The Containment Problem: Can We Actually Control AI?

The Wave is Coming. Are We Ready?

I just finished "The Coming Wave" by Mustafa Suleyman, and honestly, I’m still processing it. As someone who spends their day-to-day deep in data and architecture, it’s easy to get caught up in the "how" of what we’re building. This book forced me to take a massive step back and look at the "should."

Why this hit home for me

The reason this book has so much weight is the guy behind it. We’re not talking about a tech critic or an academic theorizing from the sidelines. This is Mustafa Suleyman—the co-founder of DeepMind and Inflection AI. He was at "ground zero" for the AI revolution. When a guy who helped build the foundation of modern AI says we have a problem, you listen.

Suleyman uses the metaphor of a "wave" because technology of this scale doesn't arrive in drips; it arrives as a tsunami that reshapes everything in its path.

My Take: The Containment Problem

The core of the book is what he calls the "containment problem". Historically, once a technology becomes powerful, cheap, and easy to use, it becomes impossible to "un-invent."

Between AI and synthetic biology, we are looking at a level of power that will be decentralized and autonomous. Suleyman is brutally honest about the fact that we currently have no plan for how to contain these tools once they’re out in the wild.

It’s not a doomsday book, but it is a wake-up call. It made me think hard about our responsibility as developers and architects. Are we just building cool tools, or are we opening Pandora’s boxes that we won't know how to close?

The Verdict: If you’re in tech—or if you just care about what the next ten years look like—this is mandatory reading. It’s a high-signal, zero-fluff look at the most transformative decade in human history.

What’s your take? Do you think "containment" is even possible at this point, or has the ship already sailed?