The drawbacks are obvious. For many classroom assignments, AI is now good enough that it can simply do your homework for you. But having AI do your homework is self-defeating. You can’t learn without doing the work.

That said, the opportunities AI may provide are numerous. Alongside its drawbacks, AI also brings the possibility for cheap and patient tutoring, on-demand explanations, extra practice problems and assistance in research.


To paraphrase Dickens, when it comes to learning, it really is both the best of times and the worst of times.

This brings me to a recent essay written by the always insightful Carl Hendrick. In it, he reviews recent findings on AI’s impact on learning, drawing an analogy between the introduction of autopilot and the degrading of pilot skill:

The thing I find really fascinating about this story and why I think it has so much import for learning and AI is that his core diagnosis was that the autopilot was in fact, not the problem. The truth is that automated function in planes saves far more lives than it endangers and I think the same will be true of self-driving cars. (I have no interest in the hackneyed claim that the machines are bad and the humans good. The machines are, on the whole, magnificent I think). The thing I think is so interesting is that when faced with an emergency, the pilots reached instinctively for more automation not less, when the right thing to do was often to drop down a level and fly. So the danger is not automation itself but over-dependency on automation and this is quite obviously what happens when we give novice learners unfettered access to chatbots.

Hendrick doesn’t blame the students for this alarming trend; he blames educational institutions themselves. Schools have been vacillating between wanting to adopt a pro-AI stance, encouraging students to use AI in their work provided they properly “cite” the results, and embracing the language of academic honesty in a vain attempt to maintain the status quo.

Obviously, this is untenable. Education will have to change because of AI.

The answer isn’t to adopt some vague AI-forward solution which simply accepts AI output as a legitimate substitute for students doing the work themselves. Instead, it will be understanding that AI, like calculators and Wikipedia, can be both an incredible tool and a debilitating crutch.

What is the Right “Amount” of AI Use for Optimal Learning?

It’s pretty clear that the solution of having AI do your homework for you will not to result in much actual learning. Enamored as I am by the possibilities unlocked by the arrival of coding agents, it’s obvious to me that vibe coding actively impairs your ability to learn to write code by hand.

Whether this trade-off is valid in the realm of productive work is a more difficult question. I suspect the answer of whether vibe coding enhances or undermines productivity depends a lot on the stage of knowledge acquisition. Beginners who haven’t built up a mental model of what code is being written are losing out on the opportunity to practice foundational skills. Experts, who can steer the process, may end up benefitting from the saved effort.

But nobody can deny that if your goal was to learn to write code by hand, getting Claude to do your work for you will result in less learning.

What’s more interesting to me is if there’s an optimal amount of, or timing for AI help. Meaning, even if too much AI is obviously bad, might learning be hindered if there is too little?

If you had asked me this question in the lead up to my 2019 book, Ultralearning, I would have said no. I was a strong advocate for doing hard problem solving and the associated mental effort it requires as a way to learn difficult skills.

However, after the research that led to my 2024 book, Get Better at Anything, I’m no longer convinced. Practicing problems is certainly an essential part of learning any skill. But the cognitive load involved can easily become too high. Getting help, whether it’s in the form of a worked example, a completion problem or just a hint to get you unstuck, can be faster for learning than floundering.

Some Tentative Guidelines

In education research, conversations about different pedagogical approaches often span decades. On that timescale, AI has only been here for the blink of an eye. Therefore, I expect it will take some time before a mature consensus on AI best-practice emerges in the research literature.

However, I’d like to share my best guess of what some of those AI guidelines might look like, absent stronger evidence-based work:

1. You can’t learn from work you don’t do yourself.

This is obvious, but it needs to be restated. While AI-assistance may help or harm learning, it’s clear that anything you get a chatbot to do in toto isn’t going to allow you to learn the underlying skill.

2. Don’t use AI on your first attempt.

For many tasks, there is only one problem. In such cases, I think it’s almost always better to attempt a solution on your own before looking up the answer. This seems particularly true in creative domains where each problem is unique. Asking an AI for help with the start of an essay, for instance, will invariably shape what direction you end up following, robbing you of the crucial experience of building your own judgement around a topic and selecting your own path to research and argue.

3. Don’t use AI on your last attempt.

What if you can’t solve the problem despite your best efforts? Assuming this is a repeated problem, the research is fairly clear that seeing a worked example or explanation or getting instruction is going to be better than continuing to struggle. Thus, getting help with AI after a failure seems prudent. But, if you’re going to benefit from learning, you need to then attempt another problem of the same type without AI help.

4. Do use AI to suggest alternative ideas and resources.

While starting with AI is likely to cause you to skip over the necessary hard thinking that results in learning, I do think AI can be useful in suggesting alternatives you might have overlooked. Here, the superhuman breadth of knowledge possessed by AI can be an enormous asset—surfacing methods, papers, books and ideas you may have missed in your first pass.

5. Do use AI to fix confusion and mistakes.

Similarly, after you’ve completed your first attempt at solving a problem, I would typically encourage looking at the expert solution for feedback. AI can not only generate expert-level solutions in many domains, but it can help point out the source of confusion and mistakes.

See-Do-Feedback (+AI)

In Get Better at Anything, I argued that much of effective learning looks like a practice loop: Seeing involves getting instruction, viewing a worked example or otherwise getting knowledge into your head to guide the search for the right answer. Doing involves practicing on your own. Feedback involves getting some relevant signal from the environment about how effective your practice was.

AI has the potential to assist with both the first and last element of that loop. What it cannot replace is the middle step—doing the work for yourself.

The post We Need Guidelines for Learning and AI appeared first on Scott H Young.


In contrast, creative skills have an aura of unlearnability to them. Either you have creative talent, or you don’t.

 But this isn’t true. Creative skills are as learnable as any other skill you’d like to master. In this essay, I’d like to show how.

The Two Components of Creative Skills

Any creative skill you’d like to learn, whether it’s painting, writing, music, design or something else, has essentially two components:

1. Technical skills.
2. Ideational skills.

Technical skills are those aspects of creativity which don’t actually require, well, creativity. Take my portrait drawing challenge as an example. Although I’m proud of what I was able to accomplish with one month’s work, there was zero creativity involved. The difficulty in this challenge was purely technical.

Technical skill is the aspect of creative expertise most impressive to non-artists. Being able to play a very difficult guitar solo flawlessly or render a compelling likeness of an image with paint or pencil is hard to do. But, strictly speaking, there isn’t much creativity required.

Ideational skills, in contrast, are more often what we refer to as creativity itself. What we think of as “creativity” is the ability to do something that is both valuable and original. Among practitioners, there’s usually a greater appreciation for these ideational skills, as technical skills often lose their aura of mysticism when you’ve spent enough time practicing them.

Great artists are usually good at both the technical and ideational components of creative skills, although some artists’s talents clearly lie more heavily on one side or the other. John Singer Sargent’s paintings are technically virtuosic—especially when you appreciate that most of his work was done alla prima, which means he did each painting largely in one go, rather than meticulously building up layers as was more common during the Renaissance:

In contrast, Vincent van Gogh is the best paradigm of an ideational genius with mediocre technical skills. He couldn’t keep up with the other artists in his Parisian atelier and eventually left. Even when he was in the mature phase of his painting, he frequently repeated the same subject multiple times because he struggled to get the details right:

Still, as much as I’m personally a fan of Sargent’s, it’s clear that Van Gogh has had far more impact as a painter, in no small part due to his original and compelling style.

With this distinction in mind, let’s look at how how to get good at each aspect of creative skills.

Achieving Technical Proficiency

Technical proficiency is the easier of the two to master. It’s hardly trivial, but technical skills are amenable to hard work and practice in a way that ideational genius often is not.

That being said, many people struggle to acquire technical proficiency in creative skills because they go about it the wrong way. All technical skills break down into three essential parts:

1. Methods.
2. Concepts.
3. Practice.

The first part is that to learn a creative skill you need to start with a method. This can be difficult because many artists have so overlearned their skills that they no longer appear to follow any particular method.

This was an aspect of portrait drawing that initially confounded me. Seeing street artists draw a good likeness in minutes or technically excellent artists render a likeness with a few flicks of their brush suggests that perhaps there isn’t a method at all, that success in art is simply an innate talent or intuition.

But this misunderstands the situation. Across many domains, researchers have found that nearly all expertise follows a similar pattern. Skills start out learned explicitly but, after more and more practice, parts of that explicit learning become automated—they become more like background intuition than a method you can easily articulate.¹

When I started my portrait drawing challenge, for instance, I didn’t turn up very good methods. So I assumed that simply putting in a lot of bulk practice with feedback would be the key to improvement. I was right that high volumes of practice are essential, but I was wrong in assuming that methods were irrelevant.


The fact is, most portrait artists use a variety of methods to gain accuracy. Some of these are perceptual, using various means to overcome optical illusions or our natural sensory limitations to correctly assess the relative sizes, angles and tonal values in an image. Others are conceptual, building up from knowledge of lighting, geometry and anatomy to infer shapes that aren’t immediately visible.

The starting point for learning any technical skill is first to learn what the methods used are, and then practice them until they become second nature.

