Information Intelligence - How to Learn Anything

How to Learn Anything

To deepen our understanding of information intelligence, let’s explore how knowledge or skills become part of ourselves—learning. There is a strong connection between mastering a skill and happiness. Whether for work or a hobby, if someone can effectively learn the necessary skills and knowledge, they are more likely to feel a sense of achievement and earn recognition both professionally and socially. We want to be good at things, and the key to becoming proficient is learning. As a middle-aged man, I recall many people teaching me specific skills—math, drawing, piano, violin, swimming, riding a bicycle, chess, solving a Rubik’s Cube, playing basketball…. However, I cannot recall anyone teaching me how to learn in general. Learning about learning itself is absent in my life. If this resonates with you, I hope the five steps of learning outlined below will provide insight.

1. Understand the Depth. To learn anything, the first step is to grasp the subject’s depth. Our brain allocates cognitive resources based on how deeply we perceive a subject; therefore, having an accurate understanding of a subject’s depth can help the brain distribute its resources effectively. This is very similar to how we assign strength based on the perceived weight of an object before lifting it – for example, we lift a water bottle casually, but engage more muscle when lifting a dumbbell. If there’s a fake dumbbell made of plastic, we might lift it dramatically because we wrongly allocate too much strength. The depth of a subject can be measured in various ways. For subjects with a clearly defined state, we can use the number of states as a measure of depth. Chess has roughly 10^43 legal states, whereas Go (Weiqi) has about 10^170 legal states. It is evident that Go is much deeper than chess. For subjects without a clear definition of states, we can use the number of layers of abstraction as the metric of depth. For example, when learning Latin and Japanese as foreign languages, we need to learn at the syllable level (word spelling and pronunciation), the sentence level (grammatical formation), and the paragraph level (speaking to articulate a point and writing an article). However, for Japanese, there is additionally a level of abstraction — the living culture behind the language, the interaction and interpretation of the most recent concepts and events in the world. These living factors do not apply to Latin, as no country currently uses it. Therefore, the depth of learning Japanese exceeds that of Latin.

Chess Openings

2. Decomposition of Learning Units. Once we understand the subject’s depth, the next step is to break it down into learning units. These units are categorised into two levels. At the macro level, learning units represent the components that need to be mastered to fully understand the subject. This involves analysing what needs to be mastered to master X, such as areas a, b, c, d, e, and so on. For example, to learn to swim, we need to become comfortable in the water, hold our breath, move the upper body, move the lower body, coordinate upper- and lower-body movements, coordinate breathing patterns, learn to dive start, and develop the skill to perform a flip turn. Once all these skills are mastered, we can swim a particular style. At the micro level, learning units are the individual components that can be practised or understood effectively. Let’s call these micro learning units “chunks.” For example, chunks for piano practice might be scales, arpeggios, or specific bars of a piece of music. When practising piano, we repeat the particular scale, arpeggio, or bar until fully mastered, then move on to the next chunk. The size of these chunks is particularly important. If a chunk is too large, it becomes too cognitively demanding for the brain to master, making internalisation difficult. Conversely, if the chunk is too small, the connections between chunks become extensive, making it harder to master the whole subject. For instance, practising each two notes as a chunk might seem simple, but even after mastering each pair, we’re unlikely to be able to play the entire piece.
Defining the two levels of learning units is particularly difficult for the learner. It is a paradox that, to accurately define the learning components (the macro units) and the chunks (the micro units), one requires subject expertise; however, as a learner, that expertise is, by definition, nonexistent. Therefore, it is necessary to seek professional advice (e.g., from a coach or teacher) to define or refine these two levels of learning units.

3. Practice until Internalised. After we break down the learning units, the next step is to focus on practising. Practising means repeatedly exercising a learning unit until it is fully mastered. The main aim of practising each unit is to internalise it. Internalising involves reaching a point at which knowledge or a skill feels effortless, stable, and automatic. This process occurs when sufficient new neural pathways form during practice, linking new knowledge or skills with existing ones. In other words, internalisation is how new knowledge or skills cease to be isolated. Instead, they become part of our cognitive framework, like new roots anchoring into the land, becoming stable and well understood.
Let’s look at some examples. When learning to solve a Rubik’s Cube, we start by breaking it down into several units (the beginner’s method, the CFOP method, the Roux method, etc.). Each method is too complex to be a single learning unit, so we break it down further. For example, in the CFOP method, the F (First Two Layers) step has 41 variations. Each of these variations is small enough to serve as a separate learning unit. When practising a specific variation (such as a twisted corner with a flipped edge), we need to repeat it many times until, when we see a particular pattern (like a twisted corner and a flipped edge at the same time), our hands can perform the pattern automatically (such as a counter-clockwise twist of the top layer, followed by a clockwise turn of the right layer). This new skill then integrates into our existing skill set for move patterns, so it becomes internalised. Another example of interlinking can be learning history. When first learning about a historical event (a learning unit), we focus on who did what at what time, which is not at the level of internalisation. When we deepen our understanding of the history event by reading more material, discussing and debating with experts, and writing down our analysis (the equivalent of practising), we start to comprehend the influence of the event on the society of that time, and how subsequent historical events are consequences of it. The cultural changes, ideological shifts, and the literature, music, and arts reflecting the historical event become clearer in our minds. The event shifts from a simple point to a web connecting many other happenings. Now, understanding of the history event becomes part of our broader understanding of the world — it is internalised.
When practising, there are a few rules to follow. The first rule is to maintain attention. In the ‘Scarcity of Attention’ section discussed earlier, we discussed how metabolically valuable attention is. When we begin practising, neuron connections form neurochemically as neurons transfer information through neurotransmitters. If we sustain attention long enough, these connections become anatomical, with dendrites and axons forming permanent pathways [1][2][3]. It’s the intensity and duration of attention that indicate how important the new information is and how likely it is that anatomical connections will form. Our brains literally think, “Wow, this new thing has been in the attention pool for so long; it must be really important, so I’d better form strong pathways to store it.” The second rule is to repeat. The learning process is “three steps forward and two steps backward.” Even with potent attention, the formation of permanent connections is unlikely without repeated practice. As early as 1885, Hermann Ebbinghaus’s experiment revealed that newly learned information is lost over time without repetition. Information loss occurs rapidly in the early stages and slows later. Therefore, we need to repeat frequently when we have just finished learning, and the frequency of repeats can decrease over time. The last rule of practising is to be patient. Physiologically, there is only a limited number of synapses that can be created at a given period of time. Therefore, no matter how hard we try, there is a finite amount of information we can absorb during that period. Learning is slow, so we need to respect the physical limits of our brains and stop believing there is such a thing as “master piano within a month”.

4. Seeking Feedback. One reason learning is difficult is that we don’t know what we don’t know. Without feedback, we might practise too large learning chunks (for example, learning all Sicilian chess opening variations at once), work on segments far beyond our ability (think of how many one-year experienced piano hobbyists are practising Clare de Lune), or even practice the wrong things (the more we practise holding a golf club the wrong way, the further we are from a good swing). Therefore, feedback from an expert can be highly valuable, as they can identify our mistakes more easily. If we had unlimited resources, the best way to practise would be to have an expert observe each move—every chess move, tennis serve, musical phrase, or ball shot—and provide immediate correction. However, having an expert available at all times isn’t practical, so we need to deliberately plan to seek professional feedback. If mastering a language is important to you now, instead of practising with a local speaker once a week, practise every day. If learning the violin is a hobby, still aim to see a teacher once a month rather than relying solely on self-teaching, which can lead to poor form.

5. Test Under Pressure. In the information technology industry, there is a type of test called stress testing. Stress testing involves deliberately adding a load—such as a large number of files to process—that is known to surpass what a system, like a group of computers performing tasks (such as a website), can handle. The aim is to identify which part of the system fails first, helping to pinpoint the weakest component (perhaps the process of writing data into physical storage). This method of testing under pressure can also be used to find the weak spot in our learning. When under pressure, only truly internalised skills or knowledge can be applied, so we can see what isn’t yet mastered. A few examples include: playing a piece of music twice as fast can reveal which phrase you’re likely to make mistakes in. Trying to solve a Rubik’s Cube as quickly as possible can show which step’s formula is the least memorised. Teaching someone the new knowledge you’ve just learned can reveal which part of the knowledge isn’t yet secure.



1. Puderbaugh, M., & Emmady, P. D. (2025). Neuroplasticity. StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing.

https://www.ncbi.nlm.nih.gov/books/NBK557811/



2. Marzola, P. (2023). Exploring the role of neuroplasticity in development, aging, and learning. Brain Sciences, 13(12), 1610.

https://doi.org/10.3390/brainsci13121610



3. Greenough, W. T., Black, J. E., & Wallace, C. S. (1987). Experience and brain development. Child Development, 58(3), 539–559.
https://doi.org/10.2307/1130197