Everyone Can See the Jobs AI Will Destroy. Almost No One Can See the Ones It Will Create.

The real problem with technology isn’t job loss. It’s that job creation is almost impossible to see while it’s happening.

Before we get started, I should explain why this article exists.

My business partner and podcast co-host, James Dwiggins, has been asking me the same question roughly every twelve minutes for the past few months.

“What about all the jobs AI is going to destroy?”

Now, James is a smart guy. He runs companies. He builds things. He has opinions about economic policy and occasionally about my diet. But lately, he has developed a mild case of what I can only describe as AI Job Anxiety, which tends to flare up whenever a new headline appears about robots replacing accountants, radiologists, or the guy who writes product descriptions for patio furniture.

So this article is essentially a public service announcement for James.

James, if you’re reading this, this is where you sit down, take a sip of tea, and let Uncle Keith explain something about how this movie tends to play out.

Because we’ve actually seen this story before. Quite a few times, as it turns out.

And every time it happens, people make the same mistake. They look at the jobs disappearing and assume that’s the whole story.

Job Loss Is Visible. Job Creation Is Usually Not.

Every time a new technology shows up, the same conversation starts.

Someone says the machines are about to eliminate millions of jobs. Someone else says technology always creates new ones. Then everyone proceeds to argue past each other like two people trying to explain gravity using different languages.

The strange thing is that both sides are usually describing real things.

Jobs do disappear when technology changes how work gets done. Entire industries sometimes vanish. But over time, the economy tends to grow new work somewhere else. The confusion usually comes from a simple problem of visibility. Job destruction happens in a room. Job creation tends to happen somewhere you’re not looking yet. And the human brain is extremely sensitive to rooms.

Imagine a factory shutting down.

A thousand people lose their jobs. You can see the building. You can count the empty parking spaces. There are interviews with workers carrying cardboard boxes full of desk photos and coffee mugs. Someone’s uncle shows up on the evening news, explaining how he gave the company 27 years of his life. It’s a very clear story. Technology took the jobs.

Now imagine the opposite process.

Across the country, 40 small companies hire five engineers each to build new software tools. A logistics startup hires a few dozen people in an office park next to a dentist and a karate studio. A medical device company quietly adds 200 jobs in a suburb nobody outside the county has ever heard of. A few thousand people go to work somewhere new.

There is no headline.

No one gathers outside the office park with a microphone and a camera crew. The jobs appear quietly, like mold growing behind drywall. The factory fire gets the attention. The humidity doesn’t.

That asymmetry does strange things to how we talk about technology. We treat the visible losses like the whole story and the invisible gains like rumors. But if you zoom out a little, the pattern tends to look different.

History Has Been Running This Experiment for a Very Long Time.

Take the automobile.

If you had been alive in 1905, the list of jobs about to disappear would have been obvious. Cities ran on horses. There were people breeding horses, feeding horses, transporting horses, and shoeing horses. There were carriage builders, saddle makers, blacksmiths, stable operators, and entire municipal systems dedicated to moving animals through crowded streets.

Then along comes a noisy metal box with an engine.

From the perspective of 1905, the math looks straightforward. The car destroys the horse economy.

And it did.

But machines have a funny habit of dragging new industries behind them like a comet tail. Gas stations appear. Mechanics. Tire factories. Auto parts suppliers. Road construction crews. Motels. Roadside diners. Suburban housing developments. Delivery trucking. Tourism. Ski resorts. Most of those things didn’t exist when the first cars showed up, rattling down dirt roads.

Even within the auto industry itself, the story didn’t unfold the way people might expect. When Ford introduced the Model T, productivity inside the factories exploded. In just six years, the number of cars produced per worker nearly tripled, while the price of a car fell from around $950 to about $440.

If you were only watching productivity, you might assume fewer workers would be needed.

Instead, the opposite happened.

Lower prices meant millions more people could afford cars. Demand exploded. Production scaled up. Employment in the auto industry grew rather than shrank.

This pattern has been studied repeatedly. Research from the McKinsey Global Institute shows that over long periods, productivity growth has often coincided with employment growth because lower costs expand demand and generate entirely new industries that were difficult to predict in advance.

The machine got more efficient. The market got bigger.

That combination shows up more often than people expect.

The Computer Did Something Similar — Quietly Enough That Most People Missed It.

