GM. This is The Great Unlock, your weekly cheat sheet for the AI revolution. We filter out the noise, red pen fake news, and give out stickers to those deserving.

Today’s deep dive: why the most expensive education experiment in American history produced nothing, what David Epstein’s new book reveals about the power of constraints, and why AI’s real gift to education is not removing limits but choosing the right ones.

Let’s unlock it.

The education system has spent decades fighting for fewer constraints. The research says that is exactly the wrong fight.

David Epstein’s new book Inside the Box: How Constraints Make Us Better opens with a thesis that should make every district leader uncomfortable. Constraints do not limit innovation. They drive it. The best solutions in history were not built in spite of limits. They were built because of them.

One of Epstein’s case studies is General Magic. A 1990s Silicon Valley startup staffed by former Apple engineers. A product that predated the smartphone by a decade. An unlimited budget. A dream team with no ceiling. They had USB. They had touchscreens. They had everything that eventually became the iPhone. They still failed. Not because of what they lacked. Because without the right limits, they were solving the wrong problem at the wrong speed with no forcing function to ship.

Epstein’s argument is precise: total freedom is paralyzing, and unlimited resources do not necessarily lead to the biggest breakthroughs. Overvaluing complete freedom can be disastrous.

The parallel to K-12 is impossible to ignore.

1. THE KANSAS CITY EXPERIMENT

$2 Billion. 12 Years. Nothing to Show For It.

In 1985, a federal judge took partial control of the Kansas City, Missouri School District. The district was failing. Test scores were plummeting. Good teachers were leaving. The judge’s solution was extraordinary: he ordered the state and district to dream without a budget. Whatever educators needed to fix the system, he would make taxpayers pay for it.

Over the next 12 years, Kansas City spent nearly $2 billion. They built 15 new schools. An Olympic-sized swimming pool with an underwater observation room. A robotics lab. A 25-acre wildlife sanctuary. A zoo. Field trips to Mexico and Senegal. Teacher salaries went up. The student-teacher ratio fell to 12 or 13 to 1, the lowest of any major district in the country. Per-pupil spending hit $11,700, the highest in the nation on a cost-of-living adjusted basis.

The results? Test scores did not rise. The black-white achievement gap did not close. Integration actually decreased. When the judge finally agreed to let the state stop making payments in 1999, there was nothing to show for the most expensive education experiment in American history.

Kansas City is General Magic with a school board. Unlimited resources. No constraints. Dream team ambitions. Same outcome.

2. THE DESIGN PROBLEM

Resources Without Design Constraints Produce Waste, Not Innovation

Epstein’s thesis is not that constraints are always good. It is that the right constraints force better decisions. Dr. Seuss wrote Green Eggs and Ham using only 50 words because a publisher bet him he could not do it. Keith Jarrett arrived at a sold-out concert in Cologne to find a broken, out-of-tune piano. He was forced to improvise around its limitations. The result was the bestselling solo piano album of all time.

The pattern repeats across every domain Epstein examines. When constraints are removed entirely, ambition drifts. Scope expands. Urgency disappears. The problem being solved becomes blurry. General Magic had the talent to build the iPhone 15 years before Apple did. Without the constraint of a shipping deadline, a limited budget, or a focused product brief, they built everything and shipped nothing.

Kansas City had the same structural problem. The judge removed every financial constraint. Educators were invited to dream. And what they dreamed of was the system they already had, but with better facilities. Olympic pools. Wildlife sanctuaries. Higher salaries. Not a single dollar was constrained toward the question that actually mattered: how do students learn?

3. THE RIGHT CONSTRAINT

AI Removes the Financial Barrier. It Does Not Remove the Design Problem.

At HierLearning, the core thesis is a deceptively simple inversion: learning should be the constant and time should be the variable. The current model does the opposite. Time is fixed. Learning adjusts to fit. The system was designed for the average, which means it was designed to fail everyone who is not.

Bloom proved in 1984 that students who receive 1:1 instruction outperform their peers by two standard deviations. We have known this for 40 years. The reason schools could not solve it was not systemic resistance. It was not a lack of will. It was financial. You cannot hire one teacher for every student. The math has never worked.

Until now.

AI removes the financial constraint on 1:1 instruction. That is the right constraint to remove. But AI does not remove the design constraint. It amplifies whatever design you already have. A system built around seat time, powered by AI, will deliver seat time faster. A system built around mastery, powered by AI, will deliver mastery at scale.

