The Mission Loop
The best way to teach is to build a game worth playing.
The mission loop is the repeating unit of a learning game. Run it once and you have a session. Run it a hundred times and you have a compounding skill. The loop forces teaching to become calibration — the learner makes a prediction, acts, compares reality to expectation, and compresses the delta into a protocol they can hand to the next player.
The Seven Steps
| Step | Name | What happens |
|---|---|---|
| 1 | Picture | Form a clear picture of current reality, desired reality, constraints, and stakes |
| 2 | Ask | Questions first — reveal goals, gaps, and priors before any answers are given |
| 3 | Mission | Bounded challenge, completable in 10–30 min, designed to cross one specific gap |
| 4 | Practice | Use AI as apprentice, mirror, simulator, or coach |
| 5 | Calibrate | Compare expected result vs actual; update prior; log the delta |
| 6 | Reduce | Compress the lesson into a mantra, checklist, protocol, or reusable prompt |
| 7 | Teach back | Explain to another — teaching proves understanding |
The loop maps to the meta-learning pipeline — but runs it end-to-end in a single session rather than across days or weeks.
| Pipeline step | Mission loop step |
|---|---|
| Awareness | Picture — see the gap |
| Capture | Ask — questions surface what you know and don't |
| Filter / Organise | Mission — bounded scope focuses attention |
| Systemize / Practice | Practice — AI as the arena |
| Reduce | Calibrate → Reduce |
| Teach | Teach back |
Four AI Roles
AI is not one thing inside a session. The role shifts with the learner's current state.
| Role | What AI does | When to activate |
|---|---|---|
| Apprentice | Assists the learner's work — drafts, researches, builds | Learner knows what to do, needs execution speed |
| Mirror | Reflects back what it observes — patterns, gaps, contradictions | Learner cannot see their own thinking clearly |
| Simulator | Creates safe environments to practice — role-play, scenario, adversarial | Learner needs reps before real stakes |
| Coach | Asks before telling — questions that reveal, not answers that inform | Learner has the map but is not moving |
The coach's job is value activation — shifting from "I should learn this" to "I need this now." The other three roles serve comprehension. Coach serves motivation. Both are required to cross a gap.
The Artifact Gate
Every mission session must produce one artifact before it closes:
- Checklist — steps compressed into a repeatable sequence
- Principle — the distilled insight in one sentence
- Protocol — the pattern encoded for reuse
- Prompt — the reusable AI instruction that reproduces the result
- Decision journal entry — prediction → action → result → updated prior
Without an artifact, the session was entertainment. The artifact is the proof that learning transferred — and the seed for the next learner's starting point. This is the Legacy Rule applied to a single session: finish and leave better maps.
The ZPD Connection
The mission works when it targets a gap in the learner's Zone of Proximal Development — the band between what they can do alone and what they can do with guidance. Too easy: no growth. Too far: no traction. The right mission is the one the learner cannot complete without AI — but can complete with it.
The calibrate step makes this explicit. When the expected result and actual result diverge, you have found the ZPD boundary precisely. That delta is the data the Teacher uses to simplify, and the Coach uses to activate.
The Stack
The mission loop and the reading-the-game cycle operate at different grains.
Mission Loop (session design)
└── Reading the Game (in-play execution)
└── VVFL (the compound loop that results)
Reading the game operates during the game — the in-play five-phase cycle of Design/Read/Act/Debrief/Compress. The mission loop operates at the session level — how you design the game that makes the in-play cycle worth running. They are not alternatives. They stack.
Context
- Navigation System — Value × Belief × Control — the three loops the mission loop runs inside
- Reading the Game — In-play execution cycle: design pictures, read at speed, act, debrief, compress
- Inner Game — Why the game is played and what makes it meaningful
- VVFL — The compound loop that results when missions align with values
- ZPD + MKO — The learning zone and who guides you through it
- Teacher Archetype — Simplifies the gap and makes it interesting (step 2 and 7)
- Coach Archetype — Activates the WHY and the WHEN (step 4, Coach role)
- Meta-Learning Pipeline — The longer arc this loop compresses into one session
Questions
What would it take to design a mission someone would volunteer for — not because they have to, but because the gap it targets is one they already feel?
- If the session produces no artifact, what does that say about the mission design rather than the learner?
- Which of the four AI roles do you reach for first — and which do you avoid? What does that reveal?
- Where in the mission loop does your own learning usually stop?