> Muad'Dib learned rapidly because his first training was in how to learn. And the first lesson of all was the basic trust that he could learn. It is shocking to find how many people do not believe they can learn, and how many more believe learning to be difficult. Muad'Dib knew that every experience carries a lesson.
-- from "The Humanity of Muad'Dib" by the Princess Irulan
The importance of "learning how to learn" has been emphasized by all of my teachers since I was in highschool, or maybe even 8th grade, decades ago.
My computer engineering professors also emphasized user centered design. For one of Google's top scientists to bring this up is an admission that they won't, or can't, design a good user experience for their tools.
I found that emphasis to be similar to two teams building from opposite starting points, and never meeting in the middle.
The issue with "learning to learn" is that it does not include the foundational skill of "how to communicate". Far too often it is not a lack of desire to learn, it is the inability to communicate what one is trying to learn. When seeking help, not only does the seeker have difficulty expressing their situation, those trying to help are not taught how to listen and will offer solutions to an issue only starting to be explained. This difficulty is then compounded by self conversation bias that negatively spins against the person seeking. That is two very high hurdles: negative self bias, and inarticulate communications while seeking guidance.
I remember the same thing. That doesn't mean they knew how to teach us to "learn how to learn". Neither does it mean that the underlying education system supported that goal.
Same goes for user-centered design. Trying to make something user-friendly is one thing, successfully doing it is another. Large organizations are especially poor at user-friendly design because the underlying structures which support that goal don't exist. Organizational science is still in its infancy.
The US education system only has one mode, and thats to survive in a slim way with overworked staff and huge classrooms. 40 kids in a math class is seen as normal.
Everything you see of its character, including emphasizing tests and practice, follows from that. Talking about good UX is miles away.
It's a problem that goes beyond the United States, overworked staff, and constraints in general, although these are legitimate concerns.
I studied in a non-US country, but the attention paid by teachers to pedagogy was virtually zero.
I mean, we had five years of English classes in high school, and by the end of high school, less than five out of 30 people in my cohort were able to string a couple of sentences together in English. And my class was made up of serious, studious young people. It seems to me that the time was not well spent, but did the teacher, a caring and generally competent person, reflect on the poor results? I highly doubt it.
“You have to learn how to learn” has been a phrase often repeated by teachers, but I don’t remember any of them emphasizing, for example, retrieval practice: you learn a skill or subject, move on to the next one, and leave it up to fate whether you remember anything from the first one.
It always surprises and saddens me that, despite having been an excellent student throughout my years of education, I remember practically nothing about 90% of the subjects I studied.
> Technology is continually reshaping industries and while many eschew learning and adopting, those who embrace it are the ones who succeed best IME.
I said it already in a reply to GP, but I'm going to say it again: I stopped caring about what people list on their resumes, your work history and education don't matter to me. I'd rather hire a hungry junior that finished a bootcamp, that has a drive and ability to absorb new things and adapt to changing environments, over somebody who's got 10 years of experience and can't do shit outside of their comfort zone.
The number of people who aren't able to learn and adapt to changing times, new tools, new ways of working, etc. is shocking.
There is a certain amount of "use it or lose it" seemingly inherent to virtually every human endeavor. But I suspect if you were to enroll in a class in any of those subjects, you'd perform radically better than a peer who'd never studied them. IOW, there's often more latent memory than we realize or can easily retrieve.
"IOW, there's often more latent memory than we realize or can easily retrieve." -
I did not find it to be true almost at all, and I tested many other people on it. When I voiced my concerns, the usual answer was, "Yes, but when you pick up a book, you will remember". And then I asked, "Try it", and the subsequent answer was, "I have to admit you are right".
The "re-absorption speed" is heavily confounded by general IQ and the kind of cognitive stimulation one receives in daily life, but the original learnings are mostly gone.
Among other things, this is why retrieval practice is important: it slows down the "forgetting rate".
