Learning

Evidence-based techniques for retaining knowledge. Research over intuition.

Quick Reference

Technique	Impact	Core Idea
Active Recall	High	Test yourself instead of re-reading
Spaced Repetition	High	Review at expanding intervals
Interleaving	Medium	Mix related topics, don’t block
Elaboration	Medium	Ask why, connect to existing knowledge
Generation	Medium	Produce answers, don’t just consume them
Feynman Technique	High	Explain simply to expose gaps
Dual Coding	Medium	Combine words and visuals
Chunking	Medium	Group into meaningful units

Active Recall

Retrieve information from memory instead of passively reviewing.

Practice:

• Close the book, write what you remember
• Flashcards: see question, produce answer from memory
• Practice tests beat re-reading notes
• Explain concepts without looking at source

Why it works: Retrieval strengthens memory traces. Each successful recall creates new access paths. Even failed attempts improve later retention.

Research: Repeated retrieval produces 400% improvement vs. studying once. (Karpicke & Roediger)

Spaced Repetition

Review material at increasing intervals: 1 day, 3 days, 1 week, 2 weeks, 1 month.

Practice:

• Use SRS software to automate scheduling
• Without software: review next day, then 3d, 7d, 14d, 30d
• Let yourself almost forget before reviewing
• Harder to recall = stronger memory formed

Why it works: Each retrieval at the edge of forgetting strengthens the memory. Massed practice (cramming) produces short-term gains, long-term losses.

The forgetting curve: Without review, we lose ~40% in days, ~90% within a month. Spaced review flattens this curve.

Tools: Anki, RemNote, Mochi, Obsidian + Spaced Repetition plugin

Interleaving

Mix related topics (ABCABC) instead of blocking (AAABBBCCC).

Practice:

• Alternate between problem types during practice
• Switch subjects within study sessions
• Create mixed practice sets for review
• Don’t do 20 of the same problem in a row

Why it works: Forces discrimination between similar concepts. Blocking lets you pattern-match without true understanding.

The catch: Interleaving feels harder and less productive. This is a feature. The difficulty improves long-term retention.

When to block: Start with blocked practice for brand-new material. Interleave once you have the basics.

Transition signals (move to interleaving when you can):

• Solve basic problems without reference material
• Explain the core concept in your own words (Feynman test)
• Recognize which technique applies to a new problem

Research: 30-40% improvement on delayed tests vs. blocking.

Elaboration

Ask “why” and “how” questions to connect new material to existing knowledge.

Practice:

• After learning a fact: “Why is this true?”
• “How does this relate to what I already know?”
• Generate your own examples
• Explain the steps in a procedure, not just the steps

Why it works: Creates semantic links to prior knowledge. More connections = more retrieval paths = better recall.

Limitation: Requires background knowledge to elaborate from. Works best when you have prior knowledge to connect to.

Generation Effect

Producing information beats passively receiving it.

Practice:

• Fill-in-the-blank exercises (cloze deletion)
• Predict the answer before revealing it
• Create your own practice questions
• Write code before looking at the solution

Why it works: Generation activates deeper encoding—semantic elaboration, distinctive processing, effortful retrieval.

Counterintuitive: Wrong answers, when corrected, can strengthen memory more than reading the right answer directly. Struggle is productive.

Feynman Technique

Explain concepts simply to expose gaps in understanding.

Steps:

1. Choose a concept you want to understand
2. Explain it in plain language, as if teaching a child
3. Identify gaps where you stumble, get vague, or resort to jargon
4. Simplify and return to source material for weak spots
5. Repeat until the explanation flows

Why it works: Combines active recall, generation, and self-explanation. Reveals illusions of competence—you can’t hide behind jargon.

Use for: Complex topics, interview prep, testing whether you actually understand something or just recognize it.

Dual Coding

Combine verbal and visual representations.

Practice:

• Sketch diagrams while reading
• Add images to flashcards
• Create mind maps from memory
• Draw system architectures, don’t just describe them

Why it works: Verbal and visual memory are separate channels. Encoding in both creates two independent retrieval paths.

Caution: Poorly designed visuals increase cognitive load. Simple > elaborate. The visual should clarify, not decorate.

Chunking

Group information into meaningful units to work within memory limits.

Practice:

• Phone numbers: 973-820-5846 not 9738205846
• Acronyms: ACID for database transactions
• Patterns: See git rebase -i as one concept, not four tokens
• Build chunks from smaller chunks (hierarchy)

Why it works: Working memory holds ~4-7 items. A chunk counts as one item regardless of internal complexity. Experts chunk more aggressively.

In practice: When learning a new domain, consciously look for patterns that can become single units. Name them.

Supporting Factors

Sleep

Sleep consolidates memory, particularly slow-wave sleep.

• Post-learning sleep stabilizes and enhances memories
• Sleep before learning prepares encoding capacity
• Naps help; all-nighters actively hurt retention
• Fatigue sabotages retention

Deliberate Practice

Not just repetition—structured practice targeting weaknesses.

