Minimally Viable Learning System

The smallest useful version of the learning system is Bear Hunter System plus Spaced Interleaved Retrieval.

BHS -> encode usable structure
SIR -> retrieve, space, interleave, and repair it

That is the minimum because it covers the two essential jobs:

1. build the structure correctly enough that knowledge has somewhere to sit;
2. retrieve the structure often enough that it becomes usable instead of decorative.

Everything else supports, diagnoses, or extends that core.

Core Thesis

A learning system becomes too heavy when the learner starts managing the system instead of learning through it.

The minimum viable version should be strong enough to produce real learning, but small enough to use when energy, time, or motivation is limited.

For now, the core is:

Prestudy -> BHS -> SIR

Prestudy prepares the first rough frame.

BHS turns information into a structured map.

SIR keeps that map alive through reconstruction, spacing, interleaving, and gap repair.

The system is complete enough when it preserves the main learning loop:

frame
-> encode
-> retrieve
-> detect gaps
-> repair

What Makes It Minimal

The minimum system keeps attention on the few questions that make the session work.

It asks:

1. Did I prepare enough to know what I am looking for?
2. Did I build a usable structure?
3. Can I reconstruct that structure without looking?
4. Did retrieval reveal a gap?
5. Did I repair the gap?

If those five questions are answered, the learning session has a spine.

That spine can be improved later with Kolbs Experiential Cycle, Marginal Gains, Focus Management: How to Enter & Recover Inside a Work Block, or The Technique Is Only as Good as the Thinking It Produces.

Those supports are upgrades. Add them when they improve the core loop instead of making the system heavier.

The Minimum Loop

Use this when the full system feels too large.

1. Prestudy: get the rough frame.
2. BHS: build the map.
3. SIR: retrieve the map.
4. Repair: fix the weak part.
5. Repeat later.

The smallest practical version:

Step	Question	Output
Prestudy	What is this mainly about?	Rough frame and questions
BHS	How does it fit together?	Usable structure
SIR	Can I reconstruct it?	Retrieval attempt
Repair	What broke?	One corrected gap
Repeat	When should I see it again?	Next retrieval point

This is enough to prevent the most common failure: consuming information that never becomes usable.

Five Dimensions As Diagnostics

The five learning dimensions improve the minimum system by diagnosing where BHS + SIR is weak.

Dimension	What It Diagnoses Inside BHS + SIR	Upgrade Direction
Deep Processing	Is BHS building real relationships, or just organized notes?	Improve comparison, chunking, prioritization, and explanation.
Retrieval	Is SIR forcing reconstruction, or just recognition?	Improve recall, spacing, interleaving, and gap repair.
Self-Regulation	Can I tell when the method is failing while I am using it?	Improve diagnosis, adjustment, and feedback loops.
Self-Management	Can I reliably create the time, focus, and environment to use the system?	Improve task setup, attention, energy, and consistency.
Mindset	How do I interpret difficulty, overload, or poor recall?	Treat friction as signal instead of evidence of inability.

This is the right relationship:

BHS + SIR = current minimum viable learning system
Five dimensions = ways to diagnose and upgrade it

Trying to improve all five dimensions separately can become too abstract. The better move is to run BHS + SIR, then ask which dimension limited the result.

Where Agentic Engineering Might Fit

Agentic Engineering is an adjacent skill system being explored.

Possible fit:

BHS -> understand the problem or architecture
SIR -> retrieve the constraints, decisions, and failure modes
Agent -> implement, review, or pressure-test
Kolbs -> reflect on the workflow

The danger is forcing agentic engineering into a study model when it may need a build model.

The shared pattern is usable structure:

• clear frame,
• good questions,
• explicit constraints,
• retrieval of relevant context,
• execution,
• feedback,
• correction.

Agentic Engineering may use BHS and SIR as support tools, but it probably needs its own workflow around planning, implementation, verification, and handoff.

Where This Knowledge Base Might Fit

The knowledge base can support the minimum learning system in three roles:

1. Archive: preserve useful syntheses.
2. Thinking partner: generate questions, frames, and next moves.
3. Retrieval surface: make old ideas easier to revisit, connect, and apply.

The open question is how much the knowledge base should participate in the BHS + SIR loop.

Possible model:

BHS creates the map
SIR tests the map
the knowledge base preserves the refined map
future sessions retrieve from the knowledge base

This becomes useful when the knowledge base changes the next question, retrieval attempt, or action. Stored ideas that never return to practice stay in archive mode.

See Knowledge Base as Thinking Partner.

Current Open Question

The live question is:

Can BHS + SIR become the stable minimum, while Agentic Engineering and the knowledge base become extensions that plug into it only when they improve usable structure?

That question should stay open.

Do not force closure too early.

The test is practical:

• Does Agentic Engineering improve the learning loop?
• Does the knowledge base improve retrieval or transfer?
• Does either one make the next session clearer?
• Or do they add management overhead?

Failure Modes

Failure	What It Looks Like	Repair
Overcompression	Reducing the system to a vague loop that loses BHS and SIR.	Return to BHS for encoding and SIR for retrieval.
Technique sprawl	Adding every useful method into the “minimum.”	Treat Kolbs, Marginal Gains, and Focus Management as upgrades.
Dashboard drift	Managing pages instead of learning.	Use the system only if it changes the next session.
Forced integration	Trying to make Agentic Engineering fit the study system before the relationship is clear.	Keep it adjacent and test overlaps.
Archive illusion	The knowledge base feels useful because it is organized, but it is not retrieved from.	Ask whether the page changes action or recall.
Dimension abstraction	Trying to improve all five dimensions separately without a real session.	Run BHS + SIR first, then diagnose the limiting dimension.

Practical Use

For a study block:

1. Prestudy enough to create the first frame.
2. Use BHS to build the structure.
3. Use SIR to retrieve and test it.
4. Repair one weak point.
5. Ask which dimension limited the result.

For an Agentic Engineering block:

1. Use BHS only if the problem needs conceptual mapping.
2. Use SIR only if important constraints or patterns need retrieval.
3. Let the agentic workflow handle planning, implementation, verification, and review.
4. Use Kolbs afterward if the workflow exposed a repeatable bottleneck.

For this knowledge base:

1. Add a page only when it creates reusable structure.
2. Retrieve from pages during real questions.
3. Link pages when the relationship changes the next move.
4. Keep asking whether the knowledge base is improving thought or only storing it.

Related

• Prestudy, BHS, and SIR: Turning Information into Usable Structure
• Bear Hunter System
• Spaced Interleaved Retrieval
• Prestudy
• Dimensions of Learning
• Knowledge Base as Thinking Partner
• Agentic Engineering