Public transcript

#541: Eric Schmidt — The Promises and Perils of AI, the Future of Warfare, Profound Revolutions o...

By: BlackNova Productions
Length: 1 h 35 min
Language: English
Text from: the video's captions
Transcribed: 14 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

Eric Schmidt joins Tim Ferriss to discuss his book The Age of AI and Our Human Future, co-written with Henry Kissinger and Daniel Huttenlocher. Schmidt explains how the collaboration began, walks through major AI milestones like AlphaGo, Halicin and GPT-3, and defines AI versus AGI, offering his own timeline of roughly 15 years for AGI. He also covers the geopolitical race with China, the dangers of AI-enabled warfare and misinformation, and the ethical and regulatory questions society must resolve within the next decade. The conversation closes with his hopes that the book prompts people to develop a philosophy for coexisting with these systems before they outstrip governance, as social media did.

Key points

Eric Schmidt co-authored The Age of AI and Our Human Future with Henry Kissinger and MIT dean Daniel Huttenlocher to frame the right questions about AI before the technology outstrips governance, as social media did.
Kissinger, at 98, argues AI marks a new epoch comparable to the Renaissance because humanity now faces a non-human intelligence similar to its own.
Three recent milestones stand out to Schmidt: AlphaGo inventing new Go strategies, MIT researchers discovering the antibiotic Halicin by screening 100 million compounds, and OpenAI's GPT-3 language model.
Schmidt defines AI as a system that gets better through learning, and distinguishes it from AGI, which would involve self-determination; he estimates AGI may be possible in roughly 15 years and one or two more breakthroughs.
He identifies four distinctive qualities of AI: it is imprecise, dynamic, emergent, and capable of learning.
The National Security Commission on AI concluded the US is slightly ahead of China, which aims for AI dominance by 2030 and graduates four times more engineers.
Schmidt warns that AI-enabled warfare could compress conflicts to milliseconds, and that an AGI arms race would be destabilizing without new diplomatic language and agreements.
He predicts everyone will have a personal AI assistant that filters information and battles misinformation, and that these systems will be regulated differently by each government and culture.
Protein folding results from DeepMind and Rosetta are Nobel-Prize-caliber discoveries happening now, and Schmidt echoes the view that AI will be to biology what math is to physics.
The book's goal is to prompt society to develop a philosophy for coexisting with AI over the next 5 to 10 years.

Questions it answers

04:29Meeting Kissinger

How did Eric Schmidt come to write a book with Henry Kissinger?

They met about 12 years ago at a Bilderberg conference; Kissinger told Schmidt his only problem with Google was that it would destroy the world. After visiting Google and giving a speech, they struck up a friendship despite very different political backgrounds, and Kissinger proposed the book as a chance to frame the right questions about AI.

Kissinger works seven days a week at 98 and mastered AI late in life.
The book aims to ask the right questions about AI before it outstrips governance, as social media did.

13:42AI milestones

What recent events convinced Schmidt that AI is a profound shift?

He points to AlphaGo inventing new Go strategies after 2,500 years of human play, MIT's discovery of the antibiotic Halicin by screening 100 million compounds, and OpenAI's GPT-3, a generative model that can speak what it knows and improves with scale.

Computers now see better than humans, enabling prediction at scale.
Trillion-parameter models cost around $100 million to build.

21:16AI vs AGI

What is the difference between AI and AGI, and when might AGI arrive?

AI is a system that gets better through learning and finds patterns humans can't see. AGI would involve self-determination and consciousness; Schmidt estimates one or two more breakthroughs and roughly 15 years, and compares extreme AGI to nuclear weapons in terms of power and need for control.

AGI optimists say 10 years, pessimists far longer or never.
Computing power doubles every 2-3 years; AGI systems will likely be few, expensive and large.

38:00AI qualities and human impact

What makes AI different from prior technologies, and what unknowns worry Schmidt?

He highlights AI's four distinctive qualities — imprecise, dynamic, emergent, capable of learning — and uses examples like a child's AI teddy bear and lonely elderly people with digital friends to show we don't yet understand the effects on bonding and formative behavior.

