UltraLearning

This is a re-post from my personal blog at https://www.michielnuyts.com/

I read ultralearning a while ago, and here are some of my notes and highlights on it. It's definitely an inspiring book and I can't wait to setup my own ultralearning project pretty soon.

Main Principles

1. Metalearning

   Draw a map, Start by learning how to learn the subject. Do research.

2. Focus

   Concentrate, remove all distractions. Get into a flow-state.

3. Directness

   Go straight ahead, learn by doing the thing you want to become good at.
   Don't trade it off for other tasks just because those are more comfortable.

4. Drill

   Attack your weakest point. Be ruthless in improving your weakest points. Break down complex skills into small parts, then master those parts and build them back together again.

5. Retrieval

   Test to learn. Test yourself before you feel confident. Push yourself to actively recall information rather than passively review it.

6. Feedback

   Don't dodge the punches. Feedback is harsh. Extract the signal from the noise so you know what to pay attention to and what to ignore.

7. Retention

   Understand what you forget and why. Learn to remeber things not just for now but forever.

8. Intuition

   Dig deep before building up. Develop your intuition through play and exploration of concepts and skills. Understand how understanding works. Deeply know things!

9. Experimentation

   Explore outside your comfort zone. True mastery comes not just from following the path trodden by others but from exploring possibilities they haven’t yet imagined.

Metalearning Broken Down

Why?

Refers to understanding your motivation to learn. If you know exactly why you want to learn a skill or subject, you can save a lot of time by focusing your project on exactly what matters most to you.

What?

Refers to the knowledge and abilities you’ll need to acquire in order to be successful. Breaking things down into concepts, facts, and procedures can enable you to map out what obstacles you’ll face and how best to overcome them.

How?

Refers to the resources, environment, and methods you’ll use when learning. Making careful choices here can make a big difference in your overall effectiveness.

Answering Why

Determine if learning a topic is likely to have the effect you want it to, before you get started.

Answering What

Once you’ve gotten a handle on why you’re learning, you can start looking at how the knowledge in your subject is structured. A good way to do this is to write down on a sheet of paper three columns with the headings “Concepts,” “Facts,” and “Procedures.” Then brainstorm all the things you’ll need to learn. It doesn’t matter if the list is perfectly complete or accurate at this stage. You can always revise it later. Your goal here is to get a rough first pass. Once you start learning, you can adjust the list if you discover that your categories aren’t quite right.

Concepts (first column)

Concepts are ideas that you need to understand in flexible ways in order for them to be useful. In general, if something needs to be understood, not just memorized, I put it into this column instead of the second column for facts.

Facts (second column)

Write down anything that needs to be memorized. Facts are anything that suffices if you can remember them at all. You don’t need to understand them too deeply, so long as you can recall them in the right situations.

Procedures (third column)

Anything that needs to be practiced. Procedures are actions that need to be performed and may not involve much conscious thinking at all. Learning to ride a bicycle, for instance, is almost all procedural and involves essentially no facts or concepts.

Underline the concepts, facts, and procedures that are going to be most challenging. This will give you a good idea what the major learning bottlenecks are going to be and can start you searching for methods and resources to overcome those difficulties.

Knowing what the bottlenecks will be can help you start to think of ways of making your study time more efficient and effective.

Often this coarse-grained analysis is enough to move on to the next phase of research. However, with more experience, you can dig deeper. You might look at some of the particular features of the concepts, facts, and procedures you’re trying to learn to find methods to master them more effectively.
If you can’t make these kinds of predictions and come up with these kinds of strategies just yet, don’t worry. This is the kind of long-term benefit of metalearning that comes from having done more projects.

Answering How

Benchmarking

The way to start any learning project is by finding the common ways in which people learn the skill or subject.

Look at the curricula used in schools to teach that subject. This could be the syllabus from a single class or the course list for an entire degree.

Syllabus from courses on Harvard, Cambridge, Udacity, ...

Do online searches for people who have previously learned that skill or use the Expert Interview Method to focus on resources available for mastering that subject.

An hour spent searching online for almost any skill should turn up courses, articles, and recommendations for how to learn it. Investing the time here can have incredible benefits because the quality of the materials you use can create orders-of-magnitude differences in your effectiveness.

The Emphasize/Exclude Method

The Emphasize/Exclude Method involves first finding areas of study that align with the goals you identified in the first part of your research. If you’re learning programming solely to make your own app, I’d focus on the inner workings of app development more than theories of computation.

