Last week a reader sent a message here on Substack asking: Why don’t you write about AI? The point they made was simple — it touches housing and many of the other issues I write about, and a lot of people believe it’s the biggest issue of our time.
My short answer is straightforward. Since relaunching this column, one of the themes has been operating in what’s known. Dozens of policy proposals cross my desk every week, and my biggest critique of most is that they assume a version of reality that just isn’t true. I’ve touched on that idea in past columns, and there are more than a few drafts sitting here on Substack that never made it to the finish line because they revolve around that same problem.
Good policy solves a problem everyday people have and starts with an operating framework that is 100% undeniable.
That’s why AI is tricky to write about through a housing lens. Most of the conversation around the technology revolves around work, which could eventually reshape housing markets in major ways. But right now there are still a lot of unknowns. That’s why I write about housing through the lens of supply and demand. Because a shortage of housing supply is the problem that exists today — and it remains unsolved. It’s the thing we can point to right now. The thing we can actively start fixing right now.
But for that reader — and for everyone else on Substack — here are my thoughts on artificial intelligence.
More than money: Work has done heavy lifting for our culture
For most of American history people have organized around a simple idea: Work is how humans earn their place.
You get a job. You contribute something useful. In return you get income, stability, dignity (though that part has eroded in some fields), and a certain level of respect from the community around you. The labor market has never been perfect, but it has been the central mechanism through which most people gain economic security and social standing.
Artificial intelligence may not destroy that system overnight. But it may slowly weaken the thing that makes it work.
That distinction matters.
Most debates about AI focus on whether the technology will eliminate jobs entirely. Economists and technologists often point out that previous waves of technological change — from mechanized farming to computers — displaced workers but eventually created new industries and new employment opportunities.
That history is real. But it may not fully apply to what’s happening now.
Past technological shifts largely automated physical labor or routine clerical work. AI is different because it targets cognitive tasks — writing, analysis, coding, design, customer support, research, and many other forms of knowledge work that expanded dramatically over the past few decades.
And unlike earlier machines, AI systems can improve themselves — or be improved using the same technology. That means the cycle of disruption could move faster and spread further across the economy than previous industrial shifts.
That doesn’t necessarily mean the United States is heading toward mass unemployment. In the near term, the bigger risk may be something subtler but just as significant: The labor market might gradually stops functioning as the primary way people gain security, status, and bargaining power.
I think you can already see hints of it happening.
Entry-level roles in some fields are becoming harder to find as companies experiment with AI-assisted productivity. In other sectors, workers may still have jobs but face weaker wages, less stability, or more competition for fewer opportunities because technology can now perform parts of their role. Entire career ladders are shrinking as fewer junior employees are needed to produce the same amount of work.
If those trends accelerate, the country could end up in a strange economic position: An economy capable of producing enormous wealth and productivity, but with fewer reliable pathways for individuals to build stable lives through work alone.
That creates a policy problem.
For generations, labor income has been the primary way economic gains are distributed across society. People work, businesses grow, wages rise, and households participate in the prosperity that follows.
But if AI allows companies to generate more output with fewer workers, that system begins to strain. Productivity rises while wage income grows modestly. Many say we’ve already observed that over the last few decades in the U.S. economy. Meanwhile, in the not-so-distant future, wealth would concentrate more heavily among the firms and investors that control the technology — if the trend continues unchecked.
At that point, the question facing policymakers isn’t just about jobs. It’s about distribution.
If labor is no longer the main channel through which people receive income and security, something else has to take its place. That could mean stronger public safety nets, expanded tax credits, broader asset ownership, or more ambitious ideas like universal basic income.
Those debates should already be taking place because it doesn’t appear AI is slowing down anytime soon.
But the deeper challenge may not be purely economic. It may be cultural.
Work has never just been about money. It’s structure, identity, and a sense of purpose. Occupations shape how people see themselves and how communities recognize contribution.
If the role of work changes dramatically, society will eventually have to answer a harder question: How do people earn dignity and standing in a system where employment is less central?
There isn’t an easy answer.
Artificial intelligence may eventually reshape large parts of the economy. But the most important shift may arrive earlier — and more quietly.
AI may not destroy work first. It may destroy the meaning of work. And we may already be on that path.
Thanks for reading In Focus! Subscribe below to get future columns and interviews delivered to your inbox.
