NewsletterBlogGlossary

Vibe Coding

What is vibe coding? An AI-driven development style where you describe intent and let LLMs generate the code.

techniques
ShareXLinkedIn

Vibe Coding — AI Glossary

Vibe coding is a software development approach where the programmer describes what they want in natural language and lets an AI model generate the actual code. Coined by Andrej Karpathy in early 2025, the term captures a style where you "fully give in to the vibes" — accepting AI-generated code without necessarily reading or understanding every line, iterating by describing bugs and desired behavior rather than editing source directly.

Why Vibe Coding Matters

Vibe coding has lowered the barrier to building functional software. Non-engineers — designers, product managers, researchers — can now prototype working applications by conversing with tools like ChatGPT, Claude Desktop, Cursor, or Replit. For experienced developers, it accelerates throwaway prototypes and internal tools where correctness standards are relaxed.

The approach has sparked legitimate debate in the engineering community. Critics point out that vibe-coded software often carries hidden bugs, security vulnerabilities, and technical debt that compound once a project grows beyond a prototype. Proponents argue it democratizes creation and that AI-generated code quality improves with each model generation. Our coverage of AI agent capabilities explores how autonomous coding tools are pushing this boundary further.

How Vibe Coding Works

A typical vibe coding session follows a tight feedback loop:

  1. Describe the desired feature or behavior in plain language
  2. Accept the AI-generated code — often without detailed review
  3. Run the code and observe the result
  4. Iterate by describing what's wrong or what to change next

This differs from agentic coding, where the AI autonomously plans, executes, and validates multi-step tasks. Vibe coding is more conversational and human-directed, though the line between the two is blurring as tools gain autonomy. The key distinction: vibe coding prioritizes speed and intuition over rigor — it works best for prototypes, personal projects, and exploratory development rather than production systems.

  • Agentic Coding: Autonomous AI-driven development where the model plans and executes entire workflows — a more structured evolution of vibe coding
  • ChatGPT: OpenAI's conversational AI, one of the most common tools used for vibe coding sessions
  • Claude Desktop: Anthropic's desktop application that enables conversational coding and file editing workflows

Want more AI insights? Subscribe to LoreAI for daily briefings.