The problem with education is not that we are learning too much. It is that we have confused remembering information with becoming capable.
There is a strange contradiction at the centre of modern education.
Outside the classroom, the world is changing almost violently fast. AI models improve in months. Entire industries appear and disappear within a few years. A teenager with a laptop can build software, publish research, start a company, make a film, learn calculus, design hardware, or reach millions of people without waiting for anyone's permission.
Inside the classroom, we can still spend weeks preparing to reproduce information on a piece of paper from memory.
That difference bothers me more every year.
Not because mathematics is useless. Not because chemistry is useless. Not because universities are useless.
Knowledge matters enormously.
But knowing something and being capable of doing something with it are not the same thing.
And I think education has spent far too long rewarding the first while claiming it produces the second.
We became extraordinarily good at passing exams
Give a student a chapter.
Give them a textbook.
Tell them exactly which portions matter.
Give them a collection of previous-year questions.
Teach them which formulas correspond to which patterns.
Test them under time pressure.
Rank everyone.
Repeat.
If the student becomes exceptionally good at this game, we call them exceptionally educated.
But what exactly have we measured?
Sometimes we have measured understanding.
Sometimes discipline.
Sometimes intelligence.
Sometimes the ability to remain focused under pressure.
And sometimes we have simply measured who became best at the examination itself.
Those things overlap, but they are not identical.
India's own National Education Policy acknowledges this problem. It explicitly calls for moving away from a culture of rote learning toward conceptual understanding, critical thinking, inquiry, application and experiential learning.
That matters because this criticism is not some rebellious argument that students should never study textbooks again. Even the people designing the system recognize that memorization cannot remain its centre of gravity.
The problem is that changing a policy document is much easier than changing an incentive system.
As long as one examination can dramatically alter someone's educational options, students will optimize for that examination.
Parents will optimize for it.
Schools will optimize for it.
Coaching institutes will optimize for it.
And eventually, childhood itself begins optimizing for it.
The problem is that changing a policy document is much easier than changing an incentive system. As long as one examination can dramatically alter someone's options, childhood itself begins optimizing for it.
The world changed. The syllabus did not change at the same speed.
The strongest defence of traditional education is that fundamentals do not expire.
That is true.
Newton's laws did not become obsolete because ChatGPT exists. Organic chemistry does not stop mattering because software engineers are paid well. Reading, mathematics and scientific reasoning remain foundational abilities.
The mistake is assuming that because fundamentals remain valuable, the way we teach and evaluate them must remain unchanged.
The labour market is moving much faster than that.
The World Economic Forum's 2025 employer survey estimated that about 39% of workers' existing skill sets would change or become outdated by 2030. AI and big data, technological literacy and cybersecurity were among the fastest-growing technical skills, while creative thinking, resilience, curiosity and lifelong learning were also rising in importance.
The OECD makes a similar point from a different direction: skill requirements are evolving faster than many policy cycles, making adaptive problem solving and lifelong learning increasingly important.
That changes what education should prepare someone for.
A school system designed around the assumption that you learn for twenty years and then apply that knowledge for the next forty makes less sense when the tools used in an industry can transform several times during a career.
The most valuable ability may increasingly be something much harder to examine:
Can you learn something you were never taught?
A school system designed around the assumption that you learn for twenty years and then apply that knowledge for the next forty makes less sense when the tools transform several times in a single decade.
Four years can be either an investment or an extraordinarily expensive waiting room
I do not believe college is inherently a waste of time.
For medicine, scientific research, law, engineering disciplines involving safety-critical physical systems and many other fields, rigorous formal education can be indispensable.
Universities can also give people brilliant professors, laboratories, collaborators, friendships and intellectual exposure that would be difficult to reproduce alone.
But "college is valuable" and "every four-year degree is automatically valuable" are very different claims.
If someone spends four years attending lectures, optimizing for grades, completing predictable assignments and leaving without ever having built, researched, sold, written, designed, experimented or worked seriously on something outside a rubric, there is a real possibility that the credential has grown more than the person has.
A degree should not be four years spent preparing to eventually begin doing real things.
The real things should begin during the education.
A computer-science student should build software.
An economics student should work with messy real datasets.
An aspiring journalist should report.
An engineer should prototype.
A designer should design for actual users.
A biologist should encounter experimentation rather than only diagrams of experiments.
Failure should happen while learning, when the cost of failure is low.
