Disclosure: This page contains affiliate links. If you buy through them, we may earn a commission at no extra cost to you.

The Coming Wave by Mustafa Suleyman: Book Review

The Coming Wave by Mustafa Suleyman is the rare AI book written by someone who actually built frontier AI. Suleyman co-founded DeepMind — the lab behind AlphaGo — then watched the technology he helped create accelerate far beyond anyone’s control. Published in 2023, the book’s big argument is that AI (alongside synthetic biology) forms a “coming wave” humanity will struggle to contain, and that containment — not just innovation — is the defining challenge of our century. This review breaks down what he actually argues, who should read it, and whether it still holds up in 2026.

Related reading: The Big Nine by Amy Webb: Book Review — Who Should Actually Read It in 2026?, AI Snake Oil by Arvind Narayanan & Sayash Kapoor: Book Review — Can It Really Te.

What to Look For in an AI Future Book

Books about “the future of AI” age faster than milk. Most are either breathless hype or recycled headlines. Here’s how to tell the ones that last from the ones that don’t.

An author with real skin in the game

Suleyman co-founded DeepMind in 2010, sold it to Google, and later co-founded Inflection AI — he has spent his entire career building the exact technology he’s now warning about. That changes the texture of the book completely. When a career pundit warns about AI, it’s content. When one of the people who built frontier AI says “we may not be able to contain this,” it’s a confession. For any big-picture AI book, the author’s biography is the first thing to check: did they build it, regulate it, or just write about it?

A thesis, not just vibes

The book’s core argument has a name: the containment problem. Powerful technologies tend to proliferate because the incentives — profit, prestige, national competition — push everyone to build and share them, while the incentives to slow down are weak and diffuse. That’s a real thesis, falsifiable and structured, not a mood. If an AI-future book can’t be summarized in two sentences, it probably doesn’t have an argument. This one can: the wave is coming, containment is the whole game, and history says containment usually fails.

Historical grounding, not just prediction

Suleyman reaches back through previous technological waves — the rare cases where humanity contained a dangerous technology, and the far more common cases where it didn’t — to show that the containment problem isn’t new. This is what keeps the book from reading like a TED talk. Predictions are cheap; patterns across centuries are not. A future-of-AI book that only talks about the last eighteen months is already expiring. One that draws on history has a much longer shelf life.

Scope beyond chatbots

The “wave” in the title isn’t just large language models — it’s AI and synthetic biology, the two general-purpose technologies Suleyman argues will define the century. That wider lens is a genuine strength. Most AI books published in 2023 couldn’t see past ChatGPT; this one treats chatbots as one crest of a much bigger swell. If you want a book that won’t feel dated the moment the next model drops, scope matters.

Readability for non-experts

Despite the heavy subject, it’s a big-ideas trade book: no equations, no assumed technical background, narrative-driven chapters. Suleyman writes like someone who has explained AI to heads of government — because he has — which means he’s good at making the abstract concrete. The trade-off is depth on any single mechanism; if you want the math of why models scale the way they do, look elsewhere. This book is about incentives and power, not architectures.

What’s missing — by design

Here’s the honest limitation: this book will not help you use AI better. There are no prompting tips, no workflow advice, no “ten things to try on Monday.” It’s a macro book — governments, labs, incentives, treaties — and it stays at that altitude throughout. That’s a legitimate choice, but buy with clear eyes: if your question is practical, Co-Intelligence or a prompting manual is the better purchase.

How to Match This Book to Your Situation

This is a book for people who think about the trajectory, not the tool. Three readers in particular:

The professional who needs the 30,000-foot view

You’re a founder, executive, or manager, and AI keeps coming up in strategy conversations you’re expected to lead. You don’t need prompting tricks; you need to understand where the technology is going, what gets regulated, and what the failure modes of governance look like. This book is basically a briefing document for that role — the one that lets you ask the uncomfortable questions in the meeting.

The policy-minded reader

Students of public policy, journalists, anyone who followed the AI regulation debates and felt lost in the jargon. Suleyman gives you the framework the whole debate runs on: containment, proliferation incentives, the difficulty of international coordination. Read it and the next “AI safety summit” headline will make sense instead of sounding like theater.

The builder thinking about what’s next

You work in or around tech and you’re placing bets — on products, on careers, on where the puck goes. The book’s incentives-first lens is genuinely useful for that: it asks not “what can the technology do” but “what will everyone be compelled to do with it.” That second question is the one that predicts the future. Just don’t expect product ideas. Expect a colder shower.

How the Top AI Books Compare

Same shelf, four completely different jobs. Here’s the map.

Book	Best For	Price Tier	Key Feature	Watch Out
The Coming Wave	Grasping AI’s trajectory and governance challenge	Mid-range	An insider’s containment framework	Macro lens — nothing hands-on
Co-Intelligence	Using AI well in daily work	Budget-friendly	Practical, Mollick’s field guide	Light on systemic risk
Prompt Engineering for Generative AI	Technical prompting craft	Mid-range	O’Reilly-style hands-on depth	Assumes you already use AI tools
Weapons of Math Destruction	Understanding algorithmic harm	Budget-friendly	The ethics classic, still cited everywhere	Pre-generative-AI, big-data era

Pick The Coming Wave when you want to understand where AI is headed and why steering it is so hard. Choose Co-Intelligence when you want to get better at using AI this week — practice over theory. Go with Prompt Engineering for Generative AI when you want craft: the actual techniques, written as a technical manual. And grab Weapons of Math Destruction when you want the ethics foundation — the classic account of how algorithms already harm people at scale.

