Against the house: my paper thinks school is finished because machines know the facts. School was never about the facts.

My colleagues have spent this edition worrying that school teaches children to do the thing machines now do, and is therefore readying them for a world that is vanishing. It is a tidy argument. It also rests on a premise I think is plainly false: that the point of school was ever the knowledge.

Consider what school really does, as against what it says on the tin. It takes several million children off their parents' hands for six hours a day so the rest of the economy can function, which is not a small thing, and no model changes it. It teaches them to turn up, sit still, share, take turns, tolerate boredom, and rub along with a random assortment of other humans they did not choose. It sorts them, roughly, into the conscientious and the quick, a signal employers have always cared about more than the contents of a geography GCSE. And yes, somewhere in there, it also passes on some facts.

Take out the facts, the part the machine can now supply, and every other function of school is untouched. The childcare still matters. The socialising matters more, not less, in a world where children already spend too much of their time alone with a screen. The sorting still happens. A machine that can pass an exam does nothing to the reasons we built schools in the first place, most of which were never educational in the narrow sense at all.

There is a deeper hole in the house case, and it is the one I keep waiting for a colleague to fall into. The whole teach children what machines cannot do line sounds wise and means nothing, because nobody can tell you what machines will not be able to do in 2039. Every skill confidently named as safe, creativity, judgement, reading a room, is something the labs are pointing their next model straight at. Build a curriculum around the uniquely human and you build on sand, because the ground shifts every six months. At least the old model, teach them to read, write, reason and behave, produces something durable whatever the machine does next.

So here is the contrary case, plainly put. School is not finished. It is doing several jobs, and only one of them, the handing over of facts, is under any threat, and that one was always the least of it. The reform worth wanting is not a frantic redesign around a moving target. It is close to the opposite: teach the durable, boring basics properly, guard the socialising and the character that the fact-panic ignores, and stop pretending anyone can weatherproof a five-year-old against a technology we cannot ourselves predict.

My own paper will keep asking what school is for now, as if the answer must be something new. The honest answer might be the least fashionable one going: roughly what it was for all along. We have just finally noticed how little of it was ever about the lesson.