What Is Vibe Coding?
Vibe Coding is a term coined by Andrej Karpathy (ex-OpenAI, ex-Tesla) that describes a way of programming where you stop writing code line by line and start describing intentions to AI.
You say what you want. The AI writes it. You review, adjust, iterate.
The concept went viral on Twitter/X and sparked debate about whether this is the future of development — or a dangerous shortcut.
The Criticism: You Stop Understanding Code
The main criticism of vibe coding is that it creates a generation of "programmers" who don't understand what they're doing — accepting AI-generated code without the ability to review or debug it.
And honestly, there's a point there.
I use AI intensively in my work. Cursor, Claude, ChatGPT. These are tools that multiply my productivity massively. But there's a huge difference between:
- A dev who uses AI as a tool: understands the generated code, can identify bugs, knows how to adapt and improve it
- A pure vibe coder: accepts everything the AI generates, hopes it works, panics when it breaks
The second profile is a problem in production.
What the Critics Are Getting Wrong
But the critics are also making a classic mistake: judging a technology by its worst use case.
Done right, vibe coding isn't "paste a prompt and accept everything." It's a fast cycle of:
- Describe what you want
- Review what was generated with a critical eye
- Iterate until you get the right result
- Understand the final code before pushing to production
When you have the technical foundation to run that cycle properly, the productivity gain is insane. I've built complete systems in days that would have taken weeks before.
How I Use It in Practice
In my projects (Laravel, Next.js, Python), the AI workflow looks like this:
Before: Think → Write → Test → Debug → Refactor (all manually)
Now: Describe → Review → Adjust → Test → Merge
The difference is that the heavy mechanical writing is out of the equation. My time goes toward architecture decisions, critical review, and adjustments that AI can't do on its own.
That's vibe coding done right.
Who It's For and Who It Isn't
Vibe coding done well is for people who already have technical foundations and want to multiply their productivity.
It's not for people learning to code. If you're still building the mental model of how code works, blindly accepting what AI generates will hurt more than help. You won't understand the errors when they show up — and they will show up.
Learn the basics first. Then use AI to accelerate.
The Future Is Hybrid
The debate between AI skeptics and vibe coders is a false dichotomy.
The devs who will stand out in the next few years aren't those who refuse AI on principle or those who use it mindlessly. They're the ones who can orchestrate AI with solid technical judgment.
Vibe coding is a new skill the market will demand. But technical foundations remain the differentiator that separates those who can do it right from those who will break production on a Friday afternoon.
Prefer Watching?
I made a short video about this on TikTok:
