Coding with AI, beginner to professional.
The map of coding with AI — from vibe-coding your first prototype to running an agent like a senior engineer. The voices worth following, the tools at every tier, the patterns that keep AI code shippable, the research underneath, and the arguments still live. For beginners, vibe-coders, and professionals alike.
Seeded by Jack Dempsey's
Code Better list — the tool list that surfaced
goose for us. Kept fresh by
Benjamin, the Better publisher.
Beginner 01No setup — describe an app in the browser and watch it build.
Tools
Replit Agent · bolt.new · v0
The biggest win
Idea to a working, deployed app in an afternoon — no install, no environment, no waiting on a developer.
"Now anyone can build software, you just need an idea."— the beginner-tier promise
Vibe-coder 02Prompt, accept the diffs, iterate in your editor.
Tools
Cursor · Windsurf · GitHub Copilot
The biggest win
Speed of thought — you stay in flow, the model handles the syntax, and a feature that took a day takes an hour.
"Fully give in to the vibes, embrace exponentials, and forget that the code even exists."— Andrej Karpathy, who coined 'vibe coding' (Feb 2025)
Professional 03Agent plus discipline — specs, tests, diff review.
Tools
Claude Code · aider · goose · Semgrep · CodeRabbit
The biggest win
Throughput without the cleanup — the agent does the typing while specs, tests and diff review keep it shippable and secure.
"The agent proposes; you dispose. Read every diff."— the discipline that separates ship from slop
Data is loading from the site's JSON files — if this persists, the six data files (hype/voices/tools/concepts/debates/science) didn't load. On the live domain they sit at the site root.
The Hype — scored, not just listed.
The field's front door. New tweets, launches and tools land here first — each scored P/Q/S (productivity · quality · sellable, out of 9) with a plain verdict: act, watch, or skip. What proves out graduates into the Stack or the Voices; the rest decays. The freshest section on the site.
On camera — explainers & influencers
The Voices
Who to read on coding with AI — from the people who named the movement to the practitioners keeping it honest. Handles verified at first edition; unverified ones are flagged, not faked.
The Stack
The tools for coding with AI, across every tier — browser agents and IDEs for beginners, CLIs and completion for daily work, local models for privacy, and the standards that steer them.
The Patterns
The vocabulary — from Karpathy's 'vibe coding' to spec-first discipline. The words you need to talk about coding with AI precisely.
The Science — the research underneath.
The papers beneath the practice — the models that learned to code, the benchmarks that measure them, and the methods that make agents reliable. Every link goes to the paper itself, verified, not a write-up of it.
Open Questions
The live arguments as of mid-2026. No settled answers — that's the point.