Reality Has Entered the Chat
Why our institutions still command while the world has learned to answer back
For most of human history, important knowledge came from above.
Not literally from the clouds, although priests did rather well out of that interpretation. But socially, institutionally, intellectually: knowledge had height. It lived in temples, courts, academies, ministries, universities, chancelleries, libraries and later in parliamentary committees with excellent catering and very limited contact with operational reality.
At the bottom were people who dealt with things.
Farmers, builders, sailors, merchants, mothers, nurses, craftsmen, mechanics, soldiers, shopkeepers. People who knew that a cow does not care about theory, that wood bends differently when damp, that patients do not read protocols before becoming ill, and that customers have an irritating habit of not behaving like policy documents.
At the top were people who dealt with abstractions.
Laws. Doctrines. Categories. Philosophies. Plans. Curricula. Procedures. Ideologies. Taxonomies. Frameworks. And, in modern times, strategic roadmaps with a mission statement so polished it can blind small animals.
Civilization needed this. Without abstraction, we would not have law, mathematics, architecture, science, accounting, engineering, contracts, maps, universities or states. A bridge does not arise from vibes. A tax system does not run on village gossip. A cathedral is not built by everyone bringing “their truth” and a slightly different stone.
Abstraction gave us scale.
But abstraction also gave us a dangerous habit: the belief that reality should obey the model.
For millennia, the prestige version of knowledge worked like this: smart people first think conceptually, then the world is instructed accordingly. The law is written, the doctrine defined, the curriculum fixed, the administrative category created, the regulation imposed. Reality is then expected to line up politely, like schoolchildren before inspection.
And when it does not?
Then reality is considered insufficiently regulated.
That, in one sentence, is much of modern government.
The old machine: simplify, classify, command
The state has always loved legibility.
A messy society is difficult to tax, police, educate, subsidize, regulate or reform. So the state simplifies. It creates categories. It names things. It maps land. It standardizes measures. It defines professions. It classifies citizens. It writes procedures.
A forest becomes timber volume.
A school becomes test scores.
A patient becomes a file.
A company becomes compliance obligations.
A citizen becomes a national register number with opinions attached.
Again: some of this is necessary. A complex society cannot run purely on local improvisation. We do not want aviation safety based on “Jan from maintenance has a good feeling about the left wing.” We do not want criminal law administered by neighborhood instinct. We do not want food safety to depend on the butcher’s aura.
But the state’s simplification instinct has a failure mode.
It starts to confuse its categories with reality.
Then the map is no longer a tool. The map becomes a prison.
This is how institutions become stupid while staffed by intelligent people. The stupidity is not always personal. It is structural. The system filters reality through categories created for administrative convenience, then punishes reality for not fitting.
Anyone who has ever dealt with bureaucracy knows the feeling. You explain the actual situation. The person behind the desk may even understand it. But the form does not. And the form, in that moment, is sovereign.
So the human says, apologetically: “The system does not allow it.”
Wonderful. We have built a civilization in which the software has constitutional authority.
The other knowledge: what reality teaches
There has always been another kind of knowledge. Lower status, often less verbal, but extremely powerful.
Practical knowledge. Craft knowledge. Tacit knowledge. Engineering knowledge. Entrepreneurial knowledge. The kind of knowledge that comes from repeated contact with reality.
The mechanic hears something wrong in an engine.
The nurse sees that a patient “doesn’t look right.”
The teacher senses the class is lost before the test results prove it.
The programmer feels the architecture becoming brittle.
The farmer knows the field by walking it, not by consulting a dashboard called Soil Excellence 2030.
This knowledge is not primitive. It is often more sophisticated than abstract knowledge because it has fewer hiding places. Reality is rude. It interrupts. It leaks, breaks, overheats, refuses, mutates, collapses, complains, underperforms and occasionally catches fire.
Theory can survive for decades in a seminar room. A bad bridge has a shorter publication cycle.
