Companies of 300 Employees, But Only 30 Humans

There are moments when you suddenly feel your age. I had one recently during a trip to Silicon Valley. Before a startup founder was due to give a presentation, we grabbed coffee together. Trying to make conversation, I asked what seemed like an innocent enough question: “How many employees do you have these days?”

His answer has stayed with me: “300, though only 30 of them are human.”

The rest? Agents, software, and intelligence.

That single remark may capture the most profound economic shift since the Industrial Revolution. Companies are no longer being built around people, but around intelligence. Human where necessary; artificial wherever possible.

Since late February, there has been a recurring theme in my conversations with business leaders. Much of it starts with Block, Jack Dorsey’s fintech company. While less familiar in Europe, Block is a major force in the U.S., spanning payment systems, digital wallets, and buy-now-pay-later services, with a workforce of more than 10,000.

Or at least it did.

A few weeks ago, Dorsey announced plans to cut 40 percent of the company’s staff: 4,000 jobs. The market’s reaction? The stock jumped 17 percent.

His message to shareholders was even more striking: “We believe Block will be significantly more valuable as a smaller and faster intelligence-native company.” And then the kicker: “I don’t think we’re too early. I think most companies are too late.”

Since then, the conversation has shifted almost entirely to one question: how do you become faster, leaner, more adaptive? And above all, what does it actually mean to be intelligence-native?

The cynical interpretation is that this is simply AI-washing: companies using artificial intelligence as a convenient justification for layoffs. But that explanation feels too simplistic, as if this were merely a communications exercise.

What if something deeper is underway?

Like a Socialist Manifesto

The signals are becoming hard to ignore. We’re seeing an explosion of new models: more capable, more autonomous, and at times genuinely unsettling. At the same time, there’s a growing stream of policy papers, suggesting that technology companies are beginning to think more seriously about the societal consequences of what they are building.

OpenAI recently published Industrial Policy for the Intelligence Age, a substantial document that reads surprisingly little like a technical white paper and far more like a political manifesto.

Its subtitle says everything: “Ideas to Keep People First.”

In other words: what happens to people in an economy where human labor is no longer as essential?

The proposals would once have seemed more at home in trade union literature than in Silicon Valley boardrooms: a four-day workweek, universal basic income, labor law reform. At times, it reads like a socialist manifesto from the heyday of organized labor.

It is almost as though the technology companies now entering this debate are sending a warning: what comes next will change everything. And if it all goes wrong, they will be able to say they saw it coming.

The pressing question is whether policymakers understand what is at stake.

Do we grasp how fundamental this shift really is? Do we understand that this is not simply another phase of digital transformation, but a redefinition of economic logic itself?

An Uncomfortable Historical Parallel

History offers an unsettling comparison.

In China, there is the concept of the Century of Humiliation, the period between 1839 and 1949, when a nation that had once been an economic powerhouse failed to adapt to the Industrial Revolution. The consequences were devastating. During the Opium Wars, foreign powers humiliated China militarily and economically. The country lost sovereignty, influence, and prosperity.

China has long since moved beyond that chapter, but the lesson remains potent.

Because what if Europe is making the same mistake today?

What if we underestimate the rise of the intelligence-native economy? What if we continue framing the future in terms of jobs, while the underlying reality shifts toward tasks, processes, and autonomous systems?

The question is not whether this transition is coming.

The question is whether we understand it, and whether we act in time.

Because if we do not, we may one day find ourselves looking back on our own version of a century of missed opportunity.