Reality
Question whether the proxy still points toward the truth — or has it drifted?
Reality is the instrument panel. Strategy sets the intention — where to position, what to bet on. Reality reads whether you got there. KPIs are not goals. They are instruments: bearing and position against strategic intentions.
Performance is second in the Tight Five — after Principles, before Platform. The sequence matters. Without principles you don't know what to measure. Without a picture of good, every feedback loop is aimless.
Dig Deeper
📄️ Commissioning
Does the output match the spec — or does it just look like it does?
📄️ Content
Is the content pipeline producing work that lands?
📄️ Crypto
Understanding value creation is crucial for assessing the potential and sustainability of projects.
📄️ Economics
KPIs for Economics and Business Cycle
🗃️ Financials
4 items
🗃️ Human Resources
1 item
📄️ Product Design
How do you know your design is getting better?
📄️ Retail Industry
Measure performance of retail stores.
🗃️ SaaS Valuation
1 item
🗃️ Software
3 items
Instruments
Two loops, one system. The inner loop forms intention — what you believe good looks like. The outer loop tests it against reality. The instruments tell you where you are, not where to point.
| Loop | Question | Without Instruments |
|---|---|---|
| Inner (intention) | What should this produce? | Building blind |
| Outer (validation) | Did it match the spec? | Feedback without signal |
| Closed (compounding) | What's better this time? | Activity without progress |
| Bearing (direction) | Where am I pointed? | Position without intention is drift |
Obstacles shift bearing temporarily — the navigator corrects by checking the northstar, not by changing it. Instruments tell you where you are. First principles tell you where to point. A metric is always a proxy: question regularly whether it still indicates the truth it was chosen to reflect.
The VVFL only compounds when the gauge reads true. Validated — against reality, not assumption. Virtuous — serving beyond self. Feedback — output changes input. Loop — each cycle starts higher.
Commissioning is the discipline: the builder never validates their own work. Syntactically correct is not functionally correct.
The Gauge
The scoreboard is an output. You can't move it directly. You win micro-moments — collisions where standards get followed, thresholds get checked, decisions get made right.
| Level | What It Is | Glory Example | Collision Example |
|---|---|---|---|
| Micro-moment | The collision you can win | — | Developer checks contrast before committing |
| Leading indicator | Signal it's compounding | — | Token resolution rate trending to 100% |
| Scoreboard | The number that moves | Lighthouse score hits 90+ | — |
| Perspective | What you chose to track | Revenue | Breakdown speed |
The Chiefs won Super Rugby by measuring work rate and breakdown intensity — metrics nobody else tracked. The tries followed. Find the unseen and underappreciated collisions that impact value creation. Pay close attention to them. The glory metrics are partly luck. The collision metrics are where you make your impact.
What you track reveals your perspective. The ability to perceive which micro-moments matter — that's the competitive advantage.
Mirrored reciprocation explains why collisions compound: go positive, go first, be constant. Every collision where you follow the standard and treat the counterparty well returns amplified. Loss aversion blocks most people from going first — but the expected value is hugely positive. The VVFL only stays virtuous when collisions are won with goodwill, not force.
Production Instruments
What reads reality in the live system. Each instrument feeds a specific VVFL station and serves a settlement bearing.
| Instrument | What It Reads | Settlement Bearing | Game Loop | Threshold |
|---|---|---|---|---|
| Commissioning API | Capability L-levels (L0-L4) | Internal | Gameplay | 188 capabilities tracked |
| Playwright + intent specs | Feature verification | Internal | Rendering | All critical paths green |
| Spec vs reality delta | Feature matrix drift from PRD | Hemisphere | Core | 0 stale readings |
| Agent receipts | Skill invocations, gate pass rate | Hemisphere | Core | Receipt count per skill |
| Sentry (planned) | Runtime errors, stack traces | External | Meta | Error rate under 5% |
| PostHog (planned) | Feature usage, activation | External | Meta | Active users per feature |
Settlement Bearings
Three boundaries where trust can break. Each bearing asks: did what arrived match what was promised?
| Bearing | Boundary | Current State | Target |
|---|---|---|---|
| Internal | Between agents within engineering | Hooks enforce (14 active). Mostly virtuous. | Composition ratio tracked per commit |
| Hemisphere | Between dream repo and eng repo | Manual handoffs. Vicious. | Automated spec-vs-reality delta |
| External | Between platform and users | No production users yet. Unknown. | First external settlement verified |
The Agent Platform (Phase 4: Settlement Bridge) wires all three. Internal is strongest. Hemisphere is the bottleneck. External is future work.
Current Reading
Snapshot from the Commissioning Dashboard. Updated 2026-03-11 — the first capability to reach L4 earns automated sync.
| Metric | Reading | Target |
|---|---|---|
| Total capabilities | 188 | — |
| Platform coverage (L1+) | 16% | 50% |
| At L4 (commissioned) | 0 | First by end of Phase 2 |
| At L3 (tested) | 10 (5%) | 15+ |
| At L2 (UI connected) | 15 (8%) | — |
| Bottleneck | L3→L4 (requires independent commissioner) | — |
Benchmark Standards
If you don't know widely accepted thresholds for "good," you have no valid context to judge reality. Show me the incentive and I'll show you the outcome.
| Minimum | Why |
|---|---|
| Industry baseline | Where does everyone start? |
| Internal baseline | Where do you start? |
| Target threshold | What does good look like? |
| Measurement method | How do you read the gauge? |
| Review cadence | How fast does the loop close? |
Questions
What is the difference between a metric that tells you where you are and a principle that tells you where to point?
- Which of your current KPIs has drifted furthest from the truth it was chosen to indicate?
- When an obstacle forces a bearing change, what is the mechanism that returns you to the northstar — and does it exist?
- If your instruments showed green while your strategy said something was wrong, which do you trust?
Context
- The North Star — The fixed reference point instruments measure against
- Commissioning — Verify the output matches the spec, not just the code
- Scoreboard — Settlement integrity: three bearings where trust can break
- Strategy — What the instruments are measuring against
- Belief System — The hull that holds when instruments deliver bad news
- Pricing — The market's instrument: price signals whether reality matches strategy
- VVFL — The feedback loop that compounds when the gauge reads true
- Control System — PID mechanics: how controllers decide how hard to correct
- Consciousness — Inner space is where better intentions form before outer expression
- Incentive Engineering — Show me the incentive, I'll show you the outcome
- Standards — Thresholds that define what "good" looks like
- Predictions — Betting on which standards win
- Peter Kaufman — Mirrored reciprocation: the mechanism behind collision compounding
- Prompt Deck PRD — Where minds collide: the instrument for winning pitch collisions
- Time + Mind PRD — Time allocation IS collision selection
- Deterministic vs Probabilistic — Instruments are deterministic gauges reading probabilistic reality
- Matrix Thinking — The grid that makes gaps visible before instruments can read them