This was a fantastic and thought-provoking piece — thank you for taking the time to go beyond the usual surface-level takes on post-labor economics.
Reading your essay made me reflect on a complementary line of thinking I’ve been exploring. Like you, I start from the premise that general-purpose AI + robotics will outperform humans in both cognitive and physical tasks, making human labor mostly obsolete. But rather than framing the future in binary terms — owners vs. everyone else — I’ve been thinking of a possible trajectory in which a free-market capitalist economy could still function under these new constraints, just in a radically transformed shape.
Let me explain my hypothesis in full.
1. Ownership is not binary, but continuous
Instead of imagining a society divided into two classes (owners vs. non-owners), I think it’s more realistic — and more aligned with historical patterns — to model ownership as a continuous distribution, likely a Pareto or power law. In this world, you’d still have ultra-owners at the top (the top 0.1% with massive capital deployed), but also a long tail of small owners operating with vastly smaller but still non-zero capital bases.
What this implies is that the scale and nature of markets each “owner” can dominate depends on the size of their capital. Large-scale owners would naturally target mass-market global products, where the economy of scale makes their capital most productive. Smaller-scale owners would find it more efficient to target hyperlocal or niche markets where their capital can provide a competitive advantage.
For example, someone with modest capital might deploy their AI+robot fleet to produce fresh Chinese meals tailored to the exact preferences of neighbors in their district, leveraging their local knowledge and personal relationships as comparative advantages.
2. Capital becomes the new axis of specialization
In this world, the scarce input is not labor but capital (financial and infrastructural), which gets deployed in the segments where it yields the best return — much like Ricardo’s comparative advantage, but now between capital owners, not laborers.
That means:
- Capital-rich actors will dominate mass-scale production of generalized goods and services.
- Capital-poor actors will focus on small-scale, high-context, or custom work (enabled by AI).
- Markets segment naturally based on capital-efficiency, with minimal friction.
- This dynamic resembles today’s startup landscape — just massively amplified by how cheap, scalable, and intelligent machines have become.
3. Even “zero-capital” individuals might climb the ladder
Here’s where it gets more speculative — but also more hopeful. Suppose you’re a person with zero capital in this world. At first glance, you’re completely excluded. But in a system of free markets and cheap AI hardware, an unusual pathway emerges.
Even today, in many high-income countries, it’s not uncommon to see people leaving fully functional appliances or electronics on the curb — older laptops, monitors, fridges, washing machines — simply because reselling them isn’t worth the time or effort. As the value of hardware continues to fall, this behavior is likely to become even more common in the kind of future we’re discussing. Capital-rich individuals will routinely upgrade to newer AI-integrated systems and discard the old ones, possibly giving them away for free.
Now imagine you’re someone starting from nothing. In that future, “waste hardware” from 2040 or 2045 could include robotic systems or compute units with immense capabilities. A model from a few years ago might be outdated for enterprise use, but for a small entrepreneur, it could be equivalent to having 100 PhDs across fields like strategy, marketing, logistics, engineering, and law — all available 24⁄7 as advisors and operators.
With that kind of capability, someone with no capital could:
- Salvage a compute unit or basic robot
- Use its AI to analyze local context and identify what small good or service is missing or desired in their immediate surroundings
- Design a microbusiness that serves, say, 5 to 20 neighbors, selling a very small number of units per month
- Automate all logistics, design, communications, and production using old hardware
Here’s the key point: in a world where basic goods (shelter, food, electricity, water, connectivity) are extremely cheap, even very low absolute income might be enough to afford a core basket of living essentials. A person running a tiny hyperlocal AI-powered operation — think selling a few units of personalized goods per month — might actually live decently, without ever formally “working” in the sense we understand it today.
This sounds absurd — almost like cheating the system. People often respond: “How could someone live off of selling just a few items a month? That’s not work!” And yet, this disbelief is historically recurrent.
If you took someone from the year 1820 and showed them that today a person can earn a living by sitting in a climate-controlled office clicking buttons on a screen for 6 hours a day, they might say: “That’s not labor. That person shouldn’t be able to afford a house and food.” If you went back to 1000 CE, a subsistence farmer would be stunned that today, someone can earn their livelihood streaming video game commentary from their bedroom or managing spreadsheets.