The second part of the Emphasize/Exclude Method is to omit or delay elements of your benchmarked curriculum that don’t align with your goals.

A good rule of thumb is that you should invest approximately 10 percent of your total expected learning time into research prior to starting. This percentage will decrease a little bit as your project scales up.

More Research or Start Learning?

Compare the marginal benefits of metalearning to regular learning. One way to do this is to spend a few hours doing more research and then spend a few hours doing more learning. After spending some time on each, do a quick assessment of the relative value of the two activities. If you feel as though the metalearning research contributed more than the hours spent on learning itself, you are likely at a point where more research is still beneficial.

Focus: Sharpen Your Knife

If I have difficult reading to do, I will often make an effort to jot down notes that reexplain hard concepts for me. I do this mostly because, while I’m writing, I’m less likely to enter into the state of reading hypnosis where I’m pantomiming the act of reading while my mind is actually elsewhere.

Drill: Attack Your Weakest Point

By identifying a rate-determining step in your learning reaction, you can isolate it and work on it specifically. Since it governs the overall competence you have with that skill, by improving at it you will improve faster than if you try to practice every aspect of the skill at once.

The rate determining step is the slowest step of a chemical reaction that determines the speed at which the overall reaction proceeds. (Improving a skill)

Finally, doing drills is hard and often uncomfortable. Teasing out the worst thing about your performance and practicing that in isolation takes guts.

By copying the parts of the skill you don’t want to drill (either from someone else or your past work), you can focus exclusively on the component you want to practice. Not only does this save a lot of time, because you need to repeat only the part you’re drilling, it also reduces your cognitive burden, meaning you can apply more focus to getting better at that one aspect

Start with a skill that you don’t have all the prerequisites for. Then, when you inevitably do poorly, you go back a step, learn one of the foundational topics, and repeat the exercise. This practice of starting too hard and learning prerequisites as they are needed can be frustrating, but it saves a lot of time learning subskills that don’t actually drive performance much.

Retrieval: Test to Learn

After reading a section from a book or sitting through a lecture, to try to write down everything you can remember. Free recall like this is often very difficult, and there will be many things missed, even if you just finished reading the text in question. However, this difficulty is also a good reason why this practice is helpful. By forcing yourself to recall the main points and arguments, you’ll be able to remember them better later.

Most students take notes by copying the main points as they encounter them. However, another strategy for taking notes is to rephrase what you’ve recorded as questions to be answered later.

Restrict yourself to one question per section of a text, thus forcing yourself to acknowledge and rephrase the main point rather than zoom in on a detail that will be largely irrelevant later.

Feedback: Don't Dodge the Punches

Sometimes the best action is just to dive straight into the hardest environment, since even if the feedback is very negative initially, it can reduce your fears of getting started on a project and allow you to adjust later if it proves too harsh to be helpful. All of these acts require self-confidence, resolve, and persistence, which is why many self-directed learning efforts ignore seeking the aggressive feedback that could generate faster results. Instead of going to the source, taking feedback directly, and using that information to learn quickly, people often choose to dodge the punches and avoid a potentially huge source of learning. Ultralearners acquire skills quickly because they seek aggressive feedback when others opt for practice that includes weaker forms of feedback or no feedback at all.

One important type of metafeedback is your learning rate. This gives you information about how fast you’re learning.

There are two ways you can use this tool. One is to decide when you should focus on the strategy you’re already using and when you should experiment with other methods. If your learning rate is slowing to a trickle, that might mean you’re hitting diminishing returns with your current approach and could benefit from different kinds of drills, difficulties, or environments. A second way you can apply metafeedback is by comparing two different study methods to see which works better.

High-intensity, rapid feedback offers informational advantages, but more often the advantage is emotional, too. Fear of receiving feedback can often hold you back more than anything. By throwing yourself into a high-intensity, rapid feedback situation, you may initially feel uncomfortable, but you’ll get over that initial aversion much faster than if you wait months or years before getting feedback. Being in such a situation also provokes you to engage in learning more aggressively than you might otherwise. Knowing that your work will be evaluated is an incredible motivator to do your best. (This is a good reason to maybe open-source learning projects like building an app, or livestream while coding.) This motivational angle for committing to high-intensity feedback may end up outweighing the informational advantage it provides.