Thing is, this has nothing to do with UX or even AI at the end of the day. Over the years I have adjusted the way I handle interviews when I'm on a hiring panel to focus on critical thinking, problem solving, and lifelong learning above all else - because they are the #1 indicator that a hire is going to be successful in a position, even if that means they need a lot more training before they become productive.
I can teach somebody who finished a 6-month coding bootcamp Go, all the internal tooling, go over the business with them, etc. if they have these skills and end up with a productive mid-level engineer who gets shit done in a few years. What I can't teach is the drive and ability to learn, that's a much longer process and if you don't already have it then I'm not prepared to develop it.
Hell, outside of looking for signs of obvious bullshit I stopped giving a shit about resumes. Your work history does me no good, your education doesn't matter to me, and your references are useless beyond making sure you aren't straight up lying to me about your employment history. Every single time I have hired somebody who has 5 years of "experience" working with technologies I bullet pointed on a JD they ended up fumbling the moment they had to do something new. Doing leetcode, pair programming sessions, take-home assignments, whiteboarding system designs, etc. for SWE positions did nothing to really improve this; for SRE/DevOps roles I tried trivia questions (how are containers implemented - like what kernel technologies do they use and what do they do, how would you go about investigating why a service is consuming 100% CPU time), throwing them at broken VM's and more take-home assignments.
AI tools only make this skillset more important - I can throw Junie, Claude Code, or Copilot and small task and end up with...an implementation. But they still fuck up, constantly, and yet again, anything that's not already been done, regularly, requires a lot of guidance from an engineer in the loop. And with the god damned death of the web thanks to AI slop being posted anywhere, the ability to find answers and reason through problems is only going to become more important when these tools fail miserably for the third time in a row.
One issue that is not discussed enough when talking about learning is mental preparation for learning. We have all had days when learning seemed easier than on other days, but we didn't pay too much attention to it, or we thought that the subject we were learning was more favorable to us, or we classified it as one of the many inexplicable or unrepeatable circumstances of life.
While we understand the importance of warming up for physical activity and recognize the need for a certain aptitude for running, weightlifting, or boxing, when it comes to more intellectual activities, we often leave things to chance: sometimes we are more alert and receptive, while at other times we are less so.
Over the years, I have found enormous benefit in practicing autogenic training, a more Western and scientific version of meditative practices that today seem to arouse the interest of those who deal with these things. I am mentally more alert, more receptive, and learning, which is always challenging, is faster.
> autogenic training, a more Western and scientific version of meditative practices
Do you have any tips for learning more and getting started? I have searched a bit, but always appreciate anecdotes of those that have found success enough to speak about it.
I will probably write a book on this topic because there is not enough material in English. But the practice is fairly simple overall and only requires time and a positive attitude. Since there is a somatic response, it is always easy to tell if it is “working", especially for the "lower" exercises.
Originally, it had to be taught by MDs, according to Dr. Schultz--the inventor of the method--and his followers, but that ship has long sailed.
It is fairly easy to find a copy of Karl Rosa's book “Autogenic Training”, a good starting point, and Luthe's in-depth multi-volume analysis of autogenic training, though I would only recommend the latter to the most avid enthusiasts.
If you need to learn how to learn then you don't know how to learn. How can you learn to learn if you don't know how to learn?
Jokes aside I'm really into learning science and make youtube videos covering learning and learning papers + an ipad app. I keep a running list of my favorite learn-to-learn resources here:
If I had to recommend only one resource it would be: The ABCs of How We Learn: 26 Scientifically Proven Approaches, How They Work, and When to Use Them by Schwartz
Even though you might be joking, I think this is a very relevant point. To take it a step further, one also has to be willing to learn to learn how to learn. I wonder if intrinsic motivation for this can be nurtured, or whether it is up to chance.
I know we're all good little rational kids here, but even rationalists need to learn about emotions. Strong emotional responses are currently holding back human advancement. If you look closely at history, it has been always thus.
