learning-notes.org

* Approach to Learning
- Maintain a growth mindset - you have the capacity to get really good at whatever you put time/effort into. Mistakes are valuable feedback and evidence you're stretching yourself.
- Have a goal in mind - something concrete you're excited by.
- Write down your big questions about the topic and (if you can) what you know about it in general.
- Both guided and unguided learning are a necessary component of growth - plan to use both techniques

* Avoiding Procrastination
- Just start - it gets much better once you start making progress
- Be mindful about what you're doing
  - Use interstitial journaling to switch contexts and make sure your attention is focused on what you're doing, clarify your next actions, and be honest about your intentions and actions
  - Be aware of your mental state and pick a class of action that matches it (e.g. something simple/physical if you're unfocused, something more intense if you're dialed in...)
  - Appreciate the difficulty of learning something new - understand that once you start(!), understanding the mountain of info gets easier
- Redirect mindlessness
  - Tell yourself you'll only work on a task for 5 minutes, then let momentum carry you beyond that.
  - Use pomodoro - 25 minutes, then 5 minute breaks, with 20-30 minute breaks every 4 pomodoros
    - Or, use third time; work for however long you want, then break for up to a third of that time. If you go back to work before the max break, the leftover time gets added to the max of the next break. This leftover gets reset when you take a break e.g. for lunch.
  - Make actions small enough to where you don't choke on them - use GTD!
  - Use the "cue" in the "cue-routine-reward-belief" habit loop to build and break habits
    - Remove the cue for bad habits (like phone or internet), and hook into existing cues to start new good habits
  - Timebox your learning sessions so you don't have to think about starting

* External Structures to Support Learning
- Creating mind maps and semantic trees
  - Do a survey of the topic as a whole and seek to understand its disparate parts
  - Get a big picture view and draw a mind map (diagram with a central concept and supporting concepts/constituent parts branching off from it, recursively) in order to understand how a topic can be broken down
  - Use this to prioritize what to learn, understand where chunks might come from, and create a learning plan
- Spaced repetition
- Zettelkasten

