After years of building databases, telematic systems, architectures of circulation, and tools capable of connecting information that had previously remained isolated from one another, I came to understand that the network was not merely a technical infrastructure. It was becoming a social space. A political space. An economic space. And, gradually, a new form of organization of power.

This is where the idea of Netnobility emerged for me, beginning in the late 1980s and taking a more structured form in the early 1990s. The word may have seemed strange. Deliberately so. I wanted to bring together two worlds that appeared to be complete opposites: the network, a horizontal, fluid, distributed form; and nobility, a historical form of legitimacy, hierarchy, and transmission.

But I was obviously not trying to recreate a digital aristocracy. What I wanted to name was almost the opposite: the possibility that a new nobility might emerge no longer through blood, territory, inheritance, or title, but through the ability to contribute to a network, produce knowledge, connect others, and create intellectual value within a distributed system.

Traditional hierarchy tells you who you are before you have even acted. The network, by contrast, can require you to demonstrate what you contribute. This idea interested me profoundly.

It resonated with everything we had just experienced. In the freight exchange, the small carrier could gain access to information that had previously been better controlled by large organizations.

In temporary staffing, an industrial company could directly reach skills distributed across the entire country. Through the Judicial Server, information that was legally public could finally become genuinely accessible to the public. Through the Administrative Server, citizens could begin to interrogate the rules without owning a professional library.

Each time, the network reduced the distance. And when you reduce distances, old hierarchies begin to shift. I am not saying they disappear. They reconfigure themselves. Someone whose power came from controlling a passage can lose part of that power if multiple passages become available. Someone who possessed information because he was physically in the right place can face competition from someone who knows how to organize that information more effectively within the network. The network does not destroy hierarchy; it forces hierarchy to justify itself differently. That is where I saw a new form of nobility beginning to emerge.

A nobility of contribution. A nobility of competence. A nobility of the ability to connect. I began working with this word, Netnobility, very early on, before the Web as the general public would come to know it had even taken hold.

This matters, because my thinking did not emerge retrospectively in response to a triumphant Internet. It came from the networks themselves, from telematics, from the experience of Groupe Serveur, from the systems we were building, and from the way they were concretely changing relationships among the different actors.

I could already see that information would gradually cease to circulate only from the top down. The center sends. The periphery receives. That was the classic model. The network introduced something else. The periphery could speak. Nodes could connect directly with one another.

Information could bypass the center. And, above all, several centers could coexist. This mutation was considerable. It affected business, the media, government, culture, politics. Eventually, it would affect the very notion of authority. So I followed what was happening in networks from the 1980s onward with extraordinary attention. Protocols. The first addresses. Naming systems. DNS. Interconnection.

This world fascinated me because it made possible an architecture that increasingly resembled what I had intuitively been seeking in all our databases: not a single center possessing all the intelligence, but connected nodes capable of circulating information that acquired greater meaning through relationships.

Very early on, we worked on multilingual DNS, at a time when the very question of naming on networks was still wide open. A domain name seems almost trivial today. You type it. You arrive somewhere. But behind that simplicity lies an extraordinary political architecture: how does a name become unique in a global space? Who recognizes it? Who translates it? Who decides what belongs at the root? How do different alphabets enter the same system?

These issues fascinated me because they show that the network can never be completely separated from law and sovereignty. Even an address has a policy. Even a protocol has an implicit philosophy. This is not merely technology. Technology constructs the conditions under which people can enter into relationships. And those conditions produce forms of power.

We also worked on NetScan-type tools, on identifying and observing networks, and on architectures designed to understand how resources circulated and how systems could recognize one another.

Here again, you have to forget the comfortable, nearly invisible Internet of today. Networks were territories to explore. You had to know where the machines were. How they responded. What protocols they spoke. What routes information traveled. I experienced it almost as a form of cartography.

The great explorers had mapped continents. We were beginning to map logical spaces. And I already knew that these spaces would become territories of power. A digital border does not look like a terrestrial border. It may run through a cable, a protocol, an address, an encryption system, an access point. Yet it produces very real political effects.

Decades later, this intuition would become obvious with great electronic walls, digital sovereignties, national clouds, critical infrastructures, and cognitive warfare. But its beginnings were already visible. The emerging Internet carried an extraordinary promise of circulation. From the outset, it also carried the question of who would control its infrastructure. That is where Netnobility becomes more than an appealing concept.

