ceases to exist, and energy (and all that is downstream of energy) becomes infinite. I have absolutely no doubt that this version of the future is possible. But what’s odd is that no one, not even the very smartest people who are so keen to bring this utopia into existence, seems to have unpacked anything beyond the first order effects of implementing such a fantasy. Even people like Tristan Harris of the Center for Human Technology, who is constantly touring the world warning us about the AI disaster scenario that he fears is right around the corner, seem to hold up the abundance utopia scenario as the happy example of what’s possible in unlikely event we get our shit together. His ability to extrapolate into the future, which appears finely honed when he makes the case against the path we’re on, seems to be completely impotent when it comes to the other path: the one he’s proposing we strive toward.
In a recent interview with Sam Harris, Tristan quoted the famous Upton Sinclair line, “It’s impossible to get a man to understand something when his salary depends on him not understanding it,” in reference to the AI CEOs who, despite the acknowledged risk, seem incapable of reversing course. But the irony there, is that Tristan himself, as the leader of the “Center For Humane Technology”, an organization built on the idea that technology is fundamentally good, appears to be bound by the very same shackles.
Paul Kingsnorth, in his recent book, Against the Machine, points this out. He says:
What is fascinating about this talk is the palpable tension between the overall message and the details it contains. The message, as you might expect from Silicon Valley, is one of cautious optimism. AI is a good thing, and can be used for our benefit. Technology as a whole can be “humane” and “aligned with humanity’s best interests.” Not all is well right now—AI is currently unsafe and needs to be reined in—but if we work harder and smarter, we can make this happen. These are the words that a mainstream audience in a rationalist culture wants to hear. Perhaps they are the only ones it is able to hear. And yet the two presenters [and here he’s referring to Tristan and his foundation partner] do a disturbingly good job of undermining their own message.
The AI optimist calls the ideal future “utopia” for a single, simple reason: because it is a future in which (they imagine) everyone can have everything they want. It is simply taken for granted that the satiation of want is a fundamental and self-justifying good. Silicon Valley elevates this notion from fact, which requires proof, to religion, which doesn’t, which is ironic given the industry’s usual zeal for data-driven decision-making. Apparently that doesn’t apply here: Getting what we want is good. Period.
We don’t have to look far, though, to see just how misguided that ideology is. In fact, it happens to be the exact topic of one of the most famous and widely read books of the last century—a book that continues to be referenced in literature and pop culture and assigned as required reading in schools to this day. Brave New World, by Aldous Huxley, is almost literally a warning about the consequences of building exactly the type of abundance society the AI people are vying for.
In Brave New World, we see the type of empty, inhuman society that emerges when its primary goals are the immediate fulfillment of all desires and elimination of all discomfort, pain, and suffering—precisely the goals the AI utopians believe to be noble and good. Huxley set out to show that while most of us imagine “oppression” as coercive, with people being forced to live a certain way against their will, it is actually much more likely that we cooperate willingly with our oppressors, who use pleasure to lure us into a permanent state of distraction and passivity.
But here’s what I find odd: We all read this book. We all agreed that its message was plain and clear, and poignant, and important to teach in our schools. So how exactly is it that we all agree to categorize Brave New World as a brilliant example of dystopia, while at the same time dedicating ourselves to building exactly the sort of society the novel warns us against?
The problem with abundance, in the sense that people are using the term, is that it can’t actually exist. Why? Because human want is not immutable. Yes, the current objects of our desire can be made disposably abundant, but that is not the same thing as providing us with everything we want, because the very things we want are downstream of what we already have, and of what’s available in the world we live in. We want things because they have value today—usually because they are not trivially available today. But the AI-utopia world is not today’s world. That’s the world of tomorrow. And giving everyone what we want today changes what we may want tomorrow.
These are what we call the second-order effects.
Material things like a beautifully crafted piece of furniture, a high-end car or an exotic vacation have value because today, those things cannot easily be obtained. But what happens when they can be? What happens when beautiful furniture is abundant—free and instantly available in any quantity, and to everyone in the world? Is there any sense of satisfaction and gratitude in possessing such furniture then? Do people still want it? When anyone who sees what you have can simply make an identical object appear in their own house instantly and for free, do you nevertheless continue to want such things? Or do our wants shift to something else that cannot be so easily obtained?
