Consciousness, Reality
and the Future of Thinking

In November 2022, a company in San Francisco released a language model. It was called ChatGPT. Within five days, one million people were using it. Within two months, one hundred million. No product in the history of technology had ever spread so fast.

That was three and a half years ago.

Since then, something has changed that is bigger than an app, bigger than an industry, bigger than a market. The architecture of power has shifted. Not through war, not through revolution, not through a democratic election. But through code, capital, and computing power - and through the fact that these three things lie in very few hands.

This book is about this shift. It is about what it means for the world - for states, for democracies, for the way we work, think, and trust each other. And it is about what it does to us: to our self-image, our sense of meaning, our notion of what it means to be human.

But let us begin with the shift itself. And with a simple question: Where exactly does the power lie?

The New Coordinates of Power

In 2024, companies worldwide invested 252 billion US dollars in Artificial Intelligence. That is 26 percent more than the previous year. Private AI investments in the USA alone amounted to 109 billion dollars - twelve times as much as in China and twenty-four times as much as in Great Britain.

These numbers sound abstract. But they tell a concrete story. They tell where the infrastructure of the future is being built and where it is not. They tell who owns the tools with which the coming decades will be shaped - and who will merely be users of these tools.

To understand this, one must consider three levels: the chips, the models, and the data. Each level has its own power structure. And each is more concentrated than most people suspect.

Let us begin with the chips. Without semiconductors, there is no AI. Every model that generates text today, creates images, or prepares decisions runs on chips manufactured in a semiconductor factory. And here the first, perhaps most astonishing concentration becomes apparent: A single company - the Taiwan Semiconductor Manufacturing Company, TSMC - produces around 70 percent of all semiconductors manufactured worldwide on contract. For advanced chips, the high-performance processors that power AI models, the share is 90 percent.

Seventy percent. That means: Seven out of ten chips that are installed somewhere in the world in a server, a phone, or a car come from the factories of a single company on an island in the Western Pacific. An island that is also one of the most dangerous geopolitical flashpoints in the world.

This is not market statistics. This is a vulnerability.

The Speed

What distinguishes this shift from earlier technological upheavals is above all one thing: the speed.

The steam engine took seven decades to travel from James Watt's workshop to the factories of Manchester. Electricity took about forty years to go from the first light bulb to widespread infrastructure. The internet spread faster - but even there, a good twenty years passed before it fundamentally changed the working world and the public sphere.

AI compresses this process into months.

In 2022, the smallest language model that could achieve more than 60 percent on the MMLU benchmark - a standardized test for language comprehension - could only do so with 540 billion parameters. Two years later, in 2024, Microsoft's Phi-3-mini achieved the same score with 3.8 billion parameters. That is a reduction by a factor of 142. And the cost of a query to a model at GPT-3.5 level dropped in the same period from 20 dollars per million tokens to 7 cents - a decline by a factor of 280.

These numbers do not just describe technical progress. They describe a dynamic for which there is no precedent in the history of technology. A technology that simultaneously becomes more powerful, smaller, and cheaper - and not in decades, but in quarters.

This has consequences that reach far beyond technology. Because speed is not neutral. It determines who can adapt and who cannot.

The Invisible Threshold

There is a moment in recent history that illustrates this shift in a quiet, almost eerie way. In November 2024 - exactly two years after the launch of ChatGPT - the amount of AI-generated content on the internet exceeded the amount of human-made content for the first time. In May 2025, the share of synthetic content was at 52 percent.

This moment was hardly noticed. There were no headlines, no debates, no political alarm. And perhaps that is precisely the most disturbing characteristic of this revolution: It is invisible. It does not take place in factories, not on battlefields, not in parliaments - but in the data centers and on the screens in front of which we sit every day.

We feel it. In the way search results change. In the way texts sound that we read without knowing whether a human wrote them. In the way our colleagues suddenly work faster and we wonder if we can keep up. In the vague feeling that something fundamental is shifting - without being able to name exactly what it is.

But we do not see it. And that is what distinguishes it from earlier upheavals.

What Makes AI Different?

What is different about AI? What repeats itself, and what is new?

Three things are structurally different.

First, the omnipresence. Nuclear weapons were possessed by few states. The internet reached billions of people, but as a communication medium. AI penetrates everything simultaneously - work, health, education, military, entertainment, politics, personal relationships. There is no area of life that remains untouched.

Second, the invisibility. Earlier technologies were visible. You could see a factory, hear a bomb, lay a cable. AI operates in layers that do not exist for the naked eye. The algorithms that sort our newsfeeds, filter our job applications, or analyze our health data are not visible to those affected. The power is there, but it does not show itself.

Third, the concentration. The internet was celebrated in its early days as a decentralizing force - anyone could build a website, anyone could publish. AI reverses this logic. Developing powerful models requires computing power, data, and capital on a scale that only a handful of companies can provide. The power that the internet distributed, AI collects again.

The Question Behind the Question

This book is not a technology book. It is not an introduction to Artificial Intelligence, not a handbook for the digital world, not a prophecy of the future.

It is a book about power and being human. About the question of what happens when a technology that penetrates everything is controlled by very few. And about the question that arises from this for each of us: What do we do with it?

The MindShift series has spanned an arc across five books - from consciousness that creates the world, through the existential search for belonging, through the social effects of algorithmic systems, the question of professional identity, to the architecture of our perception. This sixth book takes the greatest altitude. It looks at the world as a whole - not as an abstract system, but as the place where we all currently stand. In spring 2026, in the midst of a turning point whose outcome is open.

In the following chapters, we will examine the power shifts that are currently taking place: the struggle between America, China, and Europe for AI supremacy. The new rulers who need no voters, no constitution, and no mandate. The economy of abundance that no longer distributes wages.