The next part of technical proficiency is less about rote practice and more about creating a mental model of the possibilities afforded by the medium and the constraints attached. I love this little clip discussing how Steven Spielberg achieved a scene in the movie Hook where one of the Lost Boys appears to bring down his sword quickly just a few millimeters from Robin Williams’s head.

In a pre-CGI world, Spielberg achieved this effect by filming the entire scene backwards, including the dialog of the actors, so that what appears to be a near-fatal sword strike is actually the sword lifting up quickly, in reverse.

The final key to technical virtuosity is an enormous amount of practice. Both precision and fluency matter here.

Precision, in the sense of adhering to accuracy afforded by the method you’ve adopted, is necessary because it is how you learn the underlying skill. Thus, whether you’re learning a method for drawing a portrait or playing a guitar solo, your practice will be ineffective if you permit yourself to make a ton of mistakes.

Fluency matters because motor skills, like the strumming of a guitar or the flicking of a paintbrush, will be completely different if they’re executed with different rhythms or movements. 

I used to struggle with the tension between precision and fluency. In visual arts, there’s often a contrast between “tighter” more controlled styles, and “looser” more expressive ones. Sometimes it feels like I need to slow down and get more precise to improve. Other times it feels like I’d do better to just loosen up.

 But I think this distinction is a bit of an illusion. Looser “expressive” movements are simply a different kind of motor skill than tightly controlled ones. Both of these movements need practice to master, but that doesn’t negate the need for broader technical proficiency. Like many skills in life, the solution is to practice both the movements and the methods that support them.

How to Have Better Ideas

The other component of creative skills is having good, original ideas. Even this isn’t entirely removed from technical proficiency. If you want to understand the opportunities for expressing good, original ideas within a medium, you first have to understand what the medium affords and what it constrains. Creativity, then, is at least somewhat reliant on proficiency.

But it’s a mistake to simply collapse technical virtuosity down to creativity, because creativity relies on more than just being really good at the technical aspects of the craft.

Having really good, original ideas is a mixture of two different components:

1. Knowledge.
2. Judgement.

Knowledge is the most important part. There’s good evidence that creativity itself is a largely random process. What separates more creative artists from less creative ones is less the ability to generate lots of random ideas (all of us have that ability), but that the more creative artists have better quality filters and thus can judge more reliably which ideas are actually good.

Vincent van Gogh is a case in point. Although he struggled technically, he was incredibly attuned to the artistic landscape of his time. His brother was an art dealer, and Van Gogh had a lot of contact with the emerging trends in the art world. So, even though he is often billed as an “outsider” compared to his Salon-trained contemporaries, he had insider knowledge.

Similarly, of my friends who are successful serial authors, I’m more impressed by their ability to quickly ascertain which ideas are likely to be popular and engaging than their writing skill.

Generally speaking, it’s a myth that successful creatives are outsiders to their fields.² While the most creative works may come from the edges of a field, from those who are less invested in the dominant styles or paradigms, genuinely creative outsider art is practically non-existent.

This suggests that immersion in the social world that surrounds a creative endeavor is essential for truly successful creativity. That if you’re detached from the world that forms the fundamental judgements on what counts as good, original work, you’ll inevitably fail to produce something that qualifies.

The second aspect of creative idea generation is less about having creative ideas, and more about having the disposition to actually create them. Creativity, particularly original work, requires effort and risk-taking. You have to attempt something that may not work and, particularly in the case of many high-art pursuits, may not pay off financially either.

A lot of valuable creativity is suffocated not because the artist lacks original ideas, but because the pressure to conform to a particular style or genre is itself strong. As social psychologist Mihaly Csikszentmihalyi put it:

The common belief is that if creativity is rare, it is because of supply-side limitations; in other words, because there are few geniuses. The truth seems to be that the limits to creativity lie on the demand-side. If there is too little creativity, it is because both individually and collectively we cannot change our cognitive structures rapidly enough to recognize and adopt new ideas.³

Creative ideas succeed when they’re in the Goldilocks zone of originality—new enough that they’ll attract interest, but still comprehensible to the world of artistic taste.

Putting Them Together

Getting good at creative skills, of course, is easier said than done. To become an excellent painter, musician, writer or woodworker requires a ton of practice, learning, immersion and experimentation.

But even though the road may be long, it is much easier to persist in the endeavor if you have a map.

The post How to Learn Creative Skills appeared first on Scott H Young.

This raises a question: how do you find time to learn the things you want to?

In this essay, I’m sharing my approaches for finding time to devote to learning efforts. I’ll look at three different situations, as the strategies I’ve found successful differ for each:

1. The full-time learner.
2. The serious, part-time learner.
3. The person who can only devote spare moments, inconsistently.

1. Creating a Full-Time Studying Schedule

This advice applies to students who are full-time or nearly so. Typically, this includes students enrolled in school, studying full-time for an exam or otherwise tackling an intensive training program, including people tackling intensive self-directed projects akin to some of my challenges.

For these people, I strongly recommend making an explicit studying timetable. This is a schedule that blocks off the hours you’re going to devote to studying to maximize your results, given the other constraints in your life.

In terms of making the schedule (not actually sticking to it) this is probably the simplest. If you’re studying full-time, you can block in a roughly full-time work schedule, but allocate your hours to studying.

That said, there are some common pitfalls:

1. Studying too much.

Few people can reliably concentrate for more than 5 to 8 hours per day, even with high motivation. Studying schedules with 12+ hours are usually fantastical—the most likely result is that you’ll end up procrastinating and wasting time, not focusing for the entire period.

2. Forgetting to include sleep, exercise, socialization and stress-relief.

Look, I get it. When you’re studying full-time for something extremely important, it can be tempting to cut some of the other aspects of your life down in order to maximize the amount you learn. And certainly, if you have a competitive exam that you’re not sure you’ll pass, you want to be focused.

But if the studying period is longer than a week or two, certain kinds of cuts are actually counterproductive. Sleep is probably the biggest one. Sleep plays a major role in memory consolidation, so a student who only gets six hours of sleep a night to squeeze in an extra two hours of study is actually undermining their performance.

Exercise is another cognitive booster, so including a little bit—15 minutes of higher-intensity activity is sufficient if you’re super busy—is a no-brainer. I would argue some amount of social contact and stress-relief are also important, particularly when your studies will span multi-month timescales.

3. Not including enough short breaks.

The amount of breaks you need is somewhat personal, and somewhat dependent on the dominant studying activity. When I was doing intensive MIT coursework, I found that I needed fewer breaks on lecture days than on reading days, and problem-solving days required the most.

Inserting a number of 10 to 15 minute breaks is wise. Pay attention to when you start feeling fatigued or when your focus is degrading to figure out how many breaks you need for a given activity. The best kind of breaks are short walks or, if you’re a good napper, really short naps—not social media or entertainment. Save those for after.

4. Not scheduling larger chunks of time off in the evening or weekend.

This is a matter of personal preference, obviously. But I’ve found that ensuring at least some free evening time and one day off per week tends to be optimal for me. Studying more than this is actually a net-negative for my productivity.

This constraint is somewhat weaker if you have many other obligations that keep you from studying full-time. For instance, if you work another job or have other tasks and chores that keep you occupied much of the day, evenings or weekends may be your only available time to study.

_ _ _

Ultimately, studying full-time can be daunting, especially if the project isn’t one taken on out of personal interest but because of a stressful exam needed for professional advancement. However, if you approach it like a well-structured endurance event, where pacing and recovery are optimized, you’ll put yourself on a much better path for sticking with it.

2. Creating a Serious, Part-Time Studying Schedule

Studying full-time is tough, but studying part-time is arguably harder. Those who have to learn something difficult while also managing the responsibilities of a job and/or childcare face higher levels of strain with fewer hours to make progress.

However, there are good solutions here too, they simply require more planning and preparation.

The first step is to map out your mandatory commitments—times when you won’t be able to learn, such as time you’re in the office, caregiving responsibilities, etc. Remove these from your study schedule first.

Next, figure out when the largest chunks of unscheduled time exist in your calendar. For most people this will be mornings, evenings and weekends.

Then, choose a consistent studying routine that you feel would be the easiest to stick to, for the given time investment. For instance, you might decide to wake up a little earlier and study each morning from 6:00 am to 9:00 am before work. Or you might decide to stay after work and go until 8:00 pm. You might decide to study later in the evening before bed or dedicate one (or two) weekend days to the job.

All irregular time slots have trade-offs. From having attempted serious part-time learning projects in most of these slots, I can attest to some of the difficulties:

• Mornings have fewer interruptions, but they risk cutting in to sleep, particularly if you have a later chronotype.
• After work maintains continuity with the workday, but you tend to be more fatigued.
• Late evening can have more energy for many, but it often has a harder transition to start given you may have been taking time off for dinner.
• Weekends are perhaps the most free, but devoting them to studying can interfere with other social and family plans.

Regardless of how you schedule it, making a super consistent routine is crucial. So, if you’re allocating 15 hours per week on top of a full-time job by studying for 3 hours every morning before work, always study at those times. The repetition will make it more tolerable than if you try to study on an ad hoc timetable.

The same suggestions for exercise, sleep, stress relief and breaks for leisure also apply here, but it can be more complicated. Since you’re not studying full-time, ensuring sufficient hours is the priority rather than maximizing the efficiency of each hour.

