Learning to Learn

Original Article ↗ Hacker News Discussion ↗

Identifying “Foundational” Knowledge

  • Many struggle to know what foundations to learn as a novice; it’s hard to see the “root nodes” of a field from the outside.
  • Suggested heuristics:
    • Start from concrete problems and learn only what’s needed to solve them.
    • Notice concepts or procedures that recur across many problems.
    • Accept that in many domains there is no single strict foundation; “good enough for your goals” may suffice.
  • Some argue you can’t identify foundations in advance; you discover them iteratively as gaps appear.

Novice → Expert & Dunning–Kruger

  • Rough progression suggested:
    • Beginner: can’t complete projects without support.
    • Intermediate: can complete but unsure about quality.
    • Expert: has done it before in multiple ways and can explain it simply.
  • Several comments stress Dunning–Kruger: early confidence often masks large gaps; you need external feedback and harder questions to see what you don’t know.

Curriculum, Resources & the Role of Experts

  • One camp: finding the single “best” resource (e.g., a canonical textbook) is crucial; curriculum design is a big part of learning.
  • Counterpoint: “best resource” hunting easily turns into procrastination; “accurate and good enough” plus steady work usually beats endless optimization.
  • Tactics mentioned: use university reading lists, ask experts to sketch a path, study well-regarded open-source projects, and use LLMs for topic maps (but not for citations).

Practice, Projects & Interviews

  • Strong agreement that you must “do the thing”: use new skills in real tasks or projects.
  • Tension noted between learning for real-world work vs. preparing for interviews that test narrow trivia or tooling.

Efficiency vs Enjoyment

  • Some find “learning efficiency” talk sterile; they prioritize curiosity, joy, and exploration.
  • Others emphasize that real progress often feels effortful and boring; grit and drills matter, especially for math and similar fields.
  • Nuanced view:
    • Consistency can beat intensity (e.g., many easy “reps” vs. rare brutal sessions).
    • Fun, flow, and efficient methods can coexist; context (job vs hobby, time constraints) should drive the mix.

Language Learning & Pareto Ideas

  • Using frequency lists (e.g., top 800 words) is attractive, but multiple commenters say:
    • 75–80% word coverage still feels like understanding very little; comfortable reading often needs ~95–98% coverage and thousands of words.
    • Grammar, multiword expressions, and lots of input are essential.
  • Still, a few hundred–thousand high-frequency words can enable basic fluency and communication, especially in speech.

Medicine as a Case Study in Self-Learning

  • A thread explores self-studying medicine: using med-school textbook lists, anatomy/physiology texts, and clinical guidelines to better understand one’s own conditions.
  • Some enthusiasm: laypeople can significantly improve their health literacy and interactions with doctors.
  • Skepticism: full medical competence requires many years of supervised clinical experience; self-study can’t substitute for that, especially for procedures and complex chronic care.