Retention: Don't Fill a Leaky Bucket

A strategy for applying spacing, which can work better for more elaborate skills that are harder to integrate into your daily habits, is to semiregularly do refresher projects. This approach has the disadvantage of sometimes deviating quite a lot from optimal spacing; however, if you’re prepared to do a little bit of relearning to compensate, it can still be a better approach than completely giving up practice. Scheduling this kind of maintenance in advance can also be helpful, as it will remind you that learning isn’t something done once and then ignored but a process that continues for your entire life. Overlearning is a well-studied psychological phenomenon that’s fairly easy to understand: additional practice, beyond what is required to perform adequately, can increase the length of time that memories are stored.

Intuition: Dig Deep Before Building Up

Simply spending a lot of time studying something isn’t enough to create a deep intuition.

One way you can introduce this into your own efforts is to give yourself a “struggle timer” as you work on problems. When you feel like giving up and that you can’t possibly figure out the solution to a difficult problem, try setting a timer for another ten minutes to push yourself a bit further. The first advantage of this struggle period is that very often you can solve the problem you are faced with if you simply apply enough thinking to it. The second advantage is that even if you fail, you’ll be much more likely to remember the way to arrive at the solution when you encounter it. Difficulty in retrieving the correct information—even when the difficulty is caused by the information not being there—can prime you to remember information better later.

Prove Things to Understand Them

The Dunning-Kruger effect occurs when someone with inadequate understanding of a subject nonetheless believes he or she possesses more knowledge about the subject than the people who actually do.15 This can occur because when you lack knowledge about a subject, you also tend to lack the ability to assess your own abilities. It is true that the more you learn about a subject, the more questions arise. The reverse also seems to be true, in that the fewer questions you ask, the more likely you are to know less about the subject.

One way to avoid the problem of fooling yourself is simply to ask lots of questions.

Explaining things clearly and asking “dumb” questions can keep you from fooling yourself into thinking you know something you don’t.

Write down the concept or problem you want to understand at the top of a piece of paper. In the space below, explain the idea as if you had to teach it to someone else. If it’s a concept, ask yourself how you would convey the idea to someone who has never heard of it before. If it’s a problem, explain how to solve it and—crucially—why that solution procedure makes sense to you. When you get stuck, meaning your understanding fails to provide a clear answer, go back to your book, notes, teacher, or reference material to find the answer.

Experimentation: Explore Outside Your Comfort Zone

When starting to learn a new skill, often it’s sufficient simply to follow the example of someone who is further along than you. In discussing the principles of ultralearning, metalearning comes first. Understanding how a subject breaks down into different elements and seeing how others have learned it previously, thus providing an advantageous starting point. However, as your skill develops, it’s often no longer enough to simply follow the examples of others; you need to experiment and find your own path.

The first place to experiment is with the methods, materials, and resources you use to learn. This kind of experimentation is useful in helping you discover the guides and resources that work best for you. It’s important, however, that your impulse to experiment be matched with a drive to do the necessary work. A good strategy to take is to pick a resource (maybe a book, class, or method of learning) and apply it rigorously for a predetermined period of time. Once you apply yourself aggressively to that new method, you can step back and evaluate how well it is working and whether you feel it makes sense to continue with that approach or try another.

Pick some subtopic within the skill you’re trying to cultivate, spend some time learning it aggressively, and then evaluate your progress. Should you continue in that direction or pick another? There’s no “right” answer here, but there are answers that will be more useful to the specific skill you’re trying to master.

There are two advantages to doing split tests. The first is that as in scientific experiments, you will get much better information about which method works best if you limit the variation to only the factor you want to test. The second is that by solving a problem multiple ways or applying multiple solution styles to it, you will increase your breadth of expertise. Forcing yourself to try different approaches encourages experimentation outside your comfort zone.

The challenge of learning in the beginning is that you don’t know what to do. The challenge of learning in the end is that you think you already know what to do. It’s this latter difficulty that causes us to rerun old routines and old ways of solving problems that are encouraged through habit, not always because the old way is actually best. A powerful technique for pushing out of those grooves of routine is by introducing new constraints that make the old methods impossible to use. It’s practically an axiom of design that the best innovations come from working within constraints. Give a designer unlimited freedom, and the solution is usually a mess. On the other hand, creating specific constraints in how you can proceed encourages you to explore options that are less familiar to you and sharpens your underlying skills.