• Focus on skills slightly beyond current ability
• Get immediate feedback
• High concentration, not autopilot
• Quality matters more than hours logged

The “10,000 hour rule” oversimplifies. Deliberate practice explains ~26% of performance variance—significant but not everything.

Metacognition

Thinking about your thinking.

• Plan: What do I know? What’s my approach?
• Monitor: Is this working? Am I actually learning or just busy?
• Evaluate: What did I miss? What should I adjust?

The ability to accurately assess your own understanding is a skill. Develop it.

Intellectual Humility

Admitting what you don’t know signals mastery, not weakness.

Strong opinions, loosely held:

• Have conviction based on experience
• Actively seek information that would change your mind
• Update publicly when you’re wrong
• The best outcome is sometimes being proven wrong

Why it works: Pretending to know everything erodes trust. Leaders who model fallibility create environments where others grow. “I don’t know yet, but here’s how I’ll find out” is a strength.

The paradox: Expertise enables humility. Beginners overestimate their knowledge (Dunning-Kruger). Masters know how much they don’t know.

What Doesn’t Work

Technique	Problem
Re-reading	Feels productive, builds familiarity not recall
Highlighting	Passive; no retrieval or elaboration
Learning styles (VAK)	No evidence that matching “style” helps; everyone benefits from multimodal
Cramming	Works for tomorrow’s test, gone next week
Blocked practice only	Feels smoother, transfers worse
Passive video watching	Understanding is not retention

The techniques that feel easiest often work worst. Fluent reading creates an illusion of learning. Test yourself to know what you actually know.

Building a System

Minimum viable learning workflow:

1. Capture: Take notes in your own words (generation, not transcription)
2. Process: Convert notes to questions or flashcards (forces active framing)
3. Review: Follow a spaced repetition schedule
4. Test: Practice problems, teach-backs, self-quizzes
5. Sleep: Non-negotiable

Example: Learning React Query in 2 Weeks

Days 1-3: Capture — Read docs and tutorials. Take notes in your own words, not transcripts. Sketch the mental model (dual coding).

Days 2-5: Process — Convert notes to Anki cards. Frame as questions: “When does React Query refetch by default?” Force active recall framing.

Days 3-14: Review — Daily SRS review, 10-15 minutes. Let the algorithm schedule; trust the spacing.

Days 5-14: Test — Build a small project. Explain your cache strategy out loud (Feynman). Debug without docs when possible.

Throughout: Sleep — Consolidation happens overnight. No late-night cramming.

Phases overlap. You’re still capturing while processing, still processing while testing. The system is a loop, not a sequence.

The system matters less than consistency. Pick tools you’ll actually use.

Learning to Learn Faster

The meta-skill is metacognition—noticing when you’re learning vs. just busy.

Accelerators

Shorten feedback loops. Test yourself immediately after learning, not days later. Wrong answers surface gaps faster than re-reading.

Systematize, don’t improvise. Pick one SRS tool, one note system, one practice rhythm. Deciding how to learn burns time you could spend learning.

Learn domains, not just skills. Each new domain (language, instrument, sport) trains the learning process itself. The pattern—confusion → frustration → competence → fluency—becomes familiar.

The Counterintuitive Truth

What feels slow often works fast:

Feels Inefficient	Actually Effective Because
Struggle and confusion	Effortful encoding strengthens memory
Retrieval failure	Failed attempts prime later success
Interleaving topics	Forces discrimination, not pattern-match
Spacing over cramming	Edge-of-forgetting retrieval consolidates

Fluent re-reading feels productive but evaporates. The techniques that feel hardest produce the most durable learning.

The Leverage Point

Ten minutes of Anki beats two hours of “I should really review that.”

Knowing the techniques isn’t the bottleneck—running the system daily is. Build the habit; the learning compounds.

Tools

Landscape shifts fast. See awesome-fsrs and awesome-knowledge-management for current options.

Purpose	Examples
Spaced repetition	Anki, RemNote, Mochi
Connected notes	Obsidian, Logseq, Roam
Flashcard creation	Anki (with cloze), Mochi, RemNote
Coding practice	Exercism, LeetCode, project-based learning
Reading retention	Readwise (syncs highlights to SRS)

See Also

• Knowledge Design — Structuring knowledge for transfer (the teaching side of learning)
• Complexity — Chunking relates to managing cognitive load
• Learning a Programming Language
• Technical Writing
• Agent Memory
• Dunlosky et al. (2013). Improving Students’ Learning With Effective Learning Techniques. Psychological Science in the Public Interest.
• Karpicke & Roediger (2008). The Critical Importance of Retrieval for Learning. Science.
• Bjork & Bjork. Desirable Difficulties in Theory and Practice. UCLA Learning and Forgetting Lab.
• Ebbinghaus (1885). Memory: A Contribution to Experimental Psychology. (Forgetting curve)
• Roediger & Butler (2011). The Critical Role of Retrieval Practice in Long-Term Retention. Trends in Cognitive Sciences.