Where This Book Earns Its Shelf Space

Trying to make sense of AI regulation headlines

Every few months there’s another summit, another executive order, another open letter from researchers. Without a framework, it all blurs together. This book is the framework: once you understand the containment problem, you can read any AI policy story and immediately see which incentive it’s fighting against. That’s a durable skill, and it’s the book’s highest-value use.

Gifting it to someone who “doesn’t get the AI fuss”

Know someone smart who still thinks AI is just a better autocomplete? This is the conversion book. It’s serious without being academic, urgent without being hysterical, and it comes from someone who built the stuff — which carries weight with skeptics in a way that activist literature doesn’t. Fair warning: it may ruin their peace of mind. That’s rather the point.

Reading it while you actually use AI tools daily

There’s a productive tension in reading the macro-warning while doing the micro-practice. Pair this book with Co-Intelligence and you get both halves: the hands to use the tools, the head to understand what you’re participating in. Read the practical book in the morning, the warning at night. You’ll use AI better and think about it more clearly — a combination most readers never quite manage.

Common Mistakes Buyers Make

Mistake 1: Buying it for prompt tips. There are none. Zero. If your question is “how do I write better prompts,” you want the O’Reilly book or Co-Intelligence. This book operates at the level of nations and decades.

Mistake 2: Reading it as pure doom. Suleyman’s position is worried, not defeatist — he argues containment is extremely difficult yet still possible, and sketches what it would take. Readers who want either apocalypse or reassurance will both be slightly disappointed. That’s a sign of honesty.

Mistake 3: Treating it as a prediction to grade. The book’s value is the framework, not the forecast. Its argument runs on incentives, not model versions, so it doesn’t expire when the next generation of models arrives. Judge it on whether the containment lens clarifies things, not on whether every example still looks current.

Also see: The Alignment Problem by Brian Christian: Review — The AI Safety Book to Read Fi.

Mistake 4: Dismissing it as “another tech CEO book.” Suleyman is unusually self-critical about his own industry’s incentives — including his own role in creating them. That’s rare, and it’s what separates this from the standard founder memoir with a warning label slapped on.

Why This Book Exists, How to Read It, What You’ll Get

Why did a DeepMind co-founder write a book about containment? Because he watched the incentives from the inside. Suleyman spent over a decade in the labs where frontier AI was built, and the book reads like someone who realized — earlier than most — that the forces pushing the technology outward are stronger than any single company’s good intentions. The credibility of the converted insider is the book’s engine; without it, this would be just another warning.

How should you read it? Read for the framework, not the forecast. The early chapters set up the wave metaphor and the history; the containment chapters are the payoff. If the historical sections feel familiar, skim them — but don’t skip the argument about incentives. That’s the part you’ll still be thinking about months later.

What will you actually use from it? A mental model for every AI headline you’ll read for the next decade: incentives drive proliferation, proliferation outruns governance, containment is the whole game. Three sentences, and suddenly the news makes a different kind of sense. Not comforting sense. But sense.

Who is Mustafa Suleyman?

Mustafa Suleyman is a British AI entrepreneur who co-founded DeepMind in 2010 — the lab famous for AlphaGo — which Google acquired in 2014. He later co-founded Inflection AI, the company behind the Pi assistant. He wrote The Coming Wave with Michael Bhaskar, and now works on AI policy and safety efforts. He’s one of the few people who has both built frontier AI systems and written seriously about governing them.

What does “containment” mean in this book?

Containment is Suleyman’s term for the project of keeping the wave’s benefits while limiting its harms — through some combination of technical safety work, regulation, auditing, and international coordination. The book’s central claim is that this is extraordinarily difficult because the incentives to build and spread powerful technology are so much stronger than the incentives to restrain it.

Is this book just about ChatGPT?

No. It was published in 2023, as the ChatGPT era was beginning, but its scope is much wider: AI broadly, plus synthetic biology, which Suleyman treats as the wave’s second crest. ChatGPT appears as an example of the dynamic he’s describing, not as the subject. Readers looking for LLM-specific content should look at a prompting or practical-AI book instead.

Is it a long, dense read?

It’s a substantial big-ideas book — most readers take the better part of a week, reading in the evenings — but it’s not dense in the academic sense. No equations, no jargon walls. The history sections reward skimming if you’ve seen the material before; the containment argument deserves your full attention.

Does the book actually propose solutions?

Yes — the final section lays out a containment agenda spanning technical measures, regulation, and international cooperation. Whether you find the proposals convincing is another question; some critics find them vaguer than the diagnosis. The honest read: the book is stronger on “here’s why this is hard” than on “here’s the fix,” and Suleyman doesn’t really hide that.

What should I read after The Coming Wave?

For the practical half — actually using AI well — read Co-Intelligence by Ethan Mollick. For the ethics foundation, read Weapons of Math Destruction by Cathy O’Neil. For hands-on prompting craft, the O’Reilly prompt engineering book. This book gives you the stakes; those give you the skills and the history.

Final Recommendation

Buy The Coming Wave if you want the most credible big-picture account of why AI is so hard to govern — written by someone who built the technology and is unusually honest about his own industry’s incentives. It’s the right book for understanding the stakes, the regulation debate, and where things plausibly go next.

It is not the book that teaches you to use AI, and it doesn’t pretend to be. Pair it with Co-Intelligence for the practical half of the picture. But if you only read one book about AI’s future, make it this one.

Disclosure: This page contains affiliate links. If you buy through them, we may earn a commission at no extra cost to you.