Or what happens when we make access to travel abundant, so that you and everyone else can travel to any place in the world, instantly, and for free? Do Paris and Tokyo continue to be desirable destinations that people look forward to visiting? For how long can a place like Paris withstand the presence of “any single person who ever, even for a brief moment, felt like going to Paris” before it ceases to be Paris? Paris is Paris because it is Parisians—people who are both the product and the source of French culture—living there, maintaining, shaping, and flowing through it. When the vast majority of Paris becomes people visiting from elsewhere, and the vast majority of would-be Parisians aren’t even in France, but are themselves off visiting Tokyo or wherever it is they feel like going, Paris, and everywhere else in the world, no longer exists. The whole world simply becomes one mono-place, nowhere much different than anywhere else.
When this becomes reality—when access to travel becomes “abundant” do we still even want to travel? Or do the second order effects of making travel abundant undermine the entire point of travel, shifting our wants onto something else?
Or perhaps you’d say France couldn’t possibly let an unlimited number of people into Paris at once, since simply as a matter of geometry, they wouldn’t all fit, so there’d obviously need to be some kind of limit—a queue— put in place to keep things manageable. But if Paris did that, is travel still “abundant”, or is it once again scarce? If the net result of making travel abundant is that we need to make rules to keep too many people from coming, are we not just recreating scarcity? And if so, what was the point of abundance?
The implications of AI abundance aren’t limited to just physical things, either. AI plus modern pharmaceutical technology will inevitably give us the ability to mold our bodies any which way—stronger, thinner, faster, younger—but is there any satisfaction in being in incredible shape when being in shape does not require any effort, any force of will, any endurance of pain at all? When everyone around you is also in incredible shape? Or, like everything else, does giving us the thing we want eliminate the reason we had for wanting it in the first place?
Or consider the intellectual domain—are the ideas we create—an essay for example—valuable in a world where genuine insight and analysis requires no time and no effort to create? Prior to AI, if an essay existed, it meant that someone took the time and spent a portion of their finite mental energy in order to create it, and that in itself made it worth consideration. But what value should we attribute to an identical set of words on a page if we knew that no time or thought at all was required to produce them? Are we going to take the time to read something we know no one took the time to write? Or are we just going to pass it along to another AI, like the Task Rabbits we’re becoming?
The implications here also cascade into art. Say you go and see a film. Until recently, every artistic choice—the lighting, the set design, the color grading—all had to be invented, created, and painstakingly applied. They all combined to create something moving and meaningful. Wes Anderson films, for example, all have a unique color palette that is unmistakably his, and which is a big part of what his movies are. What happens then, when AI makes all those qualities abundant—as trivially accessible to the masses as applying an Instagram filter? What happens when any old student film footage can simply be fed to AI with the request that it be “reshot” exactly as if Wes Anderson had done it. Does the film emerge on the other end, now carrying the gravitas of one of Anderson’s films? Or more likely, does it become one of a thousand leeches attached to Anderson’s art, draining it of its vitality, making the remarkable unremarkable.
Wes Anderson’s films may retain their gravitas on account of the fact that they were definitively created via human effort, but everything created after the birth of AI is doomed to live in a shroud of skepticism: Yes, the film might be worth consideration if the author did indeed put in the creative work to create something moving himself, but how can we know for sure that he did? And are we really going to dedicate time to seek meaning in a piece of art if, in the back of our minds, we know that it may not have any intention or meaning behind it at all?
This is basically the story of Instagram filters. Once upon a time, those filters promised to give everyone the ability to make their photos look like those being very intentionally taken by enthusiasts who went to great lengths to achieve a look that previously only came from shooting on a vintage Lomo camera. But by making that effect abundant, Instagram made that entire look worthless, authentic or not. The sad reality is that this is about to be the story of all media. Things that are difficult and scarce—things that are inherently frictional—are what we deem valuable and worth pursuing. But make those things abundant, and they vaporize. They first become slop. Then, as people actively seek to avoid association with slop, they disappear entirely. When, after all, was the last time you saw someone put an instagram filter on a photo?
So, here I am, making the case against abundance, and naturally, the first thing people’s minds tend to do when they hear the kinds of things I’m saying right now is point to the fact that scarcity leads to suffering, and to the obvious advances in our quality of life thanks to technologically-mediated provisioning of basic human requirements like food, water, shelter, and medicine. To express doubt about whether life will be good in a world of abundance is often interpreted as equivalent to suggesting that some people simply deserve to starve or die of tuberculosis. But that’s not the right way to think about it.
