Somewhere right now, as you read this, a student is staring at a physics problem that will not yield. It is late, the desk lamp is throwing its small circle of light across a page of scribbled attempts, and the student is alone with the particular, grinding, magnificent frustration of a mind that has met its match. And a few centimetres away, on a phone or a laptop, there is a tool that could solve the problem in seconds. A large language model, an artificial intelligence, a system trained on more text than any human could read in a thousand lifetimes. The student could photograph the question, feed it to the machine, and receive, in the time it takes to exhale, a complete and polished solution. Every step explained. Every sign convention honoured. The answer at the bottom, neat as a receipt. The student could do this. And the question that defines the Olympiad in our age, the question that every student and every teacher and every parent must now answer, is whether the student should.

Let us be honest about what these machines can do, because pretending otherwise is a fool's errand, and the students of this generation are not fools. An artificial intelligence can solve most textbook problems faster than any human. It can recite the derivation of the Lorentz transformation, generate practice problems on demand, explain a concept in six different ways until one of them lands, translate a Russian problem collection into English, and check an algebraic manipulation for errors. It is, by any measure, the most powerful learning tool ever placed in a student's hands. It is patient beyond any teacher's patience, available at three in the morning, never tired, never irritated, never too busy. To pretend that it is not genuinely useful would be dishonest. To refuse to use it at all would be, in its way, as foolish as refusing to use a calculator because your grandparents did arithmetic by hand.
And yet. There is a difference between a problem being solved and a problem being learned, and this difference is the entire point of the Olympiad. When the machine solves the problem for you, the problem is gone. The answer exists, correct and complete, and you have watched it arrive, and something has been accomplished, but it was not accomplished by you. The understanding did not pass through your hands. The confusion, the false starts, the three wrong approaches that taught you why the right approach is right, the moment of seeing the structure for the first time with your own eyes, none of that happened. You received the destination without the journey, and the journey was the thing that would have changed you. The Olympiad was never, at its heart, about producing correct answers. If it were, we would have handed the medals to the machines long ago. The Olympiad is about what the struggle to find the answer does to the person doing the struggling.

There is something the machine cannot do, and it is worth naming clearly, because it is the thing that matters most. The machine cannot be confused in the way that matters. It does not experience the productive, generative, essential confusion of a mind that is in the middle of building an understanding it does not yet have. When you are stuck on a problem, genuinely stuck, your brain is doing something extraordinary. It is holding the pieces of the problem in suspension, testing them against one another, searching through everything it knows for a structure that fits. This state, this uncomfortable, effortful, sometimes maddening state, is not an obstacle to learning. It is learning. Neuroscience has a word for it, desirable difficulty, and it describes the finding that the harder your mind has to work to retrieve or construct an understanding, the more deeply that understanding is encoded, the longer it lasts, the more flexibly it can be used. The machine removes the difficulty. In doing so, it removes the very thing that builds the mind. It is the difference between being carried up a mountain and climbing it. The view from the top is the same. The person who arrives is not.
Consider what an Olympiad problem actually is. It is not an exercise in applying a known formula to a known situation. If it were, the machine would be unbeatable, because it knows every formula and has seen every situation. An Olympiad problem is a small act of creation in reverse. Someone, a problem setter, a physicist, a teacher, looked at the laws of nature and found a configuration, a setup, a way of arranging the physical world that produces a surprise, a hidden symmetry, an unexpected simplification, a place where two ideas that seem unrelated turn out to be the same idea in disguise. And then they hid that insight inside a problem and handed it to you. Solving it is not retrieval. It is discovery. It is the act of seeing, for the first time, a connection that you were not told about, that you could not have looked up, that you have to construct inside your own mind out of raw understanding and intuition and persistence. This is the act that no machine can do for you, because it is not a product that can be delivered. It is an experience that can only be undergone.

