Vibe Coding
Vibe coding is building software by describing intent in natural language and letting an AI agent write the code, judging results by behavior.
Vibe coding is a style of software development where you describe what you want in natural language, let an AI coding agent generate the implementation, and evaluate the result by how it behaves — running it, clicking through it — rather than by reading every line of code.
The term was popularized by Andrej Karpathy in early 2025 and named something real: with agentic tools like Claude Code, Cursor, and v0, the bottleneck shifted from writing code to specifying intent and verifying outcomes. By 2026, surveys put AI-generated code at roughly half of all new code at many companies, and "which agent fits your workflow" replaced "should we use AI" as the practical question.
The distinction that matters is stakes. Vibe coding shines where iteration speed beats rigor — prototypes, personal tools, UI exploration. It bites where unreviewed code carries real risk: anything handling money, auth, or other people's data. Professional agentic workflows split the difference — the AI writes most of the code, but tests, permission guardrails, and human review of the diff stay in the loop.
For the disciplined version of this workflow, start with What Is Claude Code? and Prompt Patterns for Coding Agents.
Frequently asked questions
- Who coined the term vibe coding?
- Andrej Karpathy popularized it in early 2025, describing a style of building where you 'fully give in to the vibes' — prompting an AI for code, accepting its output, and steering by running the result rather than reviewing every diff.
- Is vibe coding bad practice?
- It depends entirely on the stakes. For prototypes, internal tools, and exploration, it's a legitimate speed multiplier. For production systems, unreviewed AI code accumulates risk — security holes, hidden assumptions, unmaintainable structure — so professional workflows keep human review, tests, and permission guardrails in the loop even when the AI writes most of the code.
Related
- What Is Claude Code?A grounded explanation of Claude Code: an agentic command-line coding tool that reads files, runs commands, and works in a loop toward a goal.
- Cursor vs Claude Code vs GitHub Copilot vs Windsurf in 2026A practical, opinionated comparison of the four mainstream AI coding tools — form factor, agentic depth, model choice, and who each one is for.
- Prompt Patterns for Coding AgentsPractical prompting patterns: chaining, few-shot, context management, tool use, and output structuring.
- AI AgentAn AI agent is an LLM-driven system that pursues a goal in a loop — calling tools, observing results, iterating — instead of returning one answer.
- Agentic AIAgentic AI is the class of AI systems that act toward goals — planning, calling tools, and iterating on results — rather than only generating content.
- Best AI App Builders in 2026: v0 vs Lovable vs Bolt vs ReplitThe prompt-to-app builders compared — v0 for production UI, Lovable for full apps, Bolt for in-browser velocity, Replit for build-and-host in one place.
- v0 vs Lovable: AI App Builders Compared (2026)v0 vs Lovable — Vercel's generative UI tool vs the full-app builder. Component quality vs end-to-end apps, code ownership, and who each serves best.
- Vibe Coding in 2026: What It Is, When It Works, When It BitesAn honest guide to vibe coding — where prompt-and-accept development genuinely pays, where it accumulates risk, and the guardrails that make it professional.
- AI SlopAI slop is low-effort, mass-produced AI-generated content — fluent, generic, and unchecked — flooding feeds, search results, and codebases.