Learning Question
How can one explanation be turned into a judgment criterion for future cases?
Explanations Should Leave Something Durable
An explanation answers the current question. A judgment criterion helps answer future questions.
The difference is important for a knowledge vault. A vault should not preserve every explanation just because it was useful once. It should preserve the part that changes how the user will reason later.
When AI explains something well, the next question should often be:
What judgment criterion follows from this?This moves the conversation from understanding one case to recognizing a pattern.
What a Judgment Criterion Is
A judgment criterion is a rule, distinction, or decision test that helps evaluate future cases.
It is not always absolute. Often it is conditional:
If the question is about protocol meaning, do not reason only from the underlying transport capability.That criterion can help compare HTTP keep-alive, WebSocket, streaming, and custom protocols.
It is more useful than a bare fact because it tells the user how to think when a similar confusion appears.
The Conversion Process
A useful conversion process looks like this:
Specific case
-> explanation
-> key distinction
-> boundary
-> judgment criterion
-> future applicationFor example:
Specific case: WebSocket and HTTP both use TCP, so why are they different?
Explanation: TCP provides byte transport, while HTTP and WebSocket define different communication semantics above that transport.
Key distinction: lower-layer capability is not the same as higher-layer protocol meaning.
Boundary: this matters when comparing systems that share infrastructure but define different behavior.
Judgment criterion: when two systems use the same lower layer, identify the higher-layer rules before concluding they behave the same.
Future application: compare file upload protocols, streaming APIs, message queues, and RPC systems more carefully.
Ask AI to Extract the Criterion
AI often stops at explanation unless asked to extract the criterion.
A useful prompt is:
From this explanation, extract:
1. the specific answer
2. the key distinction
3. the boundary of the explanation
4. the judgment criterion
5. one future case where the criterion appliesThis format prevents the result from becoming either too broad or too tied to one example.
Criteria Need Boundaries
A criterion without boundaries becomes a slogan.
For example, “always separate layers” is useful but too broad. A stronger version is:
When two concepts appear similar because they share an underlying mechanism, separate the lower-level mechanism from the higher-level rules before comparing them.This criterion has a trigger condition and an action.
Good criteria often have this shape:
When [situation], check [distinction] before deciding [conclusion].or:
Do not treat [A] as [B] merely because they share [C].Criteria Should Change Behavior
A good judgment criterion should influence future action.
It may change what the user asks, where they look for evidence, how they compare options, or what they preserve in a document.
If a criterion does not change future behavior, it may be only a summary.
Ask:
- What would I do differently because I know this?
- What mistake would this prevent?
- What future confusion would this help resolve?
- What question would this make me ask earlier?
What This Chapter Does Not Claim
Not every explanation needs to become a criterion.
Some answers are temporary, local, or too small to preserve. The criterion step is useful when the explanation reveals a durable distinction or a repeated thinking pattern.
The goal is not to inflate every conversation into a principle. The goal is to recognize when a principle has appeared.
Core Mental Model
The highest-value output of an AI conversation is often not the answer. It is the judgment criterion that the answer reveals.
Specific cases are raw material. Judgment criteria are portable understanding.
Final Summary
After a good explanation, ask what future decision it should improve. That is how AI conversations become durable insight.