Today’s wrong question is this: “Can we prevent the A.I. crisis?” I’ve seen this question everywhere in the comments of my videos and headlines and in think pieces and articles and worried conversations at conferences. And people have very different ideas about what “the crisis” even means. Some think it’s A.I. replacing workers entirely. Others think it’s wealth concentrating till we’re all living in techno-feudalism. Some think it’s, well… even darker than that. (We’ll get there.)
But underneath, they’re all asking the same question: “Can we stop this?” Here’s what I need to tell you… probably not. History suggests we don’t “prevent” crises. And more importantly, we actually don’t WANT to. But, and this is crucial, we CAN shorten their pain.
A few weeks ago, I released an episode called, Who Buys Your Stuff, Robots? It wondered, if A.I replaces workers and workers lose income, who buys the stuff the robots make? The first category of “solution” to my question of “Who Buys Your Stuff, Robots?” is basically: B2B. “Josh, business to business can replace business to consumer. Businesses will just sell to other businesses. Problem solved. Boom. Done!”
Now this makes sense on the surface because B2B is big. Like really big. Almost twice the size of consumer spending in the overall economy. So I get why people think it could just replace B2C. But here’s the thing. B2B is bigger because it feeds B2C. Here’s what I mean: Steel companies sell to car manufacturers. Car manufacturers sell to you. If you can’t buy cars, manufacturers don’t need steel. Right? So the challenge with this argument is that B2B isn’t independent, it’s derivative. Follow any B2B chain far enough, you hit consumers. That’s where the value gets consumed. Remember the loop. At some point, somebody has to buy the thing, to consume it, to want it. If that final consumer doesn’t have income, the whole chain contracts backwards. B2B can’t replace B2C because it depends on it. A B2B solution is really just pushing the same problem upstream and eventually it ends up in the same place. So, we can’t solve the coming A.I. crisis with B2B. It’s a great hypothesis though!
The second category of solution is a little wilder. Something like, “Josh, A.I agents will just buy and sell and pay taxes and invest and do everything we do. The economy will just thrive without human participants.” All right, let’s dig into this one. These comments are describing a fully closed kind of machine economy, Where A.I agents are both producers and consumers. Value circulates internally. Humans just exit the loop. Now that’s not capitalism as we know it, of course, but more what we might call a “synthetic economy.” Honestly, it’s not impossible. We could technically build that.
So the question on this one isn’t “can this happen” as much as it’s “to what end?” If A.I agents are buying from other A.I agents and investing in assets owned by other A.I agents, paying taxes to governments that serve A.I agents, value has become completely detached from the reason any of us care about money in the first place. In our world, money only means something because it represents other stuff, right? Status or access to scarce resources, control over labor, fulfillment of needs. But if the robot buyers don’t actually need any of those things, the money transactions are just like bookkeeping. It’s not really an economy. It’s just like an accounting ledger, like talking to itself. The bigger question in this scenario is, “Why do we produce things at all?”
Well, we do that because humans want things. We compete for status and we buy things to try to find meaning and so on. But if machines are the only participants in the loop, what’s driving production desire? At best, it’s something like optimization. At worst, it’s just inertia. Turns out there’s a ton of sci-fi out there that explores the endpoints of this very idea. Some of you turned me on to things I’d never heard of before. And absolutely none of it is uplifting. We could design systems that exclude humans from economic participation, but why would we? This doesn’t produce any kind of world that most of us actually want to live in. So that one isn’t really a viable solution to the A.I crisis either.
The third category of solution is the darkest. It’s something like, “Josh, the rich don’t need the economy to work for everyone. They already have all the resources. Fewer people means fewer problems. And they don’t care about us anyway.” I’ve seen this one paired with depopulation theories, billionaire bunkers, all of it. Look, I get why this resonates. Wealth is concentrating. Billionaires are buying remote islands and yes, making bunkers. And it’s not just Scrooge anymore. Some tech elites do talk about reducing surplus population. “But some would rather die. If they’d rather die, then they better do it and decrease the surplus population!” It’s not crazy to think maybe this is their plan. So let’s take it seriously.
If we go down this path, we end up in a place where we ask what does a complex global civilization require to function? Because it’s that complex adaptive system that allows the ultra wealthy to live their ultra lavish lifestyles. And when we answer the question of what a complex global civilization needs to function, the logic of this solution breaks down pretty quickly because the ultra wealthy depend on systems that require a lot of other people.
Think about what it takes to maintain a billionaire lifestyle. Private jets that require global supply chains for fuel and parts and pilots and air traffic control. Luxury goods that require designers and craftspeople and logistic networks. Medical care that requires doctors and researchers and pharmaceutical supply chains. Technology that requires engineers and chip fabs and rare earth mining and electricity grids. You couldn’t run modern civilization with a hundred people and robots.
Feudal lords had all the land and weapons, but they couldn’t maintain Roman level complexity with tiny elite populations. Why? Because complex systems require large populations for specialization and redundancy and innovation. Without those things, the system simplifies. Less trade, less coordination, less resilience. Modern billionaires would face the same constraint. They might survive, but they wouldn’t be able to maintain current levels of complexity. Complex systems need the conditions that sustain complexity. Remove billions of people, you don’t get the same civilization, you get a simpler one.
So I don’t believe there’s any kind of depopulation plan, but I do think there’s a trajectory where capital optimizes for fewer workers, more wealth concentration and further isolation of elites. Not because anyone’s like planning it specifically, but because that’s where the current incentives lead. It’s not inevitable. It’s just the currently incentivized trajectory. We talked about this last week and here’s the good news. Trajectories are just numbers in a database that nobody gives a crap about. They need us more than they think. And we have more power than they seem to remember.
