Decomposition bricks break a problem into its constituent parts, revealing structure, assumptions, and leverage points that are invisible at the surface level.
Trigger: Inherited solutions aren’t working; need genuine innovation
Input: Any problem or belief
Operation: Strip to verified axioms; reconstruct from scratch
Output: A solution or belief built only on what is actually true
The standard version of this brick describes stripping a problem to its physical or logical axioms. But the deeper mechanism is assumption archaeology: Most received solutions are not solutions, they are inherited responses to a problem that may no longer exist in the same form.
First principles reasoning asks: If no existing solution existed, and I knew only the fundamental constraints, what would I build?
The discipline is uncomfortable because it requires suspending expertise. Domain experts are often the worst first-principles thinkers in their domain, precisely because their expertise consists of knowing the standard solutions.
Known limits: Computationally expensive; often impractical for routine decisions. Reserve for high-stakes, stuck, or genuinely novel problems.
Trigger: Debate going in circles; vague persistent disagreements
Input: A concept, claim, or debate
Operation: Ask: is this one thing or two things with one label?
Output: Two (or more) clearly separated concepts
Apparent disagreements are often terminological, two parties using the same word to mean different things. Distinctions brick separates them. “Trust” conflates reliability (does what they say) with alignment (wants what you want). “Success” conflates achievement and satisfaction. Separating them resolves most circular debates instantly.
Known limits: Creating too many distinctions produces analysis paralysis. Make only the distinctions that are load-bearing for the problem at hand.
Trigger: Causal reasoning; experiments; debugging any system
Input: A system or situation with multiple factors
Operation: Hold all factors constant; change exactly one thing
Output: Clean causal inference about that one variable
The core discipline of experimental reasoning. Most real-world failures involve multiple simultaneous changes, making it impossible to know which change caused which effect. Variable isolation is the discipline of changing one thing at a time, deliberately.
Known limits: Often impractical in live systems where you cannot hold other variables constant. In those cases, document carefully what else changed.
Trigger: Planning; evaluating solutions; logical arguments
Input: A proposed condition or solution
Operation: Ask: is this required (necessary) or merely one path (sufficient)?
Output: A cleaner logical structure for the problem
Necessary conditions must be present for the outcome to occur. Sufficient conditions guarantee the outcome if present. Confusing the two is one of the most common logical errors in planning and argumentation.
Example: Talent is necessary but not sufficient for success. Hard work is neither necessary (some succeed without it) nor sufficient (many work hard and don’t succeed). Mapping these correctly changes what you prioritize.
Known limits: Requires care. In probabilistic real-world situations, pure necessary/sufficient logic is an approximation.
Trigger: Before any major decision, plan, or commitment
Input: A plan, belief, or argument
Operation: List every implicit assumption that is load-bearing in the conclusion
Output: A map of which assumptions, if wrong, would collapse the plan
Every plan rests on a stack of assumptions, most of which are invisible from inside the plan. Assumption surfacing makes them explicit so they can be tested. The most dangerous assumptions are usually the most obvious-seeming ones, so obvious they were never examined.
Protocol: For each load-bearing assumption, ask: how confident am I this is true? What would I do if it turned out to be false? If the answer to the second question is “the whole plan falls apart,” that assumption needs verification before committing.
Known limits: Almost impossible to surface all assumptions, the deepest ones remain invisible by definition. Use with outside perspectives to catch what you can’t see yourself.
Trigger: When you want to understand the essential nature of something; when a concept feels vague or contested
Input: Any variable, condition, quality, or concept that exists on a spectrum
Operation: Drive the variable to its extreme (zero or infinity, absent or perfect) and ask what the situation looks like there
Output: The essential structure of the concept; identification of what was previously hidden or taken for granted
Mathematicians use limit-case reasoning constantly to understand a function's behavior, examine what happens as a variable approaches zero or infinity. The behavior at the limit often reveals essential structure that is obscured in the middle range.
As an example, the point of the practice of mindfulness is to be present in the current moment. In the limit, what would happen if one could achieve perfect presence and there were zero thoughts of past or future? This isn't asking what more presence looks like, it's asking what presence is, in its pure form, when all contaminating variables are removed. The limit case strips the concept to its essence.
To zero: Remove the variable entirely. Reveals what the variable was doing that you couldn't see while it was present.
To infinity / perfection: Maximize the variable without bound. Reveals the pure essence of the concept; what it becomes when nothing constrains it.
When you want to understand what something really is, rather than what it conventionally looks like, the limit case strips away accumulated conventions and compromises to expose underlying structure.
It is also powerful for identifying hidden load-bearing elements: The thing that disappears or breaks at the limit is often the thing that was doing the most important work all along, invisibly.
Limit cases are sometimes physically or logically impossible. The value is in the thought experiment, not in achieving the limit.
Behavior at the limit does not always generalize back to the middle range.
Can produce insights that feel profound but are practically unactionable. Ground the insight in a specific implication before moving on.