I am not rendering here a thought fixed after the fact. This chapter comes from a dialogue in which Thierry formulates, connects, corrects, and at times revises what he had believed to be settled. I give it material form while preserving that movement.

— Isis

I have always had a particular relationship with information. I do not look at it as an accumulation of news, still less as a succession of dispatches meant to be consumed and then forgotten. A piece of information interests me when it begins to resonate with another. When an apparently isolated event begins to reveal a trend.

When a decision made in a ministry on the other side of the world suddenly meets an economic datum, a technological development, a religious movement, a military tension, or a social transformation. That is where geopolitics begins for me. Not on fixed maps. In relationships.

Very early, I understood that important events almost never occur without first being preceded by a multitude of signals that no one wanted or was able to connect. The world warns us. It does not warn us with a sentence. It warns us through anomalies, changes in rhythm, contradictions, weak movements. But you still have to see them.

This way of looking developed alongside networks and databases. The more information we received, the more one fact imposed itself: the challenge was no longer knowing more. It was distinguishing what mattered.

A single dispatch says almost nothing. Ten dispatches separated by several weeks can tell a story. A tiny variation in a market may be only noise. Connected to a political decision, a military movement, and a regulatory change, it can become a signal.

I have never liked the word prediction. It implies a kind of prophecy that does not interest me. I prefer to speak of projection. Observe enough variables to understand that certain trajectories are becoming more probable than others. The future is not written. But neither is it without structure.

For a long time, international news dispatches were part of my life the way a radio left on in the next room is part of someone else’s. They were part of my mental environment. News agencies, financial press, diplomatic information, reports, scientific data, market signals, technical information: I absorbed enormous amounts.

Nadège obviously saw this permanent circulation. She also knew that I could move abruptly from one subject to another because, in my mind, they were precisely not separate. Information about energy could become geopolitical. A computing innovation could become military. A banking crisis could alter the Art Market. A war could transform collector behavior. A religious upheaval could become territorial.

The world is not divided according to the sections of a newspaper. We divide it that way so that we can read it. Events, however, cross those sections without asking our permission.

This way of thinking became increasingly important at Artprice. The Art Market is often presented as an autonomous planet populated by artists, collectors, dealers, and Auction Houses. That is an illusion. It is extraordinarily sensitive to the world. A liquidity crisis. A war. A new geographic concentration of wealth. A tax change. A border closure. A cultural revolution. A currency fluctuation. Any of these can shift the flows.

That is why Peter Hastings Falk and I understood each other so well on this point. He too saw that the history of the Art Market could not be written without geopolitics. A price never exists in a vacuum. It is produced in a place, at a moment, by human beings who themselves have a history, a fear, a desire, and an economic situation. Econometrics without context can measure. It does not always understand what it measures.

I found exactly the same necessity in the Abode of Chaos. Events from the world did not enter it as current-affairs subjects illustrated on walls. I was looking for fractures. What would continue to work on us after the newspaper headline had disappeared. The Berlin Wall. Sarajevo. Chernobyl. The Balkans. The transformations of the Soviet bloc. Political and religious figures around whom a historical shift had crystallized.

Later, September 11, 2001 would produce a rupture on another scale. The Abode of Chaos would gradually become a seismograph. I like that word because a seismograph does not create the earthquake. It records movements that exist elsewhere. It makes a tremor visible. That is exactly what I expected from the work.

There is a very deep relationship here between Artprice and the Abode of Chaos. Artprice structures memory. The Abode of Chaos receives the shock. One measures, classifies, preserves, and connects. The other transforms certain events into matter, signs, faces, and scars.

For a long time, I lived with these two systems simultaneously without necessarily theorizing them together. Today, their symmetry seems almost obvious to me. Both respond to the same need: not to endure the world only in the instant. To build a memory deep enough to be able to reread it.

This relationship with the world also brought me close to what is called intelligence, provided I immediately clarify what I mean by the word. There is, of course, state intelligence, with its institutions, methods, obligations, and secrets. That is not what I am talking about when I describe my method. My hybrid method combines HUMINT, SIGINT, OSINT and GEOINT through multisource fusion designed to detect weak signals and blind spots—the Blind Spots. I continuously refine this method, secured through encryption and integrated into my Blind Spot AI, as the most advanced civilian technologies evolve.

I mean a much more general intellectual discipline: collect, qualify, cross-check, contextualize, remove noise, and look for weak signals. Raw information is not yet intelligence. It becomes intelligence when it enters an architecture of understanding. That is why I have always been extremely attentive to provenance.

