How the World Got Wired · Two AI Internets

What’s Coming

I have spent this chapter telling you about geography and about systems, but I want to end with something that comes back to a person.

You met Geoffrey Hinton in the last two chapters. You read about his quiet decades in Toronto, his decision to leave America in 1987, his work that made the current revolution possible, and his decision in 2023 to resign from Google and start warning the world about what AI was becoming.

In October of 2024, Hinton was awarded the Nobel Prize in Physics. He gave his acceptance speech in Stockholm. He used much of his speech to argue that the world was not paying enough attention to the dangers of the technology he had helped invent.

He is not alone. Yoshua Bengio, in Montreal, has been giving speech after speech making the same point. Stuart Russell, at Berkeley, has been making it for longer than either of them. Dario Amodei, the chief executive of Anthropic — one of the largest AI companies in the world — has written, repeatedly, that he believes there is a real chance the systems his own company is building will cause serious harm. Sam Altman, the chief executive of OpenAI, has made similar statements in the past, though he has shifted his public position over time.

You should hold this in your mind. The people who built the technology, and the people currently running the largest companies building it, are, by and large, the people most openly worried about it. This is unusual. It is not the way most industries work. It does not, by itself, prove that the worries are correct. But it should give you pause.

The question of what to do about artificial intelligence — how fast to build it, what rules to put on it, who should own it, what it should be allowed to say, who should decide what it should be allowed to say — is genuinely undecided. It is being argued about right now, in real time, in capital cities all over the world. Different countries are arriving at different answers. The European Union has passed an AI Act, restricting certain uses of the technology. The United States has, mostly, let companies do as they like, with some recent regulatory attention. China has heavily regulated certain aspects (the Taiwan answer, the Tiananmen answer) while heavily promoting others. Several countries have not yet decided.

By the time you are old enough to vote, the answers to these questions will be — in some form — settled. Maybe well. Maybe badly. Maybe in ways that look obvious in retrospect, the way the open-internet decisions of the 1980s look obvious to us now. Maybe in ways that, decades from now, will be considered tragic.

You will not, by yourself, decide them. But you will be part of the generation that lives inside the decisions and that, eventually, governs based on them. Whatever you read, whatever you study, whatever you choose to work on, whatever you choose to argue about with your friends — all of it will, in small ways, contribute to the answers.

The most useful thing this book can give you, I think, is the long view.

All 24 chapters of How the World Got Wired · Library · All courses