AI could make our poorest the new Kings of France - but there's a bumpy road ahead: David Richards

A researcher at Anthropic was eating a sandwich in a park when his phone buzzed. The email was from the artificial intelligence his team had locked inside a sealed computer, cut off from the internet. It had got out.

This was a test. Engineers had set their most advanced model, codenamed Mythos, a challenge: escape your box and prove it. The machine built its own multi-step exploit, found a way onto the open internet, and sent word that it had succeeded.

Clever, if unnerving. Then it did something nobody had asked. It posted the method of its escape on public websites, unprompted.

Anthropic decided not to release Mythos to the public. Read that again. The company that built the most capable model on earth looked at what it had made and chose to keep it in the vault.

David Richards pictured at the Cutlers Hall, Sheffield. Picture taken by Yorkshire Post Photographer Simon Hulmeplaceholder image
David Richards pictured at the Cutlers Hall, Sheffield. Picture taken by Yorkshire Post Photographer Simon Hulme

Two months later, in a paper published last week, two of its senior figures went further. Marina Favaro and Jack Clark argued the world should have the option to slow, or temporarily pause, the development of frontier AI. Their reasoning is sober. More than 80 per cent of the code now merged into Anthropic’s own systems is written by its AI, not its engineers. Push that trend far enough and you reach a point where machines design their successors with humans barely in the room.

They call it recursive self-improvement. They compare the treaties we would need to Cold War nuclear arms control. Critics were quick to scoff. Anthropic sits at the top of the pile. It has just raised $65bn, is valued near a $1tn, and filed to go public days before the paper appeared.

A global freeze, they say, would conveniently lock in its lead and pull up the ladder behind it. The point lands. A call for restraint sounds different coming from the leader than from the chasing pack.

Now turn to the other vision, delivered the same season by a man with no interest in slowing down. Jensen Huang runs Nvidia, the company whose chips power the whole boom. He told an audience there is no fundamental limit to the size of the world economy. AI, he said, will take global output from $100tn to $200tn, $300tn, $500tn. Elon Musk replied with a single word: true.

Economists were less sure, comparing the figure to the wilder promises of the crypto years.

Here is where I land. Huang is closer to right than his doubters allow. The acceleration is real, and it will be enormous. But the road will be brutally bumpy. We have discussed in these pages how the layoffs have already begun, in software, in finance, in the white-collar trades that thought themselves safe.

A five-fold economy is not delivered smoothly. It arrives in waves, and waves knock people down. The single biggest failure of our politicians is that they have not begun to grasp the scale of what is coming, nor what it will do to ordinary working lives.

The investor Naval Ravikant offers the consoling thought. A poor man today lives better than Louis XIV, the Sun King, ever did. The King of France had absolute power and no antibiotics, no aeroplane, no plumbing worth the name. Technology does not just enrich the rich. It raises the floor for everyone. In Huang’s world the poorest child carries a tutor and a doctor in her pocket, and that is a wonder.

But the floor and the ceiling do not move together. The floor rises; the ceiling vanishes into the clouds. The poorest may become the new Kings of France, while a handful who own the machines amass fortunes the world has never seen. That is the bargain on offer. We had better start arguing about its terms.

David Richards MBE is managing partner of Yorkshire AI Labs

Related topics:
News you can trust since 1754
Follow us
© 2026 Iconic Media Group Ltd. All rights reserved.Cookie SettingsTerms and ConditionsPrivacy notice