Kansas City proved that removing financial constraints without changing the design produces nothing. AI will prove the same thing if districts deploy it inside the same model. The unlock is not more resources. It is better constraints. Learning before progression. Productive struggle before intervention. Mastery before movement.

4. FOUR QUESTIONS EVERY DISTRICT LEADER SHOULD ASK RIGHT NOW

You do not need to wait for the next academic year to act on this. Here are the four questions to bring to your next leadership conversation.

1. What is our spending constrained toward? Kansas City spent $2 billion unconstrained from any design thesis. Audit your budget against one question: does this expenditure build student capacity to perform independently? If the answer is “it improves the environment” or “it reduces friction” but not “it builds capacity,” you are funding the model, not the outcome.

2. Which constraints are we fighting that we should be keeping? Not every limit is a barrier. Some constraints are design features. The constraint that says a student must demonstrate mastery before moving forward is not an obstacle to progress. It is the mechanism of learning. Before you lobby to remove a constraint, ask: is this constraint protecting the learning process?

3. Are we deploying AI to remove the right constraint? AI should remove the financial barrier to 1:1 instruction. It should not remove the productive struggle that makes 1:1 instruction effective. If your AI deployment makes learning easier rather than more personalized, you are using a precision tool for a blunt job.

4. If we got unlimited funding tomorrow, would our model change? This is the Kansas City test. If the answer is “we would hire more staff and buy better technology but keep the same schedule, the same progression model, and the same accountability framework,” then the problem is not funding. The problem is design. And more money will produce exactly what it produced in Kansas City: better facilities and the same outcomes.

Sources: Epstein, D. (2026). Inside the Box: How Constraints Make Us Better. Riverhead Books. • Ciotti, P. (1998). Money and School Performance: Lessons from the Kansas City Desegregation Experiment. Cato Institute Policy Analysis No. 298.

Every budget season, the same conversation plays out in districts across the country. Principals present a list of needs. Superintendents present a list of shortfalls. The board asks: do we have enough? The answer is always no. The debate is always about how much more.

The Kansas City experiment answered that question definitively. They had enough. They had more than any district in the country. It did not matter.

The question your budget conversation should be asking is not “how much do we have?” It is “what is every dollar constrained toward?” Because a dollar constrained toward mastery-based progression is worth ten dollars spent on a better classroom without a better model inside it.

Old question: Do we have enough resources to improve outcomes?

New question: Are our resources constrained toward the right outcome?

Old question: What do we need to add?

New question: What design constraint is missing?

Old question: How do we close the funding gap?

New question: If we closed it tomorrow, would anything change?

The districts getting this right are not the best-funded ones. They are the ones where every resource decision is run through one filter: does this build the student’s capacity to perform independently, or does it make the system more comfortable for adults?

You will hear this sentence in every policy conversation this year. “We cannot innovate until we have the resources to do it.” It sounds reasonable. It sounds like a sequencing argument. First the money, then the change.

Most of the time, it is a preservation argument. First the money. Then more of the same.

Kansas City got the money. All of it. No district in American history has ever been given a more complete answer to the funding question. What they did with that answer was build a more expensive version of the system they already had. The innovation was architectural, not instructional. The buildings were extraordinary. The model inside them was identical.

Epstein’s research offers the counter-evidence from every other domain. The most innovative solutions in history came from constraints, not from their removal. The periodic table was invented under a publisher’s deadline. The bestselling piano album of all time was recorded on a broken instrument. Green Eggs and Ham used 50 words because someone said you could not write a book that way.

Innovation does not require unlimited resources. It requires a forcing function. A constraint tight enough that the only way through it is a new idea.

Here is something David Epstein’s book will not say, but every teacher already knows.

The best lessons were never the ones with the biggest budgets. They were the ones built inside the tightest constraints.

The teacher who had 28 students, one working projector, and 45 minutes to make fractions feel like something that mattered. The teacher who lost their classroom to a schedule change and taught in the hallway for a week, and those five days produced the best discussions of the semester. The teacher who was told there was no budget for new materials and built an entire unit out of newspaper clippings and a single roll of chart paper.

Constraints did not stop those teachers. Constraints made those teachers sharper. The limit was the forcing function. The creativity came from inside the box.

The shout-out this week goes to the teachers who never waited for more resources before doing better work. The ones who heard “we don’t have the budget for that” and heard it as a design challenge, not a stop sign. The ones who know, in their bones, what Epstein’s 300 pages of research confirms: the obstacle is not in the way. The obstacle is the way.

 ________________________________________________________

That's all the unlock for today. Tune in next week.

Want more? Order The Great Unlock here.