** "in my opinion" is always implied, unless a source is given **
Reading about airline crashes has radically changed how I view blame.
The way I was raised and the choices I made as an adult have given me a relatively rare point of view: people are made of humans, and humans are made of animals, and animals have limited capabilities.
I can explain someone's actions, or I can excuse someone's actions, and the difference is largely in the mind of the beholder.
Social punishment is micro and macro. On the macro it looks like shared morality and it feels like safety. On the micro it looks like emotional invalidation and it feels like danger and isolation.
first off it doesn't seem to be taught at the moment, but also I'm pretty sure that has always been the most important and foundational skill, and it seems like there might be an upper bound for what percentage of people can actually learn it.
We’re in dire need of this right now. The number of people that I work with who refuse to pick up new tools and technologies is astounding. If they _do_ try something new, they seem to avoid all but the most basic knowledge of whatever it is, and look at me crosseyed if I suggest going the slightest bit deeper (`git add -p` rather than `git add .`, for example).
I'm sure it varies a lot place to place but I've experienced much more of the opposite in the tech industry. I've heard countless times that we should switch to a different tool because it's newer, from someone who couldn't name a single specific way it's better than the existing tool. I see so much busy work at my employers and in products I use where things get changed just for the sake of change, without getting any better.
I feel like it's a mix of both, depending on how familiar thungs are. It's probably much easier for someone in tech to try moving to a new tool for no real reason than it is for them to learn a bit of CAD and a bit of electronics to make a widget (again, for no real reason).
Learning is a nontrivial skill even though it has historically been treated as such. It requires an embodied understanding of concepts from basic cognitive psychology, expertise theory, behavioral-affective psychology, metacognition, and more. Until people stop with the platitudes of "learning how to learn is important" and start teaching/training the subject directly as a skill that must be acquired, this will not change.
Simply showing a learner a few slides on spaced retrieval will not cut it.
> If we're lucky, LLMs force people to put more effort into assignments and grading and then that would help kids learn to learn as well.
I'm afraid it might be exactly opposite. Having all the knowledge at hand. all the time will lead to knowledge atrophy. Just like it already happens with ability to travel without navigation.
With how rapidly LLMs are improving, I don't know how you would construct assignments that can't be solved relatively quickly by a student feeding it into a bleeding-edge LLM. Especially since teachers often aren't PhDs and are overworked, the idea of every class of students getting handed brand new problems that aren't in the training set feels far-fetched.
I hope somebody figures this out but I don't know what the solution looks like.
I run a program for high schoolers to emphasize this skill. However, the entire K-University pipeline is designed around credentialism. Ie. do whatever you need to, cram/cheat/regurgitate, to get the rubber stamp. It's really hard to communicate the importance of self-directed education/learning how to learn when the vast majority of students' formal educational experiences tell them otherwise. Very frustrating but perhaps things are changing ...
School has two competing goals and this will never change:
1. Have the kids learn new things
2. Have the kids reach a desired level of competency
Learning happens where you are at, not where the teacher wants you to be. Every student is at a different place in understanding. It's impossible without 1-on-1 instruction to really maximize learning.
Competency is only determined via testing. Learning doesn't require testing at all, you can just speak to a student to get a good idea if they're making some progress, any progress. Competency? That basically demands a test, because it has a particular bar in mind.
Now students know they need to pass the bar, somehow, but the anxiety of that is going to cause issues with them just trying to learn. This is unfixable though, because the outside pressures demand students have some level of competency otherwise teachers are viewed as failures.
I agree. Imo, #2 is becoming more of an emphasis over time. Teacher don't have much time/energy to pursue #1. Eventually, most of them stop caring and rely on testing metrics because that's what the admins want.
It's amazing what kids can learn if they just spent a little bit of time with a 1-on-1 instructor/advisor. The anxiety you mentioned can be crippling and something I deal with regularly. Even some of the "gifted" kids (perhaps due to the expectations) have trouble avoiding the trap of overindexing on productivity/competency metrics. They're not even self aware of it, just accepts it as normal.