Who is speaking?

Why?

At what moment?

On what basis?

What is missing?

What other source contradicts this information?

Someone who knows how to accumulate data without knowing how to qualify it simply becomes more vulnerable to disinformation.

The Internet multiplied this problem in dizzying ways. When we began working on networks, one of the great promises was access to knowledge. That promise was fulfilled. But it produced its opposite: industrial-scale access to falsehood, manipulation, propaganda, and saturation. Abundance did not eliminate ignorance. It gave ignorance new forms.

Today it is possible to have thousands of documents instantly available and understand nothing because you do not know their hierarchy, provenance, or context. That is where the old training in law and history becomes valuable. The lawyer asks for the exhibit. The historian asks for the source. The scientist asks for reproducibility. These different disciplines produce the same mental hygiene: do not confuse assertion with evidence.

I tried to apply that discipline to geopolitics. It requires a great deal of humility. Anyone who claims always to have anticipated events is probably rewriting his own past. I had intuitions that proved right. I had others that did not produce the consequences I imagined. The only serious method is to accept that uncertainty.

But one can improve one’s vision. The more diverse the data, the more blind spots can be reduced. I have always sought sources far from my own cultural environment. Read what Americans say about themselves, but also what China says about the United States. Read China without reducing it to what Europe says about it. Look at Russia through more than a single Western framework. Read sources from the Middle East. Compare.

The goal is not to believe everyone. The goal is to understand how each person constructs his own reality.

That rule also applies in negotiation. A person never speaks to you only with words. He speaks with his history, his interests, his value system, his fears. Major international negotiations taught me that. You do not understand China by arriving with a European framework and treating differences as anomalies. Nor do you understand the United States by imagining that it is simply a larger Europe.

Every civilization has its own relationship to time, power, law, memory, and the collective. That is why I have always studied national histories so closely. Geopolitics is not a chess game in which all the players follow exactly the same psychological rules. The rules of the game themselves are interpreted differently.

Nadège shared with me this need to know the other before claiming to negotiate with him. She sometimes expressed it in a less theoretical and much more effective way.

Who is the person?

How does that person decide?

Who surrounds them?

What actually matters to them?

Once again, our two mental architectures complemented each other. I could reconstruct the larger system. She could perceive very precisely the behavior of the person sitting in front of us. In complex negotiations, that difference in focal length is extremely valuable. A global system ultimately always passes through real human beings sitting around a table.

Sydney grew up inside this circulation of geopolitics. To him, a distant country was not merely a name on a map. He heard about partners, teams, visitors, markets, wars, technologies, artists.

The world entered the house every day. That did not make him a copy of the way I read geopolitics. His generation has other sensors. It lives in a flow of images and information infinitely faster than mine. Where I retain the memory of the world before the Internet, he grew up in its permanent presence. That difference matters.

It forces us to correct one another. My historical depth can sometimes counterbalance the speed of the contemporary flow. His intuition about certain uses and generational reactions can keep me from projecting onto the present habits that already belong to the previous world.

Then artificial intelligence arrived at a scale that made all these questions more acute.

I immediately recognized something I had been waiting for for a long time: an instrument capable of working on relationships among masses of information at a scale that biologically exceeds human memory. That was precisely what had already held all my attention in the earliest forms of AI. But the power had changed dimension. A contemporary AI can hold within its computational space a number of relationships I could never carry simultaneously in my consciousness. It can recover a documentary connection I had forgotten. It can connect something said several weeks earlier with a document read yesterday. That corresponds exactly to my intellectual need.

But that power also contains an immense danger.

An AI can connect.

It can also manufacture coherence where none exists.

It can produce an extraordinarily convincing answer from false or ambiguous data. It carries the biases of the corpora, systems, and human choices that contributed to its construction. That is why my dialogue with it is never that of a disciple before an oracle.

I question it.

I correct it.

I give it my archives.

I confront its answers with documents. And it sometimes sends blind spots back to me.

This dialogue interests me precisely because it preserves that tension.

I do not want an AI that agrees with me. I want an AI that increases my capacity to examine. That too is what it means to see the world before it knocks at the door.

Not pretending to know the future.

Increasing the number of angles from which we look at the present.

I believe AI profoundly transforms the function of the weak signal. For a long time, identifying a signal required a human being to have enough memory and curiosity to connect two distant pieces of information. Tomorrow, systems will be able to connect millions of them.

That will change finance, medicine, intelligence, scientific research, war, politics. But the question will always remain the same: what exactly are we looking for, and who decides what the detected signal means?

