From Zero to Six Projects: What Two Months with AI Coding Tools Actually Taught Me

In November 2025, I had zero development experience. By January 2026, I'd started building six web applications using AI-assisted tools — three are making real progress, while the others await more attention. This isn't a story about becoming a developer. It's about what happens when someone with no technical background tries to understand where technology is actually going.

01.Can You Build Apps Without Coding Experience?

Let me be clear about where I started. End of November 2025. Zero coding experience. I didn't know what a component was, couldn't tell you the difference between frontend and backend, and had never opened a terminal in my life.

Two months later, I've started six projects: a book tracking platform, a skincare ingredient analyser, a community reading app, a professional document validator, and this blog you're reading now. Three are making solid progress. The others — particularly the ones designed for corporate use — need more time and attention due to their complexity.

I'm not telling you this to impress you. I'm telling you because the gap between "knowing nothing" and "making meaningful progress on real applications" has collapsed in ways I didn't expect — and that has implications for everyone.

02.Which AI Coding Tools Actually Work for Non-Developers?

The first surprise wasn't that AI could help with coding. Everyone knows that by now. The surprise was the state of the tools available today.

Claude Code in VS Code. ChatGPT's canvas. Gemini's project generation. Cursor's agent mode. These aren't experimental features buried in documentation. They're polished AI coding tools for beginners and experienced developers alike — products that understand context, remember your project structure, and can carry a conversation about what you're trying to build.

I expected these tools to exist in some limited form. I didn't expect them to work this well for someone starting from zero.

The speed of change is disorienting. When I started in late November, certain approaches were standard. By mid-December, new versions had shipped. By January, the recommended workflow had shifted again.

This isn't a complaint — it's an observation. The pace of development in this space makes it genuinely difficult to write guides that stay current for more than a few weeks. What I learned in week one was already partially obsolete by week four.

03.The Paradox of Choice

There's so much that it's paralysing.

Which framework? Next.js or Remix? Which AI tool? Claude or ChatGPT or Gemini? Which database? Supabase or Firebase or PlanetScale? Which deployment platform? Vercel or Netlify or Railway?

For someone with development experience, these are informed decisions. For someone starting from zero, they're essentially random choices made based on which YouTube tutorial you happened to watch first.

I spent more time than I'd like to admit rebuilding the same project with different tools, not because the first approach was wrong, but because I didn't have enough context to know it was right.

04.The Gap Between "Working" and "Right"

AI tools can generate code that runs. That's not the same as code that's correct, secure, or maintainable.

I built features that worked perfectly in development and broke immediately in production. I created database structures that made sense for simple cases but became nightmares when I tried to add complexity. I wrote components that functioned but violated every convention of the frameworks I was using.

The tools don't tell you when you're building technical debt. They just build what you ask for.

05.Persistence Is the Real Skill

The technology is accessible. The tutorials exist. YouTube has everything you need to learn anything you want to build.

What matters is whether you keep going when things break — and things will break constantly.

Every project I built went through phases where nothing worked, where error messages made no sense, where I spent hours on problems that had five-minute solutions I just couldn't see.

The people who succeed with these tools aren't the ones who find it easy. They're the ones who keep trying when it's not.

06.Context Is Earned, Not Given

AI tools are powerful collaborators, but they work best when you can give them context about what you're trying to achieve.

In week one, I couldn't give useful context because I didn't understand what I was doing. By week six, I could explain what I wanted, why I wanted it, and what constraints mattered.

The irony is that learning to use AI coding tools effectively required learning enough about coding to know what to ask for. The tools accelerated my learning, but they didn't replace it.

07.Why This Matters Beyond Personal Projects

We're in a period of significant global transformation. AI is reshaping work. Geopolitical structures are shifting. The skills that mattered five years ago aren't necessarily the ones that will matter five years from now.

I don't know what the implications of all this will be. But I do know that waiting on the sidelines to see how things develop feels increasingly risky.

The barrier to building things has dropped dramatically. That doesn't mean everyone needs to become a developer. But understanding what's possible — and experiencing firsthand how quickly "impossible for non-technical people" has become "achievable with persistence" — seems valuable regardless of your field.

08.The Honest Summary

Two months in, I'm building functional web applications without traditional coding skills. Not finished — building. I'm not a developer. I don't have deep technical knowledge. I couldn't whiteboard a system architecture or explain memory management.

But I can take an idea, work with AI coding tools to make meaningful progress toward a working product, and deploy it to the internet. The simpler projects work. The complex ones are getting there. That trajectory — a non-developer building websites that actually function — wasn't possible even a year ago.

If you're curious about where this technology is going, the best way to understand it is to try building something yourself. Not to become a developer — but to develop intuition about what these tools can and can't do.

The learning curve is real. The frustration is real. But the capability on the other side is also real.

The question isn't whether you have the technical background to try. It's whether you have the persistence.

---

FAQ

Do I need to learn programming to use AI coding tools?

No. The tools handle syntax and structure. What you need is clarity about what you're trying to build and persistence when things break. You won't become a developer, but you can build functional applications.

Which AI coding tool is best for beginners?

I've used Claude Code, ChatGPT, Gemini, and Cursor. Each has strengths — Claude Code for deep context understanding, Gemini for quick prototypes, Cursor for IDE integration. Start with whichever has the best free tier and switch if needed. The skills transfer across tools.

How long does it take to build something functional?

My first working project took about two weeks of evening sessions. But "working" and "production-ready" are different things. Expect months for anything complex, especially if it involves databases, authentication, or integrations.

Is the code AI generates actually good?

It works. Whether it's correct, secure, or maintainable is a different question. The tools don't tell you when you're building technical debt — they just build what you ask for. Learning to evaluate what they produce is part of the journey.