For most kids I have to go through the exercise of separating these two concerns, the learning part and the signaling part, early so they can put things in perspective.
There's also the validity of learning methods, despite what studies may claim, there's no scientific "grand theory of meta-learning", and if ideas are misapplied/misused there's a risk of falling into scientism, which would be just as harmful as economically driven credentialism. At worst it is just the austerity version of education—learn it yourself because we can't afford the school resources to teach/coach/nurture subjects.
I'd place this skill, "Learning how to Learn", with Cal Newport's notion of "Deep Work". Part of me wants to say that the latter is a precondition of the former, but I'm not sure that's the case.
A lot of what people think of as "doing stuff" relies on years or decades of training and experience. When you pay someone to maintain part of your house or to create bespoke furniture or repair your car you're not just paying for labor, you're paying for labor from someone with knowledge and equipment.
I think there's an element of curiosity that is innate, and that make a huge difference. For those so blessed, learning is fun and a welcome activity (there's an entirely separate issue of actually applying any of knowledge but at least it's a start).
That said, teachers can make all the difference in connecting with students to guide them into discovery vs. cramming facts into their heads to satisfy testing metrics.
Terrible article. Did Hassabis really say ”learning how to learn” is itself a skill, like the article paraphrases it? Surely the skill isn’t that you can learn how to learn, the skill is that you know how to learn. Just like ”learning to ride a bike” isn’t a skill; it’s something you do once, leading to a useful skill.
”Learning how to learn” sounds vaguely insightful just because of the repetition, but if you think for a bit about what it actually means it falls apart.
As a Curriculum Designer and OG dial-up Millennial I agree wholeheartedly. Too late though. Enlightenment, collaboration, and advancing human experience isn't profitable.
In other words, until one learns how to hammer a nail, it's unreasonable to assume knowledge of how to tell another to do so. AI is no exception. It's speed-running US society's final threads being severed, and okay, sigh, here we go. No, I'm not interested in fixing the problems he's identifying.
My ex had a saying from bench science..."if you're not part of the solution, you're part of the precipitant." That part. Off to go live in a van down by the river...
Join the Generalist bootcamp, it includes big picture of the world and everything, anything you ever need. Full access subscriptions at $1000.
You will Learn the following things:
Analytic philosophy,
Mathematical logic,
Pure and applied math,
Physics, CS, Systems thinking,
Engineering(Mech + electronics),
creative problem solving
And finally one art subject
Beginer Projects:
Wafer stage design.
Model nano tech projects.
Small nuclear fusion reactors.
Portable TEM machine.
Pre-req:
Just enough maturity.
You should be curious, persistence & hardworking.
We assume you will practise problem solving till you die.
Outcomes of the bootcamp:
Job guaranteed at fortune 500.
Testimonials:
We have so many happy customers working for companies having trillion dollar values.
> Muad'Dib learned rapidly because his first training was in how to learn. And the first lesson of all was the basic trust that he could learn. It is shocking to find how many people do not believe they can learn, and how many more believe learning to be difficult. Muad'Dib knew that every experience carries a lesson.
-- from "The Humanity of Muad'Dib" by the Princess Irulan
Also, The Profession by Isaak Asimov is entirely dedicated to the importance of learning.
https://classicsofsciencefiction.com/2024/04/05/profession-b...
The importance of "learning how to learn" has been emphasized by all of my teachers since I was in highschool, or maybe even 8th grade, decades ago.
My computer engineering professors also emphasized user centered design. For one of Google's top scientists to bring this up is an admission that they won't, or can't, design a good user experience for their tools.
I found that emphasis to be similar to two teams building from opposite starting points, and never meeting in the middle.