The point is: every century, the ratio between effort and quality of life improves — not evenly, and not universally, but noticeably. Tasks that once required 16 hours of backbreaking work now require 2 hours of cognitive effort. And soon, even that may be replaced with capital deployment decisions and idea generation, assisted by machines.
In this context, the idea that someone could live well while producing very little — simply because machines do all the work — is not absurd. It’s a directional prediction, consistent with a long-term historical trend.
Lastly, even though inequality might increase in relative terms, the absolute quality of life at the bottom could rise significantly. Someone in the 0.5 percentile of capital ownership — almost no wealth, by comparison — might still enjoy a standard of living that exceeds today’s global average, purely because of how cheap and abundant key goods have become.
In short, yes: the system might be unequal, but it could still be inclusive and livable — especially if low-end capital can still be productive, and if old hardware + ubiquitous AI unlock new ladders of upward mobility.
4. Capitalism becomes more accessible
In many ways, the future you and I are describing lowers the barrier to entrepreneurship. Today, launching even a small business requires hiring human talent — which is expensive, scarce, and risky. In a world of AI+robots, talent becomes software and hardware, and the main challenge is knowing how to deploy it wisely.
This means the dream of “being your own boss” or owning productive capital becomes technically feasible for far more people, even if the upper tiers of ownership remain hard to reach — as they are today.
Of course, just as today, getting from $0 to 1x basic income-equivalent might be relatively attainable, but getting to 100x or 1,000x will require outlier vision, risk tolerance, timing, and maybe luck. Not everyone will become ultra-rich, but more people could operate independently.
5. Conclusion: A restructured but still capitalist future
In sum, I think your concern about extreme wealth/power concentration is deeply valid — and UBI or Universal Basic Capital might well be required to ensure human dignity. But I’d like to suggest that the future doesn’t have to be a two-class dystopia. If AI and robots become cheap and widespread enough, and if capital is at least somewhat accessible (via markets, salvage, or open hardware), we may instead get a highly stratified but continuous landscape of capital owners — from trillion-dollar operators to local entrepreneurs powered by second-hand AI.
It would still be capitalism — just with machines doing the work, and people figuring out how to channel them intelligently.
I’d love to hear your thoughts on how this fits with or challenges your framework. Thank you again for such an excellent piece.
This was a fantastic and thought-provoking piece — thank you for taking the time to go beyond the usual surface-level takes on post-labor economics.
Reading your essay made me reflect on a complementary line of thinking I’ve been exploring. Like you, I start from the premise that general-purpose AI + robotics will outperform humans in both cognitive and physical tasks, making human labor mostly obsolete. But rather than framing the future in binary terms — owners vs. everyone else — I’ve been thinking of a possible trajectory in which a free-market capitalist economy could still function under these new constraints, just in a radically transformed shape.
Let me explain my hypothesis in full.
1. Ownership is not binary, but continuous
Instead of imagining a society divided into two classes (owners vs. non-owners), I think it’s more realistic — and more aligned with historical patterns — to model ownership as a continuous distribution, likely a Pareto or power law. In this world, you’d still have ultra-owners at the top (the top 0.1% with massive capital deployed), but also a long tail of small owners operating with vastly smaller but still non-zero capital bases.
What this implies is that the scale and nature of markets each “owner” can dominate depends on the size of their capital. Large-scale owners would naturally target mass-market global products, where the economy of scale makes their capital most productive. Smaller-scale owners would find it more efficient to target hyperlocal or niche markets where their capital can provide a competitive advantage.
For example, someone with modest capital might deploy their AI+robot fleet to produce fresh Chinese meals tailored to the exact preferences of neighbors in their district, leveraging their local knowledge and personal relationships as comparative advantages.
2. Capital becomes the new axis of specialization
In this world, the scarce input is not labor but capital (financial and infrastructural), which gets deployed in the segments where it yields the best return — much like Ricardo’s comparative advantage, but now between capital owners, not laborers.
That means:
- Capital-rich actors will dominate mass-scale production of generalized goods and services.
- Capital-poor actors will focus on small-scale, high-context, or custom work (enabled by AI).
- Markets segment naturally based on capital-efficiency, with minimal friction.
- This dynamic resembles today’s startup landscape — just massively amplified by how cheap, scalable, and intelligent machines have become.
3. Even “zero-capital” individuals might climb the ladder
Here’s where it gets more speculative — but also more hopeful. Suppose you’re a person with zero capital in this world. At first glance, you’re completely excluded. But in a system of free markets and cheap AI hardware, an unusual pathway emerges.
