The knowledge base should help decide what to think about next while preserving what has already been learned.
An archive stores finished material. A thinking partner changes the next move. It helps surface patterns, expose gaps, test frames, generate questions, and turn vague interests into concrete work.
The practical test is simple:
Did this page make the next action clearer?
If not, the system is drifting toward storage.
Core Mechanism
The knowledge base becomes a thinking partner when every interaction creates at least one of five outputs:
- a sharper question,
- a better frame,
- a visible gap,
- a next action,
- a reusable synthesis.
The page is not the endpoint. The page is a thinking surface.
Use it to externalize the current mental model, then pressure-test that model against related pages.
interest
-> question
-> relevant pages
-> pattern or contradiction
-> synthesis
-> next action
-> updated pageThis is how the knowledge base compounds. It does not grow only by adding sources. It grows when old pages start changing how new problems are approached.
Archive Mode vs Thinking Partner Mode
| Mode | Question | Output |
|---|---|---|
| Archive | Where should this go? | Stored note |
| Thinking partner | What does this change? | Updated model |
| Archive | What did this source say? | Summary |
| Thinking partner | What can I do differently now? | Action or experiment |
| Archive | What pages exist? | Index |
| Thinking partner | What pages should talk to each other? | Link, synthesis, or split |
| Archive | What do I know? | Catalog |
| Thinking partner | What am I missing? | Gap list |
Archive mode is still useful. It keeps the system legible.
But thinking partner mode is where the system becomes alive. It asks whether the current structure is producing better thinking, not just more organized material.
The Five Moves
1. Ask From A Live Problem
Start with an active pressure, not a random topic.
Good prompts:
- What am I trying to decide?
- Where am I stuck?
- What keeps recurring?
- What feels important but under-formed?
- What would make this easier to act on?
This keeps the knowledge base tied to reality. The system should serve current judgment, not only future retrieval.
2. Pull The Nearby Pages
Do not start from scratch.
Read the pages that are already adjacent to the problem. Use the graph, index, backlinks, and obvious links.
For learning questions, start with Prestudy, BHS, and SIR: Turning Information into Usable Structure, First Principles of ICS, Are You Learning, or Just Using Techniques, and Knowledge Mastery: From Recognition to Usable Knowledge.
For execution questions, start with Attention Management: Preserving Flow, Focus Management: How to Enter & Recover Inside a Work Block, Procrastination: a System Problem, and Decisional Delays.
For judgment questions, start with Applied Critical Thinking: Testing Frames, Decision Making, How to Communicate Truth Into Someone Else’s Frame, and Suicidal Empathy.
For building questions, start with Agentic Engineering, Understanding Bottleneck, and Interleaving for Complex Problem Solving.
3. Find The Tension
The useful part is usually not the page itself. It is the tension between pages.
Look for:
- a concept that appears in several places but has no clear operating model,
- two pages that imply different actions,
- a repeated failure mode,
- a missing bridge between theory and practice,
- a technique that looks correct but does not change behavior,
- an interest that keeps returning without a next action.
This is where the knowledge base stops being a library and starts acting like a mirror.
The best pages often come from tension:
- Balancing Multiple Interests: Breadth v Focus
- The Shortcut Problem
- The Technique Is Only as Good as the Thinking It Produces
- How to Unlearn Old or Bad Habits Efficiently
4. Convert Insight Into A Handle
Do not leave the output as an abstract insight.
Turn it into a handle:
- a title,
- a question,
- a checklist,
- a contrast,
- a next exercise,
- a decision rule,
- a page split,
- a journal prompt.
The handle is what makes the idea reusable.
Examples:
- “I need better study habits” becomes Prestudy, Bear Hunter System, and Spaced Interleaved Retrieval.
- “I have too many interests” becomes Balancing Multiple Interests: Breadth v Focus and Priority 0.
- “I keep consuming without changing” becomes Are You Thinking, or Just Consuming?.
- “The method looks right but nothing improves” becomes The Technique Is Only as Good as the Thinking It Produces.
A good handle lets the system call the idea back at the moment it matters.
5. File The Result Back Into The System
The output should land somewhere durable.
Use the smallest useful destination:
| Output | Destination |
|---|---|
| Passing thought | Journal |
| Active question | Journal What's On My Mind |
| Current working area | Journal Current Focus |
| Durable operating model | Wiki page |
| Personal plan | private Goals or Skills |
| Public argument | Blog draft |
| System change | README, AGENTS, or workflow page |
Do not create a page just because an idea exists.
Create a page when the idea will be reused, linked, practiced, or argued with.
What A Session Looks Like
Use this when the system feels passive.
- Pick one live question.
- Read three to six related pages.
- Write the current answer in plain language.
- Name the tension or missing piece.
- Decide whether the result belongs in Journal, a wiki page, a private template, or a blog draft.
- Add one link, one next action, or one open question.
That is enough.
The goal is not to process the whole vault. The goal is to make one part of the system more usable.
Failure Modes
| Failure | What It Looks Like | Repair |
|---|---|---|
| Knowledge hoarding | More pages, no changed behavior. | Ask what the page changes. |
| Index addiction | Reorganizing instead of thinking. | Start from a live problem. |
| Link decoration | Links exist but do not clarify the idea. | Link only when the relationship matters. |
| Permanent synthesis | Every answer becomes a polished page. | Use Journal for temporary thinking. |
| Private/public confusion | Personal planning leaks into public pages. | Keep private Goals, Skills, and Kolbs local unless intentionally publishable. |
| Agent overproduction | The LLM creates too much text. | Ask for operating model, causal chain, failure modes, practical feel, and implications. |
| Archive drift | The system stores what happened but does not guide what happens next. | End each session with a question or action. |
Practical Feel
Used well, the knowledge base should feel like returning to a room where the important threads are already laid out.
Good signs:
- the next question becomes obvious,
- old pages start explaining new problems,
- titles become handles for action,
- active interests feel alive but bounded,
- the Journal reflects current attention,
- the Index remains stable enough to navigate,
- outputs become shorter because the system already holds the context.
Warning signs:
- every session starts from zero,
- pages are polished but unused,
- the graph is interesting but not actionable,
- new sources do not update old pages,
- Journal questions stay unchanged for weeks,
- the system feels impressive but inert.
Weekly Use
Once a week, ask:
- What page changed how I acted?
- What question kept returning?
- What active thread is getting stale?
- What source, page, or experience should be synthesized next?
- What idea belongs in a blog post instead of staying internal?
The weekly review should not become a productivity ritual for its own sake.
Its job is to keep the knowledge base pointed at live thinking.