The Economics of AGI: Why Your Intuition About Scarcity is Likely Wrong

The transition to an AGI-driven economy is often framed as a binary choice between mass unemployment and utopian abundance. However, this conversation suggests that the most critical dynamics are not the immediate displacement of labor, but the long-term evolution of consumer demand and the systemic difficulty of capturing value. The non-obvious reality is that our current economic models rely on relational assumptions--the idea that humans value human-in-the-loop services--that may be fundamentally fragile. For investors, policymakers, and citizens, the advantage lies not in predicting specific job losses, but in understanding how capital will flow when the technological frontier expands. Those who prepare for a world where AI functions like electricity--a ubiquitous, commoditized layer of production--will be better positioned than those waiting for a white-collar apocalypse that may never arrive in the way we expect.

The O-Ring Trap and the Illusion of Automation

Conventional wisdom suggests that if an AI can perform 90% of a job tasks, the remaining 10% will naturally become a high-value relational haven for human labor. Alex Imas and Phil Trammell challenge this by applying the O-ring model of production. In this framework, a job is a series of interdependent tasks; if the AI can perform the other 90% of the job at a speed or quality level that the human cannot match, the human presence may actually degrade the finished product.

If you can only automate nine-tenths of the job but you can do it to a lower standard of quality than the human could do it, you might not want to automate even those nine-tenths.

-- Phil Trammell

This suggests that the human-in-the-loop sector is not a guaranteed sanctuary. If the system evolves to favor AI-native workflows--where AI agents communicate at speeds thousands of times faster than human cognition--integrating a human becomes a transaction cost burden. The relational sector might shrink, not because we stop valuing humans, but because the production flows of the future will be architected for machines.

The Hidden Dynamics of Demand Collapse

A common narrative--exemplified by recent scenario planning--is that mass white-collar automation will lead to a recessionary demand collapse. Imas and Trammell argue that this outcome requires highly improbable economic conditions. For the economy to shrink while the technological frontier expands, one must assume that human demand is strictly bounded--that we eventually hit a wall where we simply stop wanting more goods, services, or variety.

History suggests the opposite: as goods become cheaper, we do not just consume more of the same; we invent entirely new categories of consumption. The risk is not that we run out of things to want; it is that we fail to index our wealth into the capital that produces those new things.

The pessimistic framing of Moore Law is every 18 months the value of computation halves--we are just running out of uses for computation so fast that it is just sustaining Moore Law. And this is in fact literally relevant to a conversation about AI where maybe for the first time this is no longer true.

-- Alex Imas

The Indexing Imperative

The most significant systemic risk is not AI capability, but indexing friction. If AI gains accrue primarily to a handful of private firms, the average person is left with assets--like residential real estate--that are uniquely ill-suited to capture the value of an AI-driven economy.

The speakers highlight a golden window where index funds allowed broad participation in economic growth. If the future of AI becomes highly concentrated in private, non-public entities, that window may close. The strategic priority for both individuals and developing nations should be ensuring that AI infrastructure remains commoditized or, at the very least, accessible via public equity, mirroring the way electricity became a foundational, non-concentrated utility.

Key Action Items

• Prioritize Asset Exposure: Over the next 12 to 18 months, evaluate whether your portfolio is overly concentrated in legacy assets (like real estate) that do not benefit from the AI-driven expansion of the technological frontier.
• Shift from Retraining to Indexing: For developing nations or individuals, focus on strategies that provide broad exposure to the AI ecosystem (the index) rather than narrow, high-risk job-retraining programs that may be rendered obsolete by the next model iteration.
• Monitor Public vs. Private Trends: Watch the IPO pipeline of frontier AI labs. The speed at which these companies transition to public ownership is a leading indicator of whether AI wealth will be widely distributed or captured by a small cohort of private shareholders.
• Audit Relational Assumptions: If you are building a business, avoid the assumption that human touch is an inherent moat. Test whether your customers value the human element for its own sake or merely because current AI cannot yet perform the task to a sufficient quality standard.
• Prepare for Structural Volatility: Recognize that AI layoffs are often driven by social coordination--firms laying off staff to appear AI-forward to investors--rather than immediate productivity gains. This creates short-term turbulence that requires financial resilience.