By questions
Validated
Named evidence, named person, named struggle. Is this a real problem — or a projected one?
Pressure question: What would I accept as proof that this problem is real?
Setpoint
Not every positive loop is virtuous. The difference is the setpoint — a corrective loop holds a fixed target; a virtuous loop raises its own target each cycle.
06 Loops
Virtuous feedback loops.
Most loops amplify. The question is whether they amplify toward a fixed ceiling or toward a rising standard. Only the virtuous loop improves the thing it measures.
Live control loop
Compounds values-aligned gains
- Driver
- Validated positive feedback
- Setpoint
- Rising
- Outcome
- Agency — the system improves its own setpoint
- Who holds the target
- The system itself, each cycle
The Trap
Most flywheels are corrective, not virtuous.
A flywheel that grows monthly active users, refines its recommendation engine, and drives more engagement is corrective. It is optimising toward one fixed metric — engagement — and the metric itself is never questioned.
A corrective loop gives you stability. The thermostat holds the room at 21°C. The algorithm holds the feed at maximum time-on-site. The revenue target holds the team at this quarter's number. None of them ask whether the target is right. That is the definition of a corrective loop: it holds the setpoint, it does not raise it.
The virtuous loop does the thing corrective loops cannot: it looks back at the setpoint itself, asks whether that target was worth holding, and then raises it. Each cycle the system becomes the kind of system that can ask a better question. That is the only compounding that matters.
VVFL Test
Four instruments for one honest question.
The VVFL test runs in sequence. Each instrument is a gate. Miss any one and the loop can still run — but it is not virtuous.
By principles
Virtuous
The gain must align with values, not just the metric. Does this serve the system, or deplete it?
Pressure question: If this loop runs for ten years, who benefits and who pays?
By outputs
Feedback
Proof reaches back to the setpoint and raises or kills it. Did the cycle improve the starting condition?
Pressure question: What signal tells me this cycle raised the floor for the next?
By platform
Loop
The platform absorbs the lesson and lowers the next cycle's cost. Can we do it again, better?
Pressure question: What must change in the template so the next run starts higher?
Tight Five
Standards are the compounding point.
The five rungs of the loop are not equally weighted. Standards carry disproportionate leverage: one improvement to a standard propagates across every implementation that inherits it. This is the legacy rule — every cycle must leave the template stronger than it was found.
Questions
Name the real problem. Each cycle, better questions open more opportunity and reveal what the previous cycle missed.
Principles
Align every gain with values. The controller rejects gains that serve the metric while violating the system's purpose.
Protocols
Repeat the tested process. Protocols turn one successful cycle into reproducible output — captured reality you can inspect.
Standards
The compounding point. One improvement to a standard propagates across every implementation that inherits it. This is where the loop earns leverage.
Platform
Infrastructure that routes attention and lowers the cost of the next cycle. The platform is the legacy rule made durable.
The arrow points back to Questions — not to the same questions, but to better ones. Each cycle, the platform lowers the cost of finding them. Improve protocols and the platform lowers cost. Strengthen the platform and more opportunity opens. Raise resilience and the next cycle starts higher. This is not metaphor — it is mechanics.
Meaningful Endeavour
The setpoint test is ethical, not just technical.
A loop that compounds extraction is not virtuous because it grows. The distinction is what the setpoint measures, and whether raising it serves the system or depletes it. Growth is not proof of virtue. Growth in what direction is the question.
Extractive loop
- Externalises costs, internalises gains
- Depletes the substrate it depends on
- Concentrates power in the loop owner
- Optimises for extraction over flourishing
- Short-term: ends when the substrate runs out
Virtuous loop
- Internalises accountability, shares gains
- Replenishes or improves what it depends on
- Distributes agency across participants
- Optimises for flourishing over extraction
- Intergenerational: each cycle leaves a better template
The virtuous loop raises the target for the people in it, for the community it serves, and for the biosphere it depends on. The kill signal is not failure — it is when the setpoint stops rising. A loop that holds its own target has become corrective. A loop whose setpoint declines has become extractive. Both are recoverable if caught early.
Outward Gauge
Name the loop before you run it.
Classify one real loop in your own work. Name its current type, its setpoint, and the one move that would raise the setpoint. If you cannot name a move that raises the setpoint, the loop is corrective at best — and that is fine, as long as you know it.
Loop you run today: Current type (runaway / corrective / virtuous): Current setpoint: The one move that raises the setpoint: Kill signal (if the setpoint stops rising):