The issue with "learning to learn" is that it does not include the foundational skill of "how to communicate". Far too often it is not a lack of desire to learn, it is the inability to communicate what one is trying to learn. When seeking help, not only does the seeker have difficulty expressing their situation, those trying to help are not taught how to listen and will offer solutions to an issue only starting to be explained. This difficulty is then compounded by self conversation bias that negatively spins against the person seeking. That is two very high hurdles: negative self bias, and inarticulate communications while seeking guidance.
I remember the same thing. That doesn't mean they knew how to teach us to "learn how to learn". Neither does it mean that the underlying education system supported that goal.
Same goes for user-centered design. Trying to make something user-friendly is one thing, successfully doing it is another. Large organizations are especially poor at user-friendly design because the underlying structures which support that goal don't exist. Organizational science is still in its infancy.
The US education system only has one mode, and thats to survive in a slim way with overworked staff and huge classrooms. 40 kids in a math class is seen as normal.
Everything you see of its character, including emphasizing tests and practice, follows from that. Talking about good UX is miles away.
It's a problem that goes beyond the United States, overworked staff, and constraints in general, although these are legitimate concerns. I studied in a non-US country, but the attention paid by teachers to pedagogy was virtually zero.
I mean, we had five years of English classes in high school, and by the end of high school, less than five out of 30 people in my cohort were able to string a couple of sentences together in English. And my class was made up of serious, studious young people. It seems to me that the time was not well spent, but did the teacher, a caring and generally competent person, reflect on the poor results? I highly doubt it.
“You have to learn how to learn” has been a phrase often repeated by teachers, but I don’t remember any of them emphasizing, for example, retrieval practice: you learn a skill or subject, move on to the next one, and leave it up to fate whether you remember anything from the first one.
It always surprises and saddens me that, despite having been an excellent student throughout my years of education, I remember practically nothing about 90% of the subjects I studied.
> One thing we'll know for sure is you're going to have to continually learn ... throughout your career," he said.
This has been the case for literally my entire career and I assume most of the professional world for the last half century.
Technology is continually reshaping industries and while many eschew learning and adopting, those who embrace it are the ones who succeed best IME.
> Technology is continually reshaping industries and while many eschew learning and adopting, those who embrace it are the ones who succeed best IME.
I said it already in a reply to GP, but I'm going to say it again: I stopped caring about what people list on their resumes, your work history and education don't matter to me. I'd rather hire a hungry junior that finished a bootcamp, that has a drive and ability to absorb new things and adapt to changing environments, over somebody who's got 10 years of experience and can't do shit outside of their comfort zone.
The number of people who aren't able to learn and adapt to changing times, new tools, new ways of working, etc. is shocking.
There is a certain amount of "use it or lose it" seemingly inherent to virtually every human endeavor. But I suspect if you were to enroll in a class in any of those subjects, you'd perform radically better than a peer who'd never studied them. IOW, there's often more latent memory than we realize or can easily retrieve.
"IOW, there's often more latent memory than we realize or can easily retrieve." -
I did not find it to be true almost at all, and I tested many other people on it. When I voiced my concerns, the usual answer was, "Yes, but when you pick up a book, you will remember". And then I asked, "Try it", and the subsequent answer was, "I have to admit you are right".
The "re-absorption speed" is heavily confounded by general IQ and the kind of cognitive stimulation one receives in daily life, but the original learnings are mostly gone. Among other things, this is why retrieval practice is important: it slows down the "forgetting rate".
Thing is, this has nothing to do with UX or even AI at the end of the day. Over the years I have adjusted the way I handle interviews when I'm on a hiring panel to focus on critical thinking, problem solving, and lifelong learning above all else - because they are the #1 indicator that a hire is going to be successful in a position, even if that means they need a lot more training before they become productive.
I can teach somebody who finished a 6-month coding bootcamp Go, all the internal tooling, go over the business with them, etc. if they have these skills and end up with a productive mid-level engineer who gets shit done in a few years. What I can't teach is the drive and ability to learn, that's a much longer process and if you don't already have it then I'm not prepared to develop it.
