Software Engineering
How To Become Dangerously Self-Educated (with AI)
Sandeep Swadia · 7/16/2026 · 2 min read · source
Cognitive Myths & Reading Traps
- Learning Style Myth: Visual, auditory, kinesthetic preferences lack scientific evidence. Self-labeling limits learning capacity.
- Illusion of Fluency: Understanding explanation != understanding concept. Yale University study: confidence collapsed when explaining step-by-step mechanics of everyday objects (zippers, toilets).
- Highlighter Trap: Mistaking marking text for memory retention.
- Summary Trap: Creating perfect notes never read again.
- Completion Trap: Finishing book without internalizing change.
ACTOR Framework
Human = actor. AI = sidekick/companion, not shortcut.
1. Aim (A)
- Concept: Read as spy, not tourist. Focus transforms consumption to construction.
- Action: Write one-sentence mission before reading: "I am reading this book because I need to [blank]."
- AI Role (Framer): Generate three questions to carry into reading; suggest books matching specific real-world problems.
- Case Example: Lin Manuel Miranda (2008) read 800-page Alexander Hamilton biography with mission (hip hop, immigration, word-building); ignited creation of Hamilton musical.
2. Compress (C)
- Concept: Carry less, understand structure.
- Elon Musk Knowledge Tree Metaphor:
- Trunk: Load-bearing core idea.
- Branches: Chapters, major arguments.
- Leaves: Quotes, examples, stories. Avoid collecting leaves without trunk.
- Action: Write short version of key takeaway after reading.
- AI Role (Interpreter): Verify interpretation of load-bearing ideas. Identify overstatements or omissions. Applicable for complex texts (e.g., Zen and the Art of Motorcycle Maintenance, the innovators dilemma, narcissists and golemans).
3. Test (T)
- Concept: Read to find what to reject, not just agree. Disagreement drives self-discovery.
- Action: Question why passages trigger discomfort. Identify protected personal beliefs. Write notes on disagreements.
- AI Role (Sparring Partner): Challenge user interpretations, find hidden assumptions, provide counterarguments, describe failure scenarios of book advice.
- Evidence: Stanford study on death penalty showed mixed evidence polarized people further; Bill Gates writes feverishly in margins when disagreeing.
4. Own (O)
- Concept: Relive ideas in own words. Rereading creates false familiarity. Teaching moves ideas into mind.
- Action: Explain book in 1-2 paragraphs. Teach concept to someone (or wall).
- AI Role (Coach): Convert ideas into plain English, supply business analogies/examples, evaluate user's explanation.
- Evidence: Washington University in St. Louis study: active recall group retained significantly more long-term knowledge than repetitive rereading group.
5. Run (R)
- Concept: Books are software updates. MIT motto: mind and hand. Reading must change decisions/behavior.
- Action: Apply concepts to real situations.
- AI Role (Action Companion): Translate book concepts into one decision, one rule, one checklist, or one experiment.
- Case Example: Crucial Conversations book applied to target emotional safety, master internal stories, and build shared pools of meaning.