Don't use AI for life-critical decisions without human oversight.
We don't know what an AI companion teaches a child or what norms it passes on.

43:42Geopolitics and security

What did the National Security Commission on AI conclude about China and warfare?

The US is slightly ahead, but China has a national program pouring billions into AI, graduates four times more engineers, and aims for dominance by 2030. He warns that AI compresses war to milliseconds, that containment doesn't work because software leaks through ideas, and that an AGI arms race would be destabilizing without new diplomatic language.

A future cyber war could begin and end in 10 milliseconds.
Hypersonic missiles could force humans to make launch decisions in seconds.

58:00Assistants and regulation

How will society cope with misinformation and who decides the rules?

He predicts each person will have an AI assistant that knows their preferences and battles misinformation assistants, verifying who is human. He expects regulation to differ by country and culture, as with surveillance in Britain, the US and Germany, and says the internet is no longer optional so it will be regulated.

Free speech for humans doesn't mean amplification by computers.
Society, not computer scientists, should make choices like ubiquitous surveillance.

1:21:20Biology and protein folding

What recent AI discoveries matter most for science and medicine?

DeepMind and the University of Washington's Rosetta group both published protein-folding algorithms, work that would merit Nobel Prizes if done by humans. He echoes a friend's view that AI will be to biology what math is to physics, since cells and the brain remain unmodelable.

Protein folding underpins drug discovery, medicine and genetic expression.
AI approximates calculations too complicated for computers to solve exactly.

1:25:30Hopes for the book

What outcome does Schmidt hope The Age of AI produces?

He wants readers to say they must get ahead of AI and develop a philosophy for humans coexisting with these systems over the next 5-10 years. He also notes AI will mostly eliminate repetitive jobs, not creative ones like his and Tim's.

Technologists repeatedly build tools without understanding how they'll be misused.
AI is best used to watch for exceptions and alert humans, not replace judgment.

Notes

Background and the Book

Eric Schmidt's new book, The Age of AI and Our Human Future, is co-authored with Henry Kissinger and Daniel Huttenlocher, dean of MIT computer science.
Schmidt met Kissinger about 12 years ago at Bilderberg; Kissinger initially told him Google would "destroy the world," but they became friends despite political differences.
Kissinger, at 98, works seven days a week and mastered AI late in life; Schmidt calls being a workaholic the secret to longevity.
Kissinger suggested the book as an opportunity to "architect the questions" for the next 10-30 years, learning from how social media outstripped governance.

Key AI Milestones

AlphaGo: a computer beat top human Go players and invented moves unknown in the game's 2,500-year history.
Halicin: MIT researchers screened 100 million compounds to discover a new broad-scale antibiotic, the first in roughly 40 years.
GPT-3: OpenAI's transformer model can generate language; trillion-parameter models now cost around $100 million to build.
Protein folding: DeepMind and University of Washington's Rosetta both published protein-folding algorithms, discoveries Schmidt says would merit Nobel Prizes if done by humans.

Defining AI and AGI

AI is "a system that gets better through learning" — good at finding patterns humans can't see, not the killer robots of movies.
AGI (artificial general intelligence) implies self-determination; Schmidt guesses it may be 15 years away, needing one or two more breakthroughs.
Four qualities of AI: imprecise, dynamic, emergent, capable of learning.

Geopolitics and Security

The National Security Commission on AI (chaired by Schmidt) found the US slightly ahead of China, which targets AI dominance by 2030 and produces four times more engineers.
Future cyber wars could play out in 10 milliseconds; hypersonic missiles compress human decision time to seconds.
Containment doesn't work with AI because it leaks through ideas; Schmidt urges investment plus dialogue on limiting automatic war.

Society, Ethics and Regulation

AI will make information and misinformation indistinguishable; each person may need an AI assistant to filter reality.
Questions like who gets a ticket for a self-driving car's "rolling stop" show unresolved liability and ethics issues.
Surveillance acceptance varies: Britain (widely accepted), US (partial), Germany (opposed due to Stasi history).
Schmidt predicts AI will eliminate repetitive jobs, not creative ones, and that humans need meaning — augmentation beats replacement.