The Fall of Google: Illich, Counterproductivity, and the Economics of Digital Decay

When the Search Began to Fail

Sometime in the last few years the gesture started feeling different, though no one could name the week. The cursor still blinked in the familiar white field. The query still autocompleted as it always had. What had changed was quieter than the interface and harder to locate — a small drag in the movement from question to answer, the faint anticipatory tightening of the hand before clicking a result the user already half-suspected would not be the one.

When did the search begin to fail? No one can name the date.

For years the page had looked almost empty, and the emptiness was the authority — a logo, a blinking cursor, a field of white promising that nothing would come between the question and the answer except relevance. People trusted the emptiness because it concealed almost everything, and concealment still registered as benevolent.

The Slow Succession of Mild Disappointments

There was a time when the first result was the right result — not always, but often enough that the gesture of searching felt clean, a question launched into a system that answered with relevance good enough to trust the first page without performing quality control on the system’s behalf. That time has passed. The first result is now, more often than not, a paid placement, a strategic page, or a summary generated by a machine that has consumed the source without crediting it. The second result is often the same, and the user has learned to skip.

The feeling thinned. A paid result above the one you meant to click. A query rephrased because the first page looked wrong. A recipe preceded by autobiography. A business listing that still exists in the index though the business is gone. An answer summary plausible and brittle in equal measure. No single moment severe enough to force departure — only a succession of mild disappointments distributed so evenly that adaptation arrived before anger.

The Unpaid Accommodations of the Modern Searcher

The modern user has acquired habits that would have been unimaginable in the cleaner years — appending reddit, appending forum, putting the brand in quotes, ignoring the first result, scrolling past the sponsored blocks, scanning the AI summary with practiced distrust, reading it anyway, opening three tabs because no single one feels authoritative, reformulating the question in language more awkward than natural thought because the engine responds better to SEO-shaped phrasing than to how the user would actually ask.

None of these gestures is dramatic. They are unpaid accommodations by which a searcher keeps a deteriorating instrument temporarily livable, and the user who once arrived as a seeker now arrives as a worker, performing the labor of disambiguation the tool was supposed to perform on his behalf.

The retraining is partly muscular. The thumb flicks down past the sponsored block before the eye has finished reading it. The jaw sets at the AI summary. The hand opens a second tab as insurance, the way a driver checks a mirror he no longer trusts. None of it is decided; all of it has been learned.

These adaptations are practical intelligence, yes — and also a quiet report card on the environment that produced them. The user has become a connoisseur of one tool’s failure modes, trading fixes on forums like a mechanic for a car that keeps breaking: add this term, exclude that one, use quotes, try a different engine. The labor of finding has migrated from the platform to the person. The elegant part is that the person logs the migration under self-improvement.

What the user has absorbed as personal skill is, from the platform’s point of view, unpaid engineering labor at global scale. Every reformulation, every additional tab, every Reddit suffix is a signal fed back into the system, a small piece of human attention refining the very mechanism that demanded the attention in the first place. The searcher functions as a quiet contributor to the product’s training rather than as its consumer, and the product he is training was trained, in part, to require his continued training.

No line in the terms of service names the arrangement. No economic indicator tracks the hours that users spend, in aggregate, rescuing the instrument from its own decay.

Counterproductivity: What Illich Saw Coming

There is a pattern, described most precisely by Ivan Illich, that digital life has made visible with embarrassing clarity. Tools do not remain faithful to their stated purpose simply because that purpose was once real.

Beyond a certain threshold of scale and incentive, they begin undermining the very capacities they were built to serve — medicine making health harder to inhabit, schooling interfering with the learning it claimed to deliver. Illich called this counterproductivity, and he saw it coming long before the first search query was typed into a white bar.

The pattern has a shape older than Illich named it. Every institution entrusted with serving a human need has, at some point in its maturity, begun profiting from the difficulty of the need itself — hospitals from illness, schools from confusion, search engines from the distance between a question and its answer. The arrangement tends to survive because the people caught inside it are the least able to name it.

The Library After the Librarians Changed

Google arrived at that threshold so gradually that the crossing was imperceptible. The search engine once cleared the space between curiosity and finding. Now it crowds that space with obstructions while preserving the appearance of service. Visually, the old promise survives, which is what makes the experience uncanny — the tool still looks devoted to finding while behaving like an environment devoted to holding.

