# Vibe Engineering vs. Vibe Coding: Know the Difference — Pilot to Production

> Vibe coding accepts whatever AI produces. Vibe engineering uses AI while staying suspicious. One ships demos, the other ships production code.

Canonical: https://thegrowthproject.com/podcast/vibe-engineering-vs-vibe-coding/

*Pilot to Production*, the Growth Project podcast — hosted by Sam and Maya.

- Listen: https://thegrowthproject.com/podcast/vibe-engineering-vs-vibe-coding/
- Read the article: https://thegrowthproject.com/blog/vibe-engineering-vs-vibe-coding/
- Audio: https://thegrowthproject.com/audio/podcast/vibe-engineering-vs-vibe-coding.m4a?v=7c14cba1

## Transcript

**Sam:** Managers have been vibe coding forever. They tell developers what to build, test the app, never read the code.

**Maya:** That is fine. It is their job. The problem starts the day a developer does the exact same thing.

**Sam:** Welcome to Pilot to Production, from the Growth Project. I'm Sam.

**Maya:** And I'm Maya. Today: vibe coding versus vibe engineering, and why one ships demos while the other ships production code.

**Sam:** Okay, start at the start. Where does vibe coding even come from?

**Maya:** Andrej Karpathy coined it in early twenty twenty-four. Let the AI write your code, do not worry too much about understanding it, just see if it works.

**Sam:** And it went viral. Everyone was vibe coding.

**Maya:** Right. And here is the line from the post that stuck with me. Most of them should not be.

**Sam:** Harsh. Why?

**Maya:** Because vibe coding treats the AI as a black box. Input a problem, output a solution. Do not look too closely at what is inside.

**Sam:** Which, to be fair, is exactly what Karpathy was doing.

**Maya:** Exactly. He was building a one-off tool for personal use. It did not need to be maintainable. It did not need to handle edge cases. It just needed to work once.

**Sam:** So the tool is fine. The problem is where people point it.

**Maya:** That is the whole thing. The problem is not vibe coding itself. It is vibe coding the things that should be vibe engineered.

**Sam:** So define vibe engineering for me. Cleanly.

**Maya:** You are still using AI to write the code. But you are not trusting it. You are watching, you are suspicious, you know it is going to make mistakes, and your job is to catch them before they ship.

**Sam:** There is a quote in the post for this, right?

**Maya:** There is. Quote. When you are using agents to code all the time, you do not touch the code, but you just look at your screen like, I am going to catch you.

**Sam:** I am going to catch you. I love that. So the AI does the typing.

**Maya:** And you do the thinking. That is the line. You are delegating the typing, not the thinking.

**Sam:** Okay, but be concrete. What does vibe engineering actually look like at the keyboard?

**Maya:** Five habits from the post. Start with clear intent. Watch the first few lines. Question the suggestions. Understand before you merge. And know when to stop.

**Sam:** Walk me through clear intent, because everyone thinks they do this.

**Maya:** The post has a great pair. Bad prompt: add user authentication. Good prompt: implement email password auth using NextAuth, store sessions in the database, middleware for dashboard routes, rate limiting on login attempts.

**Sam:** So if your description is vague, the output is vague.

**Maya:** Every time. The AI executes what you describe.

**Sam:** And watching the first few lines, why does that matter so much?

**Maya:** Because the AI starts wrong, regularly. Wrong pattern, wrong architecture, wrong assumption. Catch it in the first three lines and you redirect. Wait until it is done and you are rewriting everything.

**Sam:** Here is my sharp question. The post says not everyone can even do this.

**Maya:** Correct. There is an experience spectrum. Juniors say, hell yeah, give me everything. Mid-levels say, this will never be good enough, my code is better. Seniors say, this is a tool, let me use it properly.

**Sam:** So juniors vibe code because?

**Maya:** They do not know what to look for. They cannot spot when the AI is wrong. Mid-levels resist because they are protective of their craft. Seniors vibe engineer because they know what good looks like.

**Sam:** Which leads to the counterintuitive bit.

**Maya:** The best results come from giving AI tools to skeptical seniors and convincing them to use them. They have the judgment to supervise.

**Sam:** Give people the first thing to do tomorrow morning.

**Maya:** Pull your last three AI-generated commits. Can you explain what the code does and why? If not, you are vibe coding. Then check your accept rate. If you take the first suggestion more than fifty percent of the time, you are not scrutinizing enough.

**Sam:** And one mindset shift?

**Maya:** Add the question pause. Before you accept, pause and ask: does this follow our patterns, what happens if this fails, what is it missing. If you cannot answer, you are vibe coding.

**Sam:** The bottom line in one breath.

**Maya:** AI tools are amplifiers. If you know what good code looks like, vibe engineering makes you faster. If you do not, vibe coding makes you dangerous.

**Sam:** This has been Pilot to Production, from the Growth Project. If your demos look great but your production code does not, that gap is judgment, and that is what we teach, at thegrowthproject.com.

**Maya:** Thanks for listening. See you next time.