This is not an argument against artificial intelligence. Let us be clear about that, because the world has enough luddites and the future belongs to those who can work alongside these tools rather than against them. The student who learns to use an AI well, who uses it to explain a concept they have wrestled with, to generate practice problems at the edge of their ability, to check a solution they have already produced by their own effort, to explore a what-if question that occurred to them, that student has a genuine advantage over the student who refuses the tool entirely. The machine is a magnificent sparring partner, a tireless tutor, an infinite library. Used well, it can accelerate learning enormously. The danger is not the tool. The danger is using the tool in the one way that guarantees no learning happens at all, which is to ask it to do the thinking for you, and then to mistake its answer for your understanding.
And there is a deeper reason why the human element remains irreplaceable, a reason that goes beyond pedagogy and into the nature of physics itself. Physics is not a collection of answers. Physics is a practice, a discipline, a way of attending to the world. It is the practice of looking at a phenomenon and asking what must be true for this to happen, of building a model, of testing the model against reality, of holding your most elegant idea up to the light and being willing to find it wrong. This practice requires something the machine does not have, which is a stake in the outcome. You care whether you understand. The confusion bothers you personally. The elegance of a good solution pleases you. The physical world surprises you, delights you, humbles you. These are not incidental emotions layered on top of the intellectual work. They are the engine of the intellectual work. They are why you keep going when the problem is hard. The machine has no stake. It does not care whether its answer is beautiful or whether it corresponds to reality. It produces, and it moves on. The caring is what makes you a physicist, and the caring cannot be automated.

So how does a student navigate this? How do you prepare for an Olympiad in an age when the answer to every question is one prompt away? You navigate it the way every generation has navigated a new tool, by deciding what the tool is for and what you are for. You use the machine to learn, never to avoid learning. You wrestle with the problem yourself first, always, for as long as you can bear it and then a little longer, because the wrestling is where the muscle is built. You let the machine explain, but you do not let it answer. You ask it to teach you the method, and then you close it and do the problem yourself, by hand, on paper, with your own mind, the way it will have to be done in the examination hall where no machine is allowed and no machine can help you. You use it to check your work after you have done your work, not to produce the work you were supposed to do. In short, you treat the machine the way a serious musician treats a recording of a great performance. You listen to it to understand what is possible. But you do not mistake listening for playing. The playing is yours. It has always been yours. And the playing is the point.
There is a vision some people have of the future, a vision in which the machines are so capable that human expertise no longer matters, in which there is no point learning to do by hand what an algorithm can do instantly. And in some domains, for some tasks, that vision may be correct. But the Olympiad is not about tasks. It is about the formation of a particular kind of mind, a mind that can look at the unfamiliar and make it familiar, that can hold complexity without collapsing it, that can persist through confusion toward clarity, that can find the hidden structure beneath the apparent chaos. These are not skills that become obsolete when a machine can perform the task. These are the skills that allow a person to direct the machine, to judge its output, to know when it is wrong, to ask it the questions that matter. The students who build these minds will not be replaced by artificial intelligence. They will be the ones who understand it, who build it, who decide what it is for. The students who skip the struggle and collect the answers will find, when the real problems arrive, the ones that have no solution manual and no prompt that quite fits, that they have nothing inside them to fall back on.

And here is the thing that I want you to hold onto, the thing that matters more than any argument about tools and methods and the future of work. The reason to struggle with a physics problem is not, in the end, to become better at solving physics problems. It is to discover what you are made of. It is to learn that you can sit with difficulty and not flee from it. It is to learn that confusion is not a verdict on your intelligence but a sign that you are at the edge of your understanding, which is the only place where growth happens. It is to learn the difference between the quick hit of an answer handed to you and the slow, earned, durable satisfaction of an understanding you built yourself. These are lessons about yourself, and they can only be learned by doing, and they will serve you in every domain of your life, long after the specific physics has faded. The machine can give you the answer. It cannot give you the person you become in the process of finding it. That person is built in the struggle, and the struggle is yours, and no technology ever invented can take it from you unless you hand it over.
So to the student reading this, the one with the desk lamp and the problem that will not yield and the machine humming quietly a few centimetres away, here is what I would say. Do not photograph the problem. Not yet. Sit with it a little longer. Let it be hard. Let it frustrate you. Try the wrong approach and learn from it, and try another. Feel the pieces of it moving in your mind, feel the moment, because it will come, when two ideas that seemed separate suddenly connect and the problem cracks open and the solution pours through, and you did that, you, with nothing but your own prepared and persistent mind. That feeling is what the machine can never give you. That feeling is what you are here for. And when you have found it, when you have built the answer out of your own understanding, then, if you like, ask the machine to check your work. It will confirm what you already know. And you will not need it to. Because you did the thing. You climbed the mountain. And the view, the view is yours, and it always will be.











































































