AI is making expertise cheap. The value moves to whatever is still scarce.
For 25 years, my company has sold skill. Businesses needed software, they couldn’t build it themselves, and they paid people who could.
That trade is how a lot of people without money made some. Learn something valuable, get hired for it, build a life on it.
Skill → Employment → Wealth
AI is putting pressure on the first link in that chain. I can see it from both sides of my own desk.
Skill used to be the way in
A valuable skill gave a person leverage. If a company wanted what you knew, it had to hire you, and it had to pay enough to keep you. The expertise came attached to a human being who could walk out the door.
That’s how a good engineer, a good analyst or a good bookkeeper could do well without starting out with anything.
What AI changes
A business can now get a growing share of that expertise from software. The first draft of the contract. The analysis that took an analyst a week. A working version of the app.
For someone who already owns a business, the chain looks different:
Capital → AI → Productivity → More capital
If you have customers, cash and a company to point AI at, you can get more done without hiring as many skilled people. The gain goes to whoever OWNS the business, and the people who used to sell that skill get paid less for it.
That could move a lot of economic power toward people who already have capital.
The same tool cuts the other way
I’m also on the other side of this.
Over the past year or so I’ve built ten working apps for myself and for Lithyem, with AI doing most of the typing. One replaced an e-signature subscription we were paying for. One went from an empty project to a live app in 71 minutes. A few years ago, each of those would have been a project with a team, a budget and a timeline.
A single person or a five-person firm can now do work that used to take an entire office.
Both things are true at once. AI can concentrate wealth, and it can hand capability to people who never had it. Which one you get depends on who controls the AI and who puts it to work best.
What’s still scarce
If intelligence and technical skill get cheap, what’s still SCARCE enough to be worth paying for?
My working list:
- Ownership. The business, the product or the equity that AI makes more productive.
- A way to reach customers. Anyone can build the app now. Getting it in front of ten thousand people is as hard as it ever was.
- Data nobody else has. Your records, your history, how the job actually gets done at your company. AI is only as good as what you give it to work with.
- Customer relationships. A client who picks up when you call took years to earn.
- Judgment. Knowing what to build, what to ask for, and when the answer AI gave you is wrong.
- Reputation. When anyone can produce polished work, people buy from the name they already trust.
- Knowing how to run AI. Getting several AI tools and your own systems working together on real work is a skill, and it’s still a rare one.
Only the last one is a technical skill. The rest are slow to build and hard to copy.
I see it in what clients pay us for. Less of it is writing code. More of it is knowing what to build and making sure it holds up once people depend on it.
What I’d look at this week
If you run a company, look at what you charge for. If part of the answer is a skill AI can now supply, that part of your price is going to come under pressure. Then go down the list and ask which of those things your company already has. That’s where I’d put AI to work first.
If you work for a company, the same list applies to you. Your skill still matters. It probably won’t carry you by itself the way it carried the last generation.
I don’t know which way this tips, toward the people with capital or toward the people who learn to use the tools. Probably both.
If the skill you sell got ten times cheaper next year, what would people still pay you for?
Related: AI Agents That Do the Work