Hell, outside of looking for signs of obvious bullshit I stopped giving a shit about resumes. Your work history does me no good, your education doesn't matter to me, and your references are useless beyond making sure you aren't straight up lying to me about your employment history. Every single time I have hired somebody who has 5 years of "experience" working with technologies I bullet pointed on a JD they ended up fumbling the moment they had to do something new. Doing leetcode, pair programming sessions, take-home assignments, whiteboarding system designs, etc. for SWE positions did nothing to really improve this; for SRE/DevOps roles I tried trivia questions (how are containers implemented - like what kernel technologies do they use and what do they do, how would you go about investigating why a service is consuming 100% CPU time), throwing them at broken VM's and more take-home assignments.
AI tools only make this skillset more important - I can throw Junie, Claude Code, or Copilot and small task and end up with...an implementation. But they still fuck up, constantly, and yet again, anything that's not already been done, regularly, requires a lot of guidance from an engineer in the loop. And with the god damned death of the web thanks to AI slop being posted anywhere, the ability to find answers and reason through problems is only going to become more important when these tools fail miserably for the third time in a row.
One issue that is not discussed enough when talking about learning is mental preparation for learning. We have all had days when learning seemed easier than on other days, but we didn't pay too much attention to it, or we thought that the subject we were learning was more favorable to us, or we classified it as one of the many inexplicable or unrepeatable circumstances of life.
While we understand the importance of warming up for physical activity and recognize the need for a certain aptitude for running, weightlifting, or boxing, when it comes to more intellectual activities, we often leave things to chance: sometimes we are more alert and receptive, while at other times we are less so.
Over the years, I have found enormous benefit in practicing autogenic training, a more Western and scientific version of meditative practices that today seem to arouse the interest of those who deal with these things. I am mentally more alert, more receptive, and learning, which is always challenging, is faster.
> autogenic training, a more Western and scientific version of meditative practices
Do you have any tips for learning more and getting started? I have searched a bit, but always appreciate anecdotes of those that have found success enough to speak about it.
I will probably write a book on this topic because there is not enough material in English. But the practice is fairly simple overall and only requires time and a positive attitude. Since there is a somatic response, it is always easy to tell if it is “working", especially for the "lower" exercises.
Originally, it had to be taught by MDs, according to Dr. Schultz--the inventor of the method--and his followers, but that ship has long sailed.
It is fairly easy to find a copy of Karl Rosa's book “Autogenic Training”, a good starting point, and Luthe's in-depth multi-volume analysis of autogenic training, though I would only recommend the latter to the most avid enthusiasts.
If you need to learn how to learn then you don't know how to learn. How can you learn to learn if you don't know how to learn?
Jokes aside I'm really into learning science and make youtube videos covering learning and learning papers + an ipad app. I keep a running list of my favorite learn-to-learn resources here:
https://www.ahmni.app/blog/learn-to-learn-resource-list
If I had to recommend only one resource it would be: The ABCs of How We Learn: 26 Scientifically Proven Approaches, How They Work, and When to Use Them by Schwartz
Even though you might be joking, I think this is a very relevant point. To take it a step further, one also has to be willing to learn to learn how to learn. I wonder if intrinsic motivation for this can be nurtured, or whether it is up to chance.
Long has been.
Best bit of career advice I ever got, back in the 90s: "Get really good at the help system".
(At the time, it was MSDN DVDs).
I know we're all good little rational kids here, but even rationalists need to learn about emotions. Strong emotional responses are currently holding back human advancement. If you look closely at history, it has been always thus.
** "in my opinion" is always implied, unless a source is given **
Reading about airline crashes has radically changed how I view blame.
The way I was raised and the choices I made as an adult have given me a relatively rare point of view: people are made of humans, and humans are made of animals, and animals have limited capabilities.