As a result, while I tend to find studying all day and all evening to be counterproductive for a full-time student (because of the average productivity drop spread over 12+ hour days), if you work full-time and only have 3 hours to study, reliably carving out those three hours will likely do more than trying to compress those hours into 90 minutes.

Ultimately, however, all of this is constrained by what you have the capacity to sustain. If you have a full-time job, small kids and other life responsibilities, a 15-hour per week studying goal may be overly ambitious. In this case, starting with a smaller amount that you can reliably do, even if you think it’s insufficient for your goal, will be better than aiming high and failing at it.

3. Recovering spare moments of time.

So far, I’ve looked at relatively serious projects. Whether it’s full-time or part-time, you’re serious about your learning goal and your drive is high. Importantly, you’re willing to prioritize your learning efforts over other things in your life by giving them dedicated time in your schedule.

Sometimes, however, such dedication isn’t possible. You may be really keen on learning something new, but you recognize that it isn’t going to take precedence over many other things that you have to do in your life. If so, how do you maximize your available time for this kind of learning?

Generally speaking, I find that dedicating time works better than trying to learn things ad hoc. This is true even if you have a lot of intrinsic motivation to learn. This is a big reason most people learn better from in-person classes than self-paced courses—the medium of instruction isn’t vastly different, but for most people, signing up for an in-person class forces them to attend at a particular time and place—effectively upgrading them to the prior strategy.

What shifts between the serious, part-time strategy and the spare-moment strategy isn’t really the amount of time dedicated (both are part-time), it’s that the latter requires you to make a decision to do some learning activity at opportune moments throughout the day.



Unfortunately, learning is effortful, and so doing some learning activity, whether that’s reading a book, watching lecture videos or practicing a skill, will usually be an uphill climb compared to other options like playing video games or watching Netflix. In other words, even if you could *in theory* capture 5 to 10 hours per week for a learning project you’re strongly motivated to pursue, most of the time you won’t. You’ll do something easier instead.

It’s for this reason that I almost always prefer the previous strategy of dedicating time. It’s much easier to commit to something that’s higher-effort when it’s habitual than if you have to choose to engage spontaneously over and over again.

The key to success with this strategy, then, tends to be finding ways to reduce the overall effort cost of engaging in learning until you’re able to pass the effort threshold regularly. The two main ways to do this are:

1. Reduce learning friction. Have your materials ready-to-go. Clearly define your next step and leave a reminder for yourself you will see immediately. Make the project always available and easy-access.
2. Reduce alternative temptations. Get rid of your TV. Delete social media from your phone. Block distracting websites or even your internet access if possible.

A third option, which is quite common, but which I don’t particularly like, is to water down the learning activity so that the learning itself becomes more appealing. This is the strategy used by a lot of educational games like Duolingo and others. While there can be some value in these apps, they often produce disappointing results for the time invested, because you’re effectively cutting yourself off from the more effective, higher-effort strategies.

Combining the Approaches

A fourth option is a hybrid between a fixed studying schedule and the spontaneous engagement just discussed. I’ve used this for projects that include both learning activities that require deep focus and ones that can be engaged in quickly with low-effort.

For instance, when learning languages, I do scheduled conversation practice and textbook study to master elements of grammar. But I also do lower-effort activities, like rewatching old television shows dubbed into the language I’m learning or working through my flashcards, at random intervals during the day.

This hybrid approach can be successful if what you’re learning genuinely involves a mixture of high- and low-effort strategies. It can help you squeeze more time out of your schedule than might be possible from dedicating chunks of time alone. The key is knowing which activities are high-effort and require dedicated chunks, and not getting frustrated when you can’t squeeze them in random ten-minute intervals throughout the day.

_ _ _

Ultimately, the studying schedule you end up sticking with will never be exactly the same as the one you plan for yourself. Real life involves unexpected interruptions, fatigue, motivation and other factors that prevent perfect adherence to a plan.

But I’ve never regretted seriously thinking out a studying plan. For learning projects that are difficult grinds, the schedule and pacing make a heavy burden feel like one you can carry. For learning projects that are fun and inspiring, the plan allows you to stop daydreaming and begin taking action. Either way, you’re far more likely to learn something if you start with a solid plan.

The post How to Find Time for Learning appeared first on Scott H Young.

1. There are behaviors, such as exercise, that clearly pass a cost-benefit test in making our lives better, and most of us would benefit from doing them more. Using some of the published data, I showed this is the case, even if you restrict the benefits to just a single measure (trading time spent exercising now to live longer later), and even if you dislike exercising.
2. Our inconsistency with these behaviors is not a problem of willpower. These are opportunities, not problems. There currently exist low-hanging fruit that could make our lives better without much cost. The difficulty isn’t one of character (which tends to encourage guilt and rationalization), but one of engineering. We need to design ways to make maintaining these behaviors consistent and stable.

Today, I want to go into some of the research on how the design of good behavioral systems actually works. But first, I have to explain a bit of what I initially got wrong.

How I Was Wrong About Habits

Habits were one of the first topics I tackled two decades ago when I first started writing. I was young and enthusiastic—and not at all versed in reading scientific papers. So most of my theories came from self-help and self-experimentation.

In particular, I subscribed to some common ideas about habits:

1. Habits are mostly about repetition and consistency.
2. A new habit can be formed in thirty days.
3. Once formed, a habit is automatic and effortless.
4. To build a habit, it’s important to maintain an uninterrupted streak. Don’t break the chain!

I now think all four of these ideas are wrong. I’m still a big believer in habits, but I think the actual data on how people sustain behavioral change is more interesting than my original ideas permitted.

Debunking My Past Beliefs

Let’s start with the last claim: that missing a single day can unravel a habit. I’m not sure where I first got this idea, but it might owe to some of the stereotype about addiction recovery—where a single mistake can throw you “off the wagon” and undermine your resolve.

Fortunately, research doesn’t back this up. In one study, they found that missing a day was associated with a little less automaticity—but not devastating relapse. This is about what you’d expect if you have the mental model that each repetition adds a little to habit strength; missing a day fails to strengthen the habit, but it isn’t fatal to the overall goal.

The belief that habits require perfect consistency might be a self-fulfilling prophecy. In other words, if you believe that you’ve completely failed when you miss a single day, that belief might cause your habit to break down rather than any underlying failure or flaw of your habits themselves.

Which leads to the third claim: that habits, once formed, operate without any continued effort. The authors of the same study also refute this idea. Even looking at the very simple behaviors they studied, researchers found that the automaticity of most behaviors plateaued at some point.

In other words, if even simple behaviors never become fully automatic, this is doubly true of more complex real-world behaviors like exercising, eating healthy or maintaining good relationships. In other words, whatever people do to sustain their behavior long-term, the reality is never so simple as a mindless habit.

This brings me to the second claim: that thirty days is all that’s required to make a habit. Even using the more modest standard of having a behavior reach it’s plateau of automaticity rather than becoming completely effortless, most behaviors required more than a month to feel maximally automatic. In this paper, the average was 66 days, and some habits required almost a year.

To be clear, I don’t think there’s anything wrong with committing to less than a year per behavior in order to make a habit stick. But we have to realize that, just as concrete can set within hours yet take months to reach full strength, many habits we set up in an initial burst of focus remain vulnerable to disruption for months afterward.

Finally, even the concept of habits as being all-important for sustaining simple, practical behaviors like exercise is something I’ve grown to question.

 While it definitely is the case that the automaticity that comes from sticking to an exercise routine helps make it feel easier, sustaining a behavior for years and decades is not primarily a matter of habits. Instead, to make a behavior stick over the long-term, we really need both new skills and changes to our self-concept.

What Really Sustains Good Behaviors

The early picture I had of habits was both overly optimistic and incomplete.

It was overly optimistic, because the real facts about habits are somewhat less compelling. Actual habits for real-world behaviors both take longer to make and rarely become completely automatic.

But in focusing only on habits, I was ignoring two of the real, durable changes that make sticking to behavior changes much easier in the long run.


The first thing we need to do is build skills. Skillful living means you haven’t just learned one solution to a problem, but many possible solutions to many possible variations on the problem.

Take healthy eating as an example. While this can be reduced to a mindless habit—following some strict dietary rules every day—it usually doesn’t become completely effortless. However, as you get better at cooking, grocery shopping and meal planning, eating healthier does get easier, even if it never follows a simple stimulus-response pattern of an automatic habit.

The second thing we must do is change our self-concept. Because so much of our conversation about important behaviors is tangled up with ideas of willpower, virtue, guilt and shame, we often downplay the importance of various behaviors. However, once we start making many of those changes, we no longer have to battle against our own cognitive dissonance, and we very often enthusiastically embrace behaviors we previously rejected.

This helps to explain the tendency of people to preach excessively about changes they’ve recently made. While this may annoy others (I’m sorry!), it stems from the pivot in your beliefs that happens once your behavior matches your knowledge.

Once you see yourself as the kind of person who does X, this self-concept forms much of the motivation to smooth over the residual effort remaining.