The best knowledge is not abstract or practical. It is both. Good theory, constantly disciplined by reality. Good practice, sharpened by concepts. The problem begins when abstraction escapes correction.
And for a long time, many institutions could escape correction.
States survived bad policies. Schools survived bad curricula. Courts survived procedural delay. Ministries survived failed reforms. Universities survived intellectual fashions. Public agencies survived user frustration. Political parties survived promises that aged like milk in the sun.
Reality grumbled, but it did not always have a loudspeaker.
Now it does.
Reality has entered the chat
The great change of our time is not simply digitalization. It is feedback.
Reality now talks back at scale.
Sensors report. Users click. Customers leave reviews. Software logs every failure. Satellites watch fields, cities and borders. Medical systems produce data. Online platforms reveal behavior. Companies run A/B tests. Engineers simulate before building. Markets transmit signals. Open-source communities debug in public. AI systems learn from vast oceans of examples.
For most of history, bottom-up knowledge was local. A craftsman knew his material. A sailor knew his waters. A merchant knew his route. But that knowledge often remained trapped in place.
Today, local signals can aggregate globally.
The world has become instrumented.
That changes the nature of knowledge. The old model asks:
What is the correct rule, doctrine or plan?
The new model asks:
What happens when this thing touches reality?
That is a revolutionary question.
It is also a deeply annoying one for people who have built careers on already knowing the answer.
The modern knowledge engine is increasingly experimental. Build, test, observe, correct. Release to a small group. Measure behavior. Compare alternatives. Learn from failure. Iterate. Update. Kill what does not work. Scale what does.
This is why so much change comes from companies, especially technology companies. Not because they are morally superior. Please. The private sector can produce its own majestic nonsense: management cults, monopoly behavior, PowerPoint cathedrals, HR theology, and products that solve problems nobody had until the product created them.
But companies, especially younger ones, are usually closer to punishment.
A bad product loses users.
A bad service loses customers.
A bad interface creates abandonment.
A bad logistics process creates cost.
A wrong engineering assumption breaks the thing.
A startup with a beautiful theory and no buyers becomes a learning experience with invoices.
The market is not morally pure. But it is often epistemically useful. It tells you, sometimes brutally, that your idea is wrong.
State systems often do not get that feedback with the same force. If a public service is slow, citizens cannot always switch. If a school system fails, children cannot wait twenty years for reform. If courts fall behind, justice becomes a historical reenactment. If regulation is absurd, businesses comply, route around it, or quietly stop trying.
The institution continues. The pain is externalized.
And because the institution survives, it may interpret survival as success.
Why the state responds to complexity with more complexity
Modern society is becoming more complex. Technology, energy, migration, healthcare, finance, climate, education, housing, AI, supply chains, aging populations — all of it is interconnected, fast-moving and difficult to understand.
Faced with this, many public institutions do what they know best.
They produce more rules.
More reporting. More permits. More procedures. More definitions. More oversight bodies. More forms. More compliance layers. More strategic frameworks. More stakeholder consultations. More “integrated approaches,” which often means nobody knows who is responsible but everyone has a meeting.
This is not because civil servants are stupid. Many are competent and overworked. The problem is that the system’s native language is control.
When a complex problem appears, the bureaucratic reflex is to specify.
But specification is not understanding.
Sometimes detailed regulation is necessary. But often it is a substitute for learning. It creates the appearance of mastery while pushing complexity downward onto citizens, companies, schools, doctors, builders, municipalities and courts.
The top produces rules. The bottom absorbs consequences.
And when the consequences become unbearable, the top produces clarifications.
There is a special place in administrative hell for clarifications.
The conservative heart of “progressive” systems
Here is one of the great ironies of modern society: many institutions that call themselves progressive are structurally conservative.
They may speak the language of change, reform, inclusion, transition and innovation. But internally they remain attached to old command structures: hierarchy, credentialism, proceduralism, legalism, committee consensus, top-down categories and suspicion of uncontrolled initiative.
