NotVibesCodedAuthor & editor
The editorial team behind NotVibesCoded
NotVibesCoded is written and reviewed by its editorial team under the standards in the editorial policy. A named author page, with full credentials, replaces this placeholder shortly.
Every article on NotVibesCoded is written or reviewed by the author before it is published, under the standards in the editorial policy. Corrections are welcome and are fixed in the open: get in touch.
Three AI coding tools, three different philosophies. How Claude Code, Cursor, and GitHub Copilot actually differ in practice, and which one fits which kind of developer.
AI-built sites ship fast and load slow. Twenty fixes, grouped by where the slowness lives: frontend, network, backend, and the process that keeps it fast.
A claim going around says models now write better code than engineers, so 'vibe coding' is dead and everything is 'precision engineering.' Wrong conclusion. The term was never about the model's skill. It's about yours.
Everyone ranks AI builders by how fast the demo looks. We ranked them by what matters: the code they leave behind. Eight tools, one rubric, zero mercy.
Agents write migrations fast, and database mistakes are the ones you can't roll back from coffee. What to verify before an AI-generated migration touches production data.
AI coding agents love inventing tokens: bgPrimaryBlue, color-blue-1, primary_hover_2. Here's a 15-minute audit that catches naming chaos, missing semantics, and accessibility failures before they fossilize in your codebase.
Every coding agent reads a context file, CLAUDE.md, AGENTS.md, Copilot instructions, and most of them are decoration. The complete guide to writing one that actually changes agent behavior, plus the prompt habits that make it pay off.
LLM evals usually mean checking text output. But when agents write frontend code, the failures are visual: broken layouts, dead interactions, accessibility regressions. Here's how to eval what users actually see.
AI-generated pull requests look confident and complete. That confidence is the hazard. A 10-minute review checklist that catches the specific ways agent-written code fails.
The demo was flawless. The launch tweet went viral. Then real users showed up. Five archetypes of vibe-coded launches, and the exact week-one failure each one earned.
A founder sent us the checkout flow their AI built. 200 lines, looked perfect in the demo. We found 14 bugs, including one that charged customers twice. Here's the full teardown.
AI can write the feature. It can't decide the feature is finished. Here's the unglamorous checklist (tests, review, edge cases, rollback plans) that separates 'demo' from 'done.'