* The Learning Process
- The goal is active and effortful engagement with the content you're learning
- A rule of thumb is that if learning is difficult, you're doing it right - most of the techniques here are used to make learning more effortful.
- Use deliberate practice - if you struggle with something, face it head on. "What would make me groan the most if I was tested on it now?"
- Take advantage of the spacing effect when scheduling sessions to do "distributed practice", where you have sessions more frequently in the beginning, then "space them out" as you improve, forcing your brain to work harder to remember the content. Distributed practice is much better than cramming!
- Engage both focused and diffuse modes of thinking by using pomodoros/third-time
- Chunks
  - Chunks allow you to be intellectually nimble, operating on a group of concepts as if they were an atomic unit
  - To form a chunk, follow all the following tips, but really concentrate on the main idea of a particular chunk, as well as the context you'd use it in
  - Observe an example, then do it yourself, then do it again and again in different contexts
- Creating Understanding
  - In order to understand something, you have to put it in your own terms. Writing an explanation of what you're learning is a good way to pin down if there's an area you don't really understand, as is the Feynman technique
  - To make "understanding in your own terms" into an action you do while reading, use "elaborative interrogation"
    - As you read content, ask yourself "why" or "how" something is true or should be done (preferably by recording the questions in some persistent fashion e.g. writing in the margins), and answer in your own words
    - You can use your source to help answer these questions initially, but your goal should be to not need it at all to answer them after you've come back to them a few times
  - If the topic you're learning is a skill or procedure you plan to put into practice, get hands on - start implementing or doing it to get a feel for it
    - Guided tutorials can make learning too "easy" - not effortful enough
    - Actively interact and experiment with tutorials - make deliberate mistakes and find out why they're wrong, ask why things are true
    - Do a "tutorial fade" - follow a tutorial verbatim and then redo the process it describes, without looking at the tutorial itself. Only use it if you get stuck, and aim to do the whole thing without looking.
- Recall, self-testing, and retrieval practice
  - We don't just want to be familiar with a piece of knowledge - we want to be able to retrieve it at will. This retrieval process is fundamentally reconstructive, not reproductive.
  - Self-testing is one of the most powerful learning techniques
    - Can take many forms - doing textbook questions, flashcards, or writing your own questions (e.g. through elaborative interrogation, above)
    - Have some way of testing your understanding of a particular source - whether this is creating something, answering questions about it, solving problems, etc.
  - Spaced repetition leverages both self-testing and the spacing effect to be tremendously effective
  - Should always at least be doing 3R (read, recite, review)
    - Read (or absorb in some form) the content
    - Reciting the information in your own words without looking at the source
    - Reviewing your recitation against the source to check for factual accuracy
  - For important content, do a full SQ3R
    - Prior to 3R, also survey and question
      - Survey: skim the text, analyze the structure of the argument, and perceive the general idea
      - Question: use the headings/subheadings to formulate questions about the content - what is it about, what knowledge do you want to get out of it, etc.
      - Also, add a terminal "Review" where you go back over the material you had trouble with, doing more cycles of recitation and review until you're comfortable with the content. Then, recite and review the passage as a whole (in whatever level of granularity is appropriate)
- Concrete Examples
  - Self-test on these when you can, they help form chunks
  - If you can access them, you can go from "can't do" to "master" rapidly by letting your brain do the work of pattern matching on a *very* high quantity (200-300) of high quality examples
  - The more abstract/difficult a subject is, the more necessary it is to work on concrete examples
- Feynman Technique - Learning by Explaining
  - You don't understand something if you can't teach it. After you finish a particular topic or chapter:
    1) Study - pick a topic and use the techniques here to try and understand it
    2) Teach - explain the topic like you're doing so to someone unfamiliar with it. Write the explanation down if you aren't actually giving it to someone. Really pretend you're giving a lecture! Include stories, analogies, simple sentences, etc.
    3) Refine - use the teaching step to identify gaps in your understanding and places where your explanation isn't simple and understandable enough
    4) Repeat and review - continue the process of studying, teaching, and refining until you're satisfied. Run it by a real person if you can to tighten it up as much as possible. Analyze if your explanation holds water, if it's simple, and if it's understandable.
- Interleaving
  - If you're studying a range of topics or chapters that share a similarity, "interleave" your studying by switching from one to another
  - e.g. do end-of-chapter problems for all chapters you're studying in random order, instead of one after another

* Self-review and advancing your skills
- Focus on process, not progress - learning is not always linear, so grade yourself on time you put in instead of what you got done
- Evaluating content you've learned
  - Square, circle, triangle
    - What *squares* with what you already know?
    - What's rolling around your head (needs more processing to internalize)? (*circle*)
    - What are you going to do differently (*delta*) as a result of this knowledge - how can you apply what you've learned?
  - When you study, you're spending time in details - take a step back and check your understanding from all altitudes/angles. Beware your initial intuition or first thought - reevaluate it to make sure it works.
- Iterate on the process - always be looking for what works and what doesn't, and update these notes accordingly. Ask about what you could be doing better - from another person/mentor if possible.
- Advancing your skills
  - Imagine buckets A -> B -> C, with all your skills residing in one of these buckets
  - A (can't do) -> B (can do with effort) C -> (can do automatically, without effort)
  - All these skills draw from the same pool of cognitive resources, so the more skills that you rely on in B, the more drained you get as they sap your cognitive resources.
  - There are three problems we face with moving skills to the right:
    1) Pile-up on B (your skills that take effort aren't getting better - they drain the resources you'd use to improve them)
       - Break the stuff in B up into smaller subskill chunks
       - If you can't take a subskill chunk to mastery within 3 sessions of 45-90 min, it's too big - try a smaller subskill
       - Do this aggressively - the longer a skill spends in B, the more practice you have at being mediocre at it
    2) Mediocre C (the things you do automatically are "good enough" but need to be better)
       - Always be evaluating the skills you have that you can do without thinking about them; if they're holding you back, move them to B and spend resources on improving them
    3) A -> C is slow
       - Simple, but not always possible to fix - we need a very high quantity (200-300) of high quality examples to let our brains subconsciously pattern match on what to do