Even today, in many high-income countries, it’s not uncommon to see people leaving fully functional appliances or electronics on the curb — older laptops, monitors, fridges, washing machines — simply because reselling them isn’t worth the time or effort. As the value of hardware continues to fall, this behavior is likely to become even more common in the kind of future we’re discussing. Capital-rich individuals will routinely upgrade to newer AI-integrated systems and discard the old ones, possibly giving them away for free.
Now imagine you’re someone starting from nothing. In that future, “waste hardware” from 2040 or 2045 could include robotic systems or compute units with immense capabilities. A model from a few years ago might be outdated for enterprise use, but for a small entrepreneur, it could be equivalent to having 100 PhDs across fields like strategy, marketing, logistics, engineering, and law — all available 24⁄7 as advisors and operators.
With that kind of capability, someone with no capital could:
- Salvage a compute unit or basic robot
- Use its AI to analyze local context and identify what small good or service is missing or desired in their immediate surroundings
- Design a microbusiness that serves, say, 5 to 20 neighbors, selling a very small number of units per month
- Automate all logistics, design, communications, and production using old hardware
Here’s the key point: in a world where basic goods (shelter, food, electricity, water, connectivity) are extremely cheap, even very low absolute income might be enough to afford a core basket of living essentials. A person running a tiny hyperlocal AI-powered operation — think selling a few units of personalized goods per month — might actually live decently, without ever formally “working” in the sense we understand it today.
This sounds absurd — almost like cheating the system. People often respond: “How could someone live off of selling just a few items a month? That’s not work!” And yet, this disbelief is historically recurrent.
If you took someone from the year 1820 and showed them that today a person can earn a living by sitting in a climate-controlled office clicking buttons on a screen for 6 hours a day, they might say: “That’s not labor. That person shouldn’t be able to afford a house and food.” If you went back to 1000 CE, a subsistence farmer would be stunned that today, someone can earn their livelihood streaming video game commentary from their bedroom or managing spreadsheets.
The point is: every century, the ratio between effort and quality of life improves — not evenly, and not universally, but noticeably. Tasks that once required 16 hours of backbreaking work now require 2 hours of cognitive effort. And soon, even that may be replaced with capital deployment decisions and idea generation, assisted by machines.
In this context, the idea that someone could live well while producing very little — simply because machines do all the work — is not absurd. It’s a directional prediction, consistent with a long-term historical trend.
Lastly, even though inequality might increase in relative terms, the absolute quality of life at the bottom could rise significantly. Someone in the 0.5 percentile of capital ownership — almost no wealth, by comparison — might still enjoy a standard of living that exceeds today’s global average, purely because of how cheap and abundant key goods have become.
In short, yes: the system might be unequal, but it could still be inclusive and livable — especially if low-end capital can still be productive, and if old hardware + ubiquitous AI unlock new ladders of upward mobility.
4. Capitalism becomes more accessible
In many ways, the future you and I are describing lowers the barrier to entrepreneurship. Today, launching even a small business requires hiring human talent — which is expensive, scarce, and risky. In a world of AI+robots, talent becomes software and hardware, and the main challenge is knowing how to deploy it wisely.
This means the dream of “being your own boss” or owning productive capital becomes technically feasible for far more people, even if the upper tiers of ownership remain hard to reach — as they are today.
Of course, just as today, getting from $0 to 1x basic income-equivalent might be relatively attainable, but getting to 100x or 1,000x will require outlier vision, risk tolerance, timing, and maybe luck. Not everyone will become ultra-rich, but more people could operate independently.
5. Conclusion: A restructured but still capitalist future
In sum, I think your concern about extreme wealth/power concentration is deeply valid — and UBI or Universal Basic Capital might well be required to ensure human dignity. But I’d like to suggest that the future doesn’t have to be a two-class dystopia. If AI and robots become cheap and widespread enough, and if capital is at least somewhat accessible (via markets, salvage, or open hardware), we may instead get a highly stratified but continuous landscape of capital owners — from trillion-dollar operators to local entrepreneurs powered by second-hand AI.
It would still be capitalism — just with machines doing the work, and people figuring out how to channel them intelligently.
I’d love to hear your thoughts on how this fits with or challenges your framework. Thank you again for such an excellent piece.