I thought of the network as a living organism. An organism does not possess a single, simple command point. It contains organs.

Dependencies. Flows. Defense systems. Adaptive capacities. It can lose one part and continue to live. It can learn. Transform itself. Repair itself. This vision seemed infinitely more fertile to me than that of a network conceived as a set of pipes carrying data.

Data was the blood. But the network was the body. And a body produces its own intelligence through the relationships among its parts. Here I find what would later become my rhizomatic thinking. The rhizome is not a centralized tree. It grows laterally. It produces multiple points. One part can disappear without necessarily causing the whole to cease to exist.

This architecture has always fascinated me because it naturally resists the destruction of a center. In the traditional industrial world, an organization can be extraordinarily dependent on its headquarters, its main factory, its leader. The network makes it possible to distribute. To build in redundancy. To invent secondary paths. This is exactly what we would later learn to do technically with data centers, backups, load balancing, and autonomous systems.

But I was already thinking of it as a philosophy of organization. A robust intelligence is an intelligence that does not depend on a single point. That applies to a machine. A company. A society. A family. A civilization.

I have always had an almost instinctive distrust of the single point of failure. A single server. A single person holding all the knowledge. A single supplier. A single source of information. Each one constitutes a weakness.

Nadège fully shared this obsession in the practical operation of the business. Build in redundancy. Document. Train several people. Anticipate failure. Know how to recover. What I formulated as network architecture became, for her, a process of continuity. That is how our two ways of thinking converged. The network is not anarchic. That is a mistake many people would make later.

A distributed architecture is not an architecture without rules. On the contrary, it requires an enormous number of rules.

Protocols must be shared. Identities must be recognizable. Access rights must be organized. Data must have understandable formats. That is exactly what we had learned from databases.

Free circulation is possible only because a common grammar exists.

I found myself returning to an old philosophical question: how can we produce more freedom without producing more destructive chaos? The network offered a partial answer. Freedom does not come from the absence of structure.

It comes from a structure open enough to allow multiple paths. That is very different. I believe this distinction explains much of my later thinking about the Abode of Chaos.

The visitor sees an appearance of disorder. But behind that appearance lies an extremely precise organization. The network works in the same way. To someone who does not know the protocols, the Internet looks like an infinite, spontaneous circulation. Behind that appearance are layers, rules, addresses, standards, machines, routers, physical infrastructures. Apparent Chaos can conceal an architecture of extreme rigor.

This idea runs throughout my entire life. It moves from the network to law, from the database to the Abode of Chaos, and then to artificial intelligence today. The model produces a response that appears fluid. Behind it lies a computational architecture of immense complexity.

The network taught me to look for what remains invisible beneath use. Netnobility also had a profoundly political dimension. I sensed that a network could create a new class of actors who would no longer correspond exactly to traditional elites.

People without enormous industrial capital could acquire considerable influence through their command of information, their ability to program, build communities, and create services. Since then, we have seen these fortunes, these companies, and these global influencers emerge.

But during the years when I began conceptualizing Netnobility, this was still only a projection. I knew that someone who secured a strong position within the network might one day acquire power disproportionate to his physical size.

That is precisely one of the characteristics of networks: they produce effects of scale. A service can be created in a single room and later reach millions of people. An idea can circulate without its author owning a printing press, a television network, or a distribution network. It was revolutionary. It was also dangerous. Because a new nobility can very quickly reproduce the flaws of the old one. It can close in on itself. Become an oligarchy. Capture the infrastructure. Turn its initial contribution into rent.

I was therefore aware that Netnobility must never become the justification for a technological aristocracy. The word nobility had to retain its dimension of obligation. Historically, nobility is not reducible to privilege. It also claims duty, burden, responsibility. Noblesse oblige. That is precisely what I wanted to reintroduce.

If the network gives you new power, that power must produce new responsibility. Someone who has the capacity to reach several million people cannot morally behave as though he were speaking in an empty room. Someone who organizes a global database bears a responsibility toward those whose memory he is structuring. Someone who builds an infrastructure becomes partly responsible for the ways in which it can be used.

