We’re About to Live Through 100 Years of Change Every Year

The 'Century-a-Year' Economy

From A Flickering Image In Soho To Intelligent Machines

In January 1926, a small group gathered in an attic laboratory at 22 Frith Street in Soho, the ground floor of which is now occupied by the legendary Bar Italia. John Logie Baird showed them something that must have seemed as much magic as new technology: a flickering, blurred moving image transmitted electrically. It was not yet modern television. The image was faint, unstable and barely convincing. But a machine had begun to capture sight and send it somewhere else.

Watching The Pace Of Change

This is a reminder of what just one technological frontier looked like a century ago. In 1926, television was not an industry. It was an experiment. Antibiotics had not yet transformed medicine. The Wright brothers had flown only 23 years earlier, and aircraft in 1926 were fragile wood-and-fabric machines with open cockpits and exposed wires and struts, flying barely faster than a car at around 100 mph. Telecommunications in 1926 relied on copper telegraph and telephone lines, with most calls manually connected by human operators, and long-distance communication remaining slow and expensive.

Fifty Years Ago, The 2026 Future Had Arrived In Fragments

Now fast forward forward fifty years.

In 1976, when I was five years old, parts of the 2026 future had clearly arrived, but only in pieces. Concorde entered commercial service and made supersonic passenger travel real, though only for a tiny sliver of the market. Queen Elizabeth II sent an email from the Royal Signals and Radar Establishment in Malvern, but email was still a demonstration, not a daily habit. Steve Wozniak completed the Apple-1, a single-board computer for hobbyists, and Apple was born from an order of 50 machines from the Byte Shop. In the same year, researchers were already experimenting with robotic grippers inspired by elephant trunks and snakes, hints of the long road to machine dexterity.

Fifty years ago, the 2026 future arrived in isolated flashes. A supersonic jet here. An email demonstration there. A hobbyist computer on a workbench. A robot gripper in a lab. These were remarkable, but they were separate. The machine that handled communication was different from the one that handled computation, which was different from the one that handled engineering, which was different from the one that handled biology. Progress was real, but it moved in silos.

That is no longer true.

Stacked Acceleration

What makes 2026 different is not simply that the technology is better. It is that technologies are beginning to stack and accelerate one another. We are seeing exponentials stacked on exponentials.

AI improves software. Software improves science. Science improves medicine. AI improves robotics. Robotics improves manufacturing. Better chips improve AI. AI helps design better chips. Biology becomes computational. Intelligence becomes a service. AI agents do work. We are on the cusp of recursive learning.

The many separate research and economic lanes and silos are collapsing into a single compounding system. We are moving towards a world in which certain domains may deliver the equivalent of 100 years of progress in a single year.

Not everywhere. Not all at once. Not in a smooth line. The technological frontier remains jagged, and capability is still ahead of deployment. But in the core engines of discovery, design, diagnosis, automation and invention, that kind of time compression is becoming noticeable.

Intelligence Was Always The Bottleneck

For most of human history, progress depended on human cognition. Brains reading papers, running experiments, writing code, designing tools, comparing results, arguing, refining and repeating. Human learning and human research were the platform, and intelligence was the bottleneck. Even when we had great machines, the speed of progress was still constrained by the speed of human thought, input, output and coordination.

Now intelligence itself is becoming industrialised. Dario Amodei, CEO of Anthropic, has said advanced AI could soon give us the equivalent of a nation of geniuses running inside a data centre, working simultaneously on science, engineering and problem-solving.

AI systems can already write code, interpret documents, analyse images, search enormous scientific spaces, generate hypotheses and increasingly act across software systems. They are not perfect, yet, but they do not need to be flawless to change the rate of progress. They only need to be good enough to remove part of the cognitive bottleneck, and then improve quickly from there.

That is what we are now experiencing.

A task that was impossible for AI last year becomes mediocre this year, useful the next, and then routine. A scientific search that once took months becomes days. Problems in mathematics and physics that once took long stretches of human effort can now be explored far faster. A software workflow that once required a team becomes something one capable person can run with agents. A diagnostic process becomes faster, cheaper and more scalable because the interpretation layer improves.

