AI and Capitalism - The Perfect Storm

Are we creating the perfect feedback loop where capitalism finally finds the workhorse that allows it to constantly produce needs? As automation scales into the territory of thinking for humankind, we could be at the precipice where there is not much room for error in terms of how to steward this technology to benefit society at large. I want to be clear here: the problem I foresee is not exactly the development of AI — it is the system in which this technology is being developed. When we ignore this larger system driven by our capitalistic agenda, we lose sight of who or what this technology is really benefitting. My fear is that it is not for humankind, including those in the billionaire class, but rather a treadmill that fuels capitalism enough to tip humanity into meaninglessness. As much as we would like to see this as an era of abundance, without understanding the effect of how AI interacts with capitalism, I see this as an era of waste. Waste of material resources, waste of human potential, waste of authentic relational building, waste of what could have been if AI were developed "right."

For this slightly long essay, there are three factors in terms of how AI and capitalism can interact with one another that I want to discuss here:

• What Are We Producing?
• The Treadmill of Efficiency
• The Meaning of Time and Time of Meaning

What Are We Producing?

Many years back when I talked about the fear of how AI can be exploited within a capitalistic society, I often get back an argument that makes a comparison between the industrial revolution and the tech revolution. The arguments laid out before me tell a familiar story where people resisted change, fearing that the industrial revolution would take away their jobs and meaning —but look, everything turned out for the better. I had unconsciously accepted this argument without much thought, but the moment I examined it consciously, something didn't sit right with me. I realized there is something fundamentally different about the need to automate manual labor versus the need to automate cognitive labor. This is a distinction we need to delineate.

For me, the distinction lies in the concreteness of how this technology can help society. Where there is vagueness - exploitation takes root. When automation first began replacing manual labor during the industrial revolution, the shift from manual to automated labor was clear. For example, the steam train was used to speed up traveling and transportation of people and goods; in the textile mills, machines took over repetitive physical tasks, increasing output per worker significantly. Each use case was singular and concrete.

The more recent shift toward automating cognitive labor has a fundamentally different flavour. This technology is being developed in a vacuum of clear purpose - there is no shared answer to who it serves or what specific problem it solves. Unlike the industrial revolution, the use case is neither singular nor concrete: the paths by which this technology benefits society are foggy, and even the question of its necessity is hard to justify. There is no cohesive vision for how it will shape society for the individual. Within the vagueness of its use cases, the arguments for and against the technology becomes hazy too.

Today's technology feels much stronger in terms of manufacturing demand for its own use rather than responding to something people actually wanted. Language models and vision models are such cases: there was little pre-existing appetite so much as manufactured one, training users to want a product not because a need was there but because a technology was built. That distinction matters: technology built to serve a clear human need looks very different from technology built to create demand for itself.

André Gorz put it well: "We consume what we don't produce, and we produce what we don't consume." This manufactured chain of dependency is exactly how the current capitalist system operates in producing needs. With AI at the helm, that speed can only accelerate, generating new needs faster than we can even register wanting them. Within that system, this pattern can turn toxic, with the benefits accruing to whoever controls the technology.

That same acceleration is what drives the next part of the problem: the manufactured demand doesn't just sell products, it raises the baseline speed at which we're expected to produce. And to meet that pace, we now have the perfect tool - AI. It doesn't just speed up our output; it's capable of outsourcing cognition itself, the actual act of thinking. That shift, and what it might mean for the evolution of humankind, is a big enough question that it deserves its own space but here, I want to focus on how AI functions as the perfect workhorse for sustaining the accelerating speed at which we're required to constantly produce.

The Treadmill of Efficiency

The idea of efficiency espoused by Frederick Taylor set into motion many of the capitalistic drivers of today.

I see, between Taylorism and the ideology of abundance, a similar thread - a basic assumption about the human condition - that if we have a perpetual supply of what we materially need, then we can finally get on with life. Minimize what we don't want, maximize what we do want. So simple. With more and more technological advancements, we can find ways to automate away "work" that is unfulfilling and free up time to finally do the work and create the things that give us meaning. With AI, we can now take this even further. By distilling and enhancing human-collated knowledge, we might be able to lower the barrier to entry for new skills, unleashing latent human creativity and allowing us to reach our full "potential."

This promise isn't new, and it has already been tested.

