The classic intractable problem of education is one of distribution. Students outnumber teachers. Always have. So the conundrum has classically been framed as a supply problem: not enough teachers, too many learners, and no good way to get what one knows into the hands of the many who need it.
For most of recorded history, the proposed fix has been a technology.
At one time it was believed that books could be the killer app. And in fairness, books have moved the needle forward an enormous amount. But they had their own limitations. Books take up a lot of space. They are usually not on the right subject, or they present a very narrow slice of what you actually need to know… Oh, and you had to be able to read to get anything out of them…which meant you needed a teacher first.
Distribution ended up being the bottleneck anyway. Even if the perfect book for your problem existed somewhere in the world, getting that knowledge to you in the moment you most needed it was a low-probability event.
The internet solved a lot of that. The space problem, for one. The internet can give you access to virtually all books and written information, and it does not need a room. But distribution still remained a problem, because you needed a computer to get to it. And you still needed to read.
Then the phone solved that. It put the unlimited library in your pocket. As memory and bandwidth increased, audio and video arrived to support the people who could not get what they needed from text. Suddenly you could pull out a computer the moment the need for knowledge raised its head. But you still had to know how to search. You had to build the query, sort the results, and judge which of the ten blue links was worth your afternoon.
And now there is AI. The last hurdle. You no longer need to know the search terms. You can ask badly, in your own words, in your own register, and get something back that is aimed at you specifically, at the moment of need, in whatever form you can take in.
So we did it.
Every barrier I just described is gone. The library is in your pocket, it will read itself aloud to you, and now it will answer in your own words at the moment you ask.
And this is where I get uneasy. Not because the technology failed. Because it worked.
The bottleneck was also a curriculum
For five hundred years, the difficulty of getting knowledge was obscured by the actual need for knowledge itself. The less glamorous, quiet work that rarely gets credited for the achievement of knowledge it promotes.
You could not retrieve an answer without first knowing what you were asking. That sounds trivial. It is not. Formulating the question is the act that turns a vague sense of not-knowing into a specific, findable gap. It requires you to monitor your own understanding well enough to locate its edge.
Students who cannot do this are not lazy or incurious; they simply have not been taught to notice the difference between confusion and ignorance.
You could not use what you found without deciding whether it fit. The book was almost never on exactly your topic. So you had to translate. You had to learn to ask what this general principle means for my particular situation, and to hold the answer against everything you already believed, and to notice when it did not square. That reconciliation is where understanding gets built. It is uncomfortable and slow and it is the whole thing.
And you could not do any of it without holding the question in mind across the gap between wanting to know and finding out. The gap was the workspace.
None of that was overhead on learning. That was the learning. Monitoring and control, enforced by circumstance rather than by design. The friction was never merely an obstacle in the path of education. For most of human history the friction was doing the teaching when no teacher was available.
Remove it, and you get something that feels exactly like learning and isn’t. Fluent, immediate, satisfying, and gone by Thursday…
What we actually discovered
Here is the part that unsettles me most.
We solved distribution, and in solving it we found out that distribution was never the intractable problem.
If teaching were fundamentally an act of moving information from a person who has it to a person who does not, the printing press should have ended the profession in 1500. It did not. The internet should have finished the job. It did not. Every generation of this technology has been announced as the thing that would finally scale expertise past the limits of human attention, and every generation has instead expanded the population of people who have access to more information than they can do anything with.
That is not a failure of the tools. The tools are extraordinary. It is evidence that we had misdiagnosed the disease.
Knowledge has to be constructed by the knower. There is no delivery mechanism for that part, and there has never been one, because it is not a delivery problem. It is a problem of learning to think about your own thinking well enough to know what you need, notice when you have not got it, and build something durable out of what you find. That work happens inside a person, and it happens far more reliably when another person is deliberately structuring it.
The uncomfortable version of the good news
The usual move here is to say that AI will not replace teachers. I do not find that interesting (maybe disheartening), and I do not think it is quite right either. Something is being replaced. The teacher-as-information-source is genuinely obsolete, and has been quietly obsolescing since Gutenberg. Good riddance. That was always the least of what teaching was.
What we have actually done is clear away, one technology at a time, everything that was not the problem. Space. Access. Literacy as a gatekeeper. Retrieval. Search. The list of things a person needs a teacher for has been narrowing for five centuries.
And what is left standing, now that the clearing is finished, is not a smaller job. It is the job. The one we were never able to see clearly, because the logistics were always in the way.
We now have to teach people how to use their powers of thought in a world that will answer any question they can manage to ask, and will never once tell them they asked the wrong one.
I do not think we are ready for that. But at least we can finally see it.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.