You Think You’re Learning. You’re Not!

Main ideas, concepts, and lessons

• Learning via “compression” vs. learning via “systems”

  • Popular explainer styles (especially on YouTube) often compress complex reality into something easier to understand (e.g., maps, concise narratives).
  • The metaphor: a map removes details (buildings→blocks, countries→shapes, movement→arrows) to reveal overall structure.
  • However, compression alone may not change behavior or produce real learning if the learner’s internal model is wrong or incomplete.

• Why understanding doesn’t automatically lead to learning

  • If people think they already understand, they may not update their beliefs.
  • Even when an explanation is clear and “understandable,” it may not transfer into corrected knowledge because learning has to pass through a mental model already in the student’s head (described as a “black box” system).

• Key example: Derek (Veritasium) physics study

  • Initial setup
    • Derek tests first-year physics students on Newton’s first and second laws of motion.
    • Pre-test average: 6 out of 26.
    • Students then watch a standard explainer video.
    • Post-test average after standard explanation: 6.3 out of 26 (almost no improvement).
  • Interpretation
    • The explainer was beautifully compressed and described by students as clear and easy to understand.
    • Yet it failed to meaningfully improve understanding—suggesting the explanation didn’t sufficiently engage learners’ existing misconceptions.
  • Intervention
    • Derek uses an alternative approach to engage students:
      • Puts two people in a video.
      • One person voices common misconceptions.
      • Misconceptions are shown/animated without explicitly labeling them as misconceptions, forcing students to work out the conflict through social interaction.
  • Feedback and results
    • Students often found it more confusing, not “easy to understand.”
    • Scores nearly doubled to 11 out of 26.

• Compression has a limitation; “friction” with existing models creates change

  • The deeper lesson is not only what was discovered, but how:
    • Derek’s process moved from assuming clarity automatically improves learning to realizing learning requires interacting with the system that produces outcomes.
  • Systems-thinking approach described:
    • Create friction with the learner’s existing model to trigger updating.
    • Rather than targeting the outcome directly, map the system producing it and find a high-leverage intervention point.

• Causal maps and leverage

  • Builds on the “map” idea:
    • Geographical maps compress spatial relationships.
    • This video argues for causal maps that compress causal relationships—“what causes what.”
  • Purpose of causal maps:
    • Help you see what affects what
    • Show where small changes can cause large downstream effects (“leverage”)
  • Speaker’s goal for their own channel:
    • Produce videos with a systems thinking perspective on important topics.
    • Provide leverage—information that helps you change something, not just understand a problem.

---

Methodology / instructions (detailed bullet format)

• Systems-thinking method for improving real outcomes (as described in the video)

  • Start with an outcome you care about (what you want to change).
  • Identify the system producing that outcome, not just the explanation or surface facts.
    • Treat the learner/observer process as a model-driven system (a “black box” that includes prior beliefs and misconceptions).
  • Map causal relationships (a causal map), focusing on:
    • Which factors cause other factors
    • What affects what
    • Where the largest downstream effects can be triggered
  • Find high-leverage intervention points, i.e. places where small changes have bigger effects downstream.
  • Design interventions that interact with existing internal models:
    • Don’t assume clarity alone will update beliefs.
    • If people already think they know, create friction with their current model so they must re-evaluate.
  • Measure whether the system has changed, not just whether the explanation seemed clear:
    • Use pre/post testing or outcome measures to detect actual learning/behavior change.

• Specific instructional approach used in the study (as a learning intervention example)

  • Use a learning presentation that:
    • Includes social interaction (two people on screen)
    • Has one participant articulate common misconceptions
    • Does not label those misconceptions as wrong
    • Lets learners resolve the conflict themselves (through cognitive engagement)
  • Expect the method to feel more confusing, but treat increased confusion as part of driving belief updating through mental effort.
  • Validate with measurable improvement (test scores increased from ~6 to ~11 in the described example).

• Video/channel strategy described by the speaker

  • Create content aimed at:
    • Showing systems and causal leverage, not only facts
  • Encourage viewers to:
    • Subscribe to the channel
    • Join on Patreon for additional causal maps
  • Emphasize that content remains free, and support is for building a team to produce more videos faster.

---

Speakers / sources featured (identified from subtitles)

• Main speaker (narrator/host) — the person speaking throughout and promoting the channel/Patreon (name not given in subtitles)
• Johnny Harris — referenced for a style of explanation using maps and storytelling
• Derek — identified through context as Derek Muller, creator associated with Veritasium
• Veritasium — the YouTube channel/source of the described PhD-related experiment and explanation style
• University of Sydney — referenced as the institution where Derek tested physics students