There is a big difference between “no one should be hungry” and “everyone should be get to live on Michelin star meals.” The former is indisputably true, but when you listen to the conversations about the world of AI abundance, it’s not the former that these people are talking about. Abundance, to Sam and Dario, doesn’t mean enough. It means “whatever we want”. Total freedom from all limits and constraints. And therefore access to medicine doesn’t mean antibiotics, vaccines, and cancer treatment, it’s plastic surgery, Ozempic, and the right to choose the height and gender of babies. Access to shelter doesn’t mean a modest apartment in a safe part of town, it’s the right to live anywhere, in any manor we choose, and to remodel our bathroom every time a Tiktok trend changes. To the AI optimists, it means a world without want.
In Brave New World, the people are all walking around popping “Soma”s, the fictional opioid designed to eliminate the pain and anxiety of a meaningless existence. We are already beginning to feel that pain and anxiety today. Without want, we will have nothing to pursue and nothing to achieve, and that meaningless existence will become our entire reality. Pursuing want imbues our lives with meaning. Absent that meaning, the thing that we want becomes meaning itself—the one thing abundance can never provide.
At this point you may be wondering, “Doesn’t AI live on computers? What does it have to do with real world things like food and real estate?” And this is, in my opinion, another place that the current discourse fails to map its goals onto their consequences. Yes: AI today is largely confined to the digital world, on account of the fact that it’s vastly cheaper and faster to use it to interact with the bits and bytes it already has direct access to. But it won’t stay that way. The only reason AI is not universally present outside of a computer today is that it’s missing an interface to the real world. A plumber still needs to get on his hands and knees under the sink to fix a leak and chefs still need to bang around in the kitchen to make food. But there is no scenario where it simply stays that way, and the line between digital and physical is already blurrier than it might seem. AI generating an image of an oil painting on a screen may not seem like much, but there are now also devices akin to 3D printers that are astonishingly good at reproducing the masterworks of history’s best oil painters, which can bring those oil paintings into the physical world. The limitation on these types of digital to analog interfaces is not the brainpower, but simply the physical form of the device the brain is loaded onto. And though that may sound like a significant limitation, at the end of the day, any brain given a physical form similar to ours as humans would then be perfectly capable of doing anything a human could. The advances made over the last decade or so at companies like Boston Dynamics should be an indication that this isn’t so much a hard limitation as a financial one. Robots are expensive, which makes iterating on them slow. But prices will fall. When they enter human salary ranges, they will become more common, which will lead to faster iteration, more capable robots, and even cheaper prices. There simply is no scenario where only knowledge work disappears, but the plumbers and welders homebuilders remain gainfully employed. You can already today just take a picture of your under-sink plumbing, upload it to Claude, and ask it how to fix a problem. The only thing that’s missing is Claude’s ability to be physically present under the sink, with access to the supplies and the appendages it needs to do the work itself. That won’t be true for long. And when that happens, it will be everywhere else too—making meals, building houses, practicing medicine, creating every bit as much abundance in the physical world as it currently does in the intellectual one.
—
For all their problems, there’s something that ancient religions seem to have understood that the modern techno-religions do not: all of them seem to recognize that the ascetic values of friction and denial are not things to be eliminated, but to be embraced as a critical part of the human experience. They are not bugs, but features. Denial, after all, is inherent in the very concept of a “pursuit” itself: what we pursue must by definition be out of reach.
In contrast, the materialist religion of today, is the opposite: a parasitic activity which uses abundance as justification to provide a small group of people with meaning by stealing it from everyone else. The people who work at Anthropic and OpenAI who tell us these fairytales about abundance and wealth do so because they themselves derive meaning out of designing these systems and machines that iron the inefficiencies out of our lives. But when they release them into the world, they transform the communities working “inefficiently” together, strip them of their ways of doing and being—the vernacular techniques, “superfluous” practices, the songs and dances (ie “culture”) that go with them—and turn their work first into a punchcard job in which field becomes factory, and then into nothing at all. And after that is done, those system designers return to work and, faced with the question of what to build next, look for ways to do it all over again—to straighten crooked paths, to improve outputs and yields by stripping out even more “inefficiencies.”
It’s a pyramid scheme.
In the end, when all the businesses that harvested others for meaning are themselves cannibalized, and perfect, efficient abundance is achieved, late capitalism and communism converge on a singularity: whether a single megacorporation is dolling out abundance it’s the government doing it, the difference is only semantic.
. . .