In 1806, Prussia lost to Napoleon at Jena. Badly.
The aftermath wasn't a military reform. It was a reform of children.
Prussian leaders had a theory: they lost because their soldiers thought for themselves. Soldiers who think for themselves hesitate. Hesitation loses battles. The solution wasn't better weapons or better tactics. It was better obedience.
So they built a system. Compulsory schooling. Age-graded classrooms. Standardized curricula. Bells that told you when to think about math and when to think about reading. A teacher at the front. Students in rows. Compliance rewarded. Deviation corrected.
It worked. Prussia became a military powerhouse. Other nations noticed.
The Import
In 1843, Horace Mann sailed to Prussia. He was the Secretary of Education in Massachusetts, and he was looking for a model. He found one.
Mann brought the Prussian system back to America. He sold it as democratic: education for all, not just the wealthy. That part was genuinely good. But the architecture he imported wasn't designed for democracy. It was designed for control.
The system spread. By the early 1900s, it was the standard model of American education. And in 1918, a teacher and scholar named Alexander Inglis wrote its operating manual.
The Six Functions
Inglis laid out six explicit purposes of modern schooling. Not hidden agendas. Not conspiracy theories. Stated objectives. They're worth reading carefully.
1. The Adjustive Function. Establish fixed habits of reaction to authority. The goal is reflexive obedience.
2. The Integrating Function. Make children as alike as possible. Predictability in a population is a management asset.
3. The Diagnostic Function. Determine each student's "proper social role." Label early. Sort accordingly.
4. The Differentiating Function. Once sorted, train each group only as far as their diagnosed role requires. No further.
5. The Selective Function. Tag the unfit (with poor grades, low placement, and social stigma) so they internalize their status.
6. The Propaedeutic Function. Cultivate a small elite to manage the system itself. The shepherds, not the sheep.
Read that list again. Notice what's missing? Learning to think. Curiosity. Creativity. Problem-solving. Those aren't features of the system. They're bugs.
The system was never broken. It was built this way.
The Machine Worked
Here's the thing nobody says: the system was a spectacular success, for the economy it was designed to serve.
The early 20th century needed rule-followers. Factory floors needed people who could execute instructions reliably, show up on time, tolerate monotony, and not ask too many questions about why. The Prussian-American model mass-produced exactly that cognitive profile.
And it worked for roughly a century.
In developmental psychology, there's a framework called the Model of Hierarchical Complexity (Commons, 2008) that measures how complexly a person thinks. Not what they know. How they reason. It defines 16 orders of complexity, from the sensorimotor reasoning of infants to the kind of cross-paradigmatic thinking that produces Einsteins.
Here's the number that should keep superintendents awake: roughly 40% of adults operate at Order 11, Systematic thinking. This is the level where you can follow and apply formal systems. Follow the procedure. Apply the framework. Execute the protocol. It's the exact cognitive tier that 120 years of Inglis-model education was designed to produce.
It's also the exact cognitive tier that AI executes with superhuman speed.
The Collision
GPT-4 doesn't "sort of" do systematic thinking. It does it better than any human alive. It follows procedures flawlessly. It applies frameworks across every domain simultaneously. It processes rule-based reasoning at a scale and speed that makes human systematic thinking look like a horse-drawn cart on a freeway.
We spent a century designing an education system to produce people who think at Order 11. And then we built machines that make Order 11 thinking worthless.
That's not a technology problem. That's a design problem.
The economy that's emerging doesn't need more systematic thinkers. It demands Order 12, Metasystematic, thinking: the ability to compare, critique, and integrate multiple systems. And Order 13, Paradigmatic, where you construct entirely new frameworks from the collision of old ones.
These are the orders where humans still hold the advantage. Where you don't just use a system but see its assumptions, its limitations, its relationship to other systems. Where you don't just follow the algorithm but ask whether the algorithm is asking the right question.
Our education system produces almost nobody who thinks this way. Not because kids can't. Because the system was designed to make sure they don't.
The Adaptive Lag
There's a concept evolutionary biologists use: adaptive lag. It describes the gap between how fast an environment changes and how fast an organism can adapt. Usually, organisms get millions of years. A slow change in climate. A gradual shift in food sources. Enough time to evolve.
We don't have that luxury. We have godlike technology running on institutions designed in 1806. The environment changed in a decade. Our schools haven't changed their fundamental architecture in a century.
David Epstein, in Range, showed that generalists, people who develop broad, flexible thinking across multiple domains, consistently outperform specialists in complex, unpredictable environments. Our system produces specialists. Narrow thinkers optimized for narrow tasks that narrow AI now does better.
Alison Gopnik's research on children's cognition makes this even more pointed. Young children naturally possess what she calls "lantern consciousness," a wide, exploratory, connection-making mode of attention. They see possibilities everywhere. Adults, by contrast, have "spotlight consciousness," focused, efficient, narrow.
School systematically converts lanterns into spotlights. That was useful when the economy rewarded focused execution. It's catastrophic when the economy demands the exact exploratory, integrative thinking we're training out of kids by third grade.
The Math Nobody's Doing
Here's the calculation every school board should run:
Take the six Inglis functions. For each one, ask: does AI do this to humans, or do humans need to do this despite AI?
Adjustive (obedience): AI doesn't need to be obedient. It's controlled by design. Training humans for obedience is training them to be replaced by the obedient machine.
Integrating (conformity): AI produces standardized outputs by default. The human premium is now divergence.
Diagnostic (sorting): AI can sort, categorize, and label at infinite scale. A human whose only skill is fitting a label has no economic value.
Differentiating (limiting): AI has no ceiling. A human artificially limited by their "diagnosed role" has been handicapped against a machine with no such limits.
Selective (stigmatizing): This never had educational value. It has less now.
Propaedeutic (elite management): The only function that even partly survives, because coordinating systems is an Order 12+ skill. But the current system only develops this in a tiny fraction of students.
Five of six functions are either obsolete or actively harmful in an AI economy. The sixth only works for the few. We trace the full history and analysis of this collision, from Prussia to GPT-4, in our whitepaper The Factory Mind: Why Our Education System Is Building Order 11 Thinkers in an Order 13 World.
From Factory to Forge
The path forward isn't adding AI tools to a Prussian classroom. That's putting a GPS on a horse-drawn cart.
The path forward is redesigning the architecture. From a factory that mass-produces compliant Order 11 thinkers to a forge that develops the Order 12 and 13 thinking AI cannot replicate.
What does that mean in practice?
It means measuring cognitive complexity, not content recall. It means designing experiences that force students to hold multiple systems in tension rather than memorize one. It means treating Gopnik's "lantern consciousness" as an asset to develop, not a disorder to medicate.
It means accepting that the system Horace Mann imported and Alexander Inglis codified was a brilliant solution to a problem that no longer exists.
Prussia needed obedient soldiers. The industrial economy needed obedient workers. Both of those needs have been automated.
The question for every superintendent, every board member, every parent reading this: What are we still building?
And what should we be building instead? For a concrete framework for what must replace the factory model, and why AI literacy alone isn't the answer, see our whitepaper Universal Basic Upgrading: Why AI Literacy Isn't Enough and What Districts Must Do Instead.