They say that the Devil's greatest trick was convincing the world that he didn't exist. Likewise I would say that the greatest trick of societal collapse was making people focus on entirely secondary problems which seem like the most catastrophic things in history.
The narrative of so-called 'AI' does many, many useful things for the technological system:
§1. It redirects people's attention towards abstract discussions of "philosophy" and the future of our civilization, so that they do not pay as much attention to more boring but nonetheless tangible issues:
§2. It justifies lay-offs that are due to the decline of cheap energy (and thus the decline in value of most companies which aren't producing anything tangible), and justifies them as being due to 'AI'. The idea that our technology is so great that it's making us obsolete is certainly a more compelling narrative than the idea that most companies simply don't make anything very important, and are incredibly bloated under their layers of bureaucracy, meaning that something necessarily has to give way. 1
§3. It plasters over the nihilistic drive of the system and the people who support it, the fact that technological progress at all costs has no teleology to it besides extracting the very same nature from which we depend on, and coercing human beings to expand the reach of the system, with some pseudo-religion of how the machines will be the new bearers of consciousness and mankind was merely a stepping stone in that 'evolution'.
It also leans onto the short attention span on our society, which is addicted to talking about the most recent thing that has grabbed everyone's attention. Because 'AI' appears so disruptive, it presents itself as the only topic which is worth discussing, because it will make everything else obsolete, but meanwhile in the real world, energy is getting more and more expensive, with disastrous consequences to follow, and so-called 'AI' can only exist because of a massive influx of energy, and the raw materials required to build it (which themselves require a lot of energy).
§4. It reinforces the need for professionals and intellectuals to tell normal people what to do and what to think. An egregious amount of articles in Substack amount to telling people how they should feel about AI, mainly that they should embrace it at the risk of "falling behind", and what the future has in stock, mainly that it's going to be great for human beings as long as you go along the game of progress. Of course there are many other parts of the internet which are increasingly "doomer", anti-AI and anti-technology, which aren't necessarily better. But the fact that so much attention is given to people who hardly do anything besides sell ideas in the first place is concerning, because it means that people are handing out their intellectual sovereignty to others who have essentially no skin in the game. What happens when an intellectual makes a prediction about 'AI' which people believe in, and which turns out to be completely wrong? Nothing to them. Those writers make a living out of selling ideas, not out of selling good ones. If they can convince a bunch of people that their topic is the most important one, they can make money and build an audience, regardless of how wrong they are.
§5. It allows all the previous narratives about technology that didn't pan out to be conveniently swept under the rug. Remember when technology was supposed to free us from tedious work, so that we may focus on relationships, happiness, relaxation and creativity? Look at how many people now tell us that we need to work incredibly hard so that we may not "fall behind" the AI adoption curve, or that human art is essentially useless now. Or remember back when we were supposed to have flying cars, perhaps even teleportation, or electricity too cheap to meter? The first two, like many gimmicks in science fiction, never panned out, and people keep bringing up solutions for "renewable" electricity, which of course never pans out either, because Reality is not interested in our narratives and our politics. Meanwhile in the real world, we keep needing more and more fossil fuels in agriculture, mining, transportation, the industrial sector, and all sorts of other things, climate change might be even worse than some of the initial models predicted, and nothing is being done about any of this because it is impossible within the technological system to go backwards.
The main ways that I have noticed AI creeping into my life are, firstly, the fact that so many AI-generated images pop up in my social media feed (whenever I go on it, which I tend to do less and less), and even in real life. And secondly, the fact that seemingly every discussion revolves around AI nowadays. I find it incredibly tiring because every time the same set of cliches are being used in every side, to the point that hardly any conversation is happening anymore. The irony about the claim that AI can replace human beings is that it is becoming more and more true, because people are becoming more and more AI: entirely mechanical, repeating the same set of cliches that they have been trained on, and also very confident in the bullshit that they spout.
I am not interested in AI because firstly, it does nothing whatsoever about the central problem concerning energy (and complexity) that every civilization faces. It doesn't matter how smart your AI becomes, it has to run on energy, and most of our energy comes from non-renewable resources. You can't technologize your way out of the fundamental problems of energy and entropy.
Secondly, I don't care about it because it doesn't help me in making the things I want to make. For me, meaningful work has an external and an internal component. Yes, there is a result that I want to accomplish, something that exists out there for others to see, but just as important is the way that it grows me and attunes me to Reality at large. An artist doesn't just learn how to draw people, he learns how to observe, how to feel others from within, how to persevere in his craft, how to be attuned to small details that most hardly pay attention to, in other words, he reaches beyond his self. This is the value of art, an experience which is beyond the usual atomized, alienated self that the modern world calcifies in us.
§1. I don't believe that AI is conscious, in fact I believe that computers are less conscious than stones (I might write more about that point), but then again, it doesn't have to be to have an impact on our society. Most people working in the system nowadays are hardly conscious to begin with, so computer programs don't have to be either to potentially replace workers in it.