Instead, we often create environments in which making a mistake costs marks, so students learn the safest possible behaviour:
do exactly what will be evaluated.
Then we wonder why people become afraid of uncertain problems.
Instead of allowing students to fail when the cost of failure is low, we make every mistake cost markstraining people to do only what is evaluated, and wonder why they fear uncertain problems.
Computer science exposes the absurdity particularly well
I see this around me constantly.
A surprising number of students want computer science without actually wanting to write software.
They want computer science because they have heard that it pays well.
That creates a strange pipeline.
Someone can spend years preparing intensely for physics, chemistry and mathematics examinations, earn an excellent rank, enter computer science, and only then seriously discover whether they enjoy programming.
Meanwhile, another teenager may have spent those same years building software, maintaining servers, contributing to open source, breaking things, debugging them at 2 a.m., learning databases, deploying applications and discovering through experience that this is what they genuinely love doing.
Yet a traditional engineering entrance examination is not designed to measure any of that.
That does not make the examination meaningless. Mathematics and scientific reasoning are legitimate indicators of academic preparation, and standardized examinations have advantages: they can be administered at enormous scale and are less dependent on subjective judgments about a student's personality or background.
But they answer a particular question.
They do not answer every question.
A physics-chemistry-mathematics examination can tell us something about academic performance in those subjects.
It cannot tell us whether someone loves building software.
We should stop pretending those are equivalent measurements.
A physics-chemistry-mathematics examination can tell us something about academic performance in those subjects. It cannot tell us whether someone loves building software.
Then there is the uncomfortable question of fairness
Whenever admissions in India are discussed, reservation enters the conversation almost immediately.
It deserves more precision than it usually gets.
India's system of affirmative action in centrally covered educational institutions includes separate provisions for Scheduled Castes, Scheduled Tribes and Other Backward Classes, with OBC benefits applying to the non-creamy layer. The Central Educational Institutions Act provides 15% reservation for SC students, 7.5% for ST students and 27% for OBC students in covered institutions; a later policy created up to 10% EWS reservation for economically weaker students who are outside the existing SC/ST/OBC reservation system.
So the argument is more complicated than saying that "half the seats are inherited by the same families forever." OBC reservation already contains a non-creamy-layer restriction, and EWS operates under a different economic framework. Open seats are also not a category reserved exclusively for candidates from the general category.
At the same time, there is a legitimate public debate over how affirmative-action systems should evolve as circumstances change: how accurately disadvantage is identified, whether benefits are reaching the people within eligible communities who need them most, how socioeconomic mobility across generations should matter, and how opportunity can be expanded without turning admissions into a zero-sum war between teenagers.
Those are difficult questions.
They deserve evidence rather than slogans from either side.
And there is an even larger problem that gets lost whenever we fight entirely over who receives the scarce seats:
Why are life-changing opportunities so scarce in the first place?
If millions of young people believe their future depends on entering a tiny collection of institutions, then allocation matters enormouslybut scarcity itself has already become a failure of the system.
The ambition should not merely be deciding who gets through the narrow gate.
It should also be building more gates.
If millions of young people believe their entire future depends on entering a tiny collection of institutions, allocation mattersbut scarcity itself is the real structural failure.
Somewhere along the way, "What do you love?" became "What pays?"
Ask a child what they want to become and, initially, the answers can be wonderfully irrational.
Astronaut.
Artist.
Scientist.
Footballer.
Inventor.
Writer.
Chef.
Then reality begins negotiating with them.
Which stream has scope?
Which degree has placements?
What package does this career get?
Which profession is safe?
Eventually, an enormous number of people converge on a small collection of respectable answers.
Engineer.
Doctor.
Finance.
Consulting.
Software.
Not necessarily because they love those things, but because those paths appear legible.
I do not think money is irrelevant. That would be easy to say and ridiculous to believe.
Financial security matters. Supporting a family matters. Economic mobility matters. Someone without a safety net cannot treat career choice like a philosophical experiment.
But there is a difference between considering money and allowing salary rankings to make your entire decision for you.
If thousands of people pursue software only because software salaries became attractive, eventually you produce a strange economy full of people who succeeded at entering their profession but never particularly wanted the profession itself.
There is an enormous human cost hidden inside that sentence.
You can earn the credential.
You can earn the salary.
You can earn the promotion.
And still wake up one day wondering why you constructed a life you do not particularly enjoy living.
