AI may genuinely transform the world. That does not mean everyone financing the transformation will make money.

Carlota Perez argues that AI is not a separate sixth technological revolution. It is a revolutionary technology within the fifth information and communications revolution. In her framework, we are approaching the end of the turning point, and the next challenge is not simply to build more technology, but to deploy AI across healthcare, education, public administration, the green transition, and global development.

Jim Chanos, the investor best known for identifying the collapse of Enron, asks a different question: even if Perez is right that a deployment period is approaching, will the investors currently buying GPUs, building data centers, assuming debt, and recognizing profits survive long enough to participate in that golden age?

The 10 Questions This Episode Sets Out to Answer

1. The more important a technology is to society, the more likely it is to become a bad investment?

2. If AI is not a new technological revolution, why does today’s capital market behave as if a new installation-period frenzy has begun?

3. Could the best AI outcome for society be the worst outcome for AI stocks?

4. Which parts of today’s AI infrastructure are the railways of the future, and which are merely future scrap metal?

5. Who ultimately pays for the “dynamic demand” Perez believes can launch a new golden age?

6. Is government the initiator of the AI deployment period—or the final buyer of the bubble?

7. What evidence would prove Perez wrong, and what evidence would prove Chanos wrong?

8. If someone is right about direction but wrong about timing, how useful are they to an ordinary investor?

9. After the bubble bursts, who actually inherits the infrastructure built during the boom?

10. If Perez and Chanos jointly managed $100, what would they own now, own after the clearing, or never own?

Three minds sit down for a ninety-minute conversation that never happened. Demis Hassabis, the scientist building the machine. Doudou, the reclusive novelist who insists no machine can save a soul unwilling to save itself. And Sebastian Mallaby, the historian who spent four years asking what culture hides inside our technology. One question on the table: does AI have a "cultural nature" — is it a mirror, or a savior? And the larger one underneath it: in the age of AI, what does a human being actually stand on?

In this episode:

• Why Silicon Valley's two loudest stories — utopia and apocalypse — might secretly be the same story.
• What AlphaGo's "move 37" revealed that a mirror never could.
• The moment the novelist tells the Nobel laureate: you can build a machine that cracks every structure in nature — but you can't build a person who knows how to die.
• Why "making up your own mind" is about to become the rarest, most expensive skill on Earth.
• What's left for us when the machine can write the poem, prove the theorem, and love you back — without ever leaving.

"The savior is faraway — because the savior doesn't exist. AI won't save anyone. The only thing that can save a person is that person waking up."