AI Feels Unprecedented. Every Revolution Did. 

That feeling of standing at the edge of something entirely new is not unique to us. The people watching Gutenberg's printing press run their first pages felt it. The mill workers displaced by steam engines felt it. The first generation to hold the internet in their hands felt it. 

The technology changes. The disorientation does not. History offers us some comfort, and we surely need it when everything around us is changing. 

In the search for context and structure in the current AI wave, it helps to turn to the past. It is fixed. It offers perspective. When we lose that perspective, we lose our ability to interpret what is unfolding before us. 

At the height of the internet boom, Michael Porter observed that in our obsession with what was new, we were forgetting to see what remained the same. That observation remains highly relevant today. The disruptive rise of AI follows a pattern that has appeared before. 

AI feels unprecedented but focusing only on what's new obscures the deeper pattern found in every major transformation. 

Five Revolutions, One Pattern 

More than five millennia ago, writing separated knowledge from the people who carried it. Writing made tacit knowledge explicit. When knowledge became physical, durable, and transferable, power shifted from those who knew to those who could write, record, and interpret.  

Power no longer had to be physically present to be felt. It could be written down, carried, and enforced at a distance. Those who could coordinate others through the written word could command at scale.  

In the 15th century, the printing press collapsed the cost of knowledge transcription, much like digitisation does for many modern products. What was once controlled by the church and court became accessible to merchants, craftsmen, and citizens. 

The Reformation, the Scientific Revolution, and the Enlightenment became inevitable once people could read, compare, and think independently, due to the access to the written word unlocked by the new cost realities. 

The Industrial Revolution began with energy. The steam engine replaced muscle, and physical production scaled, much as AI is scaling knowledge production today. 

When steam power replaced muscle, entire professions vanished. The impact was uneven, concentrating wealth among machine owners. Some lost income; others lost purpose and identity. The AI threat is similar. Threats to our identity drive uncertainty in Generative AI deployment. 

The Luddites who destroyed weaving looms understood this all too well. They were not blind to the shift but were protesting the destruction of their livelihoods. Their resistance failed not because the machines were faster, but because the structural change was already irreversible. Economists would later formalise this uneven distribution pattern as creative destruction. 

The digital revolution introduced a new challenge by giving everyone a voice and a platform. Whether a voice is worth hearing or not, it still finds an outlet and an audience. 

Yet access to that voice is mediated by the most concentrated information infrastructure in history. A handful of companies, such as Alphabet, Amazon, Apple, ByteDance, Meta, and Tencent, shape what we see, what we find when we search for, what we buy, and, increasingly, what we believe. 

Power Never Disappears. It Moves. 

And now there is AI. It appears to reason; it generates, synthesises, and reshapes how we access and create knowledge, even if that knowledge is largely reformulated rather than truly new. 

But the paradox of democratisation has intensified in the hands of a few. A small, influential group is gaining a tighter grip on information and the economy through the silent automation of decisions across society. 

History shows that these breakthroughs were never the story in themselves. They were the means through which deeper societal shifts unfolded. Across each of these moments, a consistent pattern emerges, as each wave shifts power to the next scarcity. 

From those who can write, to those who can print, to those who own machines, to those who own the algorithm, to those who control compute, to potentially those who control data, energy, and infrastructure. 

Power never disappears; it always moves to the next scarcity. 

We now understand that the digital era concentrated power in software. Whoever controlled the algorithm shaped access to attention and markets. This created the most concentrated corporate power the world has seen. 

From Software to Infrastructure and Data 

AI is changing that. What once required millions in software development is becoming more widely accessible, much like the printing press democratised the written word. 

This shift is becoming more visible as Agentic AI and Physical AI use increases. AI is increasingly making decisions, coordinating actions, and interacting with the physical world through autonomous systems. Reactive software and platforms are becoming proactive agents acting on our behalf, increasingly in the physical world. 

The centre of gravity seems to be shifting from the algorithm to the data and infrastructure. Today, it is nearly impossible to browse the internet without creating a free account or accepting tracking, as organisations race to capture and control the data you generate. Data is abundant in volume yet increasingly scarce in value. The raw material is created at unprecedented rates, but high-quality, labelled, and proprietary data, the kind that meaningfully improves AI models, remains a genuine constraint as demand from AI training outpaces supply. 

Compute capacity, energy, rare-earth materials, data centres, and cooling systems are now the new constraints. The digital crumb trails of our digitised existence are fuelling the decision logic previously embedded in the algorithm. 

The lesson from history is that power will not dissolve but move towards these and new emerging constraints. The battle for these resources is already underway, led by the same actors and closely aligned stakeholders whose dominance in software and related industries is now under pressure. 

A New Social Order 

The impact of these shifts is anything but uniform. Sectors, roles, and individuals are affected at different speeds and to different degrees. 

AI is reshaping the social and organisational hierarchy. A growing gap is emerging between those already using AI as leverage and those still waiting for clarity. 

Questions that once took years now get answered in seconds. Analyses that used to take days now take minutes. Carefully built hierarchies of expertise, experience, and knowledge are being disrupted. Sometimes for the better, sometimes not. 

Those who benefit will not necessarily be the most experienced, the most senior, or even the most knowledgeable in the traditional sense. It will not be those with the most profound investments and experience in industries being transformed. 

Those who gain will be those who collaborate best with AI, treating it as a useful but imperfect team member. They will embrace, not fight, new ways of thinking. 

The lesson is not new. The Luddites lost because the structural change was already irreversible. Weaving skills lost value regardless of resistance. Conglomerates with coordination and human-machine management skills became the real winners. It is happening again in the digital world. 

As access expands, scarcity falls, and margins shrink. Once-viable businesses and roles vanish. This is the “IKEA-isation” of knowledge work: standardised, efficient, and accessible. 

This is good news for consumers. IKEA is not only a provider of widely available, affordable flat-pack furniture; it has also helped reshape an entire industry by giving millions access to quality products at affordable prices. 

But this shift created an existential threat for organisations built on the scarcity of craftsmanship. When knowledge becomes a commodity, clinging to old models becomes harder. 

The Real Risk 

Each technological leap feels unprecedented, but deeper patterns of power shifting and reorganisation remain consistent. Misreading these moments causes people to fall behind. 

The real risk is not AI itself but misreading the structural shifts beneath. We have a stubborn tendency to mistake what is new for what is important, and what is disruptive for what is transformative. We have done this before, and we are doing it again. 

French political scientist Alexis de Tocqueville famously remarked, "Quand le passé n'éclaire plus l'avenir, l'esprit marche dans les ténèbres" (when the past no longer illuminates the future, the mind walks in darkness), to describe the profound disorientation caused by the shift from aristocracy to democracy. 

We are at risk of ignoring history’s lessons on how power will move to the next point of scarcity.  

Compute, data, and energy are the obvious candidates today. But there is a less visible constraint, the capacity to make sound judgments about which problems are worth solving, which AI outputs to trust, and when human reasoning should override the machine. Unlike compute, that cannot be scaled by investment. Unlike data, it cannot be harvested. It must be developed. 

People living through the printing press or the Industrial Revolution couldn’t see the endgame. We can't see AI's either. That’s not a failure of intelligence, but a sign of profound change. Our role is to stay oriented and ask the structural questions. 

When the past no longer illuminates the future, the mind walks in darkness. We hold the lessons of history; the key is whether we are willing to use them now to shape a future that is both informed and hopeful.