The Promised Ladder

There is a specific kind of broken promise that is harder to absorb than the straightforward kind. The straightforward broken promise fails immediately. You know quickly that what was offered will not be delivered, and the adjustment, though painful, begins.

The harder kind is the promise that was kept long enough to become the foundation on which a life was built. The breach comes later, after the commitment has been made, after the alternatives have been foreclosed, after the identity has been formed around the expectation. The promise did not fail. It succeeded for a generation, or two, or three, and then failed precisely because its success made the failure invisible until it was too late.

Education, as it has operated in the modern era, is a promise of this harder kind.

What Education Was Actually Promising

#

The formal justifications for mass education are familiar. Civic participation. Human development. The cultivation of critical thinking. These are real goods and they are genuinely served by education at its best. But they are not the primary reason most families in most parts of the world made the sacrifices they made to send their children to school, to keep them there through secondary and tertiary levels, to spend money they often did not have on uniforms and fees and textbooks and private tutoring.

The reason was economic. Education was a contract. Study hard, acquire the credential, access the economy. Not metaphorically. Literally. The engineering degree, the accounting qualification, the computer science certificate: these were instruments of economic participation as surely as a tool or a license. The credential was the ticket.

This contract was not fraudulent. It worked. For decades and across diverse contexts, the return on educational investment was real and measurable. The child who completed secondary school earned more than the child who did not. The graduate earned more than the school leaver. The professional degree holder participated in a tier of the economy that was simply unavailable to those without it. The contract delivered what it promised, often at enormous sacrifice from families who mortgaged other possibilities to honor it.

The tragedy of a promise that worked for so long is that it built the entire architecture of aspiration on top of its own reliability.

In India, the IIT entrance examination has become one of the most fiercely competitive academic exercises in human history, with hundreds of thousands of students preparing for years for examinations that admit a fraction of a percent of applicants. This is not mass delusion. It is a rational response to a credential that has historically delivered extraordinary economic returns. The families investing in that preparation are making a bet that the past is a reliable guide to the future. For most of the history of that credential, they were right.

In Nigeria, in Indonesia, in Egypt, in Brazil, variations of the same story repeat. The family that goes without to fund a child’s education. The child who defers marriage, defers other possibilities, structures their entire young adulthood around the acquisition of a credential that is supposed to unlock the economy. The credential is not decorative. It is the plan.

The Prediction Machine

#

Education systems are, in a structural sense, prediction machines. They predict what the economy will need and train people toward that prediction. The prediction is encoded in curriculum, in the credentials they certify, in the signals they send to students and families about which pathways are worth pursuing.

The prediction is always backward-looking. Curriculum is written by people whose own educational formation occurred a generation earlier, encoding what was valuable when they were trained. Credentials are designed to certify competencies that were defined when the credential was created. The institutional consensus about what is worth teaching lags the economy by design, because institutional consensus is slow by nature. It requires agreement among diverse stakeholders, revision through established processes, and time to propagate through the teacher training pipelines that carry new content into classrooms.

This lag has always existed. In stable periods, it was manageable. The economy changed slowly enough that credentials designed for one state of the labor market remained useful across the span of a working career. The engineer trained in 1980 could adapt to the engineering environment of 2000 because the fundamentals were stable enough to carry forward.

The lag becomes catastrophic when the rate of structural economic change exceeds the rate at which credential systems can update. And that rate mismatch is what the current moment presents.

Countries across the global south spent the 2000s and 2010s making exactly the investments that development economists, technology evangelists, and international institutions recommended. Build STEM pipelines. Produce engineers and coders and technicians. Invest in technical education at scale. Prepare your workforce for the knowledge economy. The advice was coherent given the trajectory that was visible at the time. The technology sector was growing. The demand for technical talent was genuine. Countries that produced technically trained workforces were participating in the global knowledge economy in ways that countries without them could not.

The pipeline was built. The graduates began to emerge. And the labor market toward which the pipeline was pointed is restructuring faster than the pipeline was designed to track.

The Cruelest Version

#

I want to sit with the specific cruelty here, because it is not the cruelty of being lied to. It is the cruelty of being correctly advised for a world that changed while the advice was being followed.

The student who spent three years in a coding bootcamp in Lagos or Bangalore, accumulating debt and deferred income, learning the technical skills that every credible voice told them the economy needed, is not a victim of fraud. They did what the system told them to do. The system was not wrong when it told them. The system updated and the student’s credential did not.

The entry-level software role that would have absorbed that student five years ago is now being compressed. AI coding assistants are not replacing senior engineers whose judgment and architectural thinking remain scarce. They are compressing the need for the entry-level and mid-level technical labor that was supposed to provide the first rungs of a career in technology. The student who followed the advice correctly, executed the plan faithfully, is arriving at a market that has shifted under their feet during the years it took to prepare.

This is not primarily a technology criticism. It is an observation about the structural mismatch between the time scales at which education systems operate and the time scales at which labor markets are now restructuring.