I can explain someone's actions, or I can excuse someone's actions, and the difference is largely in the mind of the beholder.
Social punishment is micro and macro. On the macro it looks like shared morality and it feels like safety. On the micro it looks like emotional invalidation and it feels like danger and isolation.
Leaning how to think critically is what the next genertion's most needed skill SHOULD be
first off it doesn't seem to be taught at the moment, but also I'm pretty sure that has always been the most important and foundational skill, and it seems like there might be an upper bound for what percentage of people can actually learn it.
We’re in dire need of this right now. The number of people that I work with who refuse to pick up new tools and technologies is astounding. If they _do_ try something new, they seem to avoid all but the most basic knowledge of whatever it is, and look at me crosseyed if I suggest going the slightest bit deeper (`git add -p` rather than `git add .`, for example).
I'm sure it varies a lot place to place but I've experienced much more of the opposite in the tech industry. I've heard countless times that we should switch to a different tool because it's newer, from someone who couldn't name a single specific way it's better than the existing tool. I see so much busy work at my employers and in products I use where things get changed just for the sake of change, without getting any better.
I feel like it's a mix of both, depending on how familiar thungs are. It's probably much easier for someone in tech to try moving to a new tool for no real reason than it is for them to learn a bit of CAD and a bit of electronics to make a widget (again, for no real reason).
Xlation: Classic Corporate America Condescension.
Trust me if google can do something anyone can. They are trying to "define" what "they" "want" from a "compliant workforce"
Learning is a nontrivial skill even though it has historically been treated as such. It requires an embodied understanding of concepts from basic cognitive psychology, expertise theory, behavioral-affective psychology, metacognition, and more. Until people stop with the platitudes of "learning how to learn is important" and start teaching/training the subject directly as a skill that must be acquired, this will not change.
Simply showing a learner a few slides on spaced retrieval will not cut it.
"Teach to the test" started this a while ago - memorizing and then forgetting bullet points vs engaging more deeply with a subject.
If we're lucky, LLMs force people to put more effort into assignments and grading and then that would help kids learn to learn as well.
> If we're lucky, LLMs force people to put more effort into assignments and grading and then that would help kids learn to learn as well.
I'm afraid it might be exactly opposite. Having all the knowledge at hand. all the time will lead to knowledge atrophy. Just like it already happens with ability to travel without navigation.
With how rapidly LLMs are improving, I don't know how you would construct assignments that can't be solved relatively quickly by a student feeding it into a bleeding-edge LLM. Especially since teachers often aren't PhDs and are overworked, the idea of every class of students getting handed brand new problems that aren't in the training set feels far-fetched.
I hope somebody figures this out but I don't know what the solution looks like.
I run a program for high schoolers to emphasize this skill. However, the entire K-University pipeline is designed around credentialism. Ie. do whatever you need to, cram/cheat/regurgitate, to get the rubber stamp. It's really hard to communicate the importance of self-directed education/learning how to learn when the vast majority of students' formal educational experiences tell them otherwise. Very frustrating but perhaps things are changing ...
School has two competing goals and this will never change:
1. Have the kids learn new things 2. Have the kids reach a desired level of competency
Learning happens where you are at, not where the teacher wants you to be. Every student is at a different place in understanding. It's impossible without 1-on-1 instruction to really maximize learning.
Competency is only determined via testing. Learning doesn't require testing at all, you can just speak to a student to get a good idea if they're making some progress, any progress. Competency? That basically demands a test, because it has a particular bar in mind.
Now students know they need to pass the bar, somehow, but the anxiety of that is going to cause issues with them just trying to learn. This is unfixable though, because the outside pressures demand students have some level of competency otherwise teachers are viewed as failures.
I agree. Imo, #2 is becoming more of an emphasis over time. Teacher don't have much time/energy to pursue #1. Eventually, most of them stop caring and rely on testing metrics because that's what the admins want.
