What Is a Vibe Coder? The Term That Went From Joke to Slur to Job Title in 12 Months

On February 2, 2025, Andrej Karpathy — co-founder of OpenAI, former AI lead at Tesla — fired off a tweet that he later described as a “shower of thoughts throwaway.”

“There’s a new kind of coding I call ‘vibe coding’, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists.”

One year later, the term has its own Wikipedia article, its own Upwork marketplace category, its own culture war — and one of the most schizophrenic identities in tech.

Depending on who you ask, “vibe coder” is either:

• The future of software development
• A slur
• A $200K+ job title at Amazon and Google
• A meme
• All of the above, simultaneously

Let’s talk about how we got here.

Act I: The Meme (February–April 2025)

The first few months were fun. Karpathy’s tweet resonated because it named something developers were already doing but felt weird about: using ChatGPT, Copilot, or Claude to generate code, accepting the output without fully reading it, and somehow shipping working software.

The vibe was light. People shared screenshots of Claude writing entire React apps. They made jokes about “accepting all suggestions.” The Rick Rubin comparison went viral — because Rick Rubin famously produces albums by vibing in the studio, not by playing instruments.

Nobody took it too seriously. It was a joke. A fun one.

Act II: The Slur (May–September 2025)

Then the term got weaponized.

As “vibe coding” spread beyond the early adopter bubble, traditional developers started using it as an insult. Not playfully — dismissively.

Peter Steinberger, a well-known iOS developer, said it plainly: “vibe coding is a slur.”

And in a lot of contexts, it was. Engineers reported being called “vibe coders” by managers as a putdown — even when they were doing legitimate engineering work.

The pattern was familiar. Senior developers with years of experience felt threatened by people who could ship working software without understanding the fundamentals. Junior developers who’d learned to build with AI felt dismissed by people who’d memorized algorithms they’d never use in production.

The “real programmer vs. vibe coder” debate exploded. And if you’ve been on the internet for more than fifteen minutes, you know exactly how it went.

The Eternal Holy War

The “real programmers vs. vibe coders” debate isn’t new. It’s the exact same argument that’s been happening in tech since the first compiler was written:

• Assembly programmers said C programmers weren’t “real”
• C programmers said Java programmers weren’t “real”
• Backend developers said frontend developers weren’t “real”
• “Real” developers said no-code builders weren’t “real”
• And now: traditional developers say vibe coders aren’t “real”

The same pattern is happening everywhere AI touches creative work. AI musicians vs. “real” musicians. AI artists vs. “real” artists. AI writers vs. “real” writers. It’s the same gatekeeping instinct, dressed up in a new outfit every generation.

Pat Walls put it well: “VIBE CODERS DO NOT BUILD REAL APPS. But then why do I get DMs like this literally every single day?”

The counter-argument is equally real: vibe-coded software often has genuine problems. Security vulnerabilities, spaghetti architecture, technical debt that metastasizes. 404 Media reported that freelancers are now charging $200/hour to fix vibe-coded messes, with some reporting a 300% increase in demand over six months.

So both sides have a point. Vibe coders can ship working software fast. That software can be a mess. The question isn’t which side is right — it’s what you do about it.

Act III: The Job Title (October 2025–Present)

Then something nobody expected happened: HR departments got involved.

In late 2025, “vibe coder” started showing up in actual job descriptions. Not as a joke. Not as a euphemism. As the literal job title.

Leadr.co posted a role called “AI Developer / Vibe Coder” paying $45K–$75K. Suncoast Venture Studio listed a “Vibe Coder, Applied AI & Rapid MVP Builder” at $75K–$95K. These weren’t parody postings.

But here’s where it gets truly schizophrenic: at the same time developers on Twitter were calling “vibe coder” a slur, Amazon was requiring “vibe coding and AI-assisted prototyping” as a core skill for Senior UX Designers paying $138K–$212K. Google listed “vibe coding tools” as a minimum qualification for a $265K UX Engineering Manager role. Oracle put “vibe coding practices” in the requirements for a Principal Member of Technical Staff.

The term is simultaneously an insult and a six-figure job requirement at the world’s largest companies. That’s the state of “vibe coder” in 2026.

So What Actually IS a Vibe Coder?

Strip away the culture war, and a vibe coder is someone who:

Uses AI tools as a primary development workflow. Not occasionally. Not as a novelty. As the default way they build software. Tools like Cursor, Claude Code, Copilot, Replit, v0, Bolt, or Lovable are integral to how they work.

Prioritizes shipping over architecture. The goal is working software, fast. Perfect code structure is secondary to getting something into users’ hands. This doesn’t mean the code is bad — it means the priority order is different.

Describes intent, then iterates on output. Instead of writing code line by line, they describe what they want, review what the AI produces, and iterate. The skill isn’t typing — it’s knowing what to ask for and evaluating whether the result is correct.

Operates across the full stack. Because AI tools can generate code in any language and framework, vibe coders tend to be generalists. They might build a React frontend, a Python backend, and a deployment pipeline in the same afternoon — not because they’re experts in all three, but because the AI handles the syntax and they handle the architecture and intent.

