Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI by Karen Hao

 

I am currently staying in a cottage by the beach on Coronado Island. To me, the definition of a crisis is finishing a book and not having another one lined up to start. To solve this problem, I decided to either write a book or write a review of the one I just finished. I chose the latter option.

A Book Review

“OpenAI is nothing without its people.”

That rallying cry, once scribbled on Slack channels and scrawled across hearts in an empire teetering on collapse, now echoes as the thesis of Karen Hao’s masterwork — Empire of AI: Dreams and Nightmares in Sam Altman’s OpenAI. If there’s ever been a corporate saga worthy of being adapted into an Aaron Sorkin screenplay, complete with idealism, betrayal, late-night backchanneling, and billion-dollar stakes, this is it. But Hao delivers something far more enduring: a razor-sharp meditation on power, ethics, and the myth of benevolent tech.

Imagine If Shakespeare Wrote

Succession, But for AI

Sam Altman was supposed to be the Philosopher King of Silicon Valley, a boy genius with the code of Hammurabi in one hand and GPT-4 in the other. But Hao’s chronicle opens not with triumphant fanfare, but with an ambush: