Should You Learn Physics Like Newton? Contrasting Expert and Beginner Learning Strategies

After completing the MIT Challenge I got an unusual critique. The complaint was that I shouldn’t have looked at solutions after working through problems. Great thinkers like Newton or Euler, this critic’s reasoning goes, didn’t have access to solutions and they understood the ideas better than anyone. The best process to learn something, he argued, is to prove it for yourself.

This may seem plausible, but it quickly becomes absurd if you dig deeper. Discovery is orders of magnitude more difficult than learning second-hand. Fourier fought for years to prove and win acceptance of a mathematical formula we now teach to college sophomores. Any method which takes a genius decades is likely less efficient than one which takes a clever college student a semester.

Even the scientific pioneers of today are only able to build their accomplishments after spending thousands of hours learning the theories before them. I’ve heard of only a handful of important discoveries that weren’t made by researchers who hadn’t first studied the prior work.

Despite these flaws, the critique does raise one good point: discovery is not the same as learning. Just because you can study and pass an exam, doesn’t mean you can make new scientific discoveries. Newton didn’t uncover laws of motion by getting feedback from problem sets.

Where I differ from my detractor is that I believe this is okay: discovery is different from learning and therefore requires different skills. I would even go further—learning at different levels of expertise requires different skills. Expert learners and beginners face very different challenges, and as such, should use different strategies to learn efficiently.

Mastery and the First 20 Hours

My friend Josh Kaufman recently published a book, The First 20 Hours, which document getting past the initial frustration barrier with a new skill. In contrast to the ten thousand hours supposedly required for expertise, Josh claims you can get comfortable after just twenty.

Josh correctly makes a distinction between the beginner phase of skill acquisition and mastery. Mastery isn’t just repeating the first twenty hours 500 times. Mastering a skill requires different strategies than beginner skill acquisition. When I first started learning French, each new conversation would introduce dozens of new words. Now, I have to push myself to find more nuanced ways of expressing myself instead of relying on the basics I mastered a long time ago.

Learning as a Beginner

The beginner has several advantages over the advanced learner, but also several weaknesses. Each of these weaknesses requires different strategies to overcome, and can halt progress if they’re ignored.

The first advantage for a beginner is simply the glut of learning opportunities. When you know nothing, it’s really easy to find something you haven’t learned yet.

I noticed this play out even over the small gap in skill level between my early courses in the MIT Challenge and my final courses. Vector calculus has hundreds of tutorials, explanations and guides online. Photogrammetry for machine vision has almost none.

The second advantage of being a beginner is that the biggest opportunities for performance improvement haven’t been used up yet. Understanding an obscure design pattern may help a programmer in a narrow instance, but understanding how a variable works is necessary to do almost anything.

The disadvantages of being a beginner include: a lack of context to see where knowledge learned fits into a bigger picture, low meta-knowledge on how to learn successfully in the domain, and negative feedback overwhelming the motivation to practice.

The disadvantages of a beginner are more easily hacked away than those of the expert, however, because they have more to do with finding good teachers, material and learning environments than structural difficulties.

Advanced Learning

In contrast to a beginner, advanced learners struggle under very different constraints. Now the context is known, feedback is generally positive and meta-learning strategies are already established. The difficulties are now structural.

Complacency is the biggest problem with advanced learning. In the beginning, simple participation was enough to ensure growth. Now just showing up merely maintains your skills at the past plateau. It takes deliberate practice to shift to the next level of ability.

Skill growth beyond the basics also becomes confusing for many pursuits. Going from terrible writer to passable is a straightforward process—simply write a lot and get feedback. Going from passable to great can be frustrating since that same recipe often leads to plateaus.

These shifts in obstacles from beginner to advanced necessitate different strategies for handling the different phases of learning.

Learning Efficiently as a Beginner

Beginners should take advantage of the glut of resources and unexploited big wins, but need to be careful about low-context environments, poor meta-skills and excessive negative feedback. Here are some strategies to cope:

1. Use the skill as early as possible.

You can get positive feedback and higher context just by actually using the skill you’re training in a practical environment. Early programmers should try to make something they want to build. Early language learners should try speaking to someone, or reading something they actually care about.

2. Slow down and break down.

When you don’t understand something, don’t just ignore it. Slow it down or break it down in order to understand better. This applies whether what you’re learning are new phrases or physics. Slow it down and break it into the components until you understand all the pieces.

3. Set up regular habits for practice.

If you can establish good habits to practice your skill, gains will likely come easily. In the beginning, the danger of plateauing because of habits is minimal since even a suboptimal approach will often have enormous growth compared to later phases of learning. Habits also help you combat the negative feedback bias of new pursuits.

Learning Efficiently as an Advanced Learner

For advanced learners, the biggest challenge is that the skill isn’t too challenging anymore. Meta-skills and habits tend to calcify, stunting long-term growth. Here’s how to overcome it:

1. Separate work from practice.

In the beginning, all applications of your skill are practice. For advanced learners, however, this is no longer the case. Just because you write code all day doesn’t mean that this counts as “practice” in becoming a great programmer. Chances are 95%+ of that work is not pushing your skill to a new level but repeating what you have already mastered.

There are two main ways you can separate work from practice. The first is simply to do side projects. I believe that this is the most accessible alternative for anyone who isn’t prepared to quit their job or make drastic changes immediately. It is, however, a lot more work and requires strong time-management skills to execute.

The second way you can separate them is by changing your environment. Changing jobs to one which forces you to use different skills will shift you back to the beginner phase of learning once more. The weakness of this alternative is that adaptation to a new environment usually occurs faster than you should change jobs, and that not all new job opportunities will provide the skills you truly need to work on.

2. Set a “level up” goal.

One tactic to force yourself out of complacency is to set a goal to reach an even higher level. This will force you to make those uncomfortable changes to your habits in order to reach the new goal.

As an example of this, I could enroll to take a C2 exam with my French to force myself to push up to that level. Currently, I’m under no pressure to improve dramatically, so my level stays at a comfort point. Publishing a book would have a similar effect on my writing, forcing me to reach a higher standard.

3. Learn complementary skills.

Expertise isn’t a singular proficiency but thousands of microskills and factual connections. Therefore improving expertise can be as simple as shifting to a highly related skill an practicing that.

See how the skill is compartmentalized and look for unexplored clusters of microskills that can be learned. If I wanted to improve my understanding of artificial intelligence, for example, I might decide to improve my linear algebra skills. If I wanted to get better at endurance running, I might train in swimming to practice my breathing.

There are two cautions to this approach, however. One is that learning a skill which is too far from your original mark may not be very efficient. If your goal is to write better, learning programming may be fun but it’s only weakly associated with those skills.

The second worry is in skills which have contradicting rules and conditions. If I decide to learn Italian, I’ll need to do extra work on my French to keep them separate since I may start inject Italian words into my French conversations by accident.

Navigating the Shift from Beginner to Expert

There are a couple signals you can monitor to assess your shift from the spectrum of beginner to expert. By watching for these clues, you can make adjustments to your strategy to avoid falling into to plateaus that come with advanced learning, but not burden yourself with those worries in the frustrating beginning phases.

The easiest signal to look for is simply how frustrating the activity is. If you find application of the skill to be comfortable, chances are you’ve graduated out of the beginner phase and should start considering the strategies for advanced learning.

The other signal is your growth of ability. If you’ve been practicing regularly, but your ability hasn’t grown visibly in the last few months, you should consider some of the strategies listed for advanced learners above. This can apply, even if you’re still finding the skill frustrating.

Should You Become an Expert?

I think it’s a mistake to commit to mastering every skill you start. Core career skills reap the highest rewards from mastery, but many others will still be rewarding at post-beginner levels of ability. Mastery takes 10-1000 times as much time as intermediacy does so you’ll likely only have time to truly master a handful of skills in a lifetime, even if you end up learning hundreds to a beginner level.

If you do decide to master a particular skill, one which your career will likely orbit around, then recognizing the difference in learning obstacles is important. The methods to reach a beginner or intermediate level aren’t the same as those to reach an advanced one. You’ll need to shift your strategy once the methods you used as a beginner stop working.

Read This Next
How to Be More Social
  • Trent Fowler


    I really appreciate this sort of thing. One aspect of beginner learning I’ve really been struggling with lately is successfully learning when multiple disadvantages are present.

    I’ve been making an effort to learn computer science and programming while I live in South Korea, but I don’t have peers, mentors, or any obvious source of feedback. Further, I don’t come from a computer science or math background, so my mind wasn’t in the right shape for computer programming before I started (though analytic philosophy did help somewhat).

    When I don’t understand something (which is often), there isn’t always a way for me to get an answer I can use. The legions of helpful, technical people on the internet sometimes provide answers which are as confusing as the original problem, and because of the low bandwidth of internet communication, refining an explanation is near-impossible. Motivation runs low sometimes because the environment isn’t sending me the sorts of supportive signals that other people don’t even notice, like walking into a room and seeing a friend hacking away on a ruby project.

    So far, here’s what I’m using to cope:

    1) admit that the situation is highly sub-optimal, and don’t beat yourself up when progress is slow.

    2) gather a list of toy problems which are somewhat challenging and can be solved again at a later time. I know, I know, most hackers would probably recoil at the idea of resolving a problem. But at my current low level, even writing code I’ve written before can be non-trivial, and stretches skills like code reading and debugging.

    3) When I really, really don’t understand something, all I can do is try to figure out, in the clearest terms possible, what I don’t understand. Sometimes I’ll annotate my code with phrases like “I have no idea why doing x causes error y”.

    4) If there seems to be no way for me to overcome a given obstacle, move on. Sometimes I work a little further in the book and then circle back (vertical), other times I make a lateral move into a related topic (horizontal). This is to be avoided if possible. It also suggests you might be able to order your work in a hierarchy by how much progress will require new insights. Learning how programming works requires me to have new insights at almost every step of the progress, which is both exhausting and makes my insight-having ability the primary bottleneck. Learning the basics of computer hardware is more straightforward, and is more like carefully studying a map. Studying then becomes a matter of flowing downward in the hierarchy. There is a more general principle at work here which I’m working on nailing down more formally.

  • Carl Hart

    Hi Scott,

    Which other readings (articles, books, …) – in addition to the Kaufman book – would you recommend on the topic of learning strategies?

    I have found the “Talent is Overrated” book to be quite useful for some general ideas, but I am trying to find more specific and thorough literature on this topic (more like your blogpost and Ericsson’s article on deliberate practice).

    Fantastic article, looking forward to more on this!

  • Ricardo Araujo

    Excellent article Scott.

    I also was a bit surprised with the critiques regarding you checking the solutions. I think in a situation like yours, where time is constrained, a way to quickly gather immediate feedback on your difficulties and issues in a specific subject or concept seems extremely important and obvious.

    I also feel this difference between beginner and advanced learning while I’m pursuing my goal of becoming “fluent” in psychology. Since fortunately I have several sources for a specific subject, I need to make sure I draw the line between learning from ground zero and increasing my skill level in something I’m already comfortable with. I can’t afford being redundant on my learning, so to distinguish the phase of “let me learn about this” and “I’m extremely at ease in this specific content, but I wish to go further” I make use of creativity:

    I believe that if you’re creative, then you will naturally have an easier time learning. This is because there’s positive results from being able to make new connections and thinking outside the box, but this is even more important in advanced learning. At this point, you are pretty much on the same level as many other people: you understand the concept and you know how it works in a real scenario, and you can even adapt to different settings. But…how do you advance to the next level? By being creative. This means taking yourself out of your comfort area, and by tackling new perspectives that often make you rethink the whole concept, or include completely new and complex ways of thinking that may function as subtle but important connections which allow you to get a more precise vision of things.

    I like how the steps you mentioned can actually be viewed as creative ways to advanced to the “next level”. I think creativity is (one of the) way(s) to go when you’re trying to go from “good” to “great” 🙂

  • John

    Hi Trent,

    I would STRONGLY RECOMMEND joining/starting a group of programmers. This will solve most the problems you are facing.

    Good luck 🙂

  • John


    I had a HUGE insight about progressing rapidly once we reach expert level.

    In my experience and study I have found that copying what the world-class performers do is the best approach. Some people call this stealing the best practices.

    In my humble opinion no one embodies rapid progress than the best distance runners. They continue to improve even after they have won medals and championships.

    Here are 3 observations I have made about Elite distance runners (Olympic medalists)

    1. Environment

    Simplicity/0 distractions

    These people live very simple lifestyles that are completely dedicated to running faster. For example, the best Kenyan runners live in rural towns in these old houses were they train.


    These runners live with and spend time with other high-caliber runners and coaches that are obsessed with running and training. So the social environment makes it easier for them to train hard.

    3.Surround yourself with the Best

    The best runners train in groups of 5-20 runners. These groups are HIGHLY SELECTIVE. For example, one groups from the US has the gold and silver medalists in the 10k. The same happens in Kenya where the world-record holder trains next to the Olympic gold medalist.

    4.World-class coaching

    The best 20 distance runners are probably coached by ONLY 3 top coaches. These coaches have 20-30+ years of rigorous study of running training of the best runners and study of biochemistry, physiology, and sports science. These coaches know EXPLICITLY how to train a runners from novice to a Olympian.

    5. Start young
    Most world-class runners start training seriously between 12-14 years old and they don’t peak until 28-30 years old.

    In my humble opinion distance runners and coaches have created a phenomenal deliberate practice platform. Unfortunately, like you know most skills are not as simple as running a 4 minutes mile. Most modern day skills require hundreds of mini-skills so figuring out how to train each to a high level is the challenge.

  • Pt

    Hi Scott,

    One of your best post. I was really thinking about this ‘discovery vs. learning’, this post hits the nail at this important question that has been bothering me a lot. One of the major reason my school grades suffer was because I was trying to do derive originally what I read in reference books, which caused lot of frustration, and were inefficient.
    While student learning mindlessly whizzed past me to get excellent grades!!

    I m assuming many kids in schooling system suffer a lot of this kind of confusion, assume discovery is learning and vice versa, thus there grades suffer.

  • Pt

    A typo #their grades suffer.

    No option to edit the comment for typos. 😀

  • J Tran

    Those are some really good tips, John! Unfortunately for most of us, I don’t think we can help much the ‘start young’ thing. I’m nearly 19 now and am just started getting learning the technical subjects (math, physics, chem) required for an engineering major. Imagine if I started from the age of 4 (like Mozart) and kept on practicing and improving up to the present day? I would be, as Cal Newport would say, remarkable.

    It takes years of practice to be an expert in anything we want to do well, so for a beginner like me I think coverage, and simply showing up every day would be the most important thing in this stage.

  • SJ Scott


    A great writeup explaining the difference between Kaufman’s 20 hours to learn a skill and the 10,000 hours of mastery that Gladwell talks about in Outliers.

    If you break it down right you can learn skills in a basic level very rapidly but it certainly does take quite a bit of time and effort to learn something to a master level. Yet, that to be made more efficient by learning and practicing the steps the, “right” way.

  • layman

    thanks very much for the article,am a guitar player myself and this i adapted it to my context and it was very useful.had reached a state were i just play to maintain the state but not necessarily growing…..thanks

  • Thor

    Hi Pt:
    I, like you tended to derive things more than memorize them — mainly because I can’t seem to remember important things, but if I /do/ them, then I can /re-do/ them quickly.

    It amused the pre-calc teacher when I derived the quadratic formula on the first page of every test, and it also amused the physics prof when I geometrically determined Snell’s law… I wouldn’t consider either of those excercises a great use of time, but by working around my memory block, I’ve found that it makes me much better ad synthesizing known bits to make something I can test to see if it works in real life.

    So don’t get too dismayed by others’ apparent progress… but unfortunately, there is little reward for creatively deriving things in school — at least at the lower levels; they are focused on the “early learner” instead of the people who desire mastery.

    Later, the focus shifts — if you have to figure out Snell’s law, you start to use those analytic muscles, and those who were effective at memorizing things without being able to synthesize/derive have a hard time transitioning to the new skill.

  • Hein Hundal

    Very nice post Scott. I also liked John’s comment (June 22, 2013 at 4:45 pm).

    For my vector calculus students, I wonder if doing homework counts as “Use the skill as early as possible.” I kind of want to make them apply the concepts to their own major. I will think more about this.


  • Bradley


    Great article, the information that you have provided has helped me tremendously since I started implementing it. Trent had some fantastic things to add as well. While on the topic of programming, I began jotting down ideas for a concept of an Iphone game. Since it has been about 8 years since I used any of my programming knowledge it’s pretty much all gone and as the url of my blog implies I’m back at square one. Pretty confident that my idea will go viral, been sharing my thoughts with people that I know I can trust and they think I’m onto something. I have a 3 page game description ready to go since last night and I’m still trying to add to it. Since I haven’t really played many Iphone games myself I don’t really know how complicated or advanced I will be able to make it, however I’m certainly not lacking on aspects of game play, releasable content, and replayability. Hope to hear from you soon.

    Bradley Bruce Getter “Live Your Own Life”

  • Bill Polm

    Great. Insightful as usual. And helpful. I constantly work at improving my skills at writing and watercolor and acrylic painting. Since I am way past beginner level (after many years at both), these ideas should prove very helpful.

    Incidentally, I do not fully agree with the Gladwell 10,000 required hours. Because I think the actual time necessary depends on how efficient the learner is, how focused, and smart when it comes to targetted practice.

  • Patrick Andrews (@PatrickRAndr

    “Just because you can study and pass an exam, doesn’t mean you can make new scientific discoveries. ” I couldn’t agree more. The higher education system lazily assumes these are equivalent when it selects students for research degrees based on their undergraduate results (which usually require almost no originality of thought).

  • John


    I would also add that the time it takes to reach expert level depends on whether you have a world-class coach or not.

  • Joe

    A quote from the book “thinking as science”

    “A man may have discovered some portion of truth or wisdom after spending a great deal of time and trouble in thinking it over for himself, adding thought to thought; and it may sometimes happen that he could have found it all ready to hand in a book and spared himself the trouble. But even so it is a hundred times more valuable, for he has acquired it by thinking it out for himself. For it is only when we gain our knowledge in this way that it enters as an integral part, a living member, into the whole system of our thought; that it stands in complete and firm relation with what we know, that it is understood with all that underlies it and follows from it, that it wears the color, the precise shade, the distinguishing mark, of our own way of thinking, that it comes exactly at the right time, just as we felt the need for it; that it stands fast and cannot be forgotten.”

    Excerpt From: Hazlitt, Henry. “Thinking as a Science.”

  • Adhithya

    Hi Scott,

    An enlightening article which helps one recognize when he is stuck in a plateau. Also, the emphasis on the fact that the creator’s mind is different from the problem solver’s was helpful.

  • aegrrg

    Ok, so 10 thousand hour rule is pretty dumb. Think about it. So, it will take 10k hours to be expert boxer, and 10k hours to be expert jiu-jitsu? How about the fact that jiu-jitsu contains like 10-20 times more moves? Rationally, it cant be 10k hours for all fields.

  • Eden

    I remember watching a Bill Gates interview on Youtube of him saying the same thing you are saying. The interviewer tried to draw parallels between Gates’s early life and the book called the Outliers. Gates rejected a lot of what he was saying.

  • Eden

    I remember watching a Bill Gates interview on Youtube of him saying the same thing you are saying. The interviewer tried to draw parallels between Gates’s early life and the book called the Outliers. Gates rejected a lot of what he was saying.