What We Can Learn From Outside Education

Education keeps trying to solve education from inside education. It convenes educators, studies educational research, writes educational policy, and is forever surprised when the gap it has measured for thirty years refuses to close. But almost every answer the sector is reaching for has already been worked out — not in education, but in other regulated, information-based, life-shaping sectors that took one idea seriously and built around it deliberately. The idea is that the substrate is information. Two of those sectors are worth importing whole. Medicine gives us the architecture. Business gives us the strategy.

Start with medicine, because medicine has just lived through the transition education is now entering. Moderna’s Stéphane Bancel reduces his field to a single claim: biology is, in his words, ‘only information.’ Not biochemistry, not molecular interaction — information. Your DNA is information; cancer is that information gone unstable; treatment is the act of telling the immune system which specific instability to correct. Different information for different patients with the same diagnosis, made through one shared process.

That reframe ends the analog era of medicine — the century in which, as Bancel describes it, drug development was trial and error, throwing enough at the wall to see what stuck and carrying the survivors to patients. That sentence describes the industrial era of medicine. With almost no translation, it also describes the industrial era of school.

Because education is information too. A student’s learning trajectory is information. The achievement gap is the measured distance between the information we intended to transfer and the information that actually landed. The English learner whose spoken English runs two grade levels ahead of her written English is an information-flow mismatch — and the absence of any instrument to detect that mismatch is the educational equivalent of having no biopsy, no biomarker, no early screen. The child grows. The mismatch compounds. The intervention, when it finally comes in eighth or ninth grade, costs far more than the early one would have.

Medicine’s answer was the end of average. Diagnosed by the same doctor on the same day with the same cancer, you and I now receive chemically different drugs, because the underlying instability is different in each of us. The educational parallel is exact. Two third-graders carrying the same nominal label — struggling reader, English learner, intervention candidate — need different instruction, because the information-flow mismatch underneath the label is different in each. One has strong oral comprehension and weak written production. One has the reverse. One decodes but cannot infer; one infers but cannot decode. The industrial classroom gives them all the same thing, ‘more reading,’ because it cannot see the difference. The oral-written delta is the biopsy it is missing.

Here is the move that matters most, and the one education has never made. When every patient receives a different molecule, you cannot get the drug approved — so Moderna gets the process approved instead. It proves to the regulator that the same input always yields the same output: feed in a given patient’s information, get back a given medicine, every time. Approve the process, and the products take care of themselves.

Education has spent thirty years doing the opposite — approving products. A single test, a single dashboard, a single intervention category, handed down for every district to deliver to every child. Each one standardized the product and missed the specific child who needed the specific thing. The gap did not close. Approving better products was never going to close it, because the problem was never the product. It was the absence of a process.

Two corollaries follow, and both are operational, not rhetorical. First, measurement is the substrate, not an afterthought — Moderna does not run a trial without a manufacturing process, and a school system should not run without a measurement substrate fine enough to see the individual child. Second, prevention replaces remediation as the economic logic. Catch the cancer at stage one and the odds rise while the cost falls; flag the third-grader’s oral-written delta before the literacy gap compounds and the same arithmetic holds. The instruments to do this exist now. The choice not to deploy them is the educational equivalent of declining to screen.

And the work itself gets reinvented. Bancel’s most telling design decision is that Moderna’s head of HR is also its head of digital — because in a world of AI and robotics the real question is not how to govern the tools but how to remake the work the humans do. Education’s version is the teacher’s day. You saw it in Section 3: not one adult delivering one lesson to thirty children for six hours, but one adult orchestrating many AI-augmented modes at once, doing the parts only a human can. Every district will soon carry a policy line requiring teachers to adapt instruction with AI. None has yet rewritten the workday in which that adaptation is actually possible.

Now the second sector. If medicine supplies the architecture, business supplies the strategy — and the strategic warning is blunt. AI is not a productivity revolution. It is a competitive reset. The firms that win the next decade are not the ones that adopt AI fastest; they are the ones that see earliest where value is moving and position themselves at the new control points before the equilibrium hardens.

Value moves in three waves. The first is productivity — automating and accelerating existing work. It rarely produces durable advantage, because competitors match it and the gains flow through to customers. The second is differentiation — genuinely new products and models, which build moats only when protected by proprietary data or network effects. The third is transaction-cost compression — the radical collapse of the costs of search, comparison, switching, and coordination once agents do that work at near-zero marginal cost. The third wave is where market structures themselves change, where intermediaries are bypassed, and where value migrates to whoever holds the new control points.

Map that onto schooling and the picture is uncomfortable. Wave one is the teacher with an AI tool who gets hours back and builds no lasting edge. Wave two is already running at the system’s edges, where new models redesign the cost and time structure of school around AI as the engine rather than the supplement. Wave three is the one the sector does not want to name: public education is, by design, drenched in transaction costs — adoption cycles, procurement code, compliance reporting, the information asymmetry between family and school — and every layer of intermediary exists to manage some friction below it. When agents compress that friction, the friction-management business is the one most exposed. The vendors who built quietly attractive, multi-state businesses on the friction of policy compliance are sitting precisely where the water is rising.

Business has seen this movie, and it has a moral: the losers were not the slow adopters. They were the ones who mistook efficiency for advantage — who optimized while their rivals reinvented, cut costs while their rivals captured the market. When the dust settled, the winners were the ones who understood earliest where value was moving and moved to hold it. The window for that move is open now and does not stay open. In an AI economy, advantage compounds early and value capture locks in fast.

There is one place the business framework goes silent, and in education it is the most important place: who is left out. In an agent-mediated world, the child whose family can deploy a capable agent to weigh every option on her behalf will get options. The child whose family cannot will get the default — and in an agent-mediated world the default is whatever gets assigned in the absence of an advocate. The remedy is not to slow the agents down. It is to build the public infrastructure that puts every child into the data — measured, visible, reachable by an agent acting for her — whether or not her family has the resources to advocate. That is what public infrastructure has to mean now.

Put the two lessons together and the conclusion is hard to avoid. Medicine has already shown what a regulated, information-based, life-shaping sector looks like when it takes the substrate seriously and builds deliberately: the science is real, the costs fall, the outcomes improve, and the institutions that organized around the information frame early are the ones now shaping the architecture. Education has not yet faced this. It still has districts, county offices, agencies, associations, and vendors producing better products — and the next decade will not reward better products. It will reward whoever builds the process: the measurement substrate, the early screen, the differentiated intervention, the agent-layer visibility for every child, the reinvented workday. Education does not have a Moderna yet. Where that gets built first — and which institution builds it — is what the sections that follow are about.