The Skill Set (What Actually Gets You Hired)

Based on hundreds of real job listings we track on Vibehackers, here’s what companies actually look for:

Tool Fluency

• AI coding assistants: Cursor, Claude Code, Copilot, Windsurf
• Rapid prototyping platforms: Replit, v0, Bolt, Lovable
• Agentic workflows: MCP (Model Context Protocol), RAG, tool calling
• The meta-skill: Knowing which tool to use for which problem

The Judgment Gap

This is the part that separates good vibe coders from disaster:

• Knowing when AI output is wrong (it often is)
• Understanding security implications of generated code
• Recognizing when to slow down and architect properly
• Evaluating whether the AI is solving the right problem

The gap between “accepting all AI suggestions” and “using AI as a force multiplier” is the difference between a junior vibe coder and a senior one. Experience level still matters — it just manifests differently.

Speed Without Chaos

The best vibe coders aren’t just fast — they’re fast and organized:

• Prototyping in hours, not weeks
• Iterating with tight feedback loops
• Shipping MVPs that are actually extensible
• Knowing when a prototype is ready for production hardening

Who Becomes a Vibe Coder?

The interesting thing about vibe coding as a career path is who’s entering it:

Career switchers. People from product management, design, marketing, and operations who never learned to code traditionally but can now build software with AI tools. Lenny’s Newsletter profiled a professional vibe coder at Lovable who builds internal tools and customer-facing products without a coding background.

Senior engineers who adopted AI early. Staff and principal engineers who use AI to move 5x faster on complex problems. They’re not “less technical” — they’re leveraging AI to amplify decades of experience.

Designers who ship. UX designers who go beyond Figma mockups and deliver functional prototypes. Amazon, Google, and Microsoft are all hiring for this explicitly.

PMs who build. Product managers who prototype their own ideas instead of writing PRDs and waiting for engineering allocation.

The Two-Sided Market

Here’s the punchline that nobody predicted: vibe coding created two career paths, not one.

Path 1: Vibe Coders — people who ship fast using AI tools. Roles range from $45K entry-level positions to $300K+ senior roles at FAANG companies.

Path 2: Vibe Code Cleanup Specialists — people who fix what vibe coders ship. Cleanup specialists on Fiverr and Upwork report working with 15-20 regular clients, charging $200/hour to untangle AI-generated codebases.

Both paths pay well. Both are growing. And the smartest people are positioning themselves to do both — ship fast when speed matters, then harden and refactor when stability matters.

The Rebranding Attempt

Some people have tried to escape the culture war by rebranding. Peter Steinberger — the same person who called vibe coding a slur — later proposed “agentic engineering” as the professional alternative.

Others have pushed for “AI-assisted development,” “AI-native engineering,” or “prompt-driven development.” None of these have stuck. The market chose “vibe coding” and “vibe coder,” and that’s what shows up in job titles, Upwork categories, and Wikipedia articles.

Sometimes the meme wins.

Where We Are Now (March 2026)

Here’s the state of play as we enter March 2026:

The term is permanent. With a Wikipedia entry, dedicated job boards, Upwork marketplace categories, and FAANG job descriptions, “vibe coder” isn’t going away. It’s been operationalized.

The culture war continues. And it will continue. It’s the same argument that’s happened with every technological shift in programming history. The assembly purists eventually lost. So did the “JavaScript isn’t a real language” crowd. The pattern is clear — but it takes years to play out.

The market doesn’t care about the debate. Companies are hiring vibe coders. Y Combinator reports that 25% of their W25 batch had codebases that were 95% AI-generated. An Indian vibe-coding startup called Emergent raised $70M at a $300M valuation. The money has spoken.

The identity is what you make it. Some people wear “vibe coder” as a badge of honor. Others find it reductive. Both reactions are valid. What matters is what you build.

How to Become One (If You Want the Label)

If you want to develop vibe coding skills — whether you call yourself a “vibe coder” or not:

1. Pick your AI coding tool. Cursor, Claude Code, or Copilot. Use it daily for a month. Our tools comparison covers the tradeoffs.

2. Build something real. Not a tutorial project. Something with users, or at least something you’d actually use. The getting started guide walks through the whole process.

3. Learn what AI gets wrong. Spend a week specifically looking for bugs, security issues, and architectural problems in AI-generated code. This is the skill that separates professionals from hobbyists.

4. Ship publicly. Put your work on GitHub, demo it on Twitter, show it in job applications. The vibe coder portfolio is built projects, not credentials.

5. Find the right role. Browse vibe coding jobs to see what’s actually hiring. Filter by your category — the roles span engineering, design, product, marketing, and more.

The Bottom Line

A vibe coder is someone who builds software with AI as a primary tool. That’s it. Everything else — the culture war, the slur discourse, the memes, the rebranding attempts — is noise around a simple shift in how software gets made.

The term started as a joke. It became an insult. Now it’s a job title at Amazon, Google, Microsoft, Oracle, Salesforce, and hundreds of startups.

Whether you love the label or hate it, the behavior it describes is becoming the default way software gets built. The only question is whether you’re building with it or arguing about it.