A Three-Step Plan to Living Better

Putting all of this together, we have a basic process for making improvements to the foundational aspects of your life:

1. Identify the low-hanging fruit that yield important improvements in your life.

This step is non-trivial. Yes, there are “obvious” behaviors like exercise and eating healthy. But there’s a ton of dubious practices that don’t meet this threshold. There’s also many popular fads that actively contradict decades of careful research about which sorts of actions are best.

Self-experimentation can help. But, since fully testing a behavior can take 10x-1000x longer than simply reading about it, we should be picky about what we choose to work on and start with things that have the most ample base of evidence.

2. Commit to an initial “focused” period of habit-building and experimentation.

The purpose of this phase isn’t to mechanically create a perfectly consistent habit. Rather it’s to use the period of attention to troubleshoot and smooth over the many small obstacles that prevent you from trying it out.

Consistency here helps. But even more important is self-awareness, trying to spot the little patches of friction that will cause a failure once you switch your attention to something else.

3. Make the change stable through a shift in your skills and self-image.

Assuming a behavioral change genuinely makes your life better and you’ve found a way to sustain it without excessive friction or effort, the final step is to solidify it with a shift in your skills and self-image. 

Much of this is a slow process of solving problems and changing your self-beliefs that occurs naturally as long as you keep up the behavior. But some of this can also occur at an identity level, such as deciding you’re a runner, an investor, a people person or generous.

Not every change you attempt will work through all three steps. Sometimes you’ll finish working on a change only to realize that it’s not as important to you as you had initially believed. In other cases, you’ll still believe it matters, but you’ll fail to find a good design for making it stick in your life.

But, once you’ve implemented enough of these changes, you will reach the third stage where the effect can be quite dramatic. The end result isn’t just better habits, but becoming a better version of yourself.

The post How Habits Actually Work appeared first on Scott H Young.

But, of course, we already know we should exercise. We just don’t always do it.

This is problem goes far beyond exercise. Life is full of low-hanging fruit, things that would make our lives far better than the amount of time and energy they cost us—if we did them consistently.

Seeing this gap, many of us are quick to blame willpower or laziness. We chastise those who don’t do what’s in their own best interest.

Yet we also struggle to stick with them.

Even so, we rarely apply that same scrutiny to ourselves. We’re quick to rationalize our behavior. We don’t have time. We’re too busy. We only live once! Gotta enjoy life, it’s not just about living virtuously all the time, you know?

In other words, willpower is a problem for other people. We, of course, are always making the best choices we can, given our constraints and what we want out of life.

The Problem Isn’t Willpower

The problem, as I see it, is that this framing is essentially negative. It argues “you’re a lazy person if you don’t follow a particular behavior that you admit would be good for you.” It frames problems of behavior as problems of character, and since we’re loath to admit personal weaknesses, it leads us to rationalize away our own obviously imperfect behavior.

A more useful framing, I think, would consider the problem this way:

1. There are large opportunities to make life substantially better. Instead of a negative framing, we should adopt a positive one. If you improve your health, you can enjoy extra years of illness-free life with minimal cost. If you improve your savings, you can live with less stress about money and look forward to a comfortable retirement. If you improve your relationships, you’ll experience more connection and joy, as well as better health and well-being. These are opportunities, not chores.
2. Taking advantage of those opportunities is an engineering problem, not a character problem. To take advantage of all the things researchers have learned that make for better health, wealth, relationships and happiness doesn’t require you to suddenly become a totally different person. It requires solving a behavioral engineering puzzle. Whether you exercise every day or not says much less about your character and much more about whether you’ve solved this particular design challenge in your life, as it looks now.

It is, of course, always possible that certain beneficial behaviors are oversold. Indeed, the advice space is full of hyped-up recommendations that don’t actually have much (if any) empirical backing.



It is also, of course, possible that you’ve managed to find the exact optimal point for all these sorts of habits in your current life. Perhaps the constraints you face really don’t allow for any further improvement.

But I think for the majority of people (myself included), there are clear behavioral changes that pass a sensible cost-benefit test. Most of us could reap more of those benefits by making changes to the systems that support them in our lives.

Good Habits are Like Bridges, not Virtues

Building good habits is much more like engineering a bridge than it is an act of improving your character.

Think of what has to go in place to build a bridge that won’t fall down:

1. You need to make the bridge out of a material that won’t collapse under the weight it needs to bear—including its own.
2. You need to ensure the supports the bridge rests on won’t sink into the mud.
3. You need to ensure a certain amount of flexibility, so that the forces of wind or earthquakes won’t cause the bridge to collapse.
4. And so on …

Sustaining behaviors like regular exercise, healthy eating, building savings or deepening relationships depends on similar engineering principles:

1. You need to make the behavior lightweight enough that it won’t collapse under the real pressures of your schedule and other commitments.
2. You need to ensure the environment reliably supports your behavior, making it as easy as possible to sustain.
3. You need to ensure some flexibility, so that when your life gets interrupted in various inevitable ways, you can keep it up.
4. And so on …

Adopting the design metaphor helps change the discussion around our behaviors. Instead of viewing adherence to good habits as a source of pride or shame, we can instead look at it as a design problem. Each one may pose challenging constraints, unique to your situation, but each is fundamentally solvable.

And, just as you wouldn’t salvage a defective bridge by standing under it and trying to hold it up with your arms, you wouldn’t try to salvage an unworkable habit simply by applying more willpower.

Good habits, in this sense, are important precisely because they are not a function of our character. They work because they are stable with minimal effort. We sustain any good practice because it fits into our lives without much fuss, not because we have heroic discipline and self-control.

Since we’re discussing engineering, it would help to actually discuss what the principles are, not just gesture vaguely at them. A good engineer doesn’t simply believe in science, but has actual guidelines involving forces, material strengths and so on.

In the next essay, I’ll try to summarize some of the main takeaways from years of behavioral research, showing exactly what sorts of design principles help us create sustaining—and sustainable—systems in the long-term, and what sorts of mistakes people commonly make.

The post Your Problem Isn’t Willpower appeared first on Scott H Young.

We often fail to exercise enough. The physical labor needed to get by in modern life is below what we need to be optimally healthy. So, unless we actively enjoy a sport, the tendency is to move less than we should.

Not only do we move less than we should, we also move less than we think we do. In one of my favorite studies, researchers found that when asked whether they met the recommended guidelines for physical activity (150 minutes of moderate or 75 minutes of vigorous physical activity per week), a majority of respondents said yes. But, when their activity was measured using wrist-based accelerometers, fewer than ten percent actually did.

Given this, the reality is that, for most people, the best amount of exercise to get is the amount they can practically sustain. Running every day is great, but if you can’t do that, walking is good too, and it’s a lot better than doing nothing.

Even if, practically speaking, the best exercise routine is the one you can comfortably stick with, I find this hand-waviness unsatisfying. Clearly, exercise is important. But how important?

Quantity matters. Something that improves your life by 0.1% is maybe a rounding error—you could make the case for it, but it isn’t a big deal. Something that improves your life by 50% or 100% is very important. To ignore such a big potential improvement is insanely wasteful.

Fortunately, we can do a lot better than hand-waving. We can determine not only how much exercise is optimal for health, but also give some sense of what would be ideal for you personally—given how much you actually enjoy exercising.

What’s the Optimal Amount of Exercise for Health?

Arem and colleagues pooled together data from over 600,000 people and found the following curve showing the relationship between amount of exercise and all-cause mortality:

There are a few things to note about this curve:

First, it shows diminishing returns. That means going from zero to one hour of exercise per week is much more advantageous than going from one hour to two hours of exercise per week. Getting more exercise matters most when you do it the least.

Second, benefits appear to plateau. At around five times the standard guidelines, the benefits of additional exercise appear to be completely flat. Meaning if you’re getting 750 minutes of moderate (or 375 minutes of vigorous) physical activity per week, adding more exercise isn’t associated with better health outcomes.

This is a lot of exercise. If we consider vigorous exercise to be running and moderate exercise to be a brisk walk (or, more likely if you’re getting this amount of exercise, a jog), this is the equivalent of running nearly an hour every day, or briskly walking for nearly two hours every day. Chances are, if you’re getting this much exercise, you’re doing it for reasons other than health.

Third, there’s some uncertainty as to whether the benefits merely flatten or if they actually become negative beyond this point. Directionally, it looks like the curve may bend backwards. In other words, once you reach the training volume of ultramarathoners and beyond, your health may actually get worse.

However, this is a bit misleading. There are so few people at such an extreme that the confidence intervals become broad, and we have less certainty in the measurement. It may be that exercising that much is unhealthy for you. But, as the study authors point out, we don’t actually know if this is true. All of this is to say is that even amounts of exercise that look extreme to a normal person aren’t clearly worse for health than the optimum amount, and it’s probably still healthier than the amount most of us typically get.

So, based on these data, the absolute best amount of exercise for human health is probably in the range of 3-5x the standard guidelines.

But is the Optimal for Health Actually Best for You?

The problem with this analysis is that it ignores tradeoffs. Unless you’re Bryan Johnson, you value things other than living as long as possible. How do you compare the value of an extra month of life to, say, greater professional success, deeper friendships or watching more sunsets?