If you had walked into an office in 1980 and watched computers arrive on everyone’s desks, you could have made a pretty reasonable list of jobs that were about to disappear. Typists. Secretaries taking dictation. Office machine operators. Entire departments are dedicated to filing and organizing paper.

And those jobs did shrink.

Researchers estimate that about 3.5 million jobs in the United States disappeared as personal computers and the internet spread through the economy.

Which sounds like a catastrophe.

Until you look at the other side of the ledger.

Software developers. IT specialists. Semiconductor manufacturing. Network engineers. App developers. Data analysts. Technical support. Digital marketing. E-commerce logistics. Entire industries that require computers to exist in the first place.

When researchers added up the new occupations created by that ecosystem, they found more than 19 million jobs tied to personal computing and the internet, producing a net gain of about 15.8 million jobs in the United States.

That’s the strange rhythm technology tends to follow.

It deletes jobs in clusters.
It creates jobs in ecosystems.

The Real Reason This Feels Different Has Less to Do With Economics Than With Human Attention.

The jobs that disappear are tangible. There are people standing in them. We know their names. We can point to the building where they used to work.

The jobs that appear later have a different problem.

They don’t exist yet.

There’s no building. No job title. No hiring manager. It’s difficult to worry about the unemployment rate for ski instructors before automobiles made mountain towns accessible to tourists. It’s difficult to predict app developers before smartphones existed.

Human attention tends to operate like a smoke detector. It’s tuned to sudden signals. Loud things. Immediate threats.

Job loss sets off the alarm.
Job creation behaves more like the temperature changing in a room.

It happens slowly. It spreads through the walls. You often don’t notice it until the environment has already shifted.

Productivity Doesn’t Just Eliminate Work. It Often Creates Demand.

When technology increases productivity, it usually makes something cheaper. Cheaper products tend to produce more demand. When the Model T became affordable to ordinary families, demand for cars surged.

Economists sometimes describe this as a cycle. Productivity increases output, output increases income, and income creates new demand across the economy. Over time, that demand tends to support more employment overall.

In other words, the machine makes things faster, which makes things cheaper, which makes people want more of them.

More demand tends to require more work.

Not always in the same place, but somewhere.

None of This Means the Transition Is Painless.

When technology rearranges an economy, some workers inevitably end up standing in the wrong room when the furniture moves. The research suggests that hundreds of millions of people globally may need to switch occupations as automation spreads through different industries in the coming decades.

That’s a large number of people learning new tools halfway through their careers.

But history also suggests that these transitions rarely happen overnight. The shift from horses to cars took decades. Farms and tractors overlapped for generations. Old industries fade slowly while new ones grow.

Which means the economy tends to operate less like a sudden collapse and more like a long, slightly messy renovation. Walls come down. New rooms appear. For a while, it looks like the house is falling apart, until you realize someone is building a different layout.

The Honest Answer Is Uncertainty.

We don’t know exactly what happens next.

Every technological wave rearranges the economy a little differently, and artificial intelligence will almost certainly change some industries in ways we can’t fully predict yet.

But history does have a habit of rhyming, even when it doesn’t repeat itself.

Every generation eventually becomes convinced that this time the machines will finally run out of new work for humans to do. And every generation turns out to be living inside an economy that’s inventing jobs no one could have described twenty years earlier.

So the honest answer is uncertainty.

Some jobs will disappear. Some industries will shrink. Some people will have to learn new things halfway through their careers.

But it’s also very likely that somewhere right now, in a garage, a warehouse, or a slightly depressing office park with fluorescent lights that hum faintly all day, a handful of people are building the kinds of jobs we’ll all be talking about ten years from now.

The tricky part is that you can’t see them yet.

And that’s usually how it works.

TLDR

People can easily see jobs disappearing when technology arrives. Factories close. Roles vanish. Headlines spread quickly. What’s much harder to see are the jobs that haven’t been invented yet. When automobiles arrived, the horse economy collapsed, but entire industries quietly formed around cars. The same thing happened with personal computers, which eliminated millions of jobs but created far more new ones over time. The real challenge with AI isn’t that work disappears. It’s that new work tends to appear slowly, in places nobody is looking yet.

-k

If this helped you think about AI a little differently, subscribe.

I write about the uncomfortable truths shaping real estate, business, and technology, not the ones that trend for 48 hours, but the ones that matter for the next decade.