You can earn the credential. You can earn the salary. You can earn the promotion. And still wake up one day wondering why you constructed a life you do not particularly enjoy living.
Education should create agency
I keep returning to one question:
What should an educated eighteen-year-old actually be able to do?
Not which chapters should they have completed.
Not which formulas should they be able to reproduce.
What should they be capable of?
I would want them to be able to read something difficult and understand it. To communicate an idea clearly. To reason mathematically. To distinguish evidence from confidence. To use modern tools without becoming dependent on them. To learn independently. To collaborate. To make something from nothing. To understand money. To ask good questions. To recover from being wrong. To explore several kinds of work before choosing one. To know enough science, history and philosophy to understand the civilization they have inherited.
And above everything else:
to leave education knowing how to continue educating themselves.
That does not require destroying schools.
It requires changing what success inside them looks like.
Imagine graduating with more than a marksheet.
Projects you attempted.
Things you built.
Research you conducted.
Problems you solved.
Teams you worked with.
Internships and apprenticeships.
Essays in which you developed an argument rather than reproduced one.
Experiments that failed.
Evidence of what you can actually do.
The classroom should not disappear.
It should connect to reality.
Imagine graduating with more than a marksheet: projects attempted, software built, research conducted, problems solved. Evidence of what you can actually do.
AI makes this problem impossible to ignore
For most of history, possessing information was enormously valuable because obtaining information was expensive.
Libraries mattered because books were scarce.
Lectures mattered because expert explanation was scarce.
Memorization mattered more because external knowledge was not permanently sitting inside your pocket.
That world is gone.
A student can now ask an AI system to explain thermodynamics five different ways, simulate a tutor, generate practice problems, translate research, criticize an essay or help debug a program in seconds.
That does not make learning obsolete.
It makes shallow learning easier to expose.
If a machine can recall the fact instantly, the human advantage moves further toward knowing whether the fact is relevant, whether it is correct, how it connects to other facts, what to do with it and what question should be asked next.
AI should make education more ambitious, not less.
We can finally spend less time treating students like storage devices.
So why are we still grading them as if memory were the scarce resource?
We can finally spend less time treating students like storage devices. Why are we still grading them as if memory were the scarce resource?
I am not arguing for easier education
Quite the opposite.
I think education should become harder.
Just hard in a different way.
It is easy to memorize the definition of entrepreneurship.
Try selling something.
It is easy to learn the steps of the scientific method.
Try designing an experiment whose answer you do not already know.
It is easy to study communication.
Try convincing a room full of people who disagree with you.
It is easy to answer a programming question whose solution exists in the back of a book.
Try maintaining software after actual users begin breaking it in ways you never predicted.
Reality is an extraordinarily difficult examination.
It is open-book.
You can use the internet.
You can use AI.
You can ask other people.
There is no syllabus.
And somehow, even with all those advantages, it is much harder than school.
That should tell us something.
Reality is an open-book exam: you can use the internet, AI, and ask other people. There is no syllabus. And somehow, it is vastly harder than school.
The purpose of education cannot merely be employment
There is one final trap in this entire argument.
It would be easy to replace one narrow idea of education with anotherto say schools should stop producing exam-takers and start producing employees.
That would still be too small.
Human beings are not workforce inputs.
Education should help someone understand the world, participate in society, discover what they care about, develop judgment, appreciate things that have no immediate commercial value, and become difficult to manipulate.
Poetry matters even when nobody hires you to read poetry.
History matters even when it does not increase your salary.
Pure mathematics matters before anyone knows what it will eventually be useful for.
Philosophy matters because some questions should be considered even when they cannot be monetized.
The answer to an excessively academic education system is not turning childhood into corporate training.
The answer is creating people who are both knowledgeable and capable.
The world does not need another generation trained merely to predict what answer the examiner expects.
It needs people who can encounter a problem nobody prepared them for and begin anyway.
People who can learn.
People who can build.
People who can change their minds.
People who can choose work for reasons deeper than fear.
People whose education gave them more possible lives rather than funneling them toward one approved definition of success.
Maybe that is the standard we should use.
Not:
How much did you memorize?
Not:
Which college admitted you?
Not even:
What package did you get?
But:
What can you understand, what can you create, and what kind of life are you now capable of building?
Because education was never supposed to teach us how to spend the first twenty years of life preparing for the rest of it.
It was supposed to teach us how to live the rest of it.