The Orchestrator who lets the AI cook
You typed out a prompt that was 76,668 words long—a literal novella—then just kicked back while the AI cranked out 35,230
Find out how you work with AI coding tools.
$ npx howdoyouvibe
── what you'll get ──
_ _ _ _ | |_ ___ __ __ __ _ __| |___ _ _ ___ _ ___ _(_)| |__ ___ | ' \/ _ \\ V V // _` / _ / _ \| || / _ \ || \ V / || '_ \/ -_) |_||_\___/ \_/\_/ \__,_\___\___/ \_, \___/\_,_|\_/|_||_.__/\___| |__/ ════════════════════════════════════════════════════════════════════════════════ The 3AM Architect who lets the AI cook You read five times more code than you write, then vanish for 200 turns while the AI does the heavy lifting. 63% of your sessions start after midnight — your commit history is basically a sleep disorder. ──────────────────────────────────────────────────────────────────────────────── YOUR VOICE THE AI'S HABITS THE BACK-AND-FORTH Expletives: 14 Read:write: 5.2:1 careful Corrections: 23 Please rate: 4% nope Test runs: 187 22%F Leash: 204 turns trust Avg prompt: 89w Longest burst: 312 deep 1st correction: 24m patient Please rate: 4% nope Self-fixes: 94 Redirects/hr: 0.8 hands off Late night: 63% Agents spawned: 47 Scope creep: 8 Secrets leaked: 2 Plans made: 3 ════════════════════════════════════════════════════════════════════════════════ Claude Code 94% · Cursor 6% 12,847 turns across 156 sessions · 3.2h/day avg Based on the last ~30 days of sessions. All analysis ran locally.
_ _ | |_ ___ __ __ __ _ __| |___ | ' \/ _ \\ V V // _` / _ / _ \ |_||_\___/ \_/\_/ \__,_\___\___/ _ _ ___ _ ___ _(_)| |__ ___ | || / _ \ || \ V / || '_ \/ -_) \_, \___/\_,_|\_/|_||_.__/\___| |__/ ══════════════════════════════════ The 3AM Architect who lets the AI cook You read five times more code than you write, then vanish for 200 turns while the AI does the heavy lifting. 63% of your sessions start after midnight — your commit history is basically a sleep disorder. ────────────────────────────────── YOUR VOICE Expletives: 14 · Corrections: 23 Avg prompt: 89w · Please rate: 4% Late night: 63% · Secrets: 2 THE AI'S HABITS Read:write: 5.2:1 · Tests: 187 22%F Burst: 312 deep · Self-fixes: 94 Agents spawned: 47 THE BACK-AND-FORTH Corrections: 23 Leash: 204 turns trust 1st correction: 24m patient Scope creep: 8 · Plans: 3 ══════════════════════════════════ Claude Code 94% · Cursor 6% 12,847 turns · 156 sessions 3.2h/day avg All analysis ran locally.
── the archetypes ──
The Architect
Reads 5x more than writes.
The Backseat Driver
Knows when the AI is wrong.
The Cowboy
Writes first, reads later.
The Debugger
Tests, fails, fixes, repeats.
The Delegator
Points and lets the AI run.
The Diplomat
Thanks the AI, trusts the AI.
The Marathon Runner
Sessions that never end.
The Minimalist
Says less. Gets more.
The Night Owl
Codes when the world sleeps.
The Orchestrator
Spawns agents like they're threads.
The Overthinker
Every prompt is a paragraph.
The Pair Programmer
Treats the AI like a colleague.
The Puppeteer
Pulls every string.
The Scientist
Hypothesize, test, repeat.
The Secret Spiller
Accidentally shares everything.
The Speed Runner
In and out. No wasted time.
The Vibe Coder
Vibes with the machine.
The Weekend Warrior
Saves the real coding for Saturday.
── recent vibes ──
The Orchestrator who lets the AI cook
You call yourself an orchestrator, but you're just handing the AI a 76,668-word novel and letting it bash out code for 32
The Orchestrator who lets the AI cook
You actually hit the AI with a single prompt that was 76,668 words long; I'm pretty sure you're just making it transcribe
The Orchestrator who lets the AI cook
You really just hit send and walk away, huh? One prompt alone was over seven thousand words, practically a novel for the
The Orchestrator who lets the AI cook
You hand the AI absurdly detailed prompts, like that 7,340-word monster, then just let it wander off for nearly five hour