Civilisations decline when they stop building

"The current economic realities can affect the growth trajectory of an entire generation."

It is a line I have not been able to shake off. Crises have always existed. Every generation has had its war, its recession, its oil shock. What feels different this time is the number of structural forces converging at once, on both the demand side and the supply side of the economy.

• AI is disrupting blue-collar and white-collar work at the same time.
• Asset prices are compounding faster than wages.
• Young populations are entering economies where a first home feels out of reach.
• Ageing societies are increasing fiscal pressure on everyone else.
• Geopolitical fragmentation is redrawing trade and energy flows.
• Attention economies are quietly eroding depth, focus and productivity at large.

Any one of these is manageable. All of them together is a different kind of problem.

The belief that kept the system running

Historically, capitalism survived on one fundamental belief: that effort, skill and persistence could meaningfully improve your life over time.

The system was never perfectly fair. But it was sufficiently aspirational. That distinction matters. A lot.

Economies run on capital and policy. They also run on belief.

We tend to forget the second part. If enough people believe the future can realistically get better through contribution and effort, societies stay productive, ambitious and stable. People study for years, take risks, start companies, and move cities. They delay gratification today for a better tomorrow.

When technological progress and economic gains concentrate in fewer hands, and broad-based upward mobility weakens, that belief starts to crack. And policymakers face a challenge far deeper than inflation or GDP growth.

Why this wave is different

Every major technology wave has displaced work. The loom, the tractor, the assembly line, the spreadsheet. Each time, the fear was mass unemployment. Each time, new work eventually appeared, often better paid and less physically punishing than what it replaced.

So it is fair to ask why this time should be any different. I see three reasons to take it more seriously.

The first is speed. Earlier transitions played out over a generation or more. A farmer's son could become a factory worker, and his daughter an office worker. AI capabilities are improving in months, not decades. Institutions that retrain people, such as universities, vocational systems and corporate learning, move far slower than the technology they are meant to keep up with.

The second is breadth. Industrial automation mostly hit routine physical work. AI touches analysis, writing, coding, design and customer service in one sweep, and physical AI is beginning to reach warehouses, factories and roads. Few roles sit fully outside its reach.

The third is concentration. The returns from AI flow disproportionately to those who own the models, the data, the compute and the distribution. That is a small number of companies and a small number of countries. Productivity can rise sharply while the gains stay narrow.

None of this makes a bad outcome inevitable. It makes a good outcome a matter of choices rather than defaults.

The new addiction is immediacy

The question of paramount importance becomes: how do you sustain belief in long-term progress when short-term incentives dominate everything?

Look around. Immediate consumption. Immediate validation. Immediate political cycles. Immediate shareholder expectations. Every system we have built rewards the next quarter, the next post, the next election.

The things that actually compound national strength work on a different clock. Education quality. Institutional trust. Innovation capacity. Resilience. Cultural discipline. They are long-term investments, for individuals and for nations alike. You do not see the returns in a year. Sometimes not in a decade.

That mismatch is dangerous. A society optimised for immediacy slowly loses the muscles it needs to build.

Civilisations stop building before they decline

Civilisations rarely decline because people stop consuming. They decline when they stop building.

Building roads, schools, research labs and factories. Building skills. Building institutions people trust enough to invest their lives in. Building the kind of patience that lets a young engineer spend ten years mastering something hard.

Which is why the defining policy question of the next decade may not simply be, "How do we accelerate growth?" It may be: "How do we make sure technological and economic progress still expands human opportunity at scale?"

Growth without broad opportunity is a fragile thing. It looks fine in the charts until it suddenly does not.

What a building agenda looks like

It is easy to diagnose and hard to prescribe. Still, a few principles seem clear enough to act on.

• Invest in capability, not only capacity. More data centres matter. So do more apprentices, more lab technicians and more teachers who understand the tools their students will use.
• Lower the cost of starting. Every rule, fee and delay that stands between an idea and a first customer is a small tax on belief. Removing them is cheap compared with almost any subsidy.
• Make long-term capital easier to deploy. Pension funds, sovereign funds and family offices can fund infrastructure, deep tech and manufacturing on the horizons those things actually need.
• Measure what compounds. Track skills, research output, business formation and mobility between income groups, not only quarterly growth.

None of these are new ideas. What is new is the cost of ignoring them while a general-purpose technology rearranges work underneath us.

What this means for those of us who build

I spend most of my time with people who build products, companies and teams. A few things feel clear to me.

• Leaders who invest in skills, not just headcount, will earn loyalty that no bonus buys.
• Companies that treat AI as a way to make people more capable, not just cheaper, will find better talent willing to stay.
• Founders who solve for the median person, not only the top one percent, will find markets most others ignore.

There is also a quieter responsibility. Inside every organisation, leaders decide whether AI is introduced as a threat or as a tool. They decide whether the time it saves is taken away as headcount or reinvested in harder, more valuable work. Those decisions, multiplied across thousands of companies, shape what a generation comes to believe about effort and reward.

If a generation begins to believe effort no longer translates into upward mobility, the consequences will not stay economic for long. They become social, political and civilisational.

Unprecedented times. Definitely not an easy time to be a policymaker. Perhaps not an easy time to be a leader of any kind. The test for all of us is the same: whether what we build today will still be expanding opportunity twenty years from now.