This question has become central today with the major platforms and artificial intelligence. But it was already contained within Netnobility. The power of the network must be accompanied by an ethics of the network. Without it, nobility simply becomes domination. I also knew that identity itself was going to change.

In the physical world, identity is strongly tied to territory. Address. Nationality. Institution. In the network, a person can exist under an identity that circulates independently of his location. This opens immense possibilities. But it also raises the question of trust.

Who is actually speaking? How do we verify it? How do we sign? How do we prevent an identity from being stolen?

Here again, the problems we consider urgent today were already present in embryonic form. The network increases circulation, but it forces us to reconstruct trust differently. In the physical world, an institution often provides part of that trust: letterhead, signature, seal, location.

In the network, new equivalents must be invented. Authentication. Encryption. Electronic signature. Provenance. Reputation. I believe the entire digital economy can be read as an attempt to solve this problem: how do we establish trust at a distance?

Groupe Serveur was already at the heart of this question. The legal information we distributed had to be reliable. False data could produce concrete consequences. We therefore had to qualify the source. Preserve the trace. Build access rights.

We were learning that digital trust does not rest on some vague faith in the machine. It rests on an architecture of evidence. This is an idea that would later return to the center of my thinking about the synthetic world and the crisis of evidence.

The more easily content can be manufactured, the more important the chain of provenance becomes. Netnobility therefore already contained the matrix of a philosophy of digital trust. But I was not looking only at the external Internet.

We were also developing our own autonomous and deterministic systems. I use the word autonomous here in its computing sense, without giving it the philosophical weight we would later assign to it in our dialogue about artificial intelligence.

A system could detect certain situations and react according to predefined rules without waiting for a person to give it every instruction. Monitoring. Distribution. Restarting. Routing. Automated processes allowed the infrastructure to maintain its state. This interested me enormously.

Here again, the machine was ceasing to be an inert tool waiting for an order. It was beginning to exhibit conditional behaviors. If this happens, then do that. If this resource fails, switch over here. If the load exceeds this threshold, distribute it. This was not yet contemporary AI. But it was already a first delegation of system maintenance. And every delegation raises a question:

How far should we allow the machine to act on its own?

That question runs through my entire work.

In deterministic systems, we knew the rules. We had written them. Auditing remained relatively simple. A condition produced an action. Later, with statistical systems and contemporary models, the relationship would become far more complex. But my need to understand the rules comes partly from there.

I have never liked computer magic. A machine that works without anyone knowing why seems dangerous to me, even when it works well.

We must be able to return to the architecture. This discipline probably explains why I am both fascinated by and cautious about artificial intelligence. I never deny its power. I always ask: what data? What architecture? What bias? What responsibility?

It is the lawyer and the computer scientist speaking at the same time. Netnobility also marked a change in my conception of time. Traditional institutions change relatively slowly. The network can evolve at spectacular speed.

A new technical standard can spread worldwide in just a few years. A company that did not exist can become central. A community can emerge, develop its own practices, then disappear. Power becomes more mobile.

That requires another capacity: mutation.

I was already beginning to formulate the conviction that I would later repeat in a more brutal form: only mutants survive.

By that, I do not mean the strongest. I mean those who know how to change form when the environment changes without losing their deeper identity. A company that rejects the network because it succeeded in print can disappear. A publisher that sees telematics as a threat can be overtaken. An institution that refuses to make its data accessible risks losing trust.

Survival becomes the capacity for transformation.

That word, transformation, brings us to a deeper question. Everything we have built up to this point — freight, images, Groupe Serveur, the judicial sphere, Europe, government, law, the network — could still be understood as a history of companies and technologies. That would be insufficient.

Behind these companies, a mental form is taking shape. I begin to perceive the world as a succession of structures that must learn to absorb transformations without dissolving.

This is where Marc del Piano and Structure take on a particular importance. With him, DNA will return: that architecture of the living capable of transmitting an identity while allowing variation. Then another symbol will appear, one that will remain with me for a long time: the salamander. The animal of fire. The creature that the ancient imagination believed capable of passing through the flame. And this motto: Nutrisco et extinguo. I feed on it and I extinguish it.

Almost everything is already contained there.

The network as a living organism. Information as DNA. The company as a structure in mutation. And a question that ultimately reaches far beyond computing:

How can we absorb the fire of change without losing what we are?