I believe almost anyone can learn almost anything. I believe it because I have watched it happen, including to myself. I picked up skiing at thirty-five, badly at first, then less badly. I did a Master's in artificial intelligence in my thirties while holding down a full-time job in a country that was not my own. None of that came from talent. It came from being willing to be bad at something for a while, and to stay in that uncomfortable place long enough to come out the other side.
So I have a complicated feeling about a technology whose entire promise is to remove the discomfort.
AI is extraordinary at handing you the answer. The code, the explanation, the summary, the first draft — instantly, and often well. I use it every day and I am not about to pretend that is anything but useful. But I keep circling a quieter worry, and I would rather say it out loud than tidy it away: the struggle to figure something out is not a bug. It is how we become capable. And we are building tools that are very good at removing it.
Think about how you actually learned the things you know deeply. Not the things you can recite — the things you understand, in your hands. I would bet none of them came easily. You got stuck. You tried the wrong thing. You sat with a problem long enough that something rearranged in your head, and afterwards you were different. That sitting-with, that productive frustration, is the whole mechanism. Take it away and you are left with the answer but not the understanding — which feels the same right up until the moment you need to do it yourself.
I am not making the tired argument that the new tool will rot our brains. People said it about calculators, about the printing press, about writing itself, and they were mostly wrong. The calculator did not make us worse at thinking; it freed us from arithmetic so we could think about more. The honest position is not "this is bad." It is "this is a trade, and trades have terms, and we should read them."
So here is the question I find actually useful: which struggles are worth keeping?
Some struggle is just friction, and removing it is pure gain. I do not need to handwrite a letter to value its contents. Nobody mourns the struggle of doing long division by hand. But some struggle is not friction — it is the thing itself. The struggle to make an argument is how you come to believe it. The struggle to debug your own code is how you learn to write better code. The struggle to phrase a hard idea is, quite often, the process of actually having the idea. Outsource those and you have not saved time. You have skipped the part that was the point.
I do not think there is a clean rule here, and I distrust anyone who offers one. What I have is a habit I am trying to build, in myself first. Before I let a tool do the hard part, I ask whether the hard part was where the learning lived. If it was, sometimes I do it myself anyway, slower, on purpose — not out of nostalgia, but because I would like to still be capable next year. And sometimes I let the tool take it, gladly, because it was just friction and I have better things to struggle with.
The thing I most want for my son, and honestly for myself, is not a life with less struggle. It is the judgement to choose the right struggles, and the appetite to stay in them. I worry that a world optimised to remove all discomfort will, with the kindest intentions, quietly remove that appetite too.
I do not know how this plays out. I am not even sure I am asking the question well. But I would rather build, and use, these tools as someone who is awake to the trade than as someone who only sees the convenience. The answer is cheap now. Understanding is not. I would like to keep paying for the one that lasts.