It was predicted in Andre Gorz's Farewell to the Working Class that if we could increase productivity by 5 percent a year, then by the year 2001 we could technically be looking at a 20-hour work week, or even two-hour working days. That year has come and gone. What we've found instead is the opposite: we are producing more, faster, than at any point in history, and yet the amount of work required to earn enough for a decent life - let alone to earn free time - has only grown.

That should strike us as strange. How can we be producing so much, and still have so much left to do?

Part of the answer is that productivity and reward have quietly come apart. The illusion of production and the need for forward motion has allowed us to be blindsided by where all of our hard work is actually going. Money is a good obfuscator. It lets the distance between what we produce and what we receive go unnoticed, and it hides how much of what we produce is simply waste - waste that is, ironically, necessary to keep this capitalistic system producing at all costs. It also hides the actors who has gamed the system to earn money without real tangible output. The reward is getting stronger for the latter actor and the price for a honest living is getting higher for people who want to pursue them.

We are now in an era where humans have created necessities outside of Maslow's hierarchy of needs - technology, internet access, a smartphone, a computer, apps that are now necessary for survival. The more entrenched we become in this capitalistic system, the more unnecessary needs we create within human society, needs that are now necessary for survival within that very system. Necessary, even, to be deemed human. Which brings us back to AI. It is being sold on the exact same promise Taylor made, the same promise the ideology of abundance made: this time, efficiency will finally deliver the free time it has always owed us. But if the pattern holds, AI won't shorten the treadmill - it will simply raise its speed, and raise what counts as "keeping up."

The Meaning of Time and Time of Meaning

The treadmill created by the capitalistic system has other profound side effects. One is the speeding up of changes around us, which might dramatically alter how we experience time itself. Even for me, I can feel this change somatically - always waiting to move on to the next task, never really letting time settle.

There is also something to be said for the idea that meaning can only be created in the slowness of life, when there is stability and not too much change around us. In self-psychology, there is a concept known as a selfobject experience. These are experiences that we need to maintain a coherent sense of self. For an experience to function as a selfobject, it often needs to be a stable, internalized structure - one that takes time to create. A good example is the rituals we establish, which take on significance purely through the repetitiveness of the activity itself. In other words, only within this homeostatic environment can we form the associations that give our lives meaning.

But with the speeding up of our perception of time, our ability to create stable internalized experiences will reduce. This means we will not experience the fast-paced timelines as simply time that is getting faster but as a lack of coherent experiences that allow us to feel whole. Thus as technology further speeds up our perception of time, I fear there is a real danger that we are hurtling toward a society no longer amenable to meaning-making - one where meaning can only be pursued through individual sovereignty, beyond the constraints established by society. And the effects would not stop at meaning-making. A society that can no longer form or sustain shared, repeated experiences is also one where trust erodes - trust depends on a track record of predictable, repeated experiences over time, which is exactly what a faster, less stable world makes harder to build.

One crucial aspect of stability and repetitiveness of experiences is via how we create meaning with our "work" or how we identify with it. I would argue for time to be meaningful, "work" is necessary. Since the dawn of capitalism, this word has come to mean almost exclusively a monetary exchange of labour — but "work" encompasses much more than what we've been socialized to understand. Put simply: it is any form of action we take. To wash laundry is work, to paint a painting is work, to carve wood into furniture is work, to plant and harvest a garden is work, to breastfeed is work, to walk your dog in the morning is work. In Daoist philosophy, we cannot have good without the bad, the two co-exist together. From this standpoint, could it be that when we take away the pain of "work", we essentially take away the meaning of "work" too?

I ask myself, what happens when we only accept the parts of work we want and optimize away the work we don't with AI. How does this change the way we imbue meaning into what we produce? Now the technology is here - it really puts into question, why spend more time when we can spend less. But in a similar vein to before, maybe when we spend less time on the work, we actually lose the value of work too - both in what we produce and in what the work means for us.

Concluding Thoughts

The above are simply my own opinions - something I've wanted to clarify in my own head for a while. This treadmill that capitalism has created for us is now visible in the development of AI as well. We know we need to slow AI development down, but no one seems to know exactly why we can't, or what we could actually do about it. There's an invisible force pushing the race forward, ever faster. But to end on a positive note: I think humanity has a way of forcing itself to confront a problem - even when the path there is paved with suffering - and come out the other side into a better chapter.