I sit on a school board, which means I am occasionally asked what children should be learning, and I have been giving an answer I no longer believe.
The answer was: learn to code. Learn the technical thing, because the technical thing is where the leverage is. I have said this at dinner tables, to nephews, to parents who wanted reassurance, and I said it with real confidence because for twenty five years it was simply true. It was true for me.
I now think it was half right in a way that is worse than being wrong, because half right advice gets followed.
Here is what changed my mind, and it is not a prediction about models. It is a piece of arithmetic about time. A child starting school this year finishes in 2038. That is not before the robot ramp and it is not comfortably after it. That is directly in the middle of it, in the messiest part, when the machines are capable and expensive and not yet everywhere. We are not preparing children for the age of abundance. We are preparing them for the chaos that comes first, and those need different preparation entirely.
The first thing automated is the first job
Think about how anybody becomes good at anything. A junior lawyer reads five hundred contracts and slowly develops a nose for the one clause that will cause trouble in year three. A junior developer fixes a hundred small bugs and gradually builds a map of how systems actually fail as opposed to how the documentation says they fail. A junior architect sits in meetings, mostly silent, watching how a decision gets made and who was really in the room.
None of that grunt work was ever valuable in itself. It was tuition. You paid it in tedium and it bought you judgement.
And it is exactly, precisely, the work that goes first. Read five hundred contracts, summarise, flag anomalies. That is not a hard task any more. It is a Tuesday.
So the ladder loses its bottom rungs while everybody stands around agreeing that the top of the ladder is where all the value now sits. Judgement, taste, accountability, the person who signs. Yes, correct, that is where the value sits. Now kindly explain the route up there, because the staircase has been removed and we are all just admiring the balcony.
This is not a 2038 problem, by the way. Graduate hiring is where you will see it first, and honestly, if you are in a position to hire juniors right now, that is a decision with a much longer shadow than it appears to have.
The plumber outlasts the radiologist
The second thing we have got backwards is the order.
There is an old observation in robotics, Moravec’s paradox, which says roughly that the things humans find hard are easy for machines, and the things any four year old does effortlessly are brutally hard for machines. Chess fell decades ago. Folding laundry in an unfamiliar room is still an open problem.
Which means the automation order is close to the exact inverse of the social prestige order. Structured cognitive work, the kind you needed a degree and a quiet office for, is the most exposed. Unstructured physical work in an unpredictable environment, up a ladder, under a house, in someone’s ceiling cavity with a torch in your mouth, holds out considerably longer. The electrician is safer than the analyst. This is not a comforting thought for an entire generation of parents, mine very much included, who spent decades making it clear that the point of studying hard was so you would not have to work with your hands.
I am not saying tell your child to become a plumber. I am saying the confident hierarchy underneath that advice has quietly stopped being true, and nobody has sent the memo.
What actually holds value
If you strip it back, three things survive well, and none of them is a subject.
The first is judgement under uncertainty and the willingness to own the call. Machines will generate options endlessly and beautifully. Somebody still has to choose, and be accountable for choosing, and that accountability cannot be delegated to a system because you cannot sue a model. Every automated system I have designed keeps the human veto and hands over the clerical work. Be the veto.
The second is taste, which sounds soft and is not. When production is free, the constraint moves to knowing what is worth producing. Curation becomes the scarce skill precisely because generation stopped being one.
The third is the ability to work with people who do not agree with you. This one is badly underrated and it happens to be the most durable of the lot, because it is the residue of everything abundance cannot touch. Identity, recognition, status, who gets counted. Those do not dissolve when goods get cheap. They sharpen.
The no-regret test
Here is the honest position, which is that nobody knows the timeline. Seven years, thirty years, the arithmetic is genuinely wide open.
So the only defensible strategy is the one we use at work when the roadmap is uncertain. Do the no-regret things. Ask of any skill: does this pay off if the ramp is fast, and does it also pay off if the ramp is slow? Deep judgement in a real domain passes. Being able to explain something hard to a room passes easily, in both worlds. Physical competence with real objects passes. Learning to operate machines as an amplifier rather than as a threat passes.
Memorising a syllabus so you can reproduce it under exam conditions fails both tests, and it was failing them before any of this started. We just had no particular reason to say so out loud.
And notice that not one of those is a subject you can enrol in.
The bit that gives me hope
The last few weeks of writing about this have left me fairly open on the economics and surprisingly cheerful about the children, which is not the direction I expected to travel.
Because when I look at what actually survives, judgement, taste, the ability to sit with people who disagree with you and still get somewhere, I notice these are not modern skills at all. They are the oldest ones we have. They are what education was about for roughly two thousand years before we got distracted and turned it into a sorting mechanism for jobs.
So the machines may have done something a bit funny here. In taking the executable work off our hands, they have quietly handed back the question of what a person is for, and the honest answer turns out to be older and better than the one we have been giving.
My nephew wants to know if he still has to learn his times tables. Yes, mate. Yes you do. Some things I am not budging on.
Written for KiwiGPT.co.nz · Generated, Published and Tinkered with AI by a Kiwi