Back to writing

    The Coming Age of the Solopreneur Warlord (And Why It Probably Won’t Happen)

    If you spend any amount of time on “AI Twitter” (or 60% of your LinkedIn feed), you have encountered The Narrative.

    It goes something like this:

    Until recently, to build a billion-dollar company, you needed thousands of employees, massive amounts of capital, and huge office buildings filled with mediocre coffee machines.

    But now? Now we have Generative AI. We have agents. We have automated workflows.

    The Narrative says that very soon —like, next Tuesday— a single, highly motivated 19-year-old living in a basement is going to chain together 500 AI agents, automate every function of a Fortune 500 company, and disrupt an entire industry before finishing their morning can of Monster.

    No employees. Zero marginal cost. Pure profit.

    This leads to a very specific, very terrifying vision of the future. Let’s use the term Techno-Feudalism.

    In this vision, a handful of hyper-leveraged “One-Man Empires” capture 99% of the world’s economic value because they own the best bots and the best workflows. The rest of us? We’re just sitting around, redundant, hoping one of the AI Warlords accidentally drops a digital crumb of UBI our way.

    It’s a compelling story. It hits all our primal panic buttons about relevance, survival, and terrifyingly smart nerds taking our lunch money. Perfect LinkedIn clickbait.

    But if you step back from the ledge and look at how complex systems and economies actually work, this vision of the future starts to look less like an inevitability and more like a misunderstanding of how the world works.

    Let’s dive in.

    Part 1: The Case for the AI Warlord (Why We’re Panicking)

    To understand why the panic feels so real, we have to look at the “leverage multiplier.”

    Throughout history, humans have invented tools to increase our leverage—to get more output for the same input.

    • Level 1 (The Spear): I can kill food without having to wrestle it. ( Leverage: 2x)
    • Level 10 (The Plow): One farmer can feed ten families instead of just their own. (Leverage: 10x)
    • Level 100 (The Factory): We standardize production and use steam/electricity to make things faster than human hands ever could. (Leverage: 100x)
    • Level 1,000 (The Computer/Internet): We automate information processing and distribution. (Leverage: 1,000x)

    Every jump in leverage changed the world and freaked everyone out. But AI feels different. It feels like we just jumped straight to Level 1,000,000.

    Until now, even a “10x engineer” or a brilliant CEO needed a village of regular humans to execute their vision. You needed people to write the marketing copy, answer customer support tickets, manage the payroll, and debug the CSS.

    The Techno-Feudalism thesis argues that AI agents are about to replace that village.

    Mental Visual: A stick figure wearing a crown, sitting on a throne made of GPUs. Below him, thousands of tiny robot stick figures are frantically running around doing all the work. Outside the castle walls, millions of sad, unemployed human stick figures are holding signs that say

    If one person can wield the productive capacity of 10,000 people, two things seem obvious:

    1. That person is going to get unfathomably rich.
    2. Those 10,000 people are going to be unemployed.

    If you project this out linearly, you get a world with a few trillionaire god-kings and billions of peasants. It’s a neat, tidy argument.

    It’s also almost certainly wrong.


    Part 2: Why the Economy Isn’t a Simple Subtraction Game

    The mistake some people make when thinking about the future is “ceteris paribus”—a Latin phrase economists use that means “all other things held constant.”

    The Techno-Feudalism model assumes a static world. It assumes there is a fixed amount of “work” to be done, and if AI does it all, there’s nothing left for humans.

    But the world isn’t static. It’s a wildly complex, adaptive system. When you change one massive variable (like the cost of intelligence dropping to near-zero), every other variable shifts in response.

    Here are three known system dynamics suggesting the “One-Man Empire” scenario won’t play out the way we fear.

    1. The Jevons Paradox (Or: Why We Don’t Have “Enough” Yet)

    In the 1860s, an economist named William Stanley Jevons noticed something weird. James Watt had just invented a much more efficient steam engine that used way less coal. Everyone thought, “Great, we’ll use less coal now!”

    Instead, coal consumption skyrocketed.

    Why? Because cutting the cost of steam power made steam power economically viable for thousands of new uses.

    This is the Jevons Paradox: When you increase the efficiency of a resource, you end up using more of it, not less.

    Right now, intelligence and creative labor are expensive and scarce resources. We ration them.

    • Small businesses don’t have dedicated marketing teams because it’s too expensive.
    • You don’t have a custom-coded app for managing your household chores because hiring a dev team for that is insane.
    • Niche communities don’t have high-production-value movies made about them because the ROI isn’t there.

    If AI makes the cost of executing these things near-zero, we aren’t just going to do the existing amount of work cheaper. We are going to do orders of magnitude more work.

    The “One-Man Empire” won’t just be competing to build the existing Spotify. They’ll be building the 10,000 niche versions of Spotify that couldn’t exist before.

    The demand for things we can’t even afford to imagine right now will explode.

    2. The Complexity Ceiling (Or: The Bottleneck Shifts)

    Let’s imagine our hypothetical AI Warlord, “Chad.” Chad has access to unlimited AI agents. He decides to build a competitor to Salesforce over the weekend.

    He tells his “Architect Agent” to design it. He tells 500 “Coder Agents” to build it. He tells 100 “QA Agents” to test it.

    It sounds great. But anyone who has ever managed a large project knows the hard part isn’t the typing; it’s the alignment.

    When you have infinite cheap labor, the bottleneck doesn’t disappear—it just shifts. The new bottleneck becomes orchestration, judgment, and taste.

    • Who decides which features actually matter to humans?
    • Who notices that the AI agents misunderstood the nuance of the brand voice and are now alienated customers?
    • Who integrates the 5,000 different outputs into a coherent product that doesn’t feel like a chaotic mess?

