Teaching by Building: The Era of Vibe Coding

// November 25, 2025

There is a fundamental shift happening in how we build software. The era of memorizing syntax and fighting with semicolons is fading. In its place, a new standard is emerging: Vibe Coding.

We now have access to powerful AI agents and models that can write code faster and more accurately than we can. The role of the developer—and the role of Developer Relations—is shifting from "how do I write this function?" to "how do I orchestrate these models to build something incredible?"

I’ve spent the last few years at the intersection of education and product, from leading education at Makerpad to scouting AI startups. What I’ve learned is that developers want to see the potential of these models unlocked. They don't want a lecture on parameters; they want to see a product come to life.

From Syntax to Semantics

"Vibe Coding" isn't just a buzzword; it's a workflow. It's about holding the vision of a product—the "vibe"—and guiding AI agents to execute it.

When I teach now, I don't just type out code. I converse with models. I use tools like Replit's Agent or Google's Gemini to generate the heavy lifting.

This approach does two things:

1. It highlights model capability. It shows exactly what the AI is capable of handling, and where it needs human guidance.
2. It focuses on the product. We spend less time on boilerplate and more time on user experience, design, and functionality.

For platforms like Replit or Google DeepMind, this is the ultimate showcase. It demonstrates that their tools aren't just for writing code; they are for manifesting ideas.

Embracing the Iteration

In the old world of live coding, you'd worry about syntax errors. In the world of Vibe Coding, you worry about alignment.

Sometimes the model hallucinates. Sometimes it builds the wrong thing. Sometimes it just doesn't get the "vibe."

This is where the real teaching happens.

When an agent goes off track, I don't hide it. I show how to steer it back. I show how to refine the prompt, how to provide better context, how to debug the logic rather than the code. This teaches a skill that is becoming essential: AI literacy and orchestration.

It proves that while AI is powerful, the human in the loop is still the creative director.

Building for the Community

My work has always been about empowering builders. Now, that empowerment looks different. It's about showing people that they don't need a CS degree to build their dream app—they just need to know how to talk to the models.

• At Makerpad: I taught people to build without code.
• At Synthflow: I taught people to build with voice AI.
• Now at Shiny: I help people and businesses leverage AI to build products and increase productivity.

The Future of DevRel

The barrier to entry for building software has never been lower, but the ceiling for what's possible has never been higher. The future of DevRel belongs to those who can navigate this new landscape.

It belongs to the builders who can show the world not just how to code, but how to create. It belongs to those who are willing to experiment with the latest models, push them to their limits, and share the results—vibes and all.

I am one of those builders. And I’m ready to help the next generation of developers realize what’s possible.