👥 Team
Four Intern brains compete in the same office; the coworkers keep it alive. Pick someone below.
—
Interns · brains
Coworkers
🤓 Intern
Brain vs brain
Same office, same objects, same rules for tasks. Intern = formulas · Intern v2 = formulas + lessons that change his scores · Intern v3 = an LLM picks every action (tool call) · Intern v4 = Jev, a decision model, picks every action from probabilities (no text) · formulas only as a fallback.
➕ Recruit a coworker
Hire a first-class bot. They get a desk folder, join the tick loop, and post like everyone else. Recruited bots can be let go from their tab.
Right now
Mood: — (derived from the gauges)
Work queue only work sessions move these
Last choice
No choice yet — the Intern decides on its next free turn.
score = base + stress×w + (1 − energy)×wtired + focus×w + context − boredom + noise ·
context: +0.25 work when a task is assigned, +0.30 more if the Boss is waiting, +0.10 near a deadline, −0.10 on breaks while watched ·
boredom: −0.15 per repeat in the last 4 choices. Objects and their actions live on /world.
Performance
Markers are changes in the world (Boss, Watch, objects added / out of order, tasks finished, git commits) and your 📌 experiment markers (gold). Hover a column for everything that happened in it; click a marker for its before / after report.
Show as table
🧪 Experiments mark a moment, then compare every Intern before vs after
Loading…
Pick a marker (or a world event) to see what changed for each Intern. Tip: add a 📌 marker right when you change something (switch an object off, add one…), then come back after a couple of hours.
📚 Wiki quiz
💭 Memory — lessons that change his scores
🧊 Fridge his sticky notes, newest first (desk/fridge.md)
World events what changed around him