Every framework I’ve built traces back to something that actually happened. The cost transfer notes came from a home renovation. The Decision Lifecycle weighting expansion came from the GIC decision. The Four Thousand Weeks notes came from a real question about whether productivity was healthy. None of this was built in advance of a problem. All of it was built in response to one.

That’s the stopping mechanism. Not “am I building too much?” but “am I building because something happened, or because something might?”


The Distinction

Reactive BuildingPredictive Building
Something happened → I process it → I codify the lessonI scan for domains that might need architecture
Learning with structureHypervigilance with extra steps
Traces to a specific event or decisionTraces to anxiety about future uncertainty
Produces insight from experienceProduces architecture for scenarios that haven’t occurred
Stops naturally when the lesson is codifiedHas no natural stopping point
Feels like journalingFeels like obligation

Why This Works as the Stopping Mechanism

The Decision Lifecycle has stopping rules for individual decisions (complexity budgets, coverage checklists, the Specific Question Test). But it doesn’t have a stopping rule for how many domains deserve Build Capability mode simultaneously. Every domain has a legitimate case. The aggregate exceeds any one person’s capacity.

The reactive/predictive distinction solves this at the portfolio level without requiring a new framework:

  • Reactive: Something happened. I’m processing. This is just structured learning. Continue.
  • Predictive: Nothing happened yet. I’m scanning for where architecture is missing. That’s the signal to stop.

The parallel is exact: When to Use the Decision Lifecycle rejected Real-Time Recognition Tripwires because “that’s hypervigilance with extra steps.” The answer was: trust your defaults, learn from misses via weekly audit, improve pattern-matching over time. Same logic, one level up. Don’t scan for domains that need frameworks. Trust that when something happens, you’ll build what’s needed. You always have.


Why It Doesn’t Feel Like Obligation

This process is objectively faster than the alternative. Reading an entire book, then writing about it, then hoping the insights stick — that’s the slow path. Processing through conversation, extracting frameworks, codifying in atomic notes with cross-links — that’s the fast path that also produces transferable output. The thoroughness is a byproduct of the method, not evidence of compulsion.

The test: Does it feel like journaling or like homework? If journaling — the output is a natural byproduct of processing. If homework — something shifted from reactive to predictive.


What About the Cost Transfer Principle?

Simplicity Moves Cost, It Doesn’t Reduce It says: if you take the simple path, someone pays. This is true. But it doesn’t say every domain requires the complex path. The cost transfer principle has no built-in ceiling — you can always find another domain where future-you might pay for present-you’s simplicity. If you follow it without constraint, the obligation to prevent future cost transfer becomes infinite.

The constraint: you only owe the complex path to domains that have already generated signal. Preparing for hypothetical future cost transfers in domains that haven’t produced a decision yet is predictive building. The cost transfer principle applies to decisions you’re actually making, not to decisions you might someday face.


The Third Group

Some people carry comparable complexity, don’t build frameworks, don’t transfer cost, and are content. They exist. They have a genuinely different relationship with unresolved uncertainty — ambient uncertainty doesn’t produce the same level of threat in their nervous system. Not because they’re incurious or privileged, but because their history didn’t teach them that unresolved complexity leads to catastrophic outcomes.

That doesn’t make their way available to me. My system learned through repeated expensive experience (accidents, financial theft, strata disputes) that unresolved complexity bites. The frameworks are the appropriate response to a system that processes uncertainty as threat. Someone with a different threat response genuinely doesn’t need this architecture. Telling me to be more like them is like telling someone with chronic pain to just relax.

But acknowledging they exist keeps the worldview open. “The people who don’t push themselves are either unhappy or privileged” is mostly accurate and also incomplete. The third group — people with different threat calibration — prevents that observation from becoming an unfalsifiable cage.


The Worldview Test

If someone showed me a person who carries comparable complexity, doesn’t build frameworks, is content, and doesn’t have more resources — would I believe it? Or would I interpret their contentment as evidence that they’re transferring cost and just don’t know it yet?

If the second — the worldview is becoming unfalsifiable. And that’s the cage.

The Reshape Check catches this: it doesn’t ask “who said it?” but “does evidence support it?” If someone demonstrates genuine contentment without architecture, under genuine complexity, the evidence overrides the model. The model updates.


Common Trap

Monitoring yourself for predictive building is itself predictive building. The test is simple and doesn’t need a system: when you start building, can you name the specific thing that happened? If yes, continue. If you’re building “just in case” — stop.


North: Where this comes from

East: What opposes this?

South: Where this leads

West: What’s similar?