The Best Software of the Next Decade Will Be Built by Amateurs
Millions of people are enduring problems right now that could be solved in an afternoon of development work— if they had the tools and knowledge to build the solution themselves.
A small business struggling with an inefficient process held together by memory and Post-it notes. A teacher rebuilding the same spreadsheet every semester. A social worker drowning in manual case tracking.
They all know exactly what they need. They just can’t build it. Yet.
The bottleneck in AI transformation was never capability. It’s the deployment pipeline.
For decades, we concentrated problem-solving power inside a small professional developer class. That class is essential — but it operates inside market incentives. Problems get solved when someone can make money solving them. Everything else waits.
A small business’s scheduling nightmare. A nonprofit’s donor tracking mess. A local government’s accounting disaster. These problems affect real people every day. But they’ll never be big enough markets to attract a polished product.
So they stay broken. Not because they’re hard. Because the people who understand them aren’t the people who can build the fix.
That’s starting to change.
AI is collapsing the distance between “I know what I need” and “I can build it.” Years ago, automating a workflow required software engineering skills. Today, someone with strong domain knowledge and the ability to describe their problem clearly can build things that would’ve taken a team and a budget.
The skill that matters now isn’t coding. It’s knowing what to build. Breaking messy real-world problems into components AI can act on. Prompting with clarity. Evaluating output critically. Knowing where human judgment stays in the loop.
This is the rise of the citizen developer.
And they have structural advantages professionals don’t.
When the builder is also the user, feedback is instant. Bugs get caught because the builder lives with them. Solutions stay right-sized — no one over-engineers a tool they have to use themselves every day.
A nurse with AI and domain expertise will outbuild a developer who’s never set foot in a hospital. Not on infrastructure. Not on architecture. On the thing that actually matters: solving the right problem.
A professional developer building for that nurse has to go through a slow, lossy loop — explain, interpret, build, review, iterate. Context leaks at every handoff. The result is often technically competent and functionally wrong.
The citizen developer skips all of that.
This isn’t “everyone becomes an engineer.”
It’s a new division of labor. Professionals build platforms, guardrails, and secure foundations. Citizen developers build solutions on top — fast, contextual, fitted to reality.
One skilled engineer empowering hundreds of domain experts is a compounding advantage the old model can’t touch.
Here’s what’s actually happening.
The next million software tools won’t come from startups or dev teams. They’ll come from people who got tired of waiting for someone else to fix their workflow. Doctors. Corporate managers. Small business owners.
People who’ve been sitting on solutions in their heads for years are about to have the means to build them.
We don’t have a technology problem. We have a distribution-of-power problem. And AI is one of the more credible answers to it.
The people closest to the problems are going to build things the market never would have.
That’s the citizen developer revolution. And it’s already underway.