The Friction Was The Product

Written By RJ Starr

What the higher-education debate keeps misnaming, and why its instruments cannot see it

Over six weeks in the spring of 2026, the New Yorker’s Jay Caspian Kang wrote a series on the viability of the American university system, and it opened with a small domestic question: should he keep contributing to his nine-year-old daughter’s college fund? By the end he is no longer asking about money. He is asking about something he cannot price, and the inability to price it is the most honest moment in the work. A reported series can be precise about everything except the object it is reporting on, and this one is, because the object has been misnamed. The inability to price what is being lost is not a failure of Kang’s arithmetic. It is the symptom of a category mistake that runs through the entire conversation about artificial intelligence and the university: a mistake about what the institution was ever selling.

The reflexive framing holds that college delivers information and confers a credential, and that artificial intelligence threatens both: if a model holds the knowledge and can manufacture the credentialed artifact—the essay, the problem set, the exam answer—then the delivery is redundant and the stamp is counterfeit. Freddie deBoer has already dismantled the first half of this, naming the library-card fallacy—the assumption that because information is now freely accessible, the institution that brokered access must wither. He is right that it will not. But naming the fallacy clears the ground without building on it, and leaves the real question standing: if information was never the product, what was?

The answer is friction. Earlier essays in this series have argued—in The Disappearance of Difficulty and in Access Is Not Understanding, that difficulty is not an obstacle to intellectual development but its scaffold, and that access to information is not the same as the internal labor by which a person comes to understand. What follows extends that established ground to a new site: the institution built to enforce difficulty, and what becomes of it when difficulty becomes optional. The pedagogically defensible product of the university was never merely the knowledge it transmitted or the certificate it issued. Universities sold many things: credentialing, class sorting, professional licensure, network access, cultural legitimacy. But the one good that justified the rest, the good no other institution was organized to compel, was the encounter with difficulty one would not have chosen: the structured obligation to struggle against a text, a proof, an argument that does not yield, under conditions that make avoidance costly. The reading one would have put down. The revision one would not have attempted. Information was always incidental to this; the library held it before the internet did, and the internet held it before the model did. What the institution uniquely supplied was the compulsion—the architecture that made the difficult encounter unavoidable. Artificial intelligence does not threaten the information. The information was already free. It dissolves the compulsion.

This is not a lament about rigor, and it is not nostalgia. It is a structural claim with a structural consequence, and the consequence is that almost everyone—Kang included—is watching the wrong thing disappear and calling it by the wrong name.

The Threat the Institution Cannot See

The most revealing finding in Kang’s reporting is one he registers as a surprise and then explains away. He expected administrators and education scholars to be alarmed by what artificial intelligence does to the formation of a mind. Instead he found them treating the matter almost entirely as a problem of cheating: of detection, deterrence, and academic integrity. He reached for the natural explanation, bandwidth. Cheating had become an emergency, and an emergency consumes the attention that larger questions would otherwise receive. There was no room to think about what the technology does to scholarship while the building was on fire.

It is a reasonable explanation, and it accounts for the wrong thing. Bandwidth would predict a delay—institutions that perceived the deeper threat but were forced to triage, returning to it once the fire was contained. What the reporting actually documents is not triage but stable substitution: administrators who do not appear to perceive the deeper threat at all, for whom “cheating” is not the urgent foreground of a larger danger but the danger entire. That is not a scarcity of attention. It is a property of how the attention is shaped.

An institution does not perceive a threat in the abstract. It perceives the version of the threat its existing structures are built to act on. A university is an apparatus elaborately organized around assessment, integrity enforcement, and credentialing—around procedures for adjudicating whether work is legitimate. Present that apparatus with a phenomenon that has two faces—a disciplinary face (students are submitting work that is not theirs) and an existential one (the formative encounter the work was meant to occasion is being dissolved)—and the apparatus resolves the phenomenon into the face for which it holds instruments. The disciplinary reading is not chosen over the existential one. It is the only one that comes into focus, because it is the only one the institution’s tools are constructed to render.

I will name this mechanism Tractability Capture, and offer it here as a formal construct: an institution’s perception of a threat is captured by whichever version of that threat it already possesses the means to address, and the intractable version goes not contested but unseen. The institution is not in denial, which would require having seen the threat first. It is operating exactly as its structure dictates: metabolizing an existential danger into a disciplinary one because the disciplinary one is the only one it can process, and then experiencing that translation as the whole of the problem. Tractability Capture is distinct from Meaning Dissolution, the construct developed earlier in this series: Meaning Dissolution concerns what is lost when information is transmitted without the relational context that would let it cohere. Tractability Capture concerns something prior—not what survives transmission, but what an institution is able to perceive in the first place. The first is a failure of carriage. The second is a failure of sight.

Tractability Capture is not bandwidth, and the distance between them is the entire point. Bandwidth predicts that additional resources surface the deeper question. Tractability Capture predicts the reverse: the institutions with the most developed apparatus will be the least able to see the formative threat, because their instruments are the brightest and the capture is therefore strongest. The well-funded university with a mature academic-integrity office, a detection contract, and an honor code freshly amended to name artificial intelligence is not better positioned to perceive what is being lost. It is worse positioned. Every one of those instruments renders the disciplinary face in higher resolution and leaves the existential face dark. I will state this plainly, as a prediction the framework generates and is willing to be tested against: the sophistication of an institution’s integrity response will correlate inversely with the clarity of its account of what, beneath the cheating, is actually disappearing.

The mechanism explains a second thing Kang observed but could not connect—his frustration that the leaders willing to ask what the university is for were so scarce, and that the responses he drew from faculty were so often, in his word, reactionary. They were not failing to think. They were reporting accurately from inside a structure that had already resolved the threat into its actionable form before the question of meaning could be raised. The people best equipped to enforce the disciplinary frame are, by the very same equipment, the least equipped to notice that it is the wrong frame.