I don’t pretend to be able to resolve the incommensurability of values in a human life. But we can make the analysis tractable if we ask a different question: how much do you value the time spent exercising, compared to the average value of the time/activities that exercise would replace?

Since a major reason to exercise is to live longer, we can consider this as a kind of trade-off. You exercise now, giving up something else you would have done in that time. But, as a consequence, you now live longer (on average), so you regain some of that time over the course of your life.

Considering this role of trade-offs, a better analysis would look at how much “life” you gain for each incremental bit of added physical activity, and then subtract from it however much less the exercise was worth to you than the next best use of your time.

This is a little tricky. Arem used hazard ratios, not “years of life gained”, so without knowing the baseline survival rates of the population studied, we can’t easily estimate it. Luckily, an earlier paper by Moore and colleagues did this calculation for us, albeit using a more limited data set.

With a little math⁴, we can convert Arem’s hazard ratios into years of extra life gained, using the Moore paper as an anchor:

We need to convert this into a graph that shows the additional benefit for additional exercise:⁵

On this graph, we can superimpose the “costs” of exercise. Evaluating the costs is tricky, because it depends on how much you intrinsically value (or hate) the experience of exercise.



You could imagine that exercise is purely instrumental—a means to an end. If someone waved a magic wand that let you get the benefits of exercise, but rendered you completely unconscious for the duration of the experience with no memories, you’d be indifferent to it.



In this scenario, exercise is pure experiential waste that we only do for the benefits. Under this assumption, the optimal point of exercise is between 225 and 375 minutes per week.

You can also scale this by how much you hate or enjoy exercise. Below, I’ve drawn graphs which consider the possibility that your time spent exercising is 80% as intrinsically valuable as your next-best alternative, in which case the optimal point is between 375 and 625 minutes per week.

I’ve also considered the opposite case, where exercise has negative intrinsic value, such that you’d rather give up an hour of time in your life to avoid thirty minutes of exercise. In this case, the optimal point lies between 75 and 225 minutes per week (although closer to the latter).

What we can see from this analysis is that the amount of exercise people typically get is only “rational” if we imagine that they really, really hate exercise. For instance, going from zero to a little bit of exercise is associated with roughly six times the increase in lifespan compared to the time cost of the exercise itself. That means you’d have to prefer losing nearly three and a half hours of your total lifespan for every thirty minutes spent exercising for this to be a rational choice!

I don’t know about you, but that seems pretty strong to me. Even for activities I hate, like getting a cavity filled at the dentist, I’m not sure I would actively give up much more than twice the amount of equivalent life experiences.

An additional wrinkle is that I’m assuming your exercise is moderate intensity. For a lot of people this is equivalent to brisk walking or a slow jog. For vigorous activity, the standard efficiency is roughly twice that, so this applies a fortiori. ⁶

What if You Want to Live Slow and Die Young?

One quibble might be that this analysis weights present moments equally with ones experienced in the far future. I think it’s a justifiable assumption.

When we consider financial decisions, we often use a discount rate to reduce the value of money in the far future. But I’m not sure that’s justified here, for a few reasons:

1. Money has less value in the future due to opportunity cost. Money invested now will compound, meaning the same income is worth more today than a year from now. There is no comparable mechanism with regards to lifespan.
2. Exercise has benefits beyond longevity. Exercise improves mood, cognition, sleep, stress and even your sex life. Many of those benefits occur in the near future, so a fuller analysis would need to account for these benefits that ought not to be discounted much.
3. Philosophically speaking, it’s not clear that future moments we experience are genuinely worth less than present ones. After all, when the future comes, we will experience it just the same.

Of course, it’s obvious that the fact that many of the benefits of exercising accrue in the far future while the costs of exercising are immediate does impact people’s decision making, as it does with other prudent behaviors like saving for retirement, not drinking or smoking excessively and so on. But that imprudence is usually irrational.⁷

If Exercise Matters So Much, Why Do We Struggle With It?

Exercise is a good example here, but it’s hardly unique. Life is full of actions that can dramatically increase the quality (or quantity) of our lives compared to their costs, but which we often have a hard time adopting or sustaining.

We often blame lack of information, weakness of willpower, or personal preferences as reasons we fail to capture all this low-hanging fruit.


But I tend to think the problem is more than that. Cognitively speaking, exercise belongs to a class of problems that we do not have good built-in mechanisms for solving. And thus, to successfully stick to an exercise routine year after year and reap the benefits requires constructing systems, habits, beliefs and attitudes that make up for where our instincts fail us.

There are many ways to cultivate such systems, but the basic ingredient in all of them is a certain degree of committed focus, at least for a time. Spend a little bit of time creating the system, refining and stabilizing it, and it will continue to deliver benefits throughout your entire life.

This was the goal of my year-long Foundations project, where I spent each month focusing on one of twelve different foundations for life, including fitness. The idea was that, with a combination of carefully chosen habits and deep research, I could build systems that would yield the benefits of better fitness, relationships, productivity, reading and more.

The post What’s the Ideal Amount of Exercise? appeared first on Scott H Young.

Therefore, it’s often a pleasant surprise to find that two things we had previously considered as separate phenomena are actually the same thing. The morning star is the evening star. Lambda calculus and Turing machines. The orbit of planets and the falling of an apple.

I’d like to make the case that motivation and focus are similarly equivalent. To be focused is to be motivated to persist in an activity.

Motivation as Choice

We tend to think of motivation in terms of intensity of desire. Think of a scale from zero to ten, with zero indicating having no interest in doing something and ten as having the maximum possible interest. Motivation, then, measures where we fall on this scale.

But this isn’t really the full picture. What matters with motivation isn’t simply how much we want to do something, but how much we want to do that thing, compared to all our other alternatives.

What’s more, we’re always already doing something. While we often think about the motivation to do some behavior, we’re actually talking about the motivation to stop whatever we were already doing, even if that is just sitting and staring into space, and do something else instead.

From this perspective, it’s pretty clear to me that focus and motivation both refer to the same phenomenon.

Focus is the quality of persisting on a single pursuit. This can be at small timescales, like focusing on a textbook in a quiet library while studying for an exam. But it can also be over much longer timescales, such as choosing to devote your life to painting, rather than music or sculpture or making a lot of money.

What does it mean to be focused in this way? It means that the motivation to stick to the current activity remains consistently higher than the motivation to switch to an alternative.

Lack of Focus as Lack of Motivation

A comment I’ve heard from people who self-report they have difficulty focusing is that they only have difficulty focusing on things that they don’t find motivating. In contrast, they claim, they can maintain a high degree of focus on activities that they personally find interesting.

According to my claim above, this is tautologically true. People, by definition, have a hard time focusing on things that they are not motivated to do. Focus is motivation.

My understanding of the prominent theories of ADHD, for instance, is that it is seen as deficits either in impulse control or in rewards. Put another way, with ADHD, your intentions to stick to consciously-held goals aren’t sufficiently motivating, or the tasks you’re trying to stick to are not sufficiently motivating. The mechanisms here might be different—impulse control is more strongly linked with prefrontal cortical circuitry, whereas rewards are processed in dopaminergic systems—but the effect is ultimately reducing motivation to stay on task.

To Boost Focus, Boost Relative Motivation…

Since focus is motivation, the way to improve focus is to be more motivated. Increase the relative drive you have to do the task you’re supposed to focus on, or, alternatively, reduce the relative drive to do anything else. Both work, but which is more suitable depends on the situation.

Sometimes it’s easier to boost your motivation to do a task. External carrots and sticks can help. But, generally speaking, intrinsic rewards—and costs—matter more. A task may feel important if it has big, long-range incentives for action. A task with a lot of built-in intrinsic rewards feels inherently interesting.

Generally speaking, it’s hard to sustain motivation for tasks you’re supposed to do but aren’t interested in. You can drive effort by holding a clear goal in your mind and committing to it. But effort is exhaustible, and this approach tends to result in decreasing motivation over time.

Capacity for effort is probably at least partly learned through life experiences. Learned industriousness is the idea that regular rewards for effortful action reduce the motivational cost of effort. In contrast, if trying hard typically leads to failure, the result can be learned helplessness instead.

When the activity itself supplies the rewards, our interest level is naturally high, and the driving force of effort is not needed so much. Many people who seem preternaturally able to focus are simply really interested in the things they focus on.

A key to focus, then, is to find ways to do things that are both important and interesting.

…Or Decrease Temptations

Some tasks persistently resist our interest. If you can’t boost your motivation to do something you need to focus on, you can at least reduce the appeal of alternatives.

One way to do this is to raise the costs of temptation. If you put your phone far away from you, it costs a lot of effort to get it to indulge in a distraction. If your phone is in a locked box and the only key is with someone who lives across town, it’s even harder.

Raising the cost of temptation works, but it has two additional challenges.



First, when your interest is low, there are often many possible alternatives that are more interesting, including daydreaming. This means that sustaining focus can feel like playing Whac-A-Mole, where you fight off one distraction only to slip into a different one.

Second, human behavior flows toward rewards. This means we’re loath to create future situations for ourselves that are less-than-optimally rewarding. It’s hard to lock yourself in the library to study, and it’s hard to focus on a dull textbook while you’re there.


Now this latter problem is not unsolvable. We tend to think about future impacts on motivation differently than motivation experienced in the moment. So, in some ways, pre-committing to a course of action you know will be less tempting is actually easier than avoiding those temptations when they occur. You can trick yourself out of akrasia.

But reducing distractions while trying to force yourself to do tasks you have low interest in or experience low rewards from tends not to be a stable solution. You have to cultivate an interest in the thing you’re trying to focus on (because of improved self-efficacy, curiosity, learned industriousness, etc.) or you’ll likely renege on your earlier commitment.

The post Focus is Motivation appeared first on Scott H Young.

But being totally unqualified has never stopped me before, so I’ll try to explain the answer I typically give to this question, which is that I’m skeptical of dramatic proposals to make school considerably more effective or efficient for the average student.

To be clear, that’s not because no improvement is possible. We do know some about things that work that are inconsistently applied: phonics should be taught, cognitive load should be managed, skills should be fully taught and practice should be fun and ample.

But these answers aren’t the kind that satisfy the people who ask me these questions. Instead, having had many of these conversations, I feel like the person asking already “knows” what my response should be:

Isn’t it obvious that school sucks? That we should be teaching critical thinking and problem-solving skills instead of useless facts and theories? That school should be more like real life, with real-world projects and experiments and collaboration? That there should be less of that stuffy work of sitting at a desk and memorizing things?

If you had asked me this question years ago, I probably would have agreed with you. It took reading a lot of research to convince me that this intuitively appealing idea is actually bad. Below, I’d like to explain why.

First, the Evidence

Before I get into the explanation of why these kinds of seemingly-good strategies don’t work, I should begin by pointing out that these ideas are not new. They have been tried, and they have been found wanting.

Entire books have been written pointing out the flaws in many of these strategies. I won’t be able to do the full debate justice here, but, if you’re interested, you can check out Daniel Willingham’s Why Don’t Students Like School? Greg Ashman’s The Power of Explicit Teaching and Direct Instruction or, if you want to learn more about the actual debate between proponents of both sides, try Constructivist Instruction: Success or Failure.

To briefly recap some of the evidence:

• Project Follow Through was one of the largest educational experiments ever conducted. Run in the 1970s, it compared how different teaching methodologies impact student outcomes. Direct Instruction, a method of teaching that has students sit in desks and perform extremely structured drills in unison, performed best.
• Problem-based learning tends to do worse than traditional schooling in medical education. An influential meta-analysis by Albanese and Mitchell, for instance, found that students required more time studying, had worse exam scores and ordered more unnecessary tests compared to traditionally taught students.
• Despite needing to relearn this truth every few decades, the best way to teach kids how to read has been known for centuries: break down the sound-spelling correspondence, and do lots of practice on it before moving up to authentic texts. Approaches based on skipping these drills in favor of “inspiring a love of reading” do worse.
• Practice testing and distributed practice—basically, having regular quizzes spread out over a course—are the studying methods with the best empirical support. Fancy methods like mnemonics and concept maps fare worse.
• General problem solving abilities are neither learned nor taught. While some problem-solving methods have broader applicability than others (such as the scientific process of hypothesis testing), students learn these methods better when they’re explicitly taught rather than simply giving students projects and hoping they’ll reinvent them on their own.

In short, whenever we have high-quality evidence that rigorously compares two teaching methods, the research invariably favors strong, direct instruction plus practice.⁸ Or, in other words, the exact stereotype of schooling that so many of the people asking me about school reform despise.

Your Stereotype of School is an Endangered Species

This doesn’t mean education couldn’t be better. My impression upon first encountering the Direct Instruction research was that I had never been taught this way in my entire life.

Clichés are often out of date. I went to grade school in the nineties, when the lofty aims of project-based and discovery learning were the educational orthodoxy. I spent a lot of my school years doing time-consuming projects that had us gluing and coloring and expected us to do our own research.

I’m pessimistic about real reform because the changes needed to make schools more effective are often opposite of what many people intuitively feel. For schools to teach more effectively, they should be more rigorous about carefully defining the knowledge objectives of the class, thoroughly breaking down complex skills into components, and doing lots and lots and lots of practice.

In short, a “better” school probably looks more like the stereotype of an old-fashioned schoolhouse with kids sitting at desks, drilling facts and concepts that are patiently explained by a teacher. To the extent that school becomes more like free play, project-building or acting like a scientist, it will probably be worse.

Why It’s Hard to Improve Schools

Schools face a number of practical constraints that make thorough reform difficult. Students are unmotivated. They range in background knowledge and innate ability. We care about sorting just as much as educating, so schools end up doing both.

But the real reason it’s hard to improve schools is simply that there are fundamental constraints on how the brain learns that prevent radical shortcuts.

The boring truth is that expertise in most subjects is largely a matter of having an enormous library of knowledge and skill. For example, if you want to learn a language, you need to learn a lot of words. Any method that tries to skip over the fact that there are tens of thousands of words to learn is doomed to failure. All skills are like this, it’s simply that the “atoms” of learning are usually less obvious than in languages.

When students complain about all the boring facts and skills they had to learn in school, my response is to claim that there isn’t any other type! All skills are simply an accumulation of small bits of facts, procedures and concepts.

Those small bits, in isolation, seem kind of trivial. But quantity has a quality all its own, and with enough well-integrated knowledge the result is expertise that seems almost magical to those who don’t possess it.

This means that improving education comes down to largely two different options:

First, you can increase the efficiency of the system. Efficiency here looks like the kind of factory redesign that increases product throughput—increasing the number of words learned per day, optimizing cognitive load, boosting mnemonic efficiency through spacing and retrieval—without skipping over the fundamental bottleneck in cognition.

Second, you can choose to learn different things. Given the high degree of specificity of most knowledge, the choice of what to learn can have profound consequences. But if choosing a pedagogical method is contentious, curricular choice is even more so! For every “useless” subject that reformers want to discard, there are die-hard advocates arguing that we should be putting in even more.

I believe in both of these things, and I’ve focused much of my writing career on how we can do them better, particularly outside of the typical classroom. But if you want lots of skills, there’s no way around learning a lot of stuff—including a ton of stuff that feels too obscure to be broadly useful.

What About Ed Tech?

Thus far, I’ve mostly been targeting a certain kind of questioner: someone who feels that school was maybe too boring and impractical, and who longs for the possibility that education could be more like play and less like studying.


There are reformers of all stripes, and educational technologists are another side of this debate. These are the people who champion efforts to gamify learning, carefully match teaching to each student’s ability level, develop AI-based tutoring, put an iPad in every child’s hands and so on.

In theory, these ideas are possibly useful. Drills can be boring, so wrapping them with gamification elements that reward progress and engagement might be helpful. Skills can be too hard or too easy, so adjusting difficulty automatically might be helpful. AI-tutoring, too, might help with closing Bloom’s famous 2 sigma problem.

But I’m more skeptical in practice. As Kelsey Piper writes, a lot of ed-tech games have a fairly low density of actual useful learning. I can attest to this: eager to give my son a head start on the phonetic skills involved in reading, I tried a few different iPad games with him. He mostly messed around randomly until he got the reward, largely ignoring the educational content to fixate on the cute cartoon characters.

Gamified learning is a bit like wrapping medicine in candy. Yes, it may help some students swallow some instruction they otherwise find bitter, but in practice it’s easy to pull off the candy, consume it, and throw the medicine away.

Individualized instruction aided by technology does solve some of the problems of differing ability levels. But schools aren’t just solving the cognitive problems of learning, they’re also working on motivational ones. A rigorous, but achievable, standard that applies to everyone may be more sustainable for motivation than an individually-tailored goal for each student.⁹


Similarly, while I’m hopeful that AI advances will make automated tutoring more useful, it’s still far away from the skill a teacher can provide. As someone who makes use of AI quite a bit in my own learning, I can say that it’s still relatively weak at having a good model of an individual’s skill gaps and conceptual weaknesses. It’s very much at the “better than nothing”—not the “better than teachers”—stage right now.

So, while I’m hopeful that there will be some improvements in technology around the margins, I’m skeptical of anything touted as a radical overhaul in educational process or outcomes.

What About Ultralearning?

Does this line of thinking rule out the methods I describe in Ultralearning? I don’t think so. The big distinction (and it is a big one) between my aims for that book and the aims of educational reformers is that I started with the assumption of a highly motivated learner.

The people I document in that book all began with the starting point that they were willing to work hard, even obsessively, on a project they were deeply motivated to succeed with. In such cases, the classroom structures that facilitate motivation can instead become obstacles: fixed homework assignments, mandatory lectures, exam deadlines. These things keep uninterested students going, but they may hold back the aggressively curious.

For instance, I still believe that full immersion is the best way to learn a language, provided you also augment that with a lot of the studying approaches I describe above, but it’s obviously a high-effort strategy. I’ve spoken to lots of people who have asked me for advice on how to learn a language, but relatively few that took up Vat’s and my actual strategy of avoiding speaking English.

Ultimately, I believe enhanced learning is certainly possible, and a highly motivated person can often do better than the average, and sometimes even the upper end, of what is typically seen in school. But such optimism about the possibility of learning doesn’t so easily transfer to a situation where motivation is much lower.

