For a long time, companies used the engineer like an expensive typewriter bolted to a desk.
Hand them a task. They build it. Hand them a screen. They wire it. Hand them a backend. They stand it up. Hand them a bug. They close it. Month ends. Paycheck clears. Back to the desk.
That era is over.
But not in the way most people think.
The picture in people’s heads: “AI elevated the engineer — now engineers will win.” Half right. The dangerous half is what they leave out. Because AI didn’t just elevate the engineer. It cheapened every layer of production.
First, writing code got cheap. Then design got cheap. Next in line: ad copy, creatives, targeting, A/B testing — marketing is getting cheap too. Meta Advantage+, Google Performance Max have already pointed the way: what the marketer calls “I know my audience” is migrating to the algorithm. The production front has collapsed. The marketing front is collapsing. What’s next isn’t clear, but where it ends is.
There is exactly one thing that hasn’t gotten cheap: the tap.
Attention is scarce. Distribution channels are few. There is no substitute. App Store, Play Store, Meta, Google, Stripe, AWS. As production cheapens, as marketing cheapens, value migrates upward — to the channel, the platform, the infrastructure. Scarcity attracts everything.
So the number one winner of the AI age is not the engineer. It is the platform.
What’s left for the engineer?
This is where the second divide begins. And this divide is missing from most manifestos.
The engineer is splitting in two.
On one side: the instruction engineer. Executes assigned tasks, knows syntax, works by procedure — the middle class that earned a premium for proximity to the machine. That class is done. AI already does that job, without fatigue, without complaint. That rent has been cut off.
On the other side: the founder-engineer. Finds the problem, shapes the form, converts intent into product, carries the iteration. Writing code is only one part of this person’s work — and a shrinking one. Because code got cheap. And when code got cheap, the real work came into view: what to build, why to build it, what form it should take. These questions are still expensive. Still scarce. Because AI cannot ask them. It can only answer them.
What the founder-engineer carries is not code. It is taste, direction, patience, distribution intuition, product intuition. None of that has been handed to the machine — and none of it will be anytime soon. Because AI is good at answering questions. It is not good at deciding which questions are worth asking.
This is the second winner of the AI age: the founder-engineer.
Now the equation snaps into place.
Platform on top. Doesn’t buy the product — sells the distribution. Takes its cut, holds the tap, grows on passive income.
Founder at the bottom. Finds the product, builds it, ships it. One person, three, ten at most. Midjourney makes a billion dollars in revenue with forty people. Cursor, Linear, Craft — all small cores. This is no longer the exception. It is the new model.
In the middle: a gap.
That gap is where the old mid-sized company used to sit. Large teams, departments, layers, managers, meetings. That structure is dissolving. Production got cheap, marketing got cheap, coordination got cheap — the “coordination is expensive” premise that justified the whole thing has collapsed. Coordination runs on agents now. Near-zero cost.
So is the company dying? No. It is changing what it is.
It can no longer produce, because the producers left. There is no going back to instruction people — AI absorbed them. What’s left? Capital. The company is no longer an employer. It is a buyer. No longer a paycheck writer — an equity acquirer. It finds the small outside founder, invests, acquires when it has to. The company itself is becoming a platform — a capital platform.
Three layers remain:
Platform. Founder. Capital.
Everyone in the middle is dissolving.
In this kind of equation, the language of salary starts to sound like a relic.
Why? Because salary was payment for four things:
The company carried the risk. The company put up the capital. The company opened distribution. The company brought the customer.
Today, three of those four have transferred to the platform. Distribution is on the platform. The customer is on the platform. Capital too — from Stripe Atlas to AWS credits — flows from the platform. What’s left in the company’s hands? Seed financing plus risk-absorption. That is all.
And even that — if the founder can cover it themselves — makes the company optional.
So salary is no longer payment for labor. It is the premium for avoiding risk.
For whom does salary still make sense? For the engineer who can’t yet ship a product. For the one who cannot absorb cash uncertainty. For those carrying the weight of life — family, debt, health. These are real reasons. They are not to be dismissed.
But for someone who can genuinely build? Salary is not security. It is forfeited equity. If the company captures the value that person creates, pays out ten percent, and keeps ninety — that is not a wage. It is arbitrage. The day the founder-engineer sees this clearly, they exit.
When they exit, what do they lose? The safety of a paycheck. What do they gain? Their own equity. Their own rhythm. Their own direction.
In the old age, this calculation was madness. Because they could not build the product alone. In the new age, skipping this calculation is madness. Because they can.
I am not writing this from theory.
There is an app live on iOS. The backend runs on Cloud Run. The Windows overlay sits on the desktop. The Android version is underway. I have not written a single line of code. I directed agents to do all of it. I opened screens, tracked errors, gave direction, fixed what broke, drew the form, carried the iteration.
The production gap between me and a company is gone. Three months ago it existed. Today it does not.
In 2026, I am opening the distribution front. On Apollo, LinkedIn, Google and Meta — customer research, advertising, direct contact through local field connections. Platform plus field, running together. By the end of 2026, the distribution gap between me and a company will be gone too.
By 2030? Every dollar I put into advertising will reach a more precise customer. The algorithm will have matured. The early mover’s data will have compounded. Field frequency will decrease. The moat will deepen.
Is this ambitious? Yes. Proven? No, not yet. But the direction is clear, the plan is written, the production front is already closed.
So I place a warning on myself:
Closing the production front is not success. It is the prerequisite for success. A product that works is not a product that earns. The founder who confuses these two is confusing reality with manifesto. The manifesto is compelling — but it does not bankrupt you. It misleads you quietly.
There is also platform risk. If you share your cut with the platform, you are its tenant. This is true for companies and for founders alike. The company has lawyers, lobbyists, alternatives. The founder has none. The risk is not symmetric. Not to be dismissed.
And there is a quiet cost to going alone: feedback scarcity, blind spots, the trap of believing your own thesis too fast. That belongs on the ledger too.
But none of these warnings change the direction. They calibrate the tempo. They keep the humility intact.
The new equation is forming:
Platform on top. Founder at the bottom. Middle empty.
The engineer is not a hired hand. A force that wills things into existence. But this force proves itself not through code — through direction, intuition, patience, distribution.
Salary is no longer the final word. It is a habit from the old age.
The real money, the real equity, the real weight — will not go to the one who sells hours, but to the one who ships a vision. Not to the one who writes code, but to the one who shapes form. Not to the one sitting in a department, but to the one speaking directly to the platform.
And we are still at the very beginning of this age.
I am early. Not yet proven. But I know my direction.
And I know this time, the fight is not between the engineer and the company —
It is between the mid-sized company and time.