§2. The AI obsession is overblown, but this doesn't mean that AI is entirely a scam or a bubble. I do think it is massively overinflated and cannot pay back for all of the hype that sustains it, but it has some uses. LLMs are good at working with big sets of data and extracting generic queries, especially if the answer is non-ambiguous, typically the type of stuff that requires lots of reading with a semi-technical eye in mind, like lawyer stuff. In those situations, I could possibly see LLMs have a genuine use case in, though you still need a competent person overseeing the processes and the decisions at the end of the day. Another thing they're good at is sheer brute forcing of solutions and ideas, which in some fields can be surprisingly effective, which is why we sometimes see breakthroughs even in mathematics (I haven't looked deeply into this).
§3. AI doesn't need to be good to replace most "content creators" (a dreadful word at that). AI 'art' has already shown that a lot of people don't care that much about the quality of what they "consume" (another dreadful word), so it very well might be the case that artists become phased out of an income, not because AI art is so good that it has made them obsolete, but because the average person doesn't care about what is in front of them. Personally, I do enjoy seeing the process of an artist, and artists tend to be more interesting people than the average, so in that aspect there might still be a niche for them, but it is difficult to predict how this trend will pan out because art is so unconstrained by utility, and thus incentives.
§4. It is also difficult to predict how the effects of AI getting 'better', and the economy getting worse due to declining access of cheap energy will intersect. The collapse will not be evenly distributed, meaning that the tech builders might go on for many more years to come, but then again, how long can the financial markets maintain the bubble? Hard to say. Many facets will interact all at once, not just economic and physical, but also geopolitical.
§5. AI also doesn't need to be good in domains that aren't trying to create anything, but rather destroying something. You don't need to be conscious or particularly intelligent to shoot a human being. You don't need to have great relationship skills to spoof a person and hack organizations this way. And you don't need to have an intelligent system to establish digital surveillance. I would expect AI-assisted hacks to be on a significant rise, because people are desperate for money, and AI can also lead to more breaches in security through the shoddy software that is built with it. Digital surveillance, and surveillance in general, is something I expect to arrive in the next few years, as sad as it is to say. I wish I could do something about it, but a preliminary version of Chat control has already been rolled in the EU, and I can only see it getting worse for the years to come. I don't know how much AI robots and drones might be used for physical violence, but I also wouldn't be surprised if it became more of a thing.
§6. Finally, everything seems to indicate to me that AI is entering into 'model collapse' territory. After all, where is the data to make them 5x or 20x more 'intelligent'? The internet is becoming filled with AI outputs, so is it going to use its own outputs as its "new" data? What people call an 'exponential growth' might only be the first phase of a sigmoid curve. The idea of "exponential self-improvement" misses the fact that self-improvement can only occur when you have some type of contact with Reality, which is why natural selection didn't devolve into a schizoid, chaotic mess, since organisms had to interact with Reality, at the cost of making their rate of adaptation slower. If you only have your own model of Reality as a target, you have a sort of "mental inbreeding" effect where both your model of Reality and your behavior drift away from it, but the latter (behavior) becomes really good at optimizing for what the former (the targets) incentivizes.
This sounds abstract until you realize that this is essentially what happens in school: students who become hyper-optimized to answer closed questions in artificial tests, but who are utterly stupid when it comes to real life situations. 'Model collapse' is not restricted to AI, it happens with human beings too, and we could say that societal collapse is broadly speaking a collective version of that. We are doing everything according to mainstream economic theories, or whatever new fashionable theory tells us how society works, but those theories are no longer valid because they are not in tune with Reality.
I don't know how people in the future will make sense of our times, but I think they will definitely see the 'modern world' as a kind of dark age, when literacy, attention spans and intelligence dropped significantly, when a great deal of media has been lost due to it only existing in digital form—which to be fair, is not a big loss for mankind as a whole, basically none of the videos and short form content that most people engage most of their time with is worth keeping—and when a mass psychosis afflicted people at large, making them believe that the machines they built were somehow sentient and would become our overlords, as well as focusing on those narratives over practical concerns such as having a stable supply of food and energy in the long term, having robust social networks for people to cooperate, but also find meaning and belonging in, and overall, focusing on personal happiness and enjoyment of life, over following the creed of a technological religion which asks you to sacrifice your life at the altar of Progress.
I think societal collapse will be far, far more boring than the collective imagination sees it as, though no less horrifying, resulting in mass starvation due to the unavailability of fertilizers for industrial farming, the increasing frequency of environmental disasters, and the mass violence as an attempt to change the course of society. We live in the age of the spectacle, and so there is no reason to expect that what society tells us about itself, through the use of mass media, is in any ways important or true. Just like an addict focuses on everything else besides their own addiction, our society focuses on everything besides its fundamental addiction to technology and all the problems that are downstream of it. Our dark age is not one of ignorance, but one of overwhelming facts which are always, always about the wrong things. We are blinded by the light of mere facts, and can no longer see the forest for the trees.
1 See also this piece by John Michael Greer
Go back to the list of blog posts
2026-09-26