A 67-Year-Old Vibe-Coder Is Rewriting the Rules with Generative AI

About a year ago, I had a conversation with a good friend — a 67-year-old retired network engineer — about what we now call vibe-coding. At the time, I was explaining that AI tools like ChatGPT could do more than just auto-complete lines of code. They could radically reduce the cost and complexity of software development, especially for solo builders or domain experts. He wasn’t buying it.

This friend had been investing serious money into a golf tournament application — a clever, niche idea designed for golf course operators to manage competitive play and monetize community events. I reviewed it against a few commercial offerings. His concept had teeth. But the development process? That was the bottleneck. He was paying devs to make tiny feature changes. The loop between business insight and technical implementation was just too wide.

So instead of tackling the golf app head-on, he started small. He built a wine inventory app for his son, who manages private wine collections for clients. It became his on-ramp into AI-assisted development. Over six months, he used ChatGPT, GitHub Copilot, and eventually Cursor — the IDE I nudged him toward because of its ability to ground AI on project context.

Once he adjusted his workflow, things clicked. Cursor’s AI-native IDE experience helped him stay in flow — combining AI code suggestions with real-time contextual memory across files. Copilot handled boilerplate and repetitive syntax. ChatGPT became his partner for debugging and architectural questions. And together, these tools gave him something he hadn’t had in years: momentum.

Now he’s revisiting his golf tournament app — but not to patch or tweak it. He’s rewriting it from scratch. Why? Because he’s finally seeing the gap. Not just in the code, but in the requirements. The devs he hired years ago didn’t understand golf. He does. And that domain knowledge — now that he can articulate it directly to AI — is his advantage.

That articulation process starts with ChatGPT’s voice mode. I showed him how to simply talk through what the app needed to do. The AI transformed his verbal explanations into a detailed engineering requirements document — one he could refine and expand as his thinking evolved. What used to take weeks of back-and-forth now takes hours.

This shift in his mindset is what makes vibe-coding so powerful. It’s not just about writing code faster. It’s about creating tighter alignment between the problem owner and the solution builder — even when they're the same person.

And that’s not just a personal story. That’s a signal for the enterprise.

Vibe-coding is the natural evolution of shadow IT. Business users have always found ways to bypass IT when the process slowed them down — spreadsheets, Access databases, rogue SaaS. Now, with generative AI, they’re building actual applications. But unlike traditional shadow IT, vibe-coding has the potential to integrate with enterprise governance — if we design for it. If we embrace it.

That’s the opportunity. And the risk. Because the people closest to the problem now have the tools to solve it — with or without IT. It’s up to us as leaders to decide whether we enable that safely, or watch it happen anyway.

This 67-year-old retired engineer isn’t just writing code. He’s writing the future — one vibe at a time.