The post Why I’m Skeptical About Efforts to Revolutionize Schooling appeared first on Scott H Young.


Focusing on increasing the sheer quantity of your reading is generally underrated, especially because reading comprehension and retention are strongly coupled to your background knowledge. You’ll read better if you read more. And so, to a first approximation, if you read more books on a topic, you’ll automatically read better.

I’ll give an example. Recently, I’ve been reading Martha Nussbaum’s, The Fragility of Goodness. The book has an interesting philosophical premise: the good life, indeed even being good, may be contingent on factors outside of our control. She argues the early Greek tragedians understood this, and Western philosophy has spent the last two thousand years trying to deny this possibility.

The book is interesting and fits into some of my broader reading goals of learning about what it means to live a good life, happiness and so forth.



But the book is also challenging to read. This isn’t due to faults in Nussbaum’s prose, but because she assumes the reader brings so much knowledge to the book. She draws lessons from Agamemnon and Antigone assuming the reader is already familiar with these plays. (I was not.)

While unreadable academic writing is practically a cliche, I think this stereotype masks a more fundamental problem: Many topics are deep. To write about them while also repeating every bit of background knowledge informing the conversation would be incredibly tedious. At close to 600 pages, Nussbaum’s book wouldn’t be enhanced if she expanded it to meet popular nonfiction’s assumption of readers having minimal knowledge on the topic. This means to understand The Fragility of Goodness, you need to have already read a lot.

Thus, if you want to read better, start by reading a lot. Quantity and quality reinforce each other, because as you read more, even with only partial comprehension, you build the reference points that allow you to tackle deeper and better books.

The Perils of Path Dependency

That reading better books requires reading more books sounds innocent enough—until you realize that knowledge is not a generic commodity. The more you study within a particular topic, the more adept you get at reading that topic. Indeed, if you study particular arguments, they crawl into your thoughts until you can’t easily separate them from yourself. You become better at seeing one side of a debate, but the other side becomes illegible.

I’ve become sensitive to this problem in my own work. My last book, Get Better at Anything, was originally going to be called Do the Real Thing, and I planned to focus on the research on learning transfer. 

I had encountered some interesting research on the relatively low degree of transfer of learning when researching my first book, Ultralearning. To clarify, the common finding among psychologists is that people who learn a skill tend to get good at it and skills closely related to it. Any benefits of having learned that skill drop sharply as we move to more distant skills, and the benefits learning it brings to completely unrelated skills are hard to distinguish from zero.

I thought this was fascinating and under-discussed. The implication of this, to me, was that we should spend more time doing things closer to what we want to get good at, trying to avoid artificial and substitute activities with vain hopes of transfer.

As I started researching, I got deep into some theories of learning which made almost exactly this argument: Jean Lave’s work on situated learning, Allan Collins and John Seely Brown’s work on cognitive apprenticeship and classic education reformer, John Dewey.

Only after about six months of deep research did I begin to stumble upon the critiques of this viewpoint. Essentially, the critiques offered a theory that neatly explained the findings on transfer, but rejected the takeaway that I was building my initial book around: that doing the real thing is generally best.

In particular, I was persuaded by the work of John Sweller’s cognitive load theory, Siegfried Engelmann’s Direct Instruction, John Anderson’s ACT-R theory. Together, they argue that complex skills are often best learned first in parts; that observation, not pure practice, is essential; and that reverse-engineering expertise is fraught. While direct practice still matters, it is best done in an environment with clear explanations, guided coaching and, quite often, drills that bear little resemblance to the skill you want to improve.

The result was a tricky pivot to write a completely different book from what I had initially set out to write, including a stressful six-month research detour where I forced myself to read a lot more books and papers than I’d care to recall.

The lesson I took from this, and indeed one I seem to have to learn more than once, is that knowledge is path-dependent. The books where you start your reading have an indelible influence on where you end up. It can take heroic levels of reading to undo the effects of a poorly chosen initial book.

This is the dark side of the knowledge dependency I articulated above. Because you retain and understand more when you know more, you end up being better able to understand and assimilate viewpoints that already “fit” with what you know.

When Reading More Makes You Know Less


Some topics, due to lay interest, are notorious for this problem of path-dependency.

Consider nutrition. Of all the topics I discussed in my Foundations project, nutrition spurred the most long-winded rebuttal emails. And this was despite the fact that I was, to the best of my ability, simply parroting the expert consensus. A lot of non-nutritionists, it seems, have heterodox views about diet and have a lot of “knowledge” to back it up.

How can I be so sure I’m the one who’s right? After all, I’m also a non-nutritionist who’s read a lot of books. Perhaps I’m the one with weird beliefs and my correspondents are correct in pointing out my errors.

This is certainly possible, and, based on my theory above, I wouldn’t necessarily know if I was wrong. But, while I cannot claim omniscience, I do know that at least some of my critics were wrong, because they were espousing opposite theories! Indeed, I got a handful of pointed critiques from both plant-based devotees AND carnivore-adjacent low-carbers.

I don’t want to suggest that either of these diets are necessarily unhealthy. From a practical point of view, perhaps it is best that we each simply follow whatever works for us. But a religion stops being merely personal when one begins to proselytize. If you’re not simply arguing “this is what works for me” but “this is what everyone should do”, personal anecdote stops being an effective rebuttal to accumulated scientific research.


Ironically, from a purely intellectual perspective, a lot of people who have gone deep into the rabbit hole here actually know less than they would if they hadn’t read anything at all. Meaning, if they simply had absorbed the vague popular sentiments about eating healthy (which, to be clear, are also biased), they probably would be closer to the orthodox viewpoint than after reading a ton of books.

While I may sound like a know-it-all scold here, I’m very much including myself in this critique. If you are interested in learning and knowledge, you should always be somewhat wary of the possibility that you’ll invest a lot of time and effort and end up with views that diverge even more strongly from our best estimate of “truth” than if you hadn’t begun reading in the first place.

I’ve previously made the case for believing expert orthodoxy, but my argument here doesn’t hinge upon it. It’s certainly possible to reject the standard position in a field after having carefully deliberated on all the evidence. But the conclusion you arrive at shouldn’t depend on which book you read first.

My Heuristics for Reading Better

So how can you read better books?

As already mentioned, many of the best books aren’t written for beginners. They depend on the reader already having considerable background knowledge for proper understanding. If you haven’t learned arithmetic and algebra, calculus is going to feel confusing and painful.

This means reading great classic books on your own, without a companion course or tutor, is sometimes a struggle. You simply haven’t acquired enough knowledge to fully appreciate them yet.

That doesn’t mean you can’t try. But it may take multiple re-readings and many detours to other books to squeeze all the juice out.

What counts as “better” is relative to your knowledge base. Just as the “best” ski slope isn’t necessarily the steepest, but the one that matches your abilities, the best book for you may not be the same one an extremely well-read person would reach for.

And yet “better” isn’t purely a matter of personal preference. Some books provide excellent starting points into a long and satisfying journey into a world of ideas. Other books offer a detour that will cause you to get lost in the brambles.

Some strategies that have worked for me to help sort these two apart are:

1. Start with textbooks. I’ve praised the virtues of textbooks before, and I’ll say it again. Most people who want to actually end up with knowledge (and aren’t just looking for entertainment) would do better to read a textbook rather than a general-audience nonfiction book. If textbooks are too expensive or boring for you, more popular general survey books are also usually good. For instance, most “For Dummies” books are a smarter read than most of the glossy bestsellers covering the same topics.
2. Use ChatGPT and Wikipedia. AI certainly has its weaknesses. Sycophancy is rampant, so if you ask leading questions intended to “prove” your hunch, you’ll too often get it. But, if you’re genuinely curious about whether a given point of view is standard within a field, AI doesn’t do a terrible job. (It’s a lot better than, say, Quora.) Wikipedia is also really good here too, but, because it is organized by encyclopedia topics rather than specific questions, it can take more work to find an answer.¹⁰
3. Read in projects. Don’t get one book, get ten. If you have a bunch of books on a topic, then as long as you read them relatively close together in time, you’re more likely to average out their arguments than if you read only one book and allow it to take up too much real estate in your head.
4. Look for critics. Some risk of path-dependency can be prevented by actively seeking out rebuttals to stress-test your viewpoints. This, too, can be AI-accelerated since “find me the author/papers that best argue against XYZ” is a fairly simple AI request that used to require a lot of searching.

But the methods we use are downstream of our motivations. The way to read better books is, fundamentally, to be deeply curious about what other people think, curious enough to want to spend a lot of your time working to understand them. If you’re motivated by something other than curiosity—a desire to prove a point, to be part of a “team”, or to simply be entertained—you’ll probably get what you’re looking for, it just may not be knowledge.

The post How to Read Better Books appeared first on Scott H Young.

But I’m struck by a weird asymmetry in science fiction: it’s way easier to imagine dystopias than utopias. Most worlds of science fiction are markedly worse than ours in pretty catastrophic ways. Utopian science fiction barely exists.¹¹

Some of that asymmetry may be simply that stories require conflict, and a dystopian society naturally generates conflict that makes for an interesting story. A novel where everyone’s happy and leads contented, fulfilling lives is, well, kinda boring.

