The panic is always the same. A new technology appears. It makes workers more productive. Companies realize they can do the same work with fewer people. Layoffs happen. Newspapers write about the end of work.
Then something else happens. And nobody writes about that part.
The discrepancy window
When a new technology hits, it creates a temporary gap. Companies that adopt it early can produce the same output with fewer people. For a brief moment, this looks like replacement.
The printing press put scribes out of work. The power loom destroyed hand-weaving jobs. The calculator made entire rooms of human "computers" obsolete. The sewing machine threatened tailors. Every single time, the first reaction was panic and the first move was cuts.
But here's what happens next: the technology spreads. It stops being an advantage and becomes the baseline. Every company has it. Every competitor uses it. And suddenly, doing the same output as before isn't enough anymore. Because your competitor who kept their people - and gave them the new tool - is producing more.
The window where fewer people can do the same work is real. But it's temporary. Because the system doesn't optimize for "same output, fewer people." It optimizes for "more output, always."
The bar never comes down
This is the part people miss. They see the layoffs and project them forward. "If 8 people can do what 20 did, eventually 2 people will do what 8 did, and then nobody's needed."
That's not how it works. Because the expected output doesn't stay frozen.
When calculators became standard, nobody said "great, we need fewer accountants." They said "great, now we expect monthly reports instead of quarterly. More analysis. More scenarios. More detail." The tool reduced time per calculation. The demand for calculations exploded.
When the sewing machine became standard, clothing didn't stay the same. The variety increased. The speed of fashion increased. The volume of production increased. The industry didn't shrink. It grew massively - and employed more people than hand-sewing ever did.
The printing press didn't shrink publishing. It created publishing. Before Gutenberg, there was no publishing industry. There were monks copying manuscripts. After Gutenberg, the demand for written content expanded so dramatically that the number of people working in text production grew by orders of magnitude.
The pattern is the same every time: the technology absorbs the old work, the bar rises, new work fills the gap, and the system demands more.
The greed of the system never stops
This is important to understand because it isn't a bug. It's the core mechanic of markets.
No company says "we're productive enough, let's relax." No CEO tells shareholders "we could grow, but we're comfortable." The system is structurally incapable of satisfaction. Every efficiency gain gets recycled into higher expectations.
That's why the "AI will replace everyone" narrative is wrong. Not because AI isn't powerful - it is. But because the system won't use AI to produce the same output with fewer people. It will use AI to produce dramatically more output. And producing dramatically more output requires people.
Different people, maybe. Doing different work, certainly. But people.
Software developers are a special case
Here's where it gets interesting. Most technology creates a tool that does one thing better. The power loom weaves faster. The calculator computes faster. The sewing machine stitches faster. The worker operating the machine has a clear, bounded output.
A software developer with AI is different. Their "machine" doesn't do one thing. It does everything. It's not a loom. It's a wand.
A developer with AI can perfect existing products, build new features that were never on the roadmap, create entirely new products, build internal tools that make every other department faster. The surface area of what's possible isn't bounded by the machine. It's bounded only by imagination and time.
That's why firing developers makes even less sense than firing weavers after the power loom. The weaver's machine produced more cloth. The developer's machine produces more everything. Every hour of developer time is now worth dramatically more - not in one dimension but in every direction simultaneously.
The rational response isn't fewer developers. It's pointing more developers at the infinite backlog of things that were never worth building before.
Why the short-term pain is real but misleading
I'm not dismissing the layoffs. They're real. Real people are losing real jobs right now. And the discrepancy window is painful for everyone caught in it.
But projecting the first wave of cuts into permanent mass unemployment is the same mistake people made with every previous technology. They see the destruction clearly and the creation dimly - because destruction happens fast and visibly, while creation happens slowly and across a thousand new niches nobody predicted.
Nobody in 1455 looked at the printing press and said "this will create journalism, novels, scientific publishing, advertising, and the entire education industry." They saw unemployed scribes. The scribes were real. But so was everything that came after.
The machine that never says "enough"
The system runs on greed. That's not a moral judgment. It's a mechanical observation. Markets want more. Always more. More output, more products, more efficiency, more growth.
Every tool that makes workers more productive doesn't reduce the demand for workers. It raises the bar for what workers are expected to produce. And then it demands more workers to meet the new bar.
AI is the most powerful productivity tool in history. Which means it will raise the bar higher than anything before it.
Nobody's getting replaced. Everyone's getting a new baseline.