14 Years Away From Code. Then AI Changed Everything.

I left engineering 14 years ago.

In between, I became a PM. Did project management, data analysis. Occasionally wrote Python scripts to automate things. But building a complete product from scratch? The barrier was just too high.

So I went looking for alternatives.

The No-Code Experiment

In 2024, I discovered No-Code. Bubble was my first deep dive.

“Build products without writing code?”

I spent six months mastering Bubble. Dynamic data, plugin integrations, complex workflows—I learned it all. Even shipped a working SaaS product.

I wrote 8 tutorial articles about Bubble. Search “Bubble.io” in Chinese, and my articles rank #2 on Google. Right after Bubble’s official site.

I thought I’d found the answer.

But then reality hit.

The product was done. What about the landing page? Bubble’s SEO is terrible, and it’s slow. Not suitable. Chrome extension? Bubble doesn’t support that.

And there’s a more fundamental problem: I don’t control the deployment. No error logs. No system monitoring. When something breaks, you’re just guessing.

“Switch to another tool then?”

I tried Webflow. The result? Back to Webflow 101. Its layout logic is completely different from Bubble. Everything I learned in Bubble? Useless here.

This is the truth about No-Code:

Every tool is a walled garden. Switch to another, and you start from zero.

Then Came June This Year

Claude Code launched.

I opened it with skepticism.

And then I couldn’t stop. 12+ hours a day, completely immersed.

AI doesn’t push back. Doesn’t say “that requirement doesn’t make sense.” Doesn’t schedule endless meetings. You tell it what you want, it builds it. Don’t understand something? It explains. Made a mistake? It fixes it.

Week one, I was nervous.

Started with installing VS Code. Opened it up—dark screen, walls of code.

“Do I need to understand all this?” “Or should I just let AI handle it?”

Week two, something clicked.

Not the feeling of “I can write code.” The feeling of “I know how to collaborate with AI.”

Because this is exactly what I do at work every day: define problems, break down requirements, validate outputs, coordinate across teams.

The only difference? My counterpart used to be engineers. Now it’s AI.

Six months later: three complete projects. Backend API, frontend app, Chrome extension. A working SaaS with payment processing, user authentication, PDF handling, AI analysis. Listed on both Chrome Web Store and Edge Add-ons.

Someone who left engineering 14 years ago built all of this.

The Key Difference

No-Code and AI coding solve the same problem: letting non-programmers build products.

But there’s a crucial difference.

With No-Code, you learn tools. Bubble’s logic, Webflow’s layout system, Zapier’s automation flows. Switch tools, start over.

With AI coding, you learn communication. Today it’s Claude, tomorrow it’s GPT, next month it’s something else. The core skill remains: articulate the problem clearly, break down requirements, validate results.

This skill doesn’t reset when you change tools.

“But don’t you need to manage servers?”

Used to. Not anymore. Azure CLI + GitHub Actions. Connect to GitHub, auto-deploy. Push code, and it’s live.

Final Thoughts

If you ask me what’s the biggest value of AI coding?

It’s not “writing code faster.”

It’s “enabling people who wanted to build products but couldn’t, to finally build them.” Even solo.

14 years of distance, bridged by one tool.

Still feels surreal.

If you’re thinking about building something, let’s talk.

Want to see what I built? → AI Resume Advisor