What is vibe coding?
Vibe coding is building software by prompting an AI model and accepting its code without reading it, a term Andrej Karpathy coined in an X post on 2 February 2025. The human judges the behaviour and skips the diff, and stops when “it mostly works”. Karpathy said it is “not too bad for throwaway weekend projects”; dictionaries now also use the word for any AI code generation.
Origin Andrej Karpathy coined the term on 2 February 2025 in a post on X (opens in a new tab), and Simon Willison’s “Not all AI-assisted programming is vibe coding” (opens in a new tab) of 19 March 2025 narrowed it to building software with an LLM without reviewing the code it writes.
Also on Wikipedia (opens in a new tab)
- In Polish
- Vibe coding
- Closest ladder level
- Off the ladder
- Also searched as
- vibecoding, vibe-coding, vibe code
Then $19.99 a month or $99.99 a year. Cancel from your account page. 30-day refund on a first purchase.
Why it matters.
What makes something vibe coding is that no human reads the code. If a human reviews it, the OpenSSF glossary says “it is not vibe coding”, and Simon Willison calls code you reviewed, tested and can explain “software development”. This site’s prototype policy sorts work by exposure (users, data, external effects, reversibility), and Willison calls vibe coding “grossly irresponsible” when the software is for other people. Dan Shapiro’s ladder does not place it. Our reading: it shares Level 5’s trait that no human reads the diff. They differ in what replaces the reader: in vibe coding nothing does, at Level 5 a machine-checkable verification layer does.
Vibe coding vs agentic engineering
Both put a model or an agent in charge of the typing. They differ in who answers for the result and what ends the loop.
| Criterion | Vibe coding | Agentic engineering |
|---|---|---|
| What it names | A practice: prompt, glance at the result, accept it | A discipline: directing coding agents without lowering the quality bar |
| Who writes the code | The model | Agents, which the human orchestrates and oversees |
| Who checks it | Nobody reads the diff; the human watches behaviour | The human stays accountable; tests and evals check where they can |
| What stops the loop | “It mostly works” | The quality bar is met |
| Where the name came from | Andrej Karpathy, X post, 2 Feb 2025 | Earliest use found: 2 Jun 2025 (Mario Zechner); Karpathy endorsed it on 4 Feb 2026 |
| Where it breaks | When the software is for other people: Willison calls it “grossly irresponsible” there (6 May 2026) | Drift: Willison finds the two have “started to blur” as coding agents get more reliable (6 May 2026) |
How the meaning of vibe coding changed
The word began as a description of not reading the diff. Since then it has been stretched to cover any AI code generation, so check which sense a source means.
| Date | Source | How it uses “vibe coding” |
|---|---|---|
| 2 Feb 2025 | Andrej Karpathy, X | Defines it by what it skips: “I ‘Accept All’ always, I don’t read the diffs anymore.” |
| 19 Mar 2025 | Simon Willison | Narrows it to “building software with an LLM without reviewing the code it writes”, and warns that applying it to all AI-assisted code “dilutes the term” |
| 6 Nov 2025 | Collins Dictionary, Word of the Year | Broad sense: “the use of artificial intelligence prompted by natural language to write computer code” |
| 4 Feb 2026 | Andrej Karpathy, X | Looks back: “at the time, LLM capability was low enough that you’d mostly use vibe coding for fun throwaway projects, demos and explorations.” |
| 23 Feb 2026 | Simon Willison | Keeps the original: “its original definition of coding where you pay no attention to the code at all” |
| Sep 2026 | Merriam-Webster, dictionary entry | Neutral sense: “using an artificial intelligence system to generate computer code” |
Questions about vibe coding.
Who coined vibe coding, and when?
Andrej Karpathy, in an X post timestamped 23:17 UTC on 2 February 2025. Simon Willison’s write-up of 19 March 2025 gives the date as 6 February, but the post and Merriam-Webster’s etymology note both say 2 February. Collins Dictionary named vibe coding its Word of the Year for 2025 on 6 November 2025.
What is the difference between vibe coding and agentic engineering?
The difference is what checks the code. In vibe coding nothing does: the human watches behaviour and accepts the diff unread. Agentic engineering keeps the human accountable for a quality bar, with tests and evals checking where they can. Karpathy’s published transcript of his Sequoia Ascent talk, which says an LLM cleaned it up, puts it as “Vibe coding raises the floor. Agentic engineering is about extrapolating the ceiling.”
Is all AI-assisted coding vibe coding?
No. In its original sense the term excludes code a human has reviewed. Simon Willison writes that code an LLM wrote, which you then reviewed, tested thoroughly and can explain, is “not vibe coding, it’s software development.” The word has broadened: Merriam-Webster’s dictionary entry, as captured on 18 September 2026, defines it as “using an artificial intelligence system to generate computer code”.
Is vibe coding bad?
It depends on what the software can reach. Karpathy’s original post calls it “not too bad for throwaway weekend projects”. Simon Willison wrote on 6 May 2026: “If you’re building software for other people, vibe coding is grossly irresponsible because it’s other people’s information.” This site’s prototype policy sorts work by what it can reach: users, data and external effects.
The vocabulary of AI-driven development
Every name below is defined against vibe coding: who writes the code, who checks it, and what stops the loop.
Read the long-form guide: agentic engineering vs vibe coding
Sources.
The primary sources outside this site that this page relies on.
- Post on X introducing “vibe coding” (opens in a new tab) Andrej Karpathy, X
- Not all AI-assisted programming is vibe coding (but vibe coding rocks) (opens in a new tab) Simon Willison
- Collins’ Word of the Year 2025: AI meets authenticity as society shifts (opens in a new tab) Collins Dictionary
- Post on X looking back on “vibe coding” a year later (opens in a new tab) Andrej Karpathy, X
- Writing about Agentic Engineering Patterns (opens in a new tab) Simon Willison
- Vibe coding and agentic engineering are getting closer than I’d like (opens in a new tab) Simon Willison
- Sequoia Ascent 2026 summary (opens in a new tab) Andrej Karpathy
- “vibe coding” dictionary entry (Wayback Machine capture of 18 September 2026) (opens in a new tab) Merriam-Webster
- Vibe coding (opens in a new tab) OpenSSF, OpenSSF Glossary
Keep reading.
The guides that go deeper, and the terms and comparisons next to this one.
In the docs
- Vibe coding vs agentic engineering: 21 terms comparedFree
- The autonomy ladder: which level is your workflow at?Free
- The State of Agentic Engineering, September 2026Free
- Prototype policy: graduate by risk and evidenceSubscription
- Software built outside engineeringSubscription
- The full A-Z glossarySubscription
Related terms
Comparisons
Read the guides in the same words.
Open every guide with the 7-day free trial. Each term is defined once here, and the guides use it the same way.
Then $19.99 a month or $99.99 a year. Cancel from your account page. 30-day refund on a first purchase.