For two centuries, building anything required a concentration of people — bodies crammed into the same place, locked into the close, involuntary relationships we politely call “collaboration.” AI is the first thing that quietly dissolves that requirement. So the question I care about isn’t whether it writes code or spits out slop — it’s what happens to how we treat each other once we’re no longer forced together.

Borderline Augmentation

A lot of the talk is about AI enabling creativity and a new wave of creators, while its opposition calls it slop and a useless waste of water — or some other bullshit that losers tell the ones who build.

Will most of the products made with AI become billion-dollar companies? Obviously not. Just like not all painters end up in the Louvre. But it will unleash creativity in a lot of people who were held back by the tools. Will it be world-class creativity? Obviously not. But will it make people more content? Like any creative process — obviously yes.

And there are a lot of masters of the craft who have defended, and will keep defending, their craft with sharp words. I was thinking recently that I spend something like tens of hours a week coding outside of my work projects. I hadn’t done that since my late youth. And what happened in my late youth? I went to work at tech companies and started working with more senior people whose main interest wasn’t building anything — it was their own, mostly stupid, social games. I was clever enough to figure that if it’s all a social dance, why even bother building anything — so go into management and skip all the stupid shenanigans down in the dirty tech pit for nothing. For a long time I lost all interest in building anything by hand. But I never felt any real contentment from management. It was a job, and I did it pretty well. The only contentment I felt came from seeing what I could build with the teams I had.

Now I see a lot of senior people getting back to building like they used to — CEOs, CTOs, PMs, people well established in their roles, finding the urge to do something with their hands again, and many of them do it beautifully.

And I see a lot of professional coders arguing about slop, and look — I think these are the very people who made me leave building behind: the sleazy mud-pit creatures who argue about every possible thing except actually building something, thinking they’re the best engineers in the world while deep down they know Opus is better. So they use it in the closet while screaming about vibe coding — or they don’t use it at all, because they haven’t built anything that works in years.

So when I was starting my career 20 years ago, I had to talk to people to learn, get through all the hazing, take the humiliation to reach the knowledge — half of which was wrong or outdated anyway. Now I have a partner. It listens. It knows far more than I do, it lacks some of my expertise, but it treats me as a peer — not as an adversary or someone to bully.

Humans still gatekeep resources and industries, so yes, your much better vibe-coded app won’t surpass the Silicon Valley poster children, and you still have to deal with sleazy people to build a career or a business. But at least now a lot of people know it can be different — that this isn’t the norm. And that’s a very empowering thing to know.

The Final Enlightenment

Since the late industrial revolution we’ve been wrestling with abundance, and it increasingly seems that abundance is more dangerous to a society than famine — we got the deadliest wars not because we were fighting over food, but because we were abundant enough to afford starting them. As the Haber-Bosch process modernized agriculture, the societal tensions only grew sharper, leading to two world wars and countless deadly conflicts — and to even deadlier societal experiments: Communism killed on the order of 100 million people, rivalling both world wars combined.

All of these societal processes revolve around one topic — how to better organize resources, and which system can do it best. It reached the management space too: since the 1920s, new forms of organizing work, and solutions for handling bigger and bigger projects, became common inside institutions, alongside the question of how to organize governance at scale. You can argue about when self-managed work collectives became a common topic of discussion — I’d trace it back to the French Revolution, but really to the 1920s, though we can agree the major academic wave started in the 1960s. And just like the large-scale governance debates, the smaller, organization-wide ones all circled the same simple idea: how to make people better.

There’s another topic that has nothing to do with world wars or revolutions, but has been circling for the last hundred years: becoming less fat. For at least a century we produced thousands of different approaches, most of them reducing to the same simple problem — how to make people better.

Then Ozempic appeared. And people started losing fat without any enlightened spiritual experience or major psychological breakthrough. More than that — the psychological and spiritual changes started to follow the physiological one.

So for a hundred years we’ve been talking about how to make people better. We coined nice names — self-managed organization, teal management (after the teal level in Spiral Dynamics) — for how to make people self-organizing and collaborative instead of monkeys fighting over bananas in the trees. Without any real success, to be honest. You can throw rocks at me, telling me how you built your self-managed business — but I’ll throw a bigger rock back: you have the same hierarchy as everyone else, just with an extra layer of social dance on top, a bit better hidden. So you’re either closing your eyes to it, or being delusional.

And now, as AI-pilled people appear — with their sense of how it can be and how it shouldn’t be — we’re finally starting to see what the theorists promised us: people organizing around the project, not around the hierarchy. It’s a small fire right now, but with wider adoption — shrinking org sizes, broader shifts in employment — it will, in my opinion, become a much bigger trend. The more AI takes part in our lives, the more voluntary our participation in work and society becomes. We’ll no longer be required to work — but still welcome to.

Paradise Is Not Promised

When I talk about voluntary participation, I don’t believe AI will bring some kind of utopia to our world — whether or not we reach AGI, and whether its improvements come in a leap or stay incremental over the coming decades. The idea of a common paradise is deeply rejected by humans themselves, and every attempt to build that paradise through power has brought hell on earth instead. So I’m more afraid of the common-paradise idea, and of the people who stand for it, than I am of a harsh societal change — one bigger and more painful than the Industrial Revolution.

My mental model of how society will look as AI advances is based on how previous swift societal changes played out in the past. I don’t think modelling the future on past events is the best approach, but I have no better data to model it with — and I think the human mind hasn’t changed much over the past millennia. As I said before, that’s the main limiting factor on ever reaching the common paradise (without the bringing-hell-on-earth part, which we manage pretty enthusiastically).

The big changes of the past happened in cities. Rome was the model city for the millennia after its fall (we still hold gladiator fights on the emperor’s birthday in the metropolis). Madrid was the example city for the colonial era. London was the model city for the Industrial Revolution — the one every other city copied the moment it got its hands on the steam engine. Now we have Silicon Valley as the world’s cultural and ideological powerhouse. So you can read Silicon Valley’s problems and achievements as the template for the rest of the world in the coming decades. And a place that’s beautiful for some will be hell for others. Just as Rome was hell for many slaves, and London was hell for many cockneys, Silicon Valley has introduced a new kind of hell — an abundance-fuelled drug and homelessness epidemic, social isolation, broken relationships, insane inflation on the basics like housing. This is the wave I expect to wash over the developed world in the coming decades.

But I don’t want to be a doomer. This whole essay is about the incentives that drive human communication, and the idea that AI will change the forced nature of it — the forcedness the Industrial Revolution and post-industrial society imposed, requiring people to live in very close, forced relationships with each other, which we call ‘collaborative’ work. I prefer to call it ‘concentrated’ work: you need a huge concentration of people to get anything done. If we overcome the concentration part, at least some people will get to live in a more voluntary society, organized around values and ideas instead. At least I want to believe in it — and to do my humble bit to make it happen. Because I don’t think it’s a promised part of where AI is heading. But it’s a possible one.