So those are the main varieties of crises that people seem to be imagining. None of those possibilities quite hold up under scrutiny, but all of our collective intuition, I think, still seems to be screaming, “There is a crisis coming and A.I. will have something to do with it!” I will admit: my spidey sense says the same thing.
At a macro level, humanity has seen this pattern before, multiple times across thousands of years, in every major civilization, I think. Our societies organize and then reorganize over and over. The word “collapse” often gets invoked in these situations, but to me, that word implies something that isn’t exactly happening, which is why I like this notion of reorganization better.
We’ve talked about this kind of transition before in relation to the Fourth Turning. There’s a Crisis and then a Reorganization, a Winter and then a Spring. So when people ask, “Can we prevent an A.I. crisis?: I can tell you what history suggests: probably not. The reforms we desperately need — in this case: income replacement, automation dividends, public A.I. stakes, new social contracts — these kinds of things don’t tend to happen before the disruption. They happen after. After the Triangle Fire. After the suffragists actually suffer. After the Great Depression. After the organizing conflict forces enough of us to CHOOSE.
So if we can’t preempt the crisis, what is the point of me telling you all of this? Why do I make these episodes week after week? It’s because of something that I call the Chaos Window. The Chaos Window is the period between when the old system breaks down and the new system takes hold. It’s the messy, like melty transition from winter to spring. It’s when society experiences maximum uncertainty and often maximum pain, but it’s also the season that contains maximum possibility. And here’s what history shows.
Some Chaos Windows are short and some are devastatingly long, but what makes them different isn’t luck, it’s preparation. And we’re going to go into this in more depth in next week’s episode, but for now, I’m going to summarize with a contrasting example from Chinese dynasties.
After the Han dynasty collapsed in the year 220, China fragmented for 370 years before the Tang dynasty reunified it. After the Tang collapsed in the year 907, it only took 50 years before the Song dynasty pulled it back together. Same pattern, same civilization, but one Chaos Window was seven times longer. What was the difference? After Tang, they had recent memory of unified governance. They had models, infrastructure, shared frameworks. So they didn’t spend centuries wandering. They spent 50 years. That’s 320 years of suffering AVOIDED.
The actions of the leaders in the Song Dynasty didn’t prevent the collapse. And in many ways, I don’t think that’s the goal anyway. The crisis is what allows for the change. Just like the decay of winter is what allows for the bloom of spring. We can’t avoid Winter, but we can help determine how long it lasts. The leaders of the Song dynasty did shorten the Chaos Window.
And that’s why this work matters! Not because we’re going to prevent the A.I. crisis, whatever it might look like, but because we’re pre-loading the solutions. We’re building vocabulary. We’re circulating frameworks and planting ideas. So when our Chaos Window opens, when the old system visibly breaks and everyone’s asking, “What now?” some ideas are already there.
Will they be perfect? No! But they will be ready to start testing. Is it noble to prevent suffering? Yes! Is it more noble to shorten the duration of suffering even if you know you can’t prevent it? Maybe! Because every year we shorten the chaos window is a year of fewer families destroyed and fewer lives derailed and less generational trauma created and a faster recovery for our overall society.
This isn’t small work. It is not for the faint of heart. It is civilization-level work. And it really frickin’ matters. We can do this, my friends. If you’re still here in this community saying Hello Tomorrow with me, I’m guessing this might get you a little fired up, too.
So this week’s Optimistic Rebellion is to look back in order to look forward. Think about the New Deal. It didn’t appear from nowhere in 1933. Those ideas — progressive taxation, social insurance, labor protections — they’d been circulating for decades. They’d been simmering quietly, gradually, until ready for the “suddenly.” When the Great Depression hit, the frameworks were ready to deploy. That shortened the Chaos Window. Same principle here.
“Who Buys Your Stuff, Robots?” plants a question. The Fourth Turning framework gives us context. Economic systems thinking provides us with vocabulary. Every week we are understanding the problems more and more clearly. And this allows us all to start brainstorming solutions NOW. And all of this will help us shorten our Chaos Window when it opens.
So when someone asks, “Can we prevent an A.I. crisis?” we know the answer’s probably “no,” but we also know the exact future isn’t inevitable. There are still so many choices to be made between here and there. Choices that will change the direction of the future we will actually get to experience. Next week, we’re gonna talk about the most possible high level scenarios I can currently envision about the future.
For today, our Optimistic Rebellion challenge is to start a Hello Tomorrow Brain Trust. What does that mean? Find three to five people in your network with different expertise and perspectives. Set up a monthly conversation. Could be over coffee, could be over a video call, whatever. Just to talk about what comes next. Feel free to have everybody watch or listen to the weekly Hello Tomorrow episode and tear it apart if you like. You’re not trying to solve everything. You’re building community and pre-loading ideas. So when our Chaos Window opens, we’re not starting from zero.
FDR didn’t invent his brain trust after the crash. He’d been developing it for years. And when the crisis hit, they were ready. More recently, Project 2025 did this too… of course, I’m encouraging us to move into a better future and not into some kind of retconned past, but the same principle’s there. We can do the same thing. Start the conversations now. Build the frameworks your community will need. Remember, we’re doing civilization-level work. It’s this work that will help us shorten our Chaos Window.
Keep the faith, my friends. The future is coming and together we’re gonna make it a great big beautiful one.
Original post with all source links: https://joshallan.com/2026/03/03/we-cant-prevent-the-ai-crisis/
