My wife is a vibe-coder, now what?
My wife has never written a line of code. She doesn't know what a function is. She thinks "React" is something you do after someone scares you. Last week she built her first app.
She sat down with an AI coding assistant, described what she wanted, argued with it when it got things wrong, and three hours later had a working prototype. No Stack Overflow. No tutorials. No bootcamp. Just a person with a clear idea and a tool that could translate it into software.
I watched the whole thing unfold. And honestly… It broke something in my brain.
AI is just a programming language that speaks English
I've been thinking about this for a while, and I think the framing most people use is wrong. AI isn't replacing programmers. It's a NEW programming language. Probably the most accessible one ever made. You describe what you want in plain words. The machine handles the translation to code.
But "accessible" doesn't mean "automatic." My wife didn't just say "make me an app" and got one. She spent hours refining what she wanted. She had to think through user flows, make design decisions, reject bad suggestions, and push back when the AI went in the wrong direction. She was the creative director the entire time. The AI was just the developer who happened to work for free.
The AI had zero opinions about what the app should do. It had infinite opinions about how to implement it. That's the split. Creativity and taste still live entirely with the human. The AI just removed the last barrier between "I have an idea" and "I have a working thing."
Engineers aren't coders anymore, they're managers
This is the part that genuinely excites me. I used to spend most of my day on forms and tables. Input validation. CRUD endpoints. Necessary work, but soul-crushing if you're honest about it. Now I hand that off to AI agents and move on.
My workflow has changed completely. I spin up multiple agents, each working on a different piece of the system. One is building the API layer. Another is writing tests. A third is refactoring a module I've been meaning to clean up for months. I review their output, course-correct, and focus my own time on the problems that actually require a human brain: architecture decisions, system design, figuring out what to build in the first place.
We've basically become engineering managers, except the reports are AI agents instead of people. The skill set shifted. It's less about knowing the exact syntax for a PostgreSQL upsert and more about knowing when an upsert is the right call at all. Less typing, more thinking.
And the work that's left? It's the good stuff. Architecture questions. "How do we make this scale" conversations. The creative work that got us into engineering in the first place, before we spent a decade wiring up form fields.
The gatekeeping is over, and that's fine
Some engineers feel threatened by this. I get it. We spent years learning these skills, and now someone with zero technical background can produce working software in an afternoon. But my wife's app works. It does what she needs. She saw a problem and built a solution. The fact that she used AI instead of learning JavaScript doesn't make the output less valid.
What it does mean is that the floor has risen. The baseline for "who can build software" now includes basically everyone with a clear idea. The ceiling hasn't dropped, though. Complex distributed systems, performance-critical code, security -- these still need deep expertise. AI can help, but it can't replace the judgment that comes from years of building and breaking things.
The gap between "I can make a simple app" and "I can architect a production system" is wider than ever. But the first part is no longer gated by whether you learned to code. And honestly, that's the world I want to live in.
Key takeaways
- AI is a programming language that speaks English. It removed the syntax barrier, but creativity and vision still come from the human.
- Non-technical people can now build working software, but they still need clear thinking, taste, and persistence to guide the AI.
- Software engineers have shifted from coders to managers, distributing work across AI agents and focusing on architecture, design, and the hard problems.
- The floor for who can build software has risen dramatically. The ceiling for complex engineering hasn't moved. Both of these things are good.
References: