Skip to content

Vibe Coding

Summary

Vibe coding is natural-language-driven software development where AI generates code from prompts. It democratizes building but introduces risks: security debt, cognitive debt, code bloat, hallucination, and a "lost generation" of engineers who can't debug what they didn't write.

Concept Summary
Vibe Coding Two modes: "pure" (forget code exists) vs responsible AI-assisted
Ai Slop Bloated, buggy, insecure AI-generated content
Agentic Workflows Wave 3 of AI automation: natural language describes outcome
Code Bloat AI generates 3-4x more code than humans
Cognitive Debt Understanding erodes from over-reliance on AI
Security Debt 45% of AI code has OWASP vulnerabilities
Vibe Deploying One-click deployment removing DevOps bottleneck
Hallucination LLMs predict tokens, not facts
Overconfidence Effect AI-assisted devs produce less secure code but believe it's more secure
Lost Generation AI replaces junior-level learning work
Runaway Loops Agents spiraling off-rails with vague instructions
Production Wall 6 obstacles non-coders face shipping production apps
Three Waves Of Ai Automation Chatbots → AI+automation platforms → agentic workflows

Key Entities

Entity Role
Andrej Karpathy Coined "vibe coding" (early 2025)
Swaroop A Documented 6 production obstacles as non-coder
Mahnoor Faisal Used vibe coding to learn programming
Logically Answered Critically examined security and reliability problems
Dave Farley Co-authored 150-developer AI maintainability study
Joanna Stern Independent tech journalist, human-centered approach

Key Sources

Source Type Date
Vibe Coding Explained Google Cloud article 2026-03-20
Vibe Coding Non Coder Swaroop article 2026-04-20
Vibe Coding For Learning Xda article 2026-02-16
Stop Learning N8N Nate Herk video 2026-03-21
Rise And Fall Vibe Coding Logically Answered video 2025-10-18
150 Developers Ai Study Dave Farley video 2026-01-28
Joanna Stern New Things Intro video 2026-04-22

What's Missing

  • Longitudinal studies of vibe coding impact on code quality over years
  • Economic analysis: cost comparison of vibe coding vs hiring developers
  • Regulatory/policy perspectives on AI-generated code liability
  • Success stories of production apps built entirely via vibe coding at scale