February 20, 2021


Below are some insights from the book Ultralearning by Scott H Young. In addition to the book, I recommend a follow of his instagram account, which is full of insightful cartoons on his work.

Skip to the end to see a plan of an Ultralearning Project.

Key Traits important in Ultralearning

  • Ability to focus quickly and deeply.
  • Ability to be comfortable being uncomfortable.
  • Being aware of what puts you in flow, including knowing the environment that puts you there.
  • Being aware of your general learning/working weaknesses, and where your crutches are that feel like work, but are actually procrasintation.


  • Work in spurts, whether in Pomodoro (25 minutes) or slightly longer (50 on/10 break per hour).
  • Do different things during a study period, to break up singular focus and create different challenges.
  • Jot down notes that re-explain hard concepts when reading. Do it in your own words to improve recall.
  • Take a break, like a walk or a nap, when things are feeling especially difficult. The relaxation of focus now can lead to a Eureka moment later.


  • Why, What, How of your Ultralearning project. Why do it, what should you do, and how to make it happen?
  • Why - Instrumental (in order to achieve an outside result), or intrinsic (for the sake of learning).
  • What - Concepts (ideas you need to understand), Facts (Information you can memorise), Procedures (anything you can only learn through practice, e.g. pronunciation).
  • How - Benchmarking (figuring out common learning methods - reading, internet searching, expert advice), Emphasise/Exclude (Go through Benchmarked resources and find key ones to emphasise and exclude).
  • How Much Should You Plan/Metalearn? 5-10% of total duration of project approx. Can also be done as learning slows down, so throughout project.

Direct Learning

  • Means closely tying learning to the situation you want to use it in.
  • Language learning for example. Don’t just read grammar to start a new language, speak to people and gradually notice the patterns and flow of speech. You’ll learn faster with the embarrassment that comes with failure too.
  • For more academic learning, it means testing yourself using the parameters set by the school or examination board you’re affiliated with.

When you learn something in one context and are able to transfer it to another one (Classroom - - > Real Life)

  • Can be Project-Based (in which you learn by doing the thing you want to use the skill for), immersive (diving into the context of the skill you want to learn), Overkill (testing in gradually more difficult arenas to really test yourself).
  • You need to isolate the skills to improve them individually. Drills do this.
  • Which skill would cause the greatest overall improvement if worked on?
  • E.g. Ben Franklin would improve writing by deconstructing others’ essays, re-writing their work from memory, rhyming words and synonyms to see if his vocab improved, and randomising pieces of a great work then re-ordering them. Obsessive, but it worked.
  • Drill 1: Time Slicing. Isolate a slice of time in a longer sequence of actions. e.g. a musical piece, or key language phrases, or penalties/drop-kicks.
  • Drill 2: Cognitive Components. Take one component apart and practice it. E.g. a Chinese tone repetition.
  • Drill 3: Copycat. Copying parts of skill you don’t want to drill means you can focus on the component you want to practice. E.g. Drawing based off others’ drawings mean you focus on the drawing, not the difficult framing of the picture..
  • Drill 4: The Magnifying Glass Method. Zoom in on one piece obsessively. e.g. Research a subject 10x more than you need to to develop new habits and shortcuts.
  • Drill 5: Pre-requisite Chaining. Focus on things you will struggle at early, not the super easy stuff. E.g. Speaking difficult language tenses early instead of doing what Duolingo does - repeating present tense until you’ve done 100 hours then moving into another one… You can’t have a good conversation just in the present.


Scott’s Secret weapon . Look at a number of words, say, and then try to recall them afterwards without prompting. Free recallers do much better on tests than other learners.

For long term memory to develop, we need freedom for our brains to make connections and invest ways of remembering things.

Best Method: FLASH CARDS!

Other methods: Closed-book mindmaps, pictures, writing; Self-generated problems or challenges; Writing questions for yourself to do later.


  • Most useful when it is specific and related to an action.
  • Fear of feedback is ubiquitous and damaging.
  • Dive into hardest environment to achieve most harsh and therefore accurate feedback.
  • This is one of the most important Ultralearning traits.
  • Corrective feedback is the best kind - shows your what you’re doing wrong, and how to fix it. Coaching, mentorship, teaching provides this, as well as flash card learning for specific skills.

