Quick Take
- Narration: Mikael Naramore handles the material’s mix of personal memoir, geopolitical analysis, and technical explanation with clean, authoritative delivery.
- Themes: US-China AI competition, job displacement across white and blue-collar sectors, technological power and ethical responsibility
- Mood: Urgent and lucid, with a personal undercurrent
- Verdict: A foundational text for understanding the geopolitical AI landscape, though its 2018 vantage point means some predictions now read as history.
I listened to AI Superpowers during a period when I was trying to understand the competitive landscape of artificial intelligence beyond the Silicon Valley-centric frame that dominated most of what I was reading. Kai-Fu Lee offered something none of those sources could: a view from both sides. He spent years heading Google China and built his career across American and Chinese technology institutions before founding a Beijing-based venture capital firm. That dual vantage point gives the book a structural authority that is genuinely rare in writing about technology.
The central argument is direct and, when the book was published in 2018, still contested: China had caught up to the United States in artificial intelligence at an astonishing pace, and the country was positioned to win the AI competition for reasons that had little to do with the quality of its top researchers and everything to do with data, capital, and a particular kind of implementation culture. Lee distinguishes between the age of discovery in AI, in which US researchers had genuine advantages, and the age of implementation, in which China’s ecosystem of copycat-turned-innovator entrepreneurs was uniquely suited to deploying AI at scale.
Our Take on AI Superpowers
The technical level of the book is accessible to generalist listeners. Lee does not require a computer science background to follow his argument, and the way he explains the distinction between general AI and narrow AI, and why narrow AI is where the real economic competition is happening, is among the clearest I have encountered. The history of AI that opens the book, tracing the field from its early winters through the deep learning breakthrough of 2012, serves the argument well without becoming a digression.
What distinguishes the book from most technology writing is its second half, which turns inward. Lee received a cancer diagnosis during the writing of the book, and the experience clearly recalibrated his thinking about what the AI transition will cost humanity in terms of meaning and connection, not just employment. His conclusion, that the answer to mass AI-driven job displacement is not universal basic income but a reinvestment in care work and human connection, is argued from personal experience rather than policy abstraction. That move gives the ending of the book a weight that the competitive analysis sections, however rigorous, cannot quite match.
Why Listen to AI Superpowers
Mikael Naramore’s narration serves the material without calling attention to itself. The prose in AI Superpowers moves between registers, technical, autobiographical, geopolitical, and Naramore shifts between them smoothly. The Chinese names and company names are handled with enough care that they do not become stumbling blocks, which matters in a book where Baidu, Alibaba, Tencent, and Meituan are central actors rather than background color.
Reviewers who came to the book with existing China experience found that the characterization of Chinese entrepreneurial culture was accurate in ways that the usual Western-market perspective on Chinese tech rarely captures. The gladiatorial nature of Chinese startup competition, the 9-9-6 work culture, and the way data abundance has substituted for algorithmic innovation are explained concretely rather than gesturally.
What to Watch For in AI Superpowers
The book was published in September 2018, and the world has changed substantially since then. Lee wrote before the full escalation of US-China trade friction, before the Huawei restrictions, before the AI chip export controls, and before ChatGPT reframed public understanding of where large language models stood. Some of his predictions have aged well and some have not. One reviewer noted that the book is more idealistic in tone than it would be if rewritten today, and that is accurate. Reading or listening in 2026 means treating some of the competitive analysis as historical document rather than current forecast.
The book’s treatment of AI’s impact on employment is sobering and deserves to be taken seriously, though the specific job categories and timelines Lee named in 2018 have not mapped exactly onto how the transition has actually unfolded. The underlying framework, that AI will displace white-collar as well as blue-collar work faster than most economists projected, has held up better than the specific predictions.
Who Should Listen to AI Superpowers
This audiobook suits business professionals, policy thinkers, and general readers who want a clear, authoritative account of why China became a genuine AI power and what that means for global economic competition. It is a foundational text in the literature of AI geopolitics and remains worth engaging with, provided you approach it as a 2018 perspective rather than a current map. Those looking for an up-to-date account of the AI landscape will need to supplement it with more recent sources. For listeners interested in the personal dimension of how a technology leader reckons with mortality and the human cost of the systems he helped build, the book’s final sections are unexpectedly affecting.
Frequently Asked Questions
How outdated is AI Superpowers given that it was published in 2018?
Significantly outdated on specific predictions and the competitive balance, particularly in light of US chip export restrictions, the rise of large language models, and the escalation of trade tensions. But the structural framework for understanding China’s AI ecosystem and its differences from Silicon Valley’s approach remains useful as historical and analytical context.
Do I need a technical background to follow Kai-Fu Lee’s argument?
No. Lee writes for a general audience and is careful to explain concepts like deep learning, narrow versus general AI, and supervised learning in accessible terms. The book is primarily a geopolitical and economic argument that uses technical context as scaffolding, not a technical treatise.
How does Mikael Naramore handle the Chinese company names and proper nouns throughout the narration?
With reasonable consistency and care. The pronunciation of Chinese names and companies is stable across the runtime, which is important in a book where Baidu, Alibaba, Tencent, and a range of Chinese startup names appear frequently. It does not slow or interrupt the listening experience.
What is the personal story that runs through the book, and how much of the runtime does it occupy?
Lee received a cancer diagnosis while writing the book, and this experience shapes the second half significantly. It informs his argument about what AI threatens beyond employment, specifically human connection and meaning. The personal sections are woven through the later chapters rather than forming a separate thread, and they give the book an emotional weight its purely analytical sections cannot provide.