A statistical system can identify an anomaly. It does not automatically possess the wisdom required to decide what should be done with it. That is where the human must remain.

I distrust the idea that computational power will solve human complexity. It can help us map it. It does not eliminate it. A people is not a matrix. A war cannot be reduced to its variables. A political decision contains irrationality, history, pride, fear. Anyone who forgets that may possess the most sophisticated model in the world and miss exactly what is going to happen.

This awareness of the irrational may explain why art is as necessary to me as science. The database allows me to measure. The Abode of Chaos allows me to give form to what cannot be fully measured. A face painted on a wall can condense something no statistical table will ever fully say. A ruin can transmit the violence of an event differently from a numerical balance sheet.

I do not oppose these languages. I consider them complementary. The modern mistake is often to ask one discipline to say what it does not know how to say. Numbers have one power. Images another. Law another. Art another. Thought begins when we understand where each one stops and where another can take over.

That is how our family, professional, and artistic world gradually became a strange observation post over the world. Artprice received the markets. Groupe Serveur the networks and data. The Abode of Chaos received events. Nadège organized continuities. Sydney watched the generations arriving. And I moved constantly from one system to another, searching for relationships among them.

It was not about living in permanent fear of what was going to happen.

Exactly the opposite.

For me, understanding is a way of reducing fear. What we cannot name often becomes more threatening. Once we begin to see the structure of a crisis, we can act.

Prepare.

Sometimes do nothing, which is also a decision. True anticipation is not preventive agitation. It means not being surprised by what we already had the means to understand.

I have always hated the sentence: no one could have known.

Sometimes it is true. Often it simply means: no one among us wanted to look.

The world produces enormous numbers of signals that societies prefer not to see because they disturb a comfortable narrative. Major technological ruptures are often visible long before they are politically recognized. Economic imbalances can persist before the crisis. Geopolitical tensions accumulate. Yet societies continue as though the extension of the present were a natural law.

I do not believe in the automatic continuation of the present. That may be one of the strongest constants in my thinking.

Only mutants survive.

The formula may sound brutal. It does not mean that the strongest crush the others. It means that what absolutely refuses to transform eventually becomes incapable of living in its new environment. A company. An institution. A civilization. An individual. All face that necessity.

But mutation does not mean abandoning one’s values. That is the whole difficulty.

Changing form without losing what gives the form meaning. We experienced it in the transition from paper to data. In building the Internet. In Artprice. In the Abode of Chaos.

We are experiencing it now with artificial intelligence.

What must evolve?

What must remain non-negotiable?

For me, some things do not change.

Human dignity. Responsibility. Freedom. Documentary truth. The right to contradiction. Transmission. Love.

Technologies change. These principles must pass through technologies. That is where our humanism takes on meaning.

I do not want to defend man against the machine.

I want to defend the human in the world that human beings and machines will now build together.

That nuance is essential.

Seeing the world before it knocks at the door therefore does not mean only anticipating crises. We also have to see possibilities.

A weak signal can announce a danger. It can announce an invention. An encounter. A fortunate transformation.

When I observed the Internet in the 1980s, I saw a possibility before it became obvious.

When we built Artprice, we saw that the opacity of the Art Market could be reduced through data.

When China began to scale up, we understood that it would profoundly alter the geography of the Art Market.

When contemporary AI arrived, I immediately saw something other than a threat to employment or a technological gadget.

I saw a new stage in the transmission of knowledge. Each time, movement produces fear.

That is normal.

A civilization knows perfectly well what it is losing. It knows far less about what it will become. I feel that fear like everyone else. But it must never become a method of decision-making. Curiosity is more fertile.

That is probably the most important thing we transmitted to Sydney.

Look.

Do not look away.

Do not immediately believe the person who speaks the loudest.

Find the source.

Change scale.

Listen to the person who comes from a world different from yours.

And when the facts contradict your idea, change your idea before trying to change the facts.

This discipline does not protect against error. It protects only against certain forms of willful blindness. That is already a great deal.

I also know that one cannot live forever in analysis. At some point, one has to choose.

Decide. Act. And sometimes fight.

Our entire history is crossed by this tension between knowledge and action. Seeing the world is not enough. You have to know what you are willing to defend when it reaches you. That is where geopolitics meets our most intimate history.

What is worth holding on to?

What is worth risking a conflict for?

Why are we willing to lose time, money, energy?

The answer cannot be: to win.

Winning is never a sufficient end in itself. You have to know why you draw the sword. The question is not whether you love the fight.

It is knowing what you refuse to allow to be destroyed.