What This Does Not Excuse

There is an objection here that deserves to be met before it is raised, because the strongest reader in this conversation will raise it. To reframe cheating as a structural phenomenon—as a category error in institutional perception rather than a moral failure in individual students—is to invite the charge deBoer has leveled with unusual force: that treating student dishonesty as anything other than dishonesty is a tender, therapeutic evasion, an adult abdication that indicts the assignment, the professor, and the system in order to avoid indicting the person who chose to cheat. The charge has teeth, and a structural argument that ignored it would deserve them.

So let me be exact about what Tractability Capture claims and what it does not. It is a claim about institutional perception—about why the university sees the threat it sees. It is not a claim about individual responsibility, and it does not touch the student’s. The two operate at different scales and the structural account does not soften the moral one; if anything it sharpens it. deBoer’s own thesis—that ethics are cultivated under constraint, learned through standards, penalties, and the repeated experience of being told no—is not a rival to the argument advanced here. It is a restatement of it at the moral register. Constraint is friction. The demand that a student be made to struggle honestly against a difficult thing, and be judged for refusing, is the same demand as the one this essay makes structurally: that the compelled encounter with difficulty is the irreducible good the institution exists to deliver.

The structural reading does not absolve the student. It explains why the institution can no longer deliver the constraint deBoer rightly wants. An apparatus captured by the disciplinary face of the problem will reach for detection and proctoring—for the appearance of constraint—while the actual mechanism of constraint, the friction itself, has already been removed upstream of anywhere its tools can reach. deBoer’s frustration that institutions will not simply hold the line is not evidence against Tractability Capture. It is the symptom the mechanism predicts. The line cannot be held by instruments aimed at the wrong face of the threat.

The Subtraction, Seen Twice

Strip the friction from a university and you produce an administrative efficiency. Strip it from a nineteen-year-old and you produce an arrested self. It is the same subtraction, performed on two architectures that both mistook their load-bearing wall for a wall in the way.

This is where the framework of Psychological Architecture does work the surrounding conversation cannot, because the conversation has no construct for the interior event and keeps reaching for institutional language to describe a developmental one. Consider the testimony Kang gathers, and that Jane Sloan Peters set down first. In her course on what people have endured for their convictions, students once struggled to arrive at a theme of their own; and then, through brainstorming and revision, landed on an understanding that felt personal precisely because they had been forced to fight their way to it. After the model arrived, not one of sixty students struggled. She received tidy summaries and themes that accounted for everything while meaning nothing. Kang reads this as the loss of rigor. It is something more specific and more serious. It is the disappearance of the mechanism by which a forming self differentiates.

Identity is not transmitted; it is precipitated. A self becomes distinct by encountering resistance it cannot route around, by being made to hold a difficult thing long enough that its own response to it has to take shape. The struggle Peters watched vanish was not a pedagogical nicety. It was the friction against which a particular nineteen-year-old’s particular reading of a text—and therefore some small, real increment of that person’s individuation—used to form. Remove the resistance and you do not merely get worse essays. You get a self that was never compelled to become specific, handed instead the averaged interpretation a system has determined is most acceptable, and asked to mistake it for a thought of its own.

Kang feels the magnitude of this—he reaches, instinctively, for the Bhagavad Gita and the way being made to sit with Arjuna’s paralysis changed him—but he has no apparatus to say why the friction was load-bearing, and so he can only grieve it. The grief is appropriate. It is also insufficient, because it leaves the loss looking like a matter of taste—some people were moved by hard books—rather than what it is: the removal of a developmental obligation. The encounter that individuates is not optional seasoning on an education. It is the thing an education was the occasion for. This is why the loss cannot be filed under cheating, and why no integrity policy can recover it. Cheating is a violation within the formative process. What is occurring is the quiet evaporation of the process itself.

And here the two scales resolve into one. The administrator who cannot see past cheating and the student who no longer struggles are not two problems but a single event observed at two magnifications: resistance withdrawn from a structure that had always misread resistance as the thing to eliminate. The institution removes the friction it can name—the inefficiency, the bottleneck, the labor of judgment—and calls it modernization. The student removes the friction he can feel—the difficulty, the discomfort, the labor of thought—and calls it relief. Each is responding to the same withdrawn resistance. Each had mistaken the load-bearing wall for an obstruction. Neither can see, from inside the subtraction, that the wall was holding up the building.

The Question, Relocated

Return, then, to the fund Kang cannot decide whether to keep paying into. The reason the question defeats him is that he has posed it as a question about a credential—whether the degree will still purchase what degrees once purchased—when the thing actually in jeopardy was never the credential. It was the compelled encounter with difficulty, and that encounter is now becoming optional everywhere at once: in the classroom, in the library, on the couch with the comic books and the chatbot. The degree may well survive. The friction is what is in question.

So the real question is not whether his daughter will go to college in 2035, nor whether the credential will hold its value. It is where, in a world that has made struggle avoidable in every setting that once enforced it, a nine-year-old will now be compelled to become specific—to meet the resistance against which a self takes shape. The university was, for several centuries, the place that answered this by force. It may not remain so. The urgent task is not to mourn the institution or to defend the credential. It is to recognize what the institution was actually for, so that when we build whatever follows, we do not optimize away the one thing it existed to provide. The friction was the product. Everything else was packaging—and the packaging is all anyone is arguing about.

References & Sources

Direct links are provided so that readers and respondents can consult the primary material. New Yorker columns sit behind a paywall; titles and the series path are given in full for readers with access.

Within The Artificial Era and Psychological Architecture


RJ Starr
Next

Tractability Capture