Rebecca Patterson’s recent New York Times essay uses a Jenga tower as a metaphor for the American economy in 2025. Blocks are being removed—small businesses cutting jobs, federal layoffs, consumption concentrating among the wealthy—while AI companies pile massive investments on top. Eventually, she warns, Jenga towers fall down.
She’s right about the instability. But the Jenga metaphor obscures something more important: what happens after the tower falls matters more than the fall itself. And this time, the mechanisms that historically rebuilt the tower are blocked.
The Pattern Patterson Identifies
Patterson documents real structural shifts. Small businesses, which employ 46% of American workers, are cutting staff to manage tariff costs. The federal government plans to eliminate 300,000 jobs. Meanwhile, the top 10% of households own 87% of corporate equities and are driving consumption through AI-generated wealth gains.
Most strikingly: Alphabet, Meta, Microsoft, and Amazon will spend over $380 billion on AI infrastructure in 2025 alone—more than the entire Apollo program in today’s dollars. This private-sector stimulus is currently holding the economy up.
Her question: how long can it last?
The better question: what happens when it stops?
What the Jenga Metaphor Hides
In normal economic cycles, when the tower wobbles, we have standard rebuilding mechanisms: fiscal stimulus, monetary policy, workforce retraining programs, or simply waiting for the next generation to inherit wealth and restart consumption.
These aren’t working this time. Not because of political dysfunction (though that doesn’t help), but because of structural blockages that make peaceful reset mechanisms inoperable.
The AI Hardware Trap
Patterson notes the $380 billion in AI infrastructure spending but treats it as potentially stabilizing—”if they can stay firmly in place.” This misses the key dynamic.
AI hardware depreciates fast. An Nvidia H100 GPU installed in 2024 is economically obsolete by 2027. This creates intense pressure on companies to monetize their investment within a 36-month window. They can’t wait 5-7 years for long-term research breakthroughs to pay off.
This forces a choice:
- Type A (Extension): Using AI to do things humans cannot do. Example: AlphaFold discovering protein structures. Timeline to ROI: 5-7 years.
- Type B (Substitution): Using AI to replace human labor. Example: automating data analysis, coding, customer support. Timeline to ROI: 3-6 months.
Given the depreciation clock, CFOs are structurally required to prioritize Type B. The $380 billion isn’t building new demand—it’s building infrastructure to remove the existing demand base through labor substitution.
The Training Pipeline Breaks
The impact shows up most clearly in entry-level white-collar work—the traditional path into the middle class.
Consider the “junior analyst” role across finance, consulting, tech, and corporate strategy. In 2020, juniors spent 80% of their time gathering data, formatting Excel sheets, building slides. In 2025, AI agents do that work. The remaining task—interpreting data and managing the AI—requires senior judgment.
The role hasn’t disappeared. The entry rung has. Companies are freezing junior hiring because they need seniors to supervise the AI. This breaks the training pipeline that historically moved people from middle class to upper middle class.
Federal Reserve data confirms this: hiring rates are slowing even as layoffs remain low. It’s not a wave of firings—it’s a quiet narrowing of the gate.
Why the Safety Valves Are Blocked
Historically, when technological displacement occurred, two mechanisms cushioned the blow: policy intervention and generational wealth transfer.
The Policy Wall
U.S. federal debt is now 120% of GDP. Annual interest payments exceed the defense budget. In the 1930s, when debt was 40% of GDP, there was fiscal space for a New Deal. That space no longer exists.
Any proposal for large-scale intervention—UBI, massive retraining programs—hits the wall of bond market vigilance. With inflation risks still present, the government cannot print its way through this displacement without triggering a currency crisis. The political system can offer cultural rhetoric, but it cannot offer material redistribution at the scale required.
The Inheritance Illusion
The “Great Wealth Transfer”—$84 trillion passing from Baby Boomers to younger generations—is often cited as the coming salvation. This is a statistical mirage.
Seventy percent of seniors will require long-term care. Memory care averages $150,000 per year. Medicare doesn’t cover custodial care. Medicaid requires assets to be “spent down” to near zero before it provides coverage.
For middle-class families whose wealth is locked in home equity, the math is brutal: the house is sold, the proceeds pay for the final 3-5 years of life, and the estate transfers not to children but to the private equity firms that own the care facilities.
The wealth transfer happens. It just doesn’t go to the next generation.
The Caste Crystallization
The failure of these mechanisms creates a rigid three-tier structure:
- The Neo-Aristocracy (~10%): Efficiently transfers wealth through trusts; incomes leveraged by Type A AI adoption.
- The Squeezed Middle (~60%): Faces displacement from Type B AI; expected inheritance consumed by healthcare costs.
- The Dependent Class (~30%): Already displaced by automation; reliant on a fiscally constrained state.
This isn’t a cycle. It’s a configuration.
What This Means for the Next Decade
Patterson asks how long AI companies can sustain their spending. That’s the wrong frame. The question is what happens when the displacement wave hits and the safety valves don’t work.
There are four broad scenarios:
The Productivity Miracle (Low Probability): AI breakthroughs in energy or materials generate genuine surplus—not just labor substitution but actual new wealth. This requires not just technical innovation but rapid diffusion and regulatory flexibility. Possible, but requires everything to go right.
The Controlled Correction (Most Likely): A tech recession in 2026-27 cools the AI capital expenditure boom. Labor markets soften but don’t collapse. The tower wobbles but doesn’t fall. Politics remains polarized but functional. This is the “muddle through” scenario.
The Systemic Reset (Increasingly Likely): Domestic pressure builds until institutional legitimacy fails. This doesn’t necessarily mean revolution—it could be state-level regulatory divergence, capital controls, or populist movements that enforce economic nationalism. The key feature: the federal government loses the capacity to coordinate national economic policy.
The External Release (Lower Probability, High Impact): When internal pressure becomes unmanageable and peaceful mechanisms are blocked, nations historically find release through external conflict. This isn’t inevitable—it’s just the lowest-coordination-cost option for resetting blocked equilibria. War allows governments to override bond markets, mobilize displaced labor, and justify wealth compression through “shared sacrifice.”
Why This Analysis Matters
Patterson’s Jenga metaphor implies the problem is instability. But instability is manageable if you have rebuilding tools. The real problem is that the tools are missing.
This doesn’t mean catastrophe is certain. It means the range of possible futures has narrowed, and several of those futures involve significantly more friction than institutional economists are pricing in.
The practical implication: “career upskilling” is not sufficient preparation for the next decade. Structural resilience requires understanding which mechanisms are likely to work and which are blocked. It requires looking at debt exposure, asset diversification, and geographic flexibility not as individual optimization but as protection against regime shifts.
The tower will wobble. The question isn’t whether it falls—it’s whether we can rebuild it when it does.
And right now, the rebuilding mechanisms are missing.
cafebedouin.org 