Improving feedback

  • Noise cancellation - Focus on the key metrics, not the loudest ones. e.g. in measuring the success of your blog writing, see how many people read to the end (or at least to the key info).
  • Hit the difficulty sweet spot - More information = more opportunities to learn. Adjust the difficulty level so you’re not overwhelmed, but you’ve got enough to keep an edge. Adjust the environment so you’re not able to predict when you’ll succeed and fail.
  • Metafeedback - Determine your ideal learning rate, then see when it slows (adjust methodology), and compare different study methods/structures/schedules to see which work better.
  • High-Intensity - Lots of feedback, often.

AMAZING QUOTE: “Memory is the residue of thought.” David Willingham

Forgetting Curve

  • We tend to forget things incredibly quickly after learning them. Amount of knowledge forgotten lessens over time.
  • Remembering is not a passive process. We’re engaging in a creative process of reconstruction (can be fabrications e.g. Eyewitness testimony).

Memory Systems

  • Spacing - Don’t cram, space. Spread 10 hours of learning over 10 days.
  • Good technique - print a list of words, read them over, rehearse them mentally without them in front of you.
  • Semi-regular practice and recall after you’ve learned something.
  • Proceduralisation - Procedural knowledge = knowing how (e.g. to ride a bike), vs. declarative knowledge = knowing that (e.g. Pythagorean Theorum).
  • So, learn the core concepts and make sure they make sense. Then add extra specifics as they become useful.
  • Overlearning - Practice beyond perfect - performance won’t improve, but durability of memory will. Combined with spacing and proceduralisation, and wham, you’e got quite a memory there.
  • Go a level above where you’re comfortable, and the comfortable stuff will retain as the brain needs it to.
  • Mnemonics - Create a picture in your head of the word or concept you’re trying to learn, and make it hard to forget. Recall it when needed. Powerful skill.

Richard Feynman - Famed polymath and legendary physicist.

Don’t give up on hard problems easily - Push through and see where your new limits get you.

Prove things to understand them - Make them make sense to you.

Start with a concrete example - Create an image in your head of a concept in action, so it’s less theoretical

Don’t Fool Yourself - Dunning-Kruger effect - Someone who believes they have more knowledge than they really do, because they’re ignorant of how much they lack. Ask lots of questions, even dumb ones.

The Feynman Technique - Write a concept/problem on a sheet of paper, and explain it to someone who’s never heard of it. Like a child. By re-explaining it, you will expose gaps in your own knowledge you’ve never bothered to go into.


Young uses the example of Van Gogh’s life of constant experimentation (despite near-constant criticism from friends, himself, and the artistic establishments of the day).

Growth Mindset (Carol Dweck) very important in this. Seeing your own capacity for learning as something that can be actively improved.

Tactic 1: Copy then create. Deconstruct others’ methodology and style and imitate it.

Tactic 2: Compare methods - Try two different ones over periods of time. Split testing ideas.

Tactic 3: Introduce new constraints - Never settle on the pedestal of achievement. Don’t fall into the Dunning-Kruger effect. Introduce new constraints that make the old methods impossible to use.

Tactic 4: Find your superpower in the hybrid of unrelated skills - Magic zone here.

Tactic 5: Explore extremes - Don’t play it safe.


1) Do research, get resources, plan activities

2) Schedule time (shorter chunks spread out are better). Don’t be overly ambitious. Add in more hours if you need them.

3) Execute your plan. Questions to ask if you’re slipping:

  • Metalearning - Have you researched/prepped properly?
  • Focus - are you focused?
  • Directness - are you direct enough with it?
  • Drill - Are you drilling your weakest skills?
  • Retrieval - Have I worked on my memory and recall?
  • Feedback - Am I getting good enough, honest feedback?
  • Intuition - Do I understand things, or am I just memorising them?
  • Experimentation - Do I need to branch out with new techniques?

4) Review Results

5) Choose to maintain or master what you’ve learned

Alternatives to Ultralearning

  • Low level habits - if you enjoy them (esp after a UL project), continue.
  • Formal, structured education (Can be better for community, apprenticeship, deeper immersion).

More Work
February 19, 2024
February 2, 2023

Subscribe to get updates once every few months.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.