AI Coding Assistant vs. Agentic Coding vs. Vibe Coding
New'Uses AI to code' now appears on most developer resumes and covers three very different habits. A coding assistant suggests lines as the developer types, and the developer stays in charge. Agentic coding hands a whole task to an AI agent (Claude Code, Codex, Cursor's agent mode) that edits many files and runs tests, with the developer reviewing the result. Vibe coding means accepting AI-generated code without really reading it — fine for prototypes, risky in production.
How to tell them apart on a resume
AI coding assistant
GitHub Copilot, autocomplete, inline suggestions, 'X% faster', AI chat in the IDE — help with individual lines and functions.
Agentic coding
Claude Code, Codex, Cursor agent, Devin, specs and instructions files, reviewing AI-written pull requests, tests as guardrails — delegating whole tasks.
Vibe coding
Built apps with Lovable, Bolt, Replit or v0, prototypes, 'no coding background', hackathon demos — speed over code quality.
The question that settles it
“When AI wrote code for you, how did you check it was correct before it shipped — and what is one thing you no longer let it do?”
Read the full definitions
Open the full tool for the other look-alike pairs, role profiles, and the JD decoder.