They are progressive in vocabulary and conservative in operating system.
This is especially visible in state environments. The state often wants transformation, but only if transformation behaves itself. It wants innovation that fits procurement rules. It wants local initiative that follows central templates. It wants citizen participation that confirms the framework. It wants digitalization without redesign. It wants experimentation without failure. It wants reform without upsetting the coalition of interests that made reform necessary.
So it produces managed change.
Managed change is often where change goes to become a report.
Private companies can also become conservative, especially large incumbents. Success creates antibodies against discovery. Once an organization has a profitable model, it starts protecting the model from reality. The sales department explains why customers are wrong. The legal department explains why nothing can be tried. The finance department explains why next year would be better. The brand department explains why failure would damage the narrative.
Large organizations, public or private, tend to become temples of yesterday’s knowledge.
The difference is that private temples sometimes collapse. Public temples often get renovated with more taxpayer money.
The destruction of old knowledge
The migration toward bottom-up discovery is not peaceful. It destroys things.
Not only jobs or business models. It destroys prestige, identity and certainty.
An old profession discovers that part of its expertise can be automated.
A university discipline discovers that its theories do not predict much.
A regulator discovers that the technology moved faster than the rulebook.
A school system discovers that children live in a media environment the curriculum barely understands.
A political class discovers that citizens can compare narratives in real time.
A company discovers that customers hate what the internal strategy department loved.
This is why transition is so emotionally charged.
People do not merely defend old systems because they are selfish. They defend them because those systems contain their status, language, training, morality and sense of competence.
When reality says “your model no longer works,” people often hear “your life was a mistake.”
So they resist.
They accuse the new of being irresponsible, dangerous, vulgar, simplistic, neoliberal, technocratic, populist, elitist, woke, anti-woke, insufficiently evidence-based, too evidence-based, or whatever insult the tribe keeps near the printer.
Some resistance is justified. New things can be stupid. “Disruption” has been used to excuse plenty of vandalism with venture capital. Not every old institution is obsolete. Not every startup founder is Prometheus. Sometimes he is just a man in expensive sneakers reinventing the bus, badly.
But resistance becomes pathological when it refuses contact with evidence.
Then old knowledge does not become tradition. It becomes dead weight.
The trap: denial or demolition
When knowledge systems age, societies usually fall into one of two traps.
The first is defensive preservation.
The old system insists it still works. Failures are blamed on implementation, communication, lack of funding, hostile media, bad citizens, insufficient training or the weather. The solution is always more of the same: more regulation, more resources, more authority, more enforcement.
The second trap is revolutionary amnesia.
The new reformers arrive with flamethrowers. Everything old is dismissed as obsolete. Institutional memory is treated as obstruction. Experienced practitioners are ignored. The new system starts from zero, rediscovers old problems, makes avoidable mistakes, and eventually creates its own bureaucracy.
This is how we alternate between fossils and bonfires.
Both are dumb.
Old systems often contain real knowledge. They remember failures. They encode trade-offs. They hold tacit wisdom. They know where bodies are buried, sometimes literally if we are discussing urban planning.
But old systems also accumulate nonsense. Procedures survive after their reasons disappear. Rules protect insiders. Categories become obsolete. Rituals continue because nobody remembers why they began.
The task is not preservation or destruction.
The task is intelligent migration.
Keep what still touches reality. Discard what only protects the institution from reality.
Toward learning institutions
The great institutional challenge of our time is to turn command systems into learning systems.
That sounds mild. It is not. It requires a different posture.
A command system says: we know, therefore we instruct.
A learning system says: we think, therefore we test.
A command system hides failure.
A learning system studies failure.
A command system values compliance.
A learning system values correction.
A command system fears exceptions.
A learning system asks what exceptions reveal.
A command system expands rules.
A learning system improves feedback.
This does not mean anarchy. Bottom-up discovery does not mean everyone gets to do whatever they want while shouting “innovation” into a ring light.
