Code used to be expensive. Every feature was a negotiation. Every sprint was a budget decision. Product managers ranked backlogs knowing 80% would never get built because there simply weren't enough hands.
That constraint is gone.
The first reaction is always wrong
The first thing companies will do is fire developers. It's already happening. A team of 20 ships what they shipped before with 8. CFO sees the numbers. Easy call.
And short term? They're right. Profits go up. Headcount goes down. Shareholders clap.
But this is the sugar rush before the crash. Because while you're cutting, your competitor kept their devs - and those devs are now producing 10x what they did before. They're not just doing the same work faster. They're doing work that was never possible before.
The backlog dies
Every company has a graveyard of features that "weren't worth building." Not because they were bad ideas. Because developer time was too expensive to spend on them. Legacy migrations, performance fixes, internal tooling, that redesign everyone agreed on two years ago - all sitting in a backlog column labeled "someday."
Someday is now.
When a developer can do in a day what used to take a sprint, the math changes. Those dusty backlog items suddenly have ROI. That technical debt everyone lived with because fixing it "wasn't a priority" - now it takes an afternoon. Companies that realize this will build better software than anything we've seen. Companies that just cut headcount will ship the same mediocre product with fewer people and call it efficiency.
The "everyone should code" scam
For 20 years big tech pushed one message: everyone should learn to code. Bootcamps exploded. Universities added CS programs. Governments funded coding initiatives.
This was never charity. It was supply economics.
The digitalization megatrend was far from done - still isn't - and the demand for developers was massive. Big tech needed a flood of supply to keep salaries manageable and hiring pipelines full. "Everyone should code" was a recruitment campaign disguised as social good.
It worked. The supply grew. And now AI divided the demand by 100.
Not because developers aren't needed. Because each one produces 100x more. The math that required a million developers to meet global demand now requires far fewer to produce the same output. The propaganda worked perfectly - just in time to become irrelevant.
The equalization
Here's what happens next. The companies that fired will eventually look around and notice something uncomfortable: their competitors who kept developers and pointed that 100x productivity at real problems are lapping them.
You can't re-hire fast enough when the market has moved. The talent you cut found other roles. The institutional knowledge walked out the door. And now you're competing against teams that are not just shipping faster - they're shipping things you didn't think were possible.
The hiring will come back. Not for the same roles and not at the same volume. But it will come back because the demand for software hasn't shrunk. It's grown. It always grows. What changed is the unit economics of producing it.
The real shift
The expected output of a software engineer is no longer hidden behind sprint velocity and Jira tickets. It's visible. PMs and POs can see what's possible now. And once expectations move up, they don't come back down.
This isn't a job crisis. It's a productivity revolution. The developers who survive it won't be the cheapest ones. They'll be the ones who understood that cheaper code means more ambitious software - and positioned themselves to build it.
Code is cheap now. What you build with it just got expensive.