Learning and the Subconscious Bullet Points

Summary: Your subconscious has most of your brainpower. To become a great thinker, focus on teaching your subconscious and delegating work to it. Don’t focus on conscious analysis. The subconscious isn’t the irrational part of your mind; it’s a necessary part of rational thinking. I explain the importance of the subconscious, discuss how to improve your subconscious with practice, and discuss managing your error rate.

Subconscious Computation

  • Your brain is literally a computer
      – It has a ton of computing power – more than a desktop computer today– Thinking involves computing– Over 99% of your computing power is controlled/used by your subconscious– Your subconscious does many more computations than your conscious mind
  • Your conscious mind is like a single manager trying to oversee a huge factory with thousands of workers
      – The subconscious is like the workers– Many subconscious workers can do their jobs well for months without any management– Some workers keep screwing up and can waste a lot of your manager’s precious attention
  • Subconscious ideas are called many different things, like:
      – EmotionsIntuitions– Hunches– Gut feelings– Inexplicit ideas
          – Inexplicit means not in words or symbols (like math uses symbols)– But they’re fundamentally pieces of knowledge or units of (subconscious) thought
          – That claim is using the premises and terminology of Critical Rationalism by Karl Popper– I don’t know the best way to explain this in mainstream, inductivist terminology (which views knowledge differently – I think incorrectly) but hopefully you can get the general idea– Thinking involves computation. The subconscious and conscious both do computation. The subconscious does real, actual thinking
  • Many people believe rationality involves focusing on the conscious mind
      – Emotions and intuitions are considered irrational– Explicit ideas that people can write down or speak out loud are heavily favored– This is a big mistake
  • If you want to be really smart, you can’t rely on only your conscious mind
      – More computing power enables better thinking– Your subconscious has most of your computing power– If you use your subconscious thinking power, you have a huge advantage over people who don’t– All great thinkers use a lot of subconsciouses computing power

Learning Philosophy

  • The typical goal of philosophy study is to learn some ideas in the field
      – You read Plato, Kant, Popper, whatever, and learn their ideas– Philosophy is more oriented towards teaching the history of philosophy than most fields because people struggle to agree on which philosophy ideas were actually good– In physics, more controversies get resolved, so there is more emphasis on teaching the current/correct ideas– In philosophy, people disagree more about which ideas are useful, so they’re more likely to teach multiple perspectives that disagree
  • If you want to be familiar with philosophy and its ideas, the standard approach is fine:
      – Read some books and/or take some classes– Learn the general ideas of what some famous thinkers thought– You may find this fun and interesting
  • What if you want to be a great philosopher?
      – What if you want to use philosophy to solve problems?– What if you want to win debates or guide debates to correct conclusions?– What if you want to figure out which views are true and why?– What if you want to create new and better philosophy ideas?– Then your top goal should be to improve your subconscious
          – You’ll never be great or even good without your subconscious’ help– Trying really hard to use conscious analysis will never be good enough because it relies on only like 1% of your computing power

Improving Your Subconscious

  • There are basically two ways to get a really smart subconscious
      1. Make your subconscious really smart during childhood, especially before age five
          – No one really knows how to make this happen on purpose. It’s not just luck but we have little control over it– Awful, abusive parenting lowers the chances your kid is really smart– We know some things parents should avoid doing– But we don’t know how to use good parenting to raise the chances much2. Train and improve your subconscious on purpose
          – If you’re a teenager or older, and you’re not already super smart, then this is your only option– Almost everyone needs to use this option if they want to be a great thinker– If you became somewhat smart during childhood, that means there’s less work left to do to become really smart
  • How do you improve your subconscious?
      – How do you make it a reliable, rational ally?– How do you delegate thinking effort to your subconscious?– How can you use subconscious ideas in debates or discussions? How can you communicate about them?

Practicing Skills

  • You improve your subconscious by focused practice of physical or mental skills
      – Philosophy is primarily about mental skills– Some physical skills like fast, accurate touch typing are relevant to philosophers– All physical skills involve some thinking, and many mental skills involve some physical action. Researching a topic involves not just thinking but also physically controlling a computer (or paper book) and physically typing or writing notes– Many skills are primarily physical or mental, but some are more mixed like speed reading
  • You already have helpful subconscious skills
      – Your subconscious is probably already good at physical skills like walking, controlling your hand muscles to use a fork, or typing– Your subconscious may already be good at mental skills like single digit addition as mental math, translating letters into words, or recognizing many grammar errors
          – For most incorrect sentences, any fluent English speaker will quickly and intuitively spot an error without having to consciously analyze it
              – Your experience is different than this because you read a very non-random selection of sentences. You mostly read sentences where the author didn’t notice an error. He intuitively notices most of the same errors that you do, so he’ll tend to publish writing where it’s harder for you to notice errors. This makes your grammar-error-noticing intuition/skill seem less powerful than it actually is– If you look at sentences written by someone who started learning English a month ago, you’ll easily see many errors. There’s a clear skill difference even if you use little conscious effort– You know that “Thought I philosophy about.” is wrong. It sounds wrong to your subconscious even if you don’t have a conscious thought like “the subject and verb are out of order”.
  • You can read more in my article Practice and Mastery

Automatizing Skills

  • The concept of automatization comes from Ayn Rand’s philosophy, Objectivism
      – Similar ideas have been talked about by other thinkers as well
  • Practice can lead to mastery and automatization
      – When you can do something “on autopilot” or it’s “second nature”, then your subconscious is doing most of the work– When every step feels intuitively obvious, your subconscious knows how to do it, and your conscious mind doesn’t have to guide the steps– When something seems “automatic”, that means your subconscious is doing it for you without much conscious supervision
  • When you automatize something, it frees up conscious attention for something else (such as more advanced ideas)
      – Conscious attention is a scarce, bottleneck resource
          – A key goal is to offload work from your conscious to your subconscious– You can consciously learn more basic things, practice them so your subconscious learns them, then focus your conscious attention on more intermediate or advanced things, and repeat
  • It’s common to get something 90% automatized instead of 100% automatized
      – Partially reducing the amount of work for your conscious mind is good– Partial automatization is fine if it’s a hard, important thing– Partial automatization is problematic if it’s a building block for a bunch of more advanced knowledge or if it’s something you do many times per day (such grasping an object with your hand)
  • To make progress towards being a great thinker, use this process:
      1. Learn something consciously2. Consciously refine it and try to make your knowledge high quality and low error rate3. Begin practicing and use your conscious to make adjustments as problems come up
          – In this stage, you generally do it slowly so you have time to consciously think while doing it4. Once everything is going very smoothly, practice a bunch more so your subconscious learns it and it becomes automatic
          – In this stage, you speed up. As it becomes more automatic (more computation is being done subconsciously), you should be able to go faster and should find it easier

Changing an Automatization

  • In the future, you may need to change something you already learned and automatized
      – Learning it plus changing it requires more work than learning it correctly in the first place– You should make a strong but not perfectionist effort to get things right initially before putting in a lot of practice– Frequently changing ideas you recently automatized is inefficient and indicates something is going wrong– Changing ideas you learned years ago is expected and should happen some, though hopefully not an overwhelming amount
  • To change an automatized skill/idea/knowledge, here’s the standard process:
      1. Do it slowly and more consciously2. Remember what it was like to consciously control it and what the steps or sub-parts were3. Once you’re able to do it in a non-automatic way, then you can use your conscious mind to make changes4. Practice the changes slowly and make sure you’re happy with the result5. Practice a lot more and gradually speed up (while consciously making sure you’re doing it the new way, not the old way) and make it more subconscious and automatic
  • In summary, change automated skills by turning off the automatization and taking conscious control, then making changes, then practicing in order to teach your subconscious the changed version
  • It’s possible to change automatized skills without making them conscious again then re-automatizing the new version
      – People succeed at this sometimes– But there’s currently no reliable way to do it– You can try it. Tell yourself what changes you want to make, think about them, visualize them, etc.– Maybe it will work without much effort– If it doesn’t work, you can do the full process for changing which is more effort but more reliable

Mental Skills

  • People understand practice, automatization and autopilot pretty well for physical skills
      – E.g. the concept of practicing handwriting is probably already familiar to you– Practicing sports skills is well known– People are less familiar with walking as a learning process involving lots of practice because they have little memory of doing it
  • People have less understanding about practicing mental skills, thinking processes, knowledge, ideas, etc., so I want to give some examples:
      – Practicing a foreign language is a good example
          – That uses practice to learn new words and grammar rules– People will practice with e.g. flashcards. Practicing mental skills usually doesn’t mean just sitting still and thinking– They can achieve a result where they automatically, subconsciously know what words mean– Some people stop learning early but fluency means a lot of language skills are automatic rather than requiring conscious thought, just like how you use your native language– Practicing with math worksheets can improve your mental math– Practicing with grammar worksheets can improve your intuitive, subconscious or mental grammar skill– You can practice making various types of tree diagrams, flow charts, notes, or outlines– You can practice researching or debating topics– You can practice reading news articles and looking for a specific type of error or for any error (you can also do this with books, blogs, movies or any other media on any topic)
          – You can practice with other people, so they can point out some things you missed, and you can point out some things they missed. (You can do that online at the Critical Fallibilism forum.)

Using Ideas in Your Life

  • When you learn a new idea, you need to know when and how to use it, or it won’t affect your life
      – For some ideas, especially more abstract or philosophical ideas, it’s not obvious when to use it, and you may repeatedly miss opportunities to use it– You need some triggers which your subconscious recognizes
          – Your subconscious can remind you of the idea so you can consciously apply it– Or your subconscious itself can use the idea– Reminders require less practice/effort and are suitable for ideas that you’ll rarely use– The more frequently you’ll use an idea, the more important it is to use subconscious computing power for it instead of conscious attention
  • You can practice recognizing scenarios where an idea is relevant and then using it
      – You can go through many scenarios and slowly, consciously figure out if the idea is relevant and apply it if so– After you’re good at this consciously, with more practice your subconscious can learn to do what your consciousness is already doing– The reason people keep “forgetting” to use ideas is they never practiced figuring out when (or sometimes even how) to use the idea
          – They never considered 50 scenarios and thought through in which ones the idea should be used– They never figured out clear criteria for distinguishing when the idea is relevant or not– They forget to use it because they don’t know when to use it– Practice aimed at identifying when to use an idea would involve some scenarios where you use it and some where you don’t, so you have to figure out which are which.– Doing a bunch of practice problems where the same idea is relevant every time, by design, won’t help much for figuring out which ideas to use when. Many worksheets are designed with only one type of practice problem instead of mixed problems. Both types of worksheets can be useful
  • People read ideas and think they agree, and think the idea sounds nice, but then don’t realize further learning steps are needed
      – Their plan is basically that if they are in conscious analysis mode, then their conscious mind may be able to use the idea– But the whole rest of the time, they won’t use it

Self-Improvement

  • People usually only spend around 2-4 hours per day using their conscious mind heavily
      – Doing focused thinking for more than 4 hours per day may lead to burnout– If you can’t use something subconsciously, then you won’t use it during most of the day– Your limited time window for conscious attention is precious– Your top goal for that time should be improving your subconscious
          – Focus on learning and practicing– If you use your limited conscious attention to get tasks done, then you’re missing out on scarce opportunities for self-improvement– Every time you spend conscious mental resources on something that could have been done subconsciously later, that is a lost opportunity for learning and practice– If you keep doing self-improvement and get to the point your subconscious can do something, then you can get the same task done without spending conscious attention on it directly
          – That means you’ll be able to get more total learning and practicing done in your life– Doing tasks “early” – when they require a lot of conscious effort – is expensive: it uses the resources needed for improvement and progress– Doing tasks later, when they’re subconscious (easy and automatic) is cheap because subconscious resources are plentiful– Your conscious attention is the bottleneck or limiting factor on your progress and on improving your subconscious– Using your subconscious is cheap– So you should improve your subconscious and use it a lot
              – Consciously focus mostly on learning and self-improvement, rather than having your conscious do work that could have (later) been delegated to the subconscious
  • If something is really important to do right now, then it can be worth using conscious mental resources on it
      – If this happens occasionally, it’s OK– If this happens frequently, it will be a big drain on your resources, and it will be hard to improve
  • Focus your conscious attention on powering up
      – The more you allocate conscious brainpower to self-improvement, the better off you are– If you want to be a great thinker, you need as much self-improvement as you can get– For everything other than self-improvement, delegate as much as you can to your subconscious– If your subconscious can’t do something but it isn’t urgent, you can improve your subconscious then do it, rather than using conscious attention now
  • Teach your subconscious skills that you’ll reuse frequently
      – Any kind of building block for future knowledge should be automatized in your subconscious– If you will only do something three times in your whole life, and it isn’t super important or risky, then don’t teach it to your subconscious and don’t practice it– If you will only go on a space flight to the moon once in your life, you should still practice a bunch because you won’t get any second chances and really want to avoid mistakes– But most stuff is much lower risk, in which case you should automatize anything you’ll use a lot, but rarely used stuff may not be worth delegating to your subconscious
          – Only getting your subconscious to do part of the work, not all of it, can work– People actually do basically nothing fully consciously – they use their subconscious for part of the work for pretty much everything– The issue is how much your subconscious does, not whether to use your subconscious at all
  • My article Do Primarily Easy Things – Increasing The Productivity Of Your Intellectual Labor Vs. Consumption has more information
      – I also have a podcast episode on Overreaching and Powering Up

Great Thinkers

  • To become a great thinker:
      1. Learn how to learn effectively2. Effectively learn critical thinking skills3. Study prerequisites then advanced concepts4. Use your critical thinking skills to innovate
  • More specifically, to become a great thinker:
      1. Learn the general concepts of the subconscious, practice, automatization, mastery, error rate management, integration, etc.2. Teach critical thinking skills to your subconscious3. Study prerequisites then advanced concepts while using your conscious and subconscious critical thinking skills4. Find (decisive) errors in current concepts, find unsolved problems, brainstorm solutions, and do critical thinking about your brainstormed solutions

Error Rate Management and Integration

  • Advanced ideas build on intermediate ideas which build on more basic ideas
      – Very advanced ideas can build on fifty layers of previous ideas– The more layers you build on top of an idea, the more reliable it needs to be– Metaphorically, the more levels of a building will go on top of this building block, the stronger the block needs to be because there will be more weight on it
  • When you build an idea using several previous ideas, that is called “integration” because it combines (integrates) multiple units into one new unit
      – The concept of integration comes from Ayn Rand’s philosophy, Objectivism– Advanced ideas are built through many, many integrations starting with more basic ideas– We can’t consciously think about hundreds of things at once– Integration creates more powerful ideas, which lets us do powerful thinking with only a few ideas active in our conscious mind at once
  • If you build on an error or integrate using an error, your new idea will probably be wrong
      – If you build many layers of ideas based on an error, you’ll have many wrong ideas– Before doing a bunch of building/integrating, you should carefully check for errors– High quality standards are worthwhile– Lazy shortcuts will end up wasting your effort because you’ll have to revisit them later and it’ll be more work to fix things after already building a lot on an error– Don’t be a perfectionist, but don’t ignore “small” errors– The more layers you build on an idea, the more small errors can become important
          – Great thinkers build many, many layers of ideas
  • When you use an idea, you can think of it as using all the lower level ideas (that it builds on) too
      – An error in any of those ideas can lead to an error– If you use twenty high level ideas in a thinking session, you might also be using ten thousand lower level ideas that you don’t consciously consider– You could get a wrong answer due to errors in any of those ten thousand ideas
  • The more layers you will build on an idea – the more it will be reused – the higher quality is needed
      – If an idea screws up 1% of the time, and you use it once a year and pay conscious attention to it, that’s probably fine– You can notice when it screws up and fix it– If a low level idea in your subconscious has a 1% error rate, that could result in multiple errors per day– And if an error occurs in your subconscious, you often won’t know what/where the error is– Searching for subconscious errors can be a lot of work– Changing an idea that many other ideas already built on is also a lot of extra work, because the changes may be relevant to those other ideas and require that those other ideas also be changed
  • People are usually satisfied with a 95% success on practice worksheets
      – Schools will grade that an “A”– But if you want to use that skill as a subconscious building block for many layers of more advanced ideas, then that isn’t good enough– If an idea will be subconsciously used hundreds of times per day, then it needs an error rate way below 1% so that it doesn’t cause daily problems
  • The need for a low error rate in frequently reused ideas comes up with computer software
      – Low level software functions are used by higher level functions, sometimes in twenty or more layers– A single call to a high level function can result in thousands of calls to lower level functions– The low level functions need to be extraordinarily reliable. A 1% error rate would be really bad and would make building up twenty more layers impossible– When there’s a bug, people usually search in the most recent several layers of code– Finding bugs in low level functions can be very difficult and time consuming
  • When you learn, and something goes wrong, you’ll usually search for an error in the idea you’re currently focusing on and a few previous layers of ideas
      – If the error isn’t there, people often just get stuck and end up giving up– People don’t realize they’re often stuck due to an error in an earlier layer– People are bad at looking through, evaluating and changing ideas that are ten or more layers below the surface– It’s harder to revisit old layers with a subconscious than with computer code
  • As you learn, you’re creating many layers of ideas, just as software developers build many layers of functions
      – To build to advanced ideas, you need to make the building blocks really good
          – Building blocks need a low error rate (the more it will be used – the more layers will go above it – then the lower the error rate that’s needed)– Building blocks need to be automatized (taught to your subconscious)– In software, if something is used a lot, it’s strongly preferred to make it fully automatic. Something that’s used less often, or very difficult to automate, may require manual actions by a user
  • People often consciously forget how their older ideas work
      – This makes fixing lower level errors hard– It makes introspection hard– It helps to keep notes on your learning processes– It helps to consciously use ideas periodically to refresh them in your memory– Ideas from childhood are the hardest to remember or have useful notes about
          – What did I put in my mind when I was an infant? I forget. I don’t know– It’s possible to change things and solve problems in other ways besides what I’ve been explaining, but it’s harder. Think of it as requiring creative work-arounds, which is less reliable than having a clear process to follow
  • To keep your error rate low, you need to clearly understand your ideas before being satisfied
      – Don’t put up with some level of confusion– You can try learning the next few levels despite some confusion. That often helps. Just make a note that the partially confused idea is incomplete and must be revisited reasonably soon– If you don’t know what you mean in a clear way, your idea isn’t good enough to build on and become a great, advanced thinker– And you need to be able to evaluate the correctness of your ideas yourself– You need to know how to judge your ideas and need to judge them

Perspectives on Learning

  • There are many broad ways to think of learning
      – Learning can be seen as dealing with a succession of problems, and progressing to better problems– Learning can be seen as evolving your ideas– Learning can be seen as automatization and integration
  • You can also think of learning as a process of improving your ability to evaluate ideas
      – If you can’t tell what’s correct or incorrect, then you have more to learn about this– You’re ready to begin learning a topic when you know enough prior ideas that you can find errors and judge what’s right or wrong– If you can’t tell what’s good, you can’t practice– Practice requires trying something repeatedly, and trying to repeat what you did when it was correct, and trying to make adjustments when it’s incorrect– Practice requires that you can tell the difference between success and failure– If you don’t clearly know what success or failure look like, you can’t practice productively, because you won’t know what to do more or less of, or what to change

Concluding Summary

Your brain is a computer. Your subconscious has most of the computing power. Great thinkers use that subconscious computing power effectively (everyone uses it some). The best way to use your conscious computing power is for learning new things and teaching them to your subconscious. Delegate things to your subconscious which you’ll use many times.

Learning involves practice. When you practice, there are two main stopping points. People often stop practicing when they’re able to succeed using conscious attention. But for important knowledge that you’ll use frequently or which serves as a building block for other knowledge, you should keep practicing to reach the second stopping point: when your subconscious is good at it (this is generally when it becomes fast, easy, reliable, intuitive, easy-to-remember and possible to do on autopilot).

Just reading books and essays, and listening to podcasts, isn’t enough to learn. You must do focused practice. You must break what you’re learning down into manageable parts, figure out success and failure criteria, and practice until you’re successful while doing it mostly subconsciously (mostly on autopilot). A way to test that your subconscious is doing most of the work is trying to do it when you’re very mentally tired. In other words, try doing it when you’re not able to do much conscious thinking. Conscious skills degrade much more from fatigue than subconscious automatizations do. Mentally tired people are low on conscious mental resources but usually still have plenty of subconscious mental resources left.

For the ideas which you’ll reuse a lot or where mistakes are very costly – the ones where practice and subconscious skill are most important – you also need them to be highly reliable. You need a very low error rate. The more an frequently an idea will be used – or the more costly an error is – the more you need a success rate like 99.9999% not just 95% or 99%. High standards are important if you hope to build many more layers of knowledge above what you’re doing.

One of the main reasons people don’t get very far with philosophy is they find it hard because they’re trying to do it all with their scarce, precious conscious attention. They try to focus really hard and do things consciously, which is a huge burden, and which never lets them build up very many layers of knowledge. You can’t reach advanced philosophical knowledge by focusing just on your conscious mind. You must get beginner and intermediate knowledge into your subconscious before you can do advanced philosophy. (There is more than one good path to advanced knowledge. Any single, specific idea could be skipped and you could still succeed. But you need a path, not just to skip straight to complicated stuff. And only a few paths actually work well; most random paths won’t work.)

Learn in Small Chunks

To make progress as a philosopher, you should focus on learning things that you can successfully automatize within a week. Finishing the same day you start, in a few hours, is even better. Break your learning goals into small enough parts to finish things and reach success in a short time frame.

If that’s hard for you, begin by practicing learning itself. Learn something non-philosophical that’s easy to learn, just so you can finish a learning process quickly and successfully with no distractions due to the subject matter. Don’t try to learn two things (e.g. economics and how to automatize new skills) at the same time. You can start by working on the process for practicing and automatizing physical skills, then do mental skills, and then use that process to learn basic, intermediate and finally advanced philosophy. You can become an advanced philosopher while always focusing on an immediate sub-goal that you can finish in under a week (while also having a job or some other activities, not doing philosophy full time).

Put another way, all of philosophy can be divided into parts that take less than 20 hours to learn both consciously and subconsciously. If you can’t finish learning projects that quickly, you need to break them into smaller chunks, or choose goals that are more similar to what you’re already successful at (avoid skipping steps; instead do things that directly build off what you already know), or improve your learning process.

It’s OK if sometimes projects fail or take too long because you can’t plan and organize perfectly and you don’t fully know what learning something will be like before you learn it. Also, don’t count yourself as not being done just because you didn’t succeed yet. Failing on time is often better than extending a project. You can recognize and accept the failure, then plan to try again later. You should generally do at least one other thing successfully before trying again. And when you try again, you should have a new plan. The new plan may be similar to the old plan, but should have at least one change designed to fix what went wrong before. If you have trouble figuring out what went wrong, you can work on something else or you can try dividing the project into smaller chunks (that way, if a chunk fails, there’s a smaller area to search for an error).

There is some unpredictability, but most or your projects should succeed on time – if that isn’t happening, then you need to make changes to your approach. If you aren’t frequently finishing steps/sub-goals successfully, it’s very hard to tell if your learning is actually working. You need to use steps that you understand so that you can evaluate success and failure yourself. (It’s partially possible to use a coach or tutor to do some evaluations for you and help guide your learning projects, but you need to mostly understand things yourself.) In order to evaluate success and failure, you need to work on topics which are adjacent to knowledge you already have. Your existing knowledge that you learned successfully is areas where you can judge confidently and reliably success and failure, and where your subconscious can do most of the work. You need to build step by step from that existing knowledge. If you start losing confidence, feeling confused or failing more often, that’s a sign you’re skipping steps and should work on stuff closer to what you’re already good at.

The fastest and most effective way to learn knowledge very different from what you currently know is by many small, fast, successful steps (which is sometimes called “gradualism”), not by skipping steps. (It may be different for an infant or young child who doesn’t already have a substantial base of existing knowledge to build on. Very early learning isn’t understood as well as later learning and has some differences.)


I also made a video discussing this article.