    Chad can’t just press a button and walk away. He has to act as the conductor of an incredibly complex, high-speed orchestra.

    Mental Visual: One sweaty, stressed-out stick figure trying to conduct a symphony where every musician is playing a different song at 500 mph. The stick figure is tangling himself in baton strings. It’s not an empire; it’s a circus.

    The complexity of managing an agent army scales exponentially faster than the output they produce. Humans will still be needed as the “glue” in these systems—the contextualizers, the relationship builders, the final arbiters of quality.

    3. Hyper-Competition Erodes the Throne

    Feudalism works because the resource is scarce and easy to defend. In the middle ages, that was land. You had the castle, you had the knights, you controlled the land.

    The AI revolution is doing the exact opposite. It’s taking a scarce resource (intelligence/skill) and making it abundant and permissionless.

    If Chad builds a billion-dollar company in his basement on Tuesday, what happens on Wednesday?

    On Wednesday, 50,000 other people look at Chad and say, “Oh, I can do that too.” And because the barriers to entry are now near-zero, they do.

    The “One-Man Empire” theory assumes that the first person to figure out the workflow gets to keep the monopoly forever. But in reality, economic gravity kicks in.

    When everyone has superpowers, nobody has superpowers.

    If everyone can spin up a hyper-efficient company instantly, the competition becomes absolutely ferocious. Profit margins get competed down to almost nothing. The only way to win isn’t just to have AI—everyone has AI—it’s to have better human insight, better branding, better trust, better niche understanding.

    We won’t see a few monolithic kings; we’re more likely to see a Cambodian jungle of millions of highly specialized, fiercely competitive micro-entities, constantly evolving and eating each other.


    Part 4: The “Compute Warlord” Objection (Or: Who Owns the Factory?)

    At this point, the pessimist in your brain might be screaming:

    “Okay, sure. Maybe the software gets democratized. But what about the hardware? You can’t run a God-Tier AI on a MacBook. You need a data center the size of Rhode Island that consumes enough electricity to power Argentina. And only three companies can afford that. Won’t they be the Emperors?”

    This is the Hardware Moat argument.

    It suggests that the future isn’t a meritocracy of ideas; it’s a tyranny of GPU ownership. If Sam Altman or Satya Nadella controls the only brain big enough to run the economy, aren’t the rest of us just digital sharecroppers?

    It’s a terrifying thought. But it misses a crucial distinction in how infrastructure works.

    1. The “Electricity” Model

    Imagine it’s 1910. You want to start a factory. To do that, you need massive amounts of electricity.

    Building a power plant is impossibly expensive. Only a few tycoons—the “Electricity Warlords”—can afford to build the dams and burn the coal to generate the juice.

    Did the Electricity Warlords end up owning every business in America? Did the CEO of ConEd also become the CEO of General Motors, the local bakery, and your toaster?

    No.

    Why? Because electricity is a commodity.

    The business model of a power plant isn’t to hoard the electricity; it’s to sell it to as many people as possible. If they keep the electricity for themselves, they go bankrupt. They need you to build a toaster. They need you to open a factory.

    The owners of the Supercomputers (the Compute Warlords) are in the same boat. They are spending hundreds of billions of dollars to build these “Intelligence Power Plants.” The only way they make that money back is by renting that intelligence out to the rest of us for pennies on the dollar.

    They don’t want to be the economy. They want to power the economy.

    2. Training vs. Inference (The “Recipe” vs. The “Cake”)

    There is a massive difference between creating an AI model and using one.

    Training (creating the model) is the expensive part. That’s the Supercomputer part. That’s the “Manhattan Project” level effort.

    Inference (running the model) is becoming shockingly cheap.

    Once the model exists, it can be copied, distilled, and run on much smaller hardware. We are already seeing “local models” that run on a gaming laptop that are almost as smart as the supercomputer models from two years ago.

    The “Compute Warlords” might have the edge on the absolute cutting frontier (the smartest possible thought), but for 99% of economic activity, the “good enough” intelligence will be cheap, abundant, and available everywhere.

    3. The Economic Suicide of Exclusion

    Let’s play out the nightmare scenario. Let’s say one company—call them OmniCorp—achieves AGI. They have the Supercomputer. They lock the doors. They don’t let anyone else use it. They use their AI to produce all the goods and services in the world.

    Who buys them?

    If OmniCorp automates everyone out of a job and refuses to share the “means of production,” nobody has any money. If nobody has money, nobody buys OmniCorp’s products. The economy collapses, and OmniCorp’s stock goes to zero.

    Capitalism is a loop. Producers need consumers. The “Compute Warlords” have a selfish, survival-based incentive to ensure that the rest of us remain economically viable, creative participants.

    They don’t want to rule a wasteland. They want to sell API credits to a thriving civilization.

    The Symbiotic Future

    The doom-mongers are right about one thing: The structure of work is going to change radically.

    The era of the “mid-level information processor”—the person whose job is to take data from one spreadsheet and move it to another—is probably ending. That is scary, and it will cause real short-term pain and disruption that we need to take seriously as a society.

    But the idea that this leads to a permanent landed gentry of AI owners and a permanent underclass of useless humans ignores history and economics.

    We aren’t heading for Feudalism 2.0. We’re heading for something weirder. A world where the barrier to creation is zero, where hyper-niche demands are met instantly, and where the definition of “work” shifts from execution to direction, taste, and humanity.

    It’s going to be a chaotic ride. But don’t buy your serf outfit just yet.

    Want more thinking like this?

    Occasional updates on next-gen AI, self-learning systems, and what we're building. Unsubscribe anytime.