The Economy Can’t Absorb All This AI Productivity

With AI, everyone gets more done in the same time. A developer builds more software, and a content producer makes more content. Someone in customer support can handle ten customers at once where they used to handle one. But the economy can't absorb all this productivity. You can handle ten customers now, but the nine extra customers aren't there.

Say you work in customer support for a company that sells accounting software to small businesses. There are ten of you, and each of you answers one chat at a time. Now AI looks up the customer's account and writes the first answer, and each of you can handle ten chats at once. But the company still has the same customers, and they don't ask more questions. Nobody opens an extra chat because the answers come faster. The company could try to win customers from its competitors, but the competitors use the same AI.

A company that gets ten times as much software from its developers still sells to the same customers at the same prices. The New York Times reported that about 560,000 new apps came out on the App Store in the first half of 2026, twice as many as a year earlier. Downloads grew only 2% in that time. People who make content with AI write for readers who don't have more time to read than before. I wrote about that in Everyone Is Marketing With AI, and Nobody Is Listening.

In the statistics, productivity is output per hour worked, and output is the goods and services people pay for. If you could handle ten chats but only one customer writes, the statistics see no change. They only see a change when the company handles the same chats with fewer hours, which means it lets people go. That leads to the problem I wrote about in Can AI Make Us So Productive That It Destroys the Economy?

In the US, productivity grew 2.1% over the past year, according to the Bureau of Labor Statistics. The average since 1947 is 2.2%. Each of us gets far more done with AI, and for the whole economy it looks like any other year.

If everyone gets ten times as much done, the economy needs ten times as many customers to absorb it, or customers with ten times as much money. Even in a good year it grows a few percent. The US economy grew at an annual rate of 3.4% in the third quarter, according to the Bureau of Economic Analysis. ING estimates that about a third of US growth this year comes from AI itself, from the money companies spend on AI technology and data centers. In Germany, where I live, the economy grew 1.3% over the past year.

So the ten people in support answer the same chats faster and then wait for the next one. Or the company sees that one person can handle all the chats and lets the other nine go.

What I'm building

Delegate tasks. Get software.

Give Vroni a GitHub issue, bug report, spec, or rough idea. It reads the repo, plans the change, writes code, runs checks, and works toward a review-ready pull request.

Take a look at vroni.com