The future stops arriving in isolated inventions and starts arriving as system-wide acceleration.

Computational Biology And Medicine

Gene editing has crossed the threshold. Healthcare and biology may be where progress becomes most visceral and, literally, life-changing.

In 2023, the FDA approved Casgevy, the first FDA-approved therapy using CRISPR/Cas9 genome editing, for sickle cell disease. That was not just another drug approval. It was a threshold moment. We moved from talking about gene editing as a powerful scientific tool to seeing it recognised as an approved therapeutic intervention in patients.

In early 2026, the FDA cleared Life Biosciences to begin a first-in-human trial of partial epigenetic reprogramming, an approach intended to reverse cellular ageing markers in the eye. This is not immortality. It is not age reversal across the whole body. It is an early, tightly constrained clinical test. But it is still a line in the sand. Regulators are now prepared to let human trials begin on therapies explicitly aimed at rejuvenation biology itself. Once that door opens, capital, talent and experimentation tend to follow.

Compounding Medicine

A century ago, medicine was still basic. Penicillin had not yet been discovered, and cancer treatment in 1926 was largely confined to surgery and early radiation.

Today, medicine is becoming computational across the whole stack: earlier detection, better diagnostics, more precise imaging, biomarker-led treatment selection, faster drug discovery and more targeted therapies.

And the pace is now accelerating even within a single business cycle. In just the last five years, DeepMind’s AlphaFold has moved protein structure prediction from a decades-long scientific bottleneck to a system that can predict structures at scale with accuracy competitive with experiment, and the AlphaFold database now contains over 200 million predicted structures. In parallel, gene editing has moved from promise to approved therapy, and the first human trial of partial epigenetic reprogramming has now been cleared.

Cancer is one part of that story. Multi-cancer blood tests aim not just to detect a cancer signal, but to predict where in the body it may be coming from. The results so far are mixed, which matters because it stops the story becoming simplistic. Biology is hard, but the direction is clear: we are seeing more ways to detect disease earlier, diagnose it more precisely and develop more effective treatments. We are not at the point where cancer is simply solved, but we are seeing a growing stream of credible breakthroughs and powerful new treatments across very different cancer types, something that would have been hard to imagine even a few years ago.

This is why the health story matters so much to my century-in-a-year thesis. AI is beginning to accelerate the search process in biology itself, helping move from detection to diagnosis, from diagnosis to target, and from target to treatment more quickly than human research could manage.

AGI Changes The Rate Of Change

This brings us to AGI, which I have written about in previous editions of this newsletter.

Ray Kurzweil forecasts human-level AI arriving around 2029. Demis Hassabis has said that AGI-like systems could begin to emerge in the next five to ten years, and has described AGI as potentially having ten times the impact of the Industrial Revolution at ten times the speed. Dario Amodei has argued for very short timelines to highly capable AI, far shorter than the old mid-century consensus.

Their views and timelines may differ, but they rhyme, and crucially there is a continuum, not a sudden switch being flipped one morning. Long before anyone agrees that AGI has officially arrived, we are likely to live through a period in which AI capabilities move steadily closer to it, and even that approach will be enough to change science, medicine, business and productivity in profound ways.

You do not need to believe the most aggressive timeline to see the implication.

The Shape Of Progress Has Changed

The point is not simply that 2026 is more advanced than 1926. Of course it is. The point is that the shape and pace of progress has changed and will it will continue to accelerate.

In 1926, Baird’s blurred television image hinted that a new medium might one day exist. In 1976, Concorde, email, the Apple-1 and early robotics hinted at different futures, but those futures still arrived in fragments. In 2026, those strands are converging.

The issue is no longer capability alone. It is implementation.

The frontier is racing ahead. Deployment is lagging behind. That gap is where the next decade will be won or lost, by individuals, companies and countries alike.

Years That Contain Centuries

The twentieth century gave us a hundred years of extraordinary change. The next phase may give us something stranger, and much faster - years that contain centuries of change compared to the past.

The question is no longer whether the world will change faster.

It is at what pace and whether we can implement, absorb and act on that change quickly enough to remain productive, competitive and relevant.

What is your plan over the next year?

Thank you for reading.