It's amazing what kids can learn if they just spent a little bit of time with a 1-on-1 instructor/advisor. The anxiety you mentioned can be crippling and something I deal with regularly. Even some of the "gifted" kids (perhaps due to the expectations) have trouble avoiding the trap of overindexing on productivity/competency metrics. They're not even self aware of it, just accepts it as normal.
For most kids I have to go through the exercise of separating these two concerns, the learning part and the signaling part, early so they can put things in perspective.
There's also the validity of learning methods, despite what studies may claim, there's no scientific "grand theory of meta-learning", and if ideas are misapplied/misused there's a risk of falling into scientism, which would be just as harmful as economically driven credentialism. At worst it is just the austerity version of education—learn it yourself because we can't afford the school resources to teach/coach/nurture subjects.
I'd place this skill, "Learning how to Learn", with Cal Newport's notion of "Deep Work". Part of me wants to say that the latter is a precondition of the former, but I'm not sure that's the case.
This is every generation's .ort needed skill
One of my favorite pieces on this topic is this talk "Stop Treading Water: Learning to Learn":
https://www.youtube.com/watch?v=j0XmixCsWjs
Google's top AI scientist says 'learning how to learn' will be next generation's most needed skill
https://techxplore.com/news/2025-09-google-ai-scientist-gene...
I think all knowlwdge "work" is bust. Selling knowlwdge will be like selling knitted socks.
People who do stuff will make money
A lot of what people think of as "doing stuff" relies on years or decades of training and experience. When you pay someone to maintain part of your house or to create bespoke furniture or repair your car you're not just paying for labor, you're paying for labor from someone with knowledge and equipment.
Feels like a bit of a useless cliche. Being able to learn has far more to do with iq than any pedagogical technique.
I think there's an element of curiosity that is innate, and that make a huge difference. For those so blessed, learning is fun and a welcome activity (there's an entirely separate issue of actually applying any of knowledge but at least it's a start).
That said, teachers can make all the difference in connecting with students to guide them into discovery vs. cramming facts into their heads to satisfy testing metrics.
Learning how to learn is the most important skill to acquire in school. If I may add, learning how to learn effectively is a lifetime journey.
Terrible article. Did Hassabis really say ”learning how to learn” is itself a skill, like the article paraphrases it? Surely the skill isn’t that you can learn how to learn, the skill is that you know how to learn. Just like ”learning to ride a bike” isn’t a skill; it’s something you do once, leading to a useful skill.
”Learning how to learn” sounds vaguely insightful just because of the repetition, but if you think for a bit about what it actually means it falls apart.
Imo it's simple, just learn by doing and making stuff.
As a Curriculum Designer and OG dial-up Millennial I agree wholeheartedly. Too late though. Enlightenment, collaboration, and advancing human experience isn't profitable.
In other words, until one learns how to hammer a nail, it's unreasonable to assume knowledge of how to tell another to do so. AI is no exception. It's speed-running US society's final threads being severed, and okay, sigh, here we go. No, I'm not interested in fixing the problems he's identifying.
My ex had a saying from bench science..."if you're not part of the solution, you're part of the precipitant." That part. Off to go live in a van down by the river...
/s !
Future internet road maps be like:
Join the Generalist bootcamp, it includes big picture of the world and everything, anything you ever need. Full access subscriptions at $1000.
You will Learn the following things:
Analytic philosophy, Mathematical logic, Pure and applied math, Physics, CS, Systems thinking, Engineering(Mech + electronics), creative problem solving And finally one art subject
Beginer Projects: Wafer stage design. Model nano tech projects. Small nuclear fusion reactors. Portable TEM machine.
Pre-req: Just enough maturity. You should be curious, persistence & hardworking. We assume you will practise problem solving till you die.
Outcomes of the bootcamp: Job guaranteed at fortune 500.
Testimonials: We have so many happy customers working for companies having trillion dollar values.
Horror of horrors. Please no.