The page has been redesigned, over years, to answer the question without sending the visitor to the source — summaries, snippets, panels, and generated previews interposing themselves between the seeker and the shelf, each a small act of cannibalism performed on the ecosystem that produced the knowledge being summarized. The person who wrote the article sees his work paraphrased in a box. The visitor reads the box. The click never arrives. The knowledge was consumed, and the path to its origin was quietly severed, with the severance presented as convenience.

The creator economy, so called, rests on Google’s willingness to index. A blog, a portfolio, a small business — each depends on being found, and being found depends on compliance with a set of optimization requirements that shift without warning and are enforced by an algorithm whose criteria are proprietary.

So the creator shows up as a supplicant who has been told he is a customer. The terms change at the platform’s discretion. His only recourse, when they change against him, is to comply or to vanish from the map.

The casualties are specific and largely invisible — the independent food blog that once ranked first for its recipes and now lives eight results deep beneath aggregators that scraped it, the specialist health site written by actual clinicians that has been demoted below wellness content produced for engagement, the small-town newspaper whose obituaries have quietly stopped surfacing because obituaries rarely generate advertising impressions. The knowledge remains intact on the servers that host it. The path to the knowledge has been rearranged. A generation of writers built their livelihoods on the assumption that the map would hold, and the map was redrawn without warning, and no one who depended on the old map received a notice.

Borges imagined a total library before search engines existed.

In “The Library of Babel,” every possible book already exists — truth and every refutation, every delirious falsehood, every perfect page hidden in an astronomical mass of uselessness. The librarians suffer not because information is absent but because totality makes relevance miraculous and nearly impossible to secure. Google emerged as something stranger than that library: a library whose custodians learned to profit from the searcher’s inability to reach the shelf directly.

Borges imagined the total library. Google built something stranger: a library that profits from the distance between the question and the shelf.

The Lock-In of Learned Incompetence

There is a particular cruelty in a tool that makes the user responsible for its own failures.

The cruelty is not sadistic; it is structural. When the search returns poor results, the instinct — trained by years of use — bends toward reformulating the query rather than blaming the engine. The user assumes the fault is his: he searched badly, phrased the question wrong, failed to use the right operator. The engine has taught its users to attribute its declining quality to their own insufficiency, and the attribution is a form of lock-in more durable than any contract.

Google’s deepest power lived in the default rather than in the homepage — the quiet structural fact of being the place where questions go when the person asking has not yet decided to go anywhere in particular.

Defaults rarely register as commands; they register as the absence of a problem. A default that quietly stops working is one of the most profitable arrangements in the modern economy: the worse the product, the more labor it extracts, and the more labor the user sinks in, the higher the cost of leaving — because the fluency he has built only works on this exact decay, and nowhere else.

He searched badly, phrased it wrong, failed to use the right operator. The engine taught him to blame himself.

Illich described this pattern in medicine: the patient treated for years by a system that has made his health harder to maintain is also the patient most dependent on that system, because his body has been organized around its interventions.

Google’s users occupy the analogous position.

They have been trained to search badly — to add qualifiers, to distrust the first result, to do the platform’s quality control unpaid — and the training is the cage. It transfers nowhere. The competence is bonded to the very fault that produced it.

Remove the decay and the expertise evaporates overnight. This is the strange genius of the arrangement: a tool that converts its own failure into a skill, and the skill into a reason to stay.

Something subtler erodes as well. The capacity for finding — a capacity older than any search engine, tied to memory, to library habit, to the discipline of following a citation toward its source — has been partially ceded to an instrument that now handles it poorly. The user, asked where to look for something, increasingly cannot answer without mentioning Google, even when Google has stopped being a reliable way to find it. The instrument absorbed the verb, and when the instrument declined, the verb declined with it, and a generation of searchers lost fluency in a skill their grandparents, equipped only with card catalogs and encyclopedias, had possessed without noticing.

Eleven Minutes, Four Reformulations

The answer, when it comes, comes from a stranger in a 2019 forum post who had no stake in being found — and it is almost right, which lately counts as a win.

Eleven minutes. Four reformulations. He closed the tabs without quite registering the cost, because the cost has no line item — only a memory gone vague, of a time when the first result was the right result, when the white field answered the question instead of standing between him and it. The cursor still blinks, patient as ever. It has learned to look like help.