Good bottom-up systems need structure. Science has methods. Engineering has standards. Markets need law. Medicine needs ethics. Software needs architecture. Aviation needs regulation. Nuclear plants should not be managed like a hackathon, unless the hackathon includes evacuation zones.
The point is not to abolish top-down structure. The point is to make top-down structure corrigible.
Correctable. Testable. Revisable. Exposed to feedback.
That is the missing feature in many modern institutions.
How to transition without burning the furniture
So how do we move from obsolete knowledge to better knowledge without either freezing in denial or smashing everything and starting again with the confidence of a toddler holding a hammer?
A few principles help.
First, do knowledge archaeology before reform.
Before changing a system, find out what it actually knows. Not what the official chart says. What it really knows. Who solves problems? Which informal workarounds keep things functioning? Which rules exist because something once went terribly wrong? Which procedures are meaningful, and which are institutional barnacles?
Every organization has grey-haired routers: people who know how things actually move. Ignore them and reform will fail beautifully.
Second, experiment in parallel.
Do not replace an entire system with one grand reform designed by people who have never had to operate it. Create bounded experiments. Regulatory sandboxes. Pilot courts. Experimental schools. Local energy markets. Alternative procurement tracks. Limited-scope digital procedures.
Let models compete against reality before declaring victory.
Third, put sunset clauses on rules.
A regulation should not live forever just because nobody has the courage to kill it. Rules should expire unless they prove usefulness. Review should ask: did this solve the problem, what did it cost, what behavior did it distort, and who now benefits from keeping it alive?
Fourth, regulate outcomes more and methods less.
If the goal is safety, emissions reduction, accessibility, reliability or fairness, define the outcome clearly and allow multiple ways to achieve it. Micromanaging the method freezes old knowledge into law. Outcome-based systems leave room for discovery.
Fifth, use red teams.
Before launching a policy, ask competent critics to break it. Not symbolic consultation. Real attack. How will this be gamed? What incentives does it create? What happens at scale? What will practitioners do to survive it? What will citizens experience? Where does the system fail?
Every major policy should crash in simulation before it crashes into reality.
Sixth, treat citizens and practitioners as sensors.
A complaint is not merely negativity. It is data with emotion attached. A teacher’s frustration, a nurse’s workaround, a builder’s delay, a parent’s story, a small company’s compliance nightmare — these are signals. Institutions that ignore them become blind.
Seventh, separate values from mechanisms.
This is crucial. A person can support cleaner energy and still criticize a stupid energy policy. One can support better education and still reject a fashionable curriculum. One can support safety and still oppose absurd regulation. One can support social justice and still question bureaucratic machinery built in its name.
Modern politics constantly confuses criticism of the method with betrayal of the goal. That kills learning.
A serious society must be able to say: we share the aim; now let us fight honestly about what actually works.
The new humility
The old intellectual posture was architectural.
We design the order, then reality must fit.
The new posture is ecological.
We enter a system we do not fully understand. We observe. We test. We adapt. We intervene carefully. We learn from feedback. We remain suspicious of our own models.
This is not anti-intellectual. It is more intellectually demanding than the old arrogance. It requires theory without priesthood, expertise without immunity, authority without infallibility, and institutions without the childish need to be right all the time.
It also requires emotional maturity, which may explain the shortage.
The future belongs neither to bureaucrats nor to disruptors. It belongs to systems that can learn.
States must learn faster. Companies must remember better. Universities must reconnect theory to reality. Regulators must become more experimental. Citizens must become more than complaint generators. Experts must accept correction. Politicians must stop treating complexity as a communications problem.
Reality has entered the chat.
It is not polite. It does not respect credentials. It does not wait for committee approval. It has no interest in whether your framework was inclusive, evidence-based, stakeholder-aligned or printed on recycled paper.
It simply replies:
No.
Try again.
Look closer.
You missed something.
Your model is wrong.
That answer is annoying.
It is also where knowledge begins.