Recently, I’ve been thinking the reason for this negative skew may be different: perhaps human beings just aren’t designed to view any world as utopian. We’re both problem-centered in our focus, making us ignore the things that are going well, and our drives and needs are frequently misaligned so, even when we get what we want, we don’t always get what we need.

Why are We So Rich, Yet Everyone Feels Poor?

I’ve spent a lot of time thinking about Scott Alexander’s essay on the vibecession. 

To summarize briefly: Americans think the economy sucks. Yet, by almost all measures, it is doing fantastically well. Alexander tries to diagnose why economists and everyday people disagree to such a large extent.

The knee-jerk reaction is that the economists are just wrong. But Alexander goes to great lengths to explain why this probably isn’t true. In most ways economists can measure, people aren’t doing too badly: unemployment is down, inflation is under control, and, for Americans at least, they’re much, much richer than almost everywhere else in the world, both in terms of the very top and also for most of the middle.

I know it’s fashionable in certain circles to dismiss economists entirely as being an apologia for laissez-faire capitalism and right-wing ideologues. But I think this criticism mistakes a certain subset of economists (who are currently dwindling in number) for the field as a whole. There are plenty of left-leaning economists, and the professional incentives hardly align with suppressing pessimism about the state of the economy for everyday people.

So, it seems likely to me that if there was a really good case for why the economy is actually terrible and the vibecession were justified, economists would endorse it.

My personal guess is that pessimism about the economy can probably be explained by an overall increasing negativity in news media. When people are asked about their personal situation, they tend to respond more positively, but when asked about “the economy”, the question expands to other people, and respondents are more likely to rely on what they hear from news reports.

News has gotten a lot more negative. But that simply bumps the question up further: why is the news so pessimistic? Why does our society feel increasingly dystopian?

You Get What You Want, But You Can’t Always Get What You Need

Media pessimism is probably best explained by consumer preference. People are more likely to read about bad news.


This doesn’t mean people want bad things to be happening in the world. It’s simply that if everything is good, we don’t need to hear about it. Our evolved preference is to seek out information when there are potential threats.

This makes an asymmetry in news production that’s similar to the one we see in science fiction. Feel-good stories are the minority. When they do appear, they’re often kind of boring. Feel-bad or feel-scared stories dominate the headlines because these drive reader interest.

We have always wanted to hear bad news, but the vibecession is recent. The reason we feel particularly pessimistic about the world today is that media has become increasingly optimized to give us what we want. The hypercompetitive, algorithm-driven media environment is simply a much more desirable product, from the perspective of consumer preferences, than the more boring newspaper era.

This isn’t limited to news media pessimism. Our society has gotten much better at servicing nearly all of our most basic drives:

• Hunger. Food today is more abundant, convenient, palatable and cheaper than ever before. Nearly all of our diet-related health problems come from eating too much, not from nutritional deficiencies or toxic food additives.
• Entertainment. Infinite feeds of short-form video, algorithmically tailored to our personal interests. Our dwindling ability to read books and think deeply is directly related to the always-on faucet of easy entertainment.
• Romance. Swipe right, swipe left. An infinite sea of potential mates at our fingertips, allowing us either to indulge forever in short-term hookups or scrutinize endlessly to find the perfect partner, who, of course, doesn’t actually exist.
• Physical ease. Our physical environment requires increasingly less effort. The result is that we don’t move enough and are sicker as a result.

It’s easy to blame “capitalism” or big corporations or some other nefarious force for all of this. After all, if the big tech companies didn’t insist on making phones so addictive, maybe we would read more books. And if industrial food producers didn’t make such ultra-processed junk, we wouldn’t have so many health problems.

But the companies are all locked into the same bind. If Meta doesn’t make the most addictive social media website, they’ll be outcompeted in the attentional marketplace by ByteDance or YouTube or some new start-up that will give people what they desire.



The real fault isn’t in the companies, but in ourselves.

Victims of Our Own Success

As we’ve become richer and more technologically advanced, we’ve become better at delivering the things people want—but what we want isn’t always what we need.

It’s hard to describe our current moment as a utopia with a straight face. Indeed, I have a lot of fears and misgivings about the direction society is taking. I’m worried about democracy, war, warming, divisiveness and the possibility that the robots will eventually kill us all.

But if we define a utopia as a society that gives people an abundance of the things they want, then, at least compared to nearly all actually-existing societies, we’re living in it! 

True, there are still problems that might be solved in some glorified Star Trek future that don’t exist today: teleportation, robot butlers, cures for aging, and world peace. But if we avoid speculative futures, it’s pretty clear that we’re closer to the utopia of human desires than we’ve ever been in the past.

Yet, if we look at human flourishing, the kinds of things we need to be psychologically fulfilled, the picture doesn’t look so good. Rates of depression and anxiety have skyrocketed. People are anxious, fearful, inattentive and unhappy.

One story you can tell about this trend is that technology is to blame. The rise of smartphones, social media and easy entertainment have glued us to screens rather than real friends, hobbies and time for reflection. If so, the problem would be not that we live in a world without those things (friends, hobbies and time for reflection still exist, after all), but that our world has too much the things people desire, and they crowd out the things we actually need.

Stoicism, Social Control and Ozempic

If this diagnosis is correct, what can we do about it? I see only three options, all of which have serious drawbacks.

The first is prudence. Cultivate virtue. Resist vice. Delete social media from your phone. Buy only veggies and whole grains for your pantry. Don’t drink, gamble or own a television.

I’m sympathetic to this path. And, given that this is the only factor that we can, as individuals, realistically change, it’s usually the one I spend more time advocating. It requires recognizing that there has always been a tug-of-war in the human soul between the steeds of passion and reason, but in today’s world the passions are being pulled by a rocket ship.

But, if “have more willpower” was a universal solution, we’d have adopted it by now. Losing weight by eating fewer calories than you burn has always worked. But sustaining weight loss through willpower alone has astonishingly low success rates.

I don’t think this failure of willpower means we should stop trying. But if we’re in an escalating arms race with technology increasingly able to satisfy our most basic drives, that effect isn’t going to be balanced with increasing willpower by all but the saintly few.

The second option is regulation. Tax sodas. Ban smartphones from school. Force social media platforms to moderate content and change their algorithms.

This path has the advantage of not relying on the limits of human willpower. If ultra-processed foods become illegal, we won’t need to use self-discipline to avoid eating them.

But I’m also skeptical that social control can ever provide the full solution in a democratic society.

In Julia Belluz and Kevin Hall’s book Food Intelligence, the authors go to great lengths to describe how the hyperpalatable and over-processed food environment is making us fat and sick. But, when they get to the section of what to do about this, it becomes clear that the solutions that would actually work would be draconian.

Banning soda, junk food, fast food and anything ultra-processed would work, but outside of a few health-obsessed foodies, I doubt it could reach a critical mass of acceptance in society. If prohibition didn’t work for alcohol, it’s hard to imagine a scenario where it would work for donuts.

The kinds of social control we’re more likely to get are those where only a minority is impacted. Smartphone bans in school are popular because kids don’t get a vote. Smartphone bans for grown ups, in contrast, feel like a violation of civil rights.

Similarly, my country, Canada, is considering banning smoking permanently for those born after a certain year. I understand the impulse. But it’s only popular because most people don’t smoke at all. Alcohol also causes many social harms, but I see no similar proposals for banning booze, because teetotalers are in the minority.

Ultimately, I have a hard time understanding the people who scoff at the idea of individuals applying willpower and habit changes to avoid vice, yet seem enthusiastic about solving the problem with government regulation and control. The same people who don’t have enough willpower to avoid overeating and obsessive phone usage are, supposedly, the same people who will vote for a policy to prevent them from accessing the objects of their desires.¹²

So, if willpower is weak and social regulation requires collective willpower, what does that leave us with?

The third option seems to be: use technology to modify our brains to change what we actually want.

Until recently, this would have been the premise of a (probably dystopian) science fiction novel. But with the arrival of GLP-1 agonists, it’s the world we’re now living in. 

GLP-1 drugs work not by speeding up your metabolism or preventing you from accumulating fat but by changing how much you want to eat in the first place. Users experience reduced “food noise” and cravings for junk food. Grocery stores are apparently even going to need to increase the amount of produce they stock, as a new group of shoppers suddenly feel the urge to buy kale and broccoli.

The effects don’t seem to be entirely limited to eating behaviors either, with some early reports suggesting the medications may help with other kinds of addictive behaviors such as drug abuse or alcoholism.

GLP-1 drugs seem to work well with few major side-effects, but they’re hardly the only example. ADHD cases have exploded, in part due to loosening diagnostic criteria. But also because inattentiveness is a spectrum and paying attention is harder than it used to be. Medications may make it easier for people to focus and ignore distractions. Unfortunately, the side-effects of these drugs are probably worse than GLP-1s, but it’s possible to imagine a new focus wonder drug of the future that has fewer side-effects and more widespread adoption.

I find it hard to get enthused about a future where we create an abundance of human vices, and cure that abundance by creating drugs that make us desire them less. It sounds, well, dystopian. But I suspect that this will end up being the path humanity follows, if only because it is easiest.

The post What if We’re Already Living in Utopia? appeared first on Scott H Young.