The scope of civil rights activism has naturally expanded into digital rights, ethics, and AI safety. What is AI going to do to our future? How will it change the balance between work and life? What skills should we be developing? How do we maximize quality of life and preserve liberty through one of the biggest technological shifts in history?

The simplest definition of post-labor economics is an economic regime where human labor input is no longer a binding constraint on economic output. That might sound abstract, but it captures something real: for all of human history, labor has been a binding constraint on anything humans want to do. Whether building Rome or a bridge, reaching the moon or building a nuclear reactor, everything has depended on human labor input—the mind, the body, the hands. All of what follows is still speculative, grounded in models, data, and trends, but the pattern is worth taking seriously.

The economic baseline definition of labor is compensated human activity that produces valuable transformation—taking a piece of information, a person, or a piece of matter and making it more valuable. Cutting down a tree, milling it into timber, and building a house with it is one kind of transformation. A university is another: taking someone less educated and making them more informed is a service transformation.

From a first-principles physics standpoint, labor is the expenditure of energy to process information or manipulate matter, yielding an economic state change. That distinction matters because what people actually pay for is the outcome, not human input specifically.

Four basic categories produce that valuable change: cognition, dexterity, strength, and the social-emotional component. Cognition is the ability to plan, think, and solve problems. Dexterity is fine manipulation, as in mechanics or surgery. Strength has already been largely automated away by heavy equipment and power tools. The social-emotional component covers authenticity, empathy, trust, and accountability. The open question is what happens once humans are no longer the only thing capable of all four.

The basic litmus test for when labor substitution becomes economically rational is: better, faster, cheaper, safer. It becomes economically irrational to use a human when capital equipment is better, faster, cheaper, and safer across the board. The tractor is a clear example—once it arrived, nobody in a modern economy farms by hand; if you do it, it’s a hobby, not a job.

A job isn’t one monolithic role—it’s a collection of discrete tasks. What technology does is unbundle those tasks, breaking them into smaller pieces and repackaging them, either giving them to humans still required for the job or handing them off to a machine.

Artificial intelligence and robotics are still fundamentally a type of automation, defined economically as labor-saving technology. It started with the printing press over five centuries ago, which enhanced human output by a factor of roughly 150 to 400—meaning it took that much less time to produce each finished volume. Automation isn’t about making things 20 or 50 percent more efficient; it’s about producing economically valuable output hundreds, thousands, or millions of times cheaper. Electric lights are another example: candlelight once cost roughly 200 to 1,000 times as much per day to illuminate a house as electric light does today.

AI and robotics represent the latest form of automation. Positing a future state where human labor becomes economically irrational is intrinsically an argument that automation will be capable of dislocating and subsuming those human tasks. Humanity has been encroaching on human labor for a long time—the Ottoman Empire banned the printing press for over two centuries to protect the jobs of scribes, and restraining that technological advance is one of the reasons the Ottoman Empire eventually fell behind and ceased to be an empire.

AI satisfies the classic definition of a general purpose technology, alongside steam, electricity, the wheel, and metallurgy: pervasiveness across most sectors, continuous improvement, and innovation spillovers. Healthcare, law, government, and administration are already being touched by it. Large language models specifically offer continuous improvement, rapidly evolving and becoming functionally cheaper over time, with new abilities emerging as they scale.

Electricity offers a useful intuition for continuous improvement. Its early uses weren’t obvious—it took time to figure out lighting, then electromagnets, electric motors, diodes, resistors, integrated circuits, and eventually artificial intelligence. Electricity is arguably the king of general purpose technologies, and AI is a downstream consequence of it; new uses for electricity are still being discovered more than a century after it was first harnessed.

The third component of a general purpose technology is innovation spillovers, or network effects. A technology that is both pervasive and constantly improving eventually transforms a value stream—the successive economic transactions that turn raw material, whether raw experience or stone and clay, into something useful. AI hasn’t yet fully transformed value streams, but it’s starting to. Many people already use chatbots for personal health, therapy, classwork, or family matters—things that change how people run their lives, including detailed personal research into health issues that would otherwise have required hundreds of hours and significant expense.

Rapid innovation brings real challenges. The first is deflation and demonetization. Inflation is the usual concern, but deflation—falling prices—is arguably more relevant here, because technology is fundamentally deflationary: it lets us do more with less effort, and money is a proxy for effort. Technology increases the carrying capacity of the planet. Without the Haber-Bosch process, which fixes nitrogen in the soil, the planet’s carrying capacity would be roughly two to four billion people, meaning a large share of the world’s current population would not be sustainable at all.

Demonetization describes the situation where one good or service replaces others so completely that those other transactions simply never happen. People save enormous amounts of money using AI for things they would otherwise have paid experts for—transactions that never occur, and never show up anywhere as economic activity.

On the other side is the deflationary death spiral. Companies use AI or automation to cut costs; when everyone does that, it reduces demand for goods and services offered by other people, which means those companies make less money, hire fewer people, and pay out fewer wages, leading to less aggregate demand. To survive, companies lay people off and automate more, and the cycle continues.

A market economy only strictly requires two things to clear: a supply of goods and demand to consume them. Labor is not technically required for capitalism to function, but humans have always supplied the strength, dexterity, cognition, and empathy that made supply and demand meet. Household income drives 70 to 80 percent of GDP, and wages drive roughly 82 percent of household income. If AI and robotics make it economically irrational to hire people, household income drops, spending drops, jobs dry up, and the tax base dries up with them—80 to 85 percent of U.S. federal revenue comes from income and payroll taxes, so the government’s finances are exposed to the same shift.

There are three sources of household income: wages, transfers such as welfare, SNAP, Social Security, and public education, and capital, such as dividends, equity, and rent. These three buckets are the keystone of post-labor economics. If wages go away, household income has to be replaced by transfers and capital. Universal basic income is one example of leaning on transfers. But if wages go away, the government also has to replace income and payroll tax revenue with something else—wealth taxes, value-added taxes, or corporate taxes among the options.

The other lever is the capital bucket. Public capital already exists in places like the Alaska Permanent Fund or Norway’s sovereign wealth fund. Alaska’s fund takes oil revenues and capitalizes an endowment that pays dividends to citizens; Norway’s fund holds roughly two trillion dollars. Public capital can fund government directly or distribute money straight to families. Private capital—stocks, bonds, rental properties—matters too. A post-labor framework would create more on-ramps for people to accumulate capital through vehicles like Employee Stock Ownership Plans or Employee Ownership Trusts. When employees become part-owners, household wealth increases substantially.

Combining these programs, it’s plausible to push median household income toward roughly 140,000 dollars a year. Layered with the deflation that automation itself produces, effective spending power could reach something like 300,000 dollars a year without traditional employment. This isn’t socialism or communism—it’s a way of keeping the system functioning if human labor input is substantially automated away.

There will always be exceptions. Humans are biologically wired to place higher value on embodied experience with real people—someone might watch a concert from home for a fraction of the cost, but will still pay a premium to be there in person for the human provenance of the experience. Human presence, accountability, and provenance are likely to remain permanent fixtures of the economy given our neurobiology and legal systems; in legal theory, you still need a “throat to choke”—someone who can be held accountable, or fired, when something goes wrong.

Automation is good—it’s why we pursue it. But it demands serious thought about the macroeconomics: what happens when everyone makes the individually rational microeconomic choice to use AI instead of a human?

Questions and Answers

One attendee asked what would make AI companies spread out their capital to the rest of the world as they build up their user base. The honest answer is that, presently, not much does. Henry Ford famously realized his own employees couldn’t afford his cars, so he raised wages so they could—this is fundamentally a demand-constrained economy, and in the long run it’s rational for the wealthy to cooperate with a system that keeps circulation going. What’s needed are incentive structures and on-ramps for capital, like sovereign wealth funds or employee ownership, that are attractive enough to sell themselves.

Another attendee asked why this isn’t socialism or communism. The intent behind post-labor economics is to be politically viable and capitalist-friendly. Communism involves abolishing private property, which is a difficult sell in America; the framework instead leans on things already proven to work, like the Alaska Permanent Fund, which was created by a Republican governor. The goal is for everyone to become a capitalist, with income coming from the things people own and are entitled to, rather than solely from exchanging labor for wages.

A third question raised companies like Anthropic paying employees even as they automate their own jobs. Companies generally want to automate away jobs; they don’t hire out of charity, and market efficiency is protected under the prevailing economic order. To align incentives, some companies may choose to pay people to automate themselves out of a task so the company itself can stay competitive—the automation flywheel is already picking up speed.

Someone asked what happens if society becomes dependent on automation and the infrastructure goes down—can we revert? Hunger is a powerful motivator, so yes, in principle, though it would be painful. Society has been dependent on technology and the division of labor for a very long time already. The diversity of human thought includes preppers and planners preparing for worst-case scenarios, and the broader goal is to prevent the kind of social collapse that could occur if unemployment spiked toward 40 percent with no solutions in place.

A final question concerned security—how this economy deals with AI overstepping or failing. Humans are usually the weakest link. The interlocks and safety checks for full autonomy haven’t been fully developed yet, but the market incentive to create autonomous agents is enormous. Every failure becomes something the broader system learns from: a failure mode gets found, gets fixed, and the process moves on. In aggregate, that’s how trust gets built—by finding and solving vulnerabilities as they appear.

Humanity and technology form a complex adaptive system. We change technology, and technology changes us back. Nobody can fully predict the final shape it will take, but the economic incentive behind it is immense, so it’s going to keep moving forward regardless.