Bliki: Vibe Coding

2026-05-21 · Source: Martin Fowler · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

Vibe coding, a term coined by Andrej Karpathy in February 2025, describes building software applications by iteratively prompting a Large Language Model (LLM) for changes and trying them out, crucially without ever reviewing the generated code. This technique enables individuals without programming knowledge to create applications. While useful for rapid development of prototypes or disposable software for limited audiences, vibe coding presents significant limitations. These include issues with maintainability due to low code quality, potential correctness problems stemming from LLM hallucinations, and serious security risks, such as exposing sensitive information or credentials. It differs from "Agentic Programming," where developers actively review and manage LLM-generated code, emphasizing the "forget that the code even exists" aspect as both its utility and primary constraint.

Key takeaway

For non-technical professionals or developers needing quick prototypes, vibe coding offers a path to rapidly create disposable software. However, if your application handles sensitive data, will be widely used, or requires long-term maintainability, you must avoid this approach. Be aware of the significant security vulnerabilities, low code quality, and potential for incorrect behavior that arise from not reviewing LLM-generated code. Prioritize traditional development or agentic programming for critical systems.

Key insights

Vibe coding is building software via LLM prompts without code review, suitable for disposable applications.

Principles

• Code review is critical for quality and security.
• LLM-generated code requires skepticism.
• Separate concepts despite semantic diffusion.

Method

Iteratively prompt an LLM for software features and changes, test outputs, and accept all modifications without inspecting the underlying code.

In practice

• Build personal, throwaway weekend projects.
• Rapidly prototype simple web applications.
• Create tools for a close group of collaborators.

Topics

• Vibe Coding
• LLM Application Development
• Agentic Programming
• Software Security
• Code Quality
• Rapid Prototyping

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Student, Software Engineer, AI Engineer

Related on AIssential

• AI tools for hearing difficulties — helpful or harmful for language learning? — AI assistive technology presents a dilemma for language learners with hearing difficulties, balancing aid with skill development.
• How Enterprise Marketing Teams Are Actually Using AI Agents in 2026 — Enterprise marketing is shifting from human-assisted AI tools to autonomous AI agents that redesign entire workflows.
• [AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models — Microsoft's MAI models showcase a transparent, full-stack AI strategy emphasizing clean data and local agent execution.
• Are we getting better at explaining things because of our interaction with LLMs? — Crafting clear LLM prompts may enhance human explanatory and communication abilities.
• OpenAI expands Codex with role-specific plugins to build a general-purpose app for non-developers — OpenAI is transforming Codex into a general-purpose, role-specific AI work app, expanding beyond coding for non-developers.
• How to Use Claude Managed Agents? — Anthropic's Claude Managed Agents simplifies AI agent deployment by providing a hosted, stateful, and secure execution environment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Martin Fowler.