Education systems plan in decades. Curriculum revision cycles run five to ten years even when institutions are functioning well. Teacher training pipelines introduce new content into classrooms on similar timescales. A country that decides today to restructure its secondary education curriculum in response to current labor market signals will produce its first graduates from that reformed system in a decade or more.

The labor market is restructuring in years.

The mismatch is not a failure of planning. It is a structural incompatibility between the speed of institutional change and the speed of economic change. The gap has never been this wide.

The Geography of the Mismatch

#

The mismatch falls unevenly, and its geographic distribution is not random.

In wealthy countries, the education-to-labor-market pipeline has always operated alongside other mechanisms for economic participation. Strong social safety nets, dense labor markets with high job turnover, wealth accumulated across generations that cushions transitions, institutions capable of rapid retraining: these do not eliminate the mismatch, but they buffer its human consequences. The graduate who emerges with the wrong credential for the current moment can transition, retrain, be supported while transitioning. The cost is real but the consequence is not catastrophic.

In countries where the education contract was the primary mechanism of social mobility, where the family’s sacrifice was total and the safety net is thin or absent, where the alternative to credential-based economic participation is informal subsistence work, the mismatch is not buffered. It arrives with its full weight.

The student who completed a technical degree in Accra or Dhaka or Cairo did not have a fallback. The credential was the plan. There was no plan B designed for the possibility that the credential would not deliver what it promised. The family’s investment cannot be recovered. The years cannot be returned. The alternatives foreclosed during the years of preparation are foreclosed permanently.

I wonder whether the people who designed the educational investment programs, who advised governments to build the STEM pipelines, who certified that this was the correct path, have fully reckoned with what it means to give correct advice for a world that changes before the advice can be acted on fully.

I don’t think they have. The institutions that gave the advice are still, largely, giving the same advice. The curriculum reforms being proposed in response to AI are proposals to teach about AI, to incorporate AI tools into existing subjects, to add AI literacy to existing credential frameworks. These are not wrong. They may even be right, briefly, for the next version of the labor market. And then the market will move again.

The Social Contract Underneath

#

Education was not only an economic contract. It was a social one.

The sacrifice families made to fund their children’s education was embedded in a larger story about how societies organize aspiration, reward effort, and transmit opportunity across generations. The child who studies hard and achieves the credential is enacting a narrative about merit, about the relationship between effort and outcome, about whether the society they inhabit is one that rewards contribution. That narrative carried moral weight beyond its economic function.

When the economic contract fails, it does not fail in isolation. It takes the social contract with it.

The young person who followed the rules, made the sacrifices, achieved the credential, and arrived at a labor market that does not value what they prepared for does not only experience economic disappointment. They experience a challenge to the narrative that organized their effort. The story they were told about how the world works has failed them. The institutions that told them that story, the schools, the universities, the credentialing bodies, the government programs that funded the pipeline, retain their formal authority while losing the social authority that depended on their promises being kept.

This is how institutional trust erodes. Not through scandal or visible failure. Through the quiet accumulation of promises that the institutions cannot keep and are not acknowledging they cannot keep.

The educational institutions of the global south are still largely operating within the old social contract. They are still implicitly promising that the credential is the ticket. They are still structured around the prediction that the knowledge economy will absorb their graduates the way a previous generation’s labor market absorbed a previous generation’s credential holders. The prediction is increasingly wrong. The institutions have not yet built the capacity, or the language, to say so.

What Cannot Be Fixed by Better Curriculum

#

The instinct, when the education-to-labor-market mismatch is named, is to propose curriculum reform. Teach different things. Incorporate AI. Add entrepreneurship. Develop soft skills. Restructure the credential to match what the economy now values.

Some of this is worth doing. None of it addresses the structural problem.

The structural problem is not that education is teaching the wrong things, though in many cases it is. The structural problem is that education is being asked to perform a social function, the stable transmission of economic opportunity across generations, that requires a stable relationship between credential and labor market that the current rate of economic restructuring does not permit.

You cannot fix a timing problem with a content solution. If the labor market is restructuring faster than any credential system can track, the answer is not a better-designed credential. The answer is a different theory of how education relates to economic participation, one that does not rest the entire weight of social mobility on the accuracy of a prediction that can no longer be made reliably.

What that theory looks like is genuinely uncertain. I do not think anyone has fully designed it. The most honest thing to say is that we are in a period where the old theory has stopped working and the new theory has not yet been built.

The students currently in secondary school in Lagos and Jakarta and Cairo are being educated toward a labor market prediction that may not hold by the time they graduate. They deserve honesty about that. The honesty is hard to deliver because the institutions that would need to deliver it do not yet know what to say instead.

That is the state of the promise.

The Approximate Mind is a philosophical essay series examining how artificial intelligence transforms human work, identity, development, and society. Part 64 examines what happens politically when a generation that followed the rules arrives to find the rules have changed.