There is a phrase people use when they want to describe uneven recovery… the K-shaped economy.
One part of society rises. Another falls behind. The image is meant to capture divergence.
But I do not think that framework is sufficient anymore.
What we are living through is not simply a K-shaped economy. It is increasingly an L-shaped economy for a very large portion of the public. And the distinction matters.
A K implies that while outcomes diverge, both sides are still on some trajectory. One may be better than the other, but each remains in motion. An L is different. An L is a drop followed by stagnation. It is the removal of a credible path back. It is not hardship. It is managed stagnation…
A condition where one part of the economy is being organized, quietly, into a permanent flatline while another part accelerates away.
That is what I believe is happening to a growing share of people.
I say that as someone who is not standing still. I am slammed every waking hour of the day. Even augmented by multiple AIs, I still cannot keep up with the volume of what is now possible, what is now expected, and what is now required just to remain competitive.
That is not a normal signal. That is not what a stable economy feels like. That is not what a healthy broad-based expansion feels like. That is not what a familiar business cycle feels like.
It feels like a system where capability itself is splitting.
The Mechanism
Every generation worries about automation. Readers have heard “this time is different” enough times to reflexively discount it. So it is worth being precise about what is actually different.
Prior technological transitions distributed their leverage through physical goods. A loom, a tractor, a personal computer — these tools were expensive, but once purchased, they diffused into the broader economy at a predictable rate. Productivity gains took time to concentrate and eventually spread, because the tools themselves were finite objects that workers and small firms could acquire, learn, and deploy.
Intelligence is not like that.
Intelligence itself is becoming an unevenly distributed production layer.
It is priced. It is gated. It is metered. It compounds faster at the top because the people already operating at scale can afford more of it, integrate it more deeply, and extract more output per unit of it than anyone else. The gap does not close as adoption spreads. It widens, because the return on each additional unit of machine intelligence is not linear — it scales with the capital, data, networks, and systems already in place to absorb it.
And here is the part that is rarely said out loud… even the people on the upper branch are not coasting. They are running a Red Queen's race — sprinting at compounding speed just to hold position. That is not a complaint. It is a diagnostic. Because if staying in place at the top now requires that velocity, then the entry speed for anyone trying to climb onto that branch has been reset to something no ordinary worker can generate.
The treadmill at the top is what makes the flatline at the bottom permanent.
On one side are people and firms who can combine software, capital, automation, global labor access, and now intelligence amplification. They are not just working harder. They are compounding faster. Their output per hour is rising. Their responsiveness is rising. Their optionality is rising. Their reach is rising.
On the other side are people whose labor, bargaining power, and resilience are being systematically eroded. Not always because they are lazy. Not always because they made bad choices. Often, the rules of the environment are changing faster than a normal person can realistically adapt to them.
This is where the old economic metaphors break down.
A classic K-shaped story still assumes the lower branch has shape. It still suggests there is a path, however painful, that remains legible. But what if that is no longer true for millions of people? What if the real story is that a large segment falls, then stays pinned there while the upper layer accelerates away?
That is an L.
The Diffusion Problem
The standard reassurance, when anyone raises concerns about technological concentration, is that diffusion will eventually handle it. Tools get cheaper. Access widens. Productivity gains spread. The bottom catches up. This has been true enough times in history that it functions almost as a reflex.
But the reflex depends on an assumption that is worth naming: that when technology reaches the lower branch, it arrives as empowerment.
That is not what is happening.
Machine intelligence is diffusing downward. It is just not diffusing as a tool the bottom branch gets to wield. It is diffusing as the infrastructure that optimizes them. Gig-work dispatch algorithms. Automated credit scoring. Predictive hiring filters. Scheduling software that compresses labor into the smallest billable units. Collections systems that never sleep. Surveillance layers embedded in every platform a precarious worker has to use to earn a living.
This is diffusion. It is simply diffusion in the wrong direction.
The upper branch receives machine intelligence as leverage. The lower branch receives it as management. The same technology, deployed on opposite sides of the same economy, performs opposite functions — multiplying the capability of one group while tightening the operating conditions of the other.
This is why the classical diffusion argument no longer rescues us. It assumes a single vector — tools moving from expensive to cheap, from scarce to common, from elite to ordinary. But when the economic topology has already split, diffusion does not flow across the gap. It flows along each branch separately. On one branch, it accelerates acceleration. On the other it makes stagnation more efficient.
The gains are not merely unequal. They are self-reinforcing. The people with access to capital, networks, data, and machine intelligence are operating on a curve that ordinary workers cannot see, much less catch, while the same underlying technology quietly becomes the substrate of the flatline they are expected to endure.
This is not just about income inequality, though that is part of it. It is not just about job displacement, though that is part of it, too. It is about the emergence of a different economic topology altogether.
The old world assumed a broad middle could convert effort into stability. Show up. Build skill. Work hard. Gain experience. Move up. Save. Buy a home. Raise a family. Retire with some dignity.
That social contract was already fraying.
Now we are entering a phase where effort alone is no longer enough, because leverage is no longer distributed in any remotely normal way. A single highly capable person with the right tools can now do the work that used to require a small team. A small team can do what once required an entire department. A company with enough software and AI scaffolding can outmaneuver incumbents that are still structured for an older era.
That does not mean everyone on the outside is doomed. But it does mean we should stop pretending the baseline environment is ordinary.
It is not ordinary when one part of the population is drowning in demand, execution pressure, and amplified output while another part is increasingly locked into precarity, debt, and shrinking room to recover.
It is not ordinary when intelligence itself is becoming an unevenly distributed production layer.
It is not ordinary when the answer to “work harder” is met by a system in which the people already ahead can multiply their effort through machines, while everyone else is told to simply be more resilient.
The Unreality Problem
This is why so many people feel disoriented.
They can sense that something is off, but the language they are given does not quite fit the lived reality. They are told the economy is strong while their own experience feels brittle. They are told innovation is creating opportunity, while the number of doors that lead to durable security seems to keep shrinking.
People can endure pain more easily than they can endure unreality.
If you tell them this is just another cycle, but what they are actually experiencing is a structural downgrading of their long-term position, they will eventually stop trusting not just the analysis, but the institutions delivering it.
And they should.
The public is not crazy for feeling that the ladder is being pulled up. In many cases, it is.
The challenge is that this process does not always announce itself dramatically. It happens in layers. First, through pricing pressure. Then through housing. Then through education. Then through healthcare. Then, through labor market compression. Then, through the silent expectation that every worker must now compete not only with other humans, but with software, offshore labor, and machine-augmented operators who can move at a radically different speed.
By the time the pattern is obvious, much of the damage is already structural.
Why the Letter Matters
This is why I do not think the K-shape metaphor is strong enough anymore.
A K still leaves too much room for false reassurance. It implies divergence, but not entrapment. It implies inequality, but not immobilization. It still carries the scent of temporary imbalance.
An L is harsher, but more honest.
It describes what happens when a drop is followed not by recovery, but by endurance. When a society no longer fails people episodically, but organizes itself in a way that leaves a growing number with no realistic path back to stability.
And if we are honest, that is what many people are staring at right now.
The cruel irony is that some of the same forces driving this split are extraordinary. AI, automation, software leverage, global communications, distributed coordination, low-cost intelligence — these are remarkable tools. In the right hands, they can expand human capacity and unlock things that used to be impossible.
But tools do not distribute themselves ethically. And when extraordinary leverage enters a system that is already unequal, it does not level the field. It magnifies the existing gradient.
This is what many people still do not fully grasp.
This is not merely a story of the rich getting richer. It is a story of the structurally advantaged becoming machine-amplified while a large portion of the public is left in a state that is neither collapse nor recovery, but managed stagnation.
That is the L.
The upper tier may still experience something closer to acceleration, even abundance. But the bottom and much of the middle are not simply lagging. They are being pressed into a flatter future, where the margin for error disappears, and the path back becomes increasingly theoretical.
Once you see that, a lot of modern life starts to make more sense. The anxiety. The exhaustion. The resentment. The strange split-screen feeling where one person is talking about an unprecedented opportunity while another is wondering how they are supposed to survive the next five years.
Those are not contradictory realities. They are the two faces of the same economy.
So no, I do not think this is a normal K-shaped economy.
I think it is becoming something more dangerous…
an L-shaped economy for the many, and an exponential one for the few.
The Diagnosis Problem
The reason naming this correctly matters is not rhetorical. It is practical.
Every policy response, every institutional reform, every piece of public commentary, every individual life strategy is downstream of how the problem is framed. If you believe this is a K-shape, you look for ways to steepen the lower branch — retraining, credentialing, adaptation, resilience. Those are the interventions that assume motion. They assume a trajectory.
If it is an L, those interventions are not merely insufficient. They are a category error. You cannot retrain your way out of a structure that has removed the path back. You cannot out-adapt an environment where the leverage gap is widening faster than any individual can close it.
Retraining assumes the problem is the worker. It is not. A worker retrained today is a worker who has acquired, at their own expense and on their own time, a slightly upgraded position on the same flatline. They have not crossed onto the upper branch. They cannot, because the upper branch is not defined by skill. It is defined by leverage — capital, data, networks, ownership, access to compounding systems — and none of those are things a training program can deliver. You cannot retrain your way across a structural gap. You can only retrain your way into a marginally more optimized version of the stagnation you were already in.
Misdiagnosis here is not a minor analytical error. It is the reason the official language and the public reality keep drifting further apart. It is the reason trust is eroding in exactly the institutions that should be helping people interpret what is happening to them.
When mobility dies for enough people, the issue is no longer simply economic.
It becomes civilizational.
And no civilization has ever survived long by telling a large share of its people that their flatline is a recovery.