The Future of Work in the Age of Artificial General Intelligence
The traditional narrative surrounding artificial intelligence and its impact on jobs often paints a bleak picture: robots are coming for everything, and only the most creative, uniquely human tasks will remain. However, a new paper by Pascual Restrepo, an associate professor of economics at Yale University, challenges this assumption. His research suggests that the future of work may not be as dire as many fear, but it also raises complex questions about economic inequality and the role of human labor.
Restrepo’s working paper, published by the National Bureau of Economic Research, argues that most human work will not be automated in an era of artificial general intelligence (AGI). The reason isn’t that AI lacks the capability to perform these tasks, but rather that much of what people do for a living simply isn’t important enough to justify replacing with machines.
“The model opens up the intriguing possibility that much of today’s work may not be essential for future growth and may never be automated,” Restrepo writes in his paper titled We Won’t Be Missed: Work and Growth in the AGI World. “Instead, compute may be directed toward bottleneck work critical for future progress—such as reducing existential risks, defending against asteroids, or mastering fusion energy—leaving large parts of the labor market unchanged.”
Not Obsolete—Just Irrelevant
Restrepo’s central argument is that AGI does not render human skills obsolete; instead, it revalues them. In this new economic landscape, the scarcity isn’t skilled labor or intelligence—it’s compute. This means that skills are valued based on the opportunity cost of the computational resources needed to replicate them.
“In fact, if compute and human skill are the only scarce resources, average wages are higher in a post-AGI world. On the other hand, labor’s relative role shrinks.”
His analysis assumes that compute will be allocated to areas most valuable for economic growth, leaving less important jobs to be filled by humans.
Two Kinds of Work in the AI Economy
Restrepo distinguishes between two types of work in the AI economy: “bottleneck” work and “supplementary” work. Bottleneck work includes tasks essential for economic growth, such as producing energy, maintaining infrastructure, advancing science, and national security. Supplementary work, on the other hand, consists of roles the economy can do without and still expand, such as arts and crafts, customer support, hospitality, design, academic research, and even the work of professional economists.
In Restrepo’s framework, the economy will automate every bottleneck task using compute, but supplementary work may be ignored by AI. This could mean that jobs like baristas and novelists might survive largely intact—not because of any special human magic, but because the cost of replicating them with compute would be too high when AI has more pressing challenges.
Bottleneck work, according to Restrepo, sounds almost like science fiction: reducing existential risks, defending against asteroids, or mastering fusion energy. Meanwhile, socially intensive work, such as hospitality, live performances, and entertainment, will remain human because it is costly to replicate with compute and not essential for future growth.
Surviving Automation Is Not the Same as Sharing in Growth
However, there is a sobering message in Restrepo’s paper. Surviving automation and sharing in economic growth are two different things. In an AGI world, wages would become decoupled from GDP. Today, as the economy grows, workers typically benefit through rising wages and improved living standards. But in the post-AGI economy he models, that link breaks. Once AI handles all the tasks essential for growth, economic expansion is driven entirely by adding computational resources.
Human work, whether essential or supplementary, is valued not by its contribution to growth, but by what it would cost to replace it with compute. That ceiling, in the long run, is low.
Labor’s Share of GDP Goes to Zero
One of the paper’s starkest findings is that labor’s share of GDP converges to zero. Total computational resources in the economy could eventually reach 10⁵⁴ floating-point operations per second. The computing power of all human brains combined is roughly 10¹⁸ flops.
In an economy where wages are anchored to what compute would cost to replicate human work, human labor becomes economically marginal—not worthless, but negligibly small relative to the overall pie. “Most income will accrue to owners of computing resources,” the paper concludes.
This raises a crucial question: who owns the compute? Restrepo notes that one approach is to redistribute these gains through universal income. Another is to treat compute as a public resource—akin to land or natural capital—and distribute its returns broadly.
Two Modes of Automation
The paper also explores the path to this future, identifying two modes of automation. In a “compute-binding” transition, AI adoption is constrained by available hardware, allowing for gradual adjustment and continuous wage paths. Workers have time to reallocate.
In an “algorithm-binding” transition—the current moment, with AI capabilities advancing in sudden leaps—the picture is more destabilizing. “Inequality may rise sharply: workers whose tasks cannot yet be automated enjoy large temporary wage premiums, while others face sudden wage declines as theirs are,” he writes.
This mirrors what is happening in the trades today, with electricians, plumbers, and HVAC technicians commanding strong premiums, especially on data-center construction. Construction workers on data center projects currently earn an average of about $81,800 annually—roughly 32% more than those on non-data center builds.
We Won’t Be Poorer—but We May Not Be Richer Either
Restrepo offers one piece of reassurance: workers as a group are not made worse off by the transition. Because AGI expands what the economy can produce, total labor income in the post-AGI world—across all workers—is higher than in the pre-AGI baseline.
The arrival of AI cannot make us collectively poorer, the paper argues, because we could always retreat to a no-AI zone and produce exactly as we did before. The fact that we don’t means the new arrangement is better in aggregate.
But that collective gain is cold comfort if it is concentrated at the top of the income distribution—among the companies, investors, and nations that own the data centers.
‘We Won’t Be Missed’
The paper’s title, borrowed from its closing argument, captures the existential wager of the AGI economy. “Historically, work provided not only income but also recognition that one’s efforts improved society’s well-being,” Restrepo writes. “Work gave people the sense that they would be missed. In an AGI world, that connection is severed.”
Today, if half the workforce stopped showing up, the economy would collapse. In the AGI world, we would not be missed.
For Restrepo, the message is not one of despair, but of clear-eyed reckoning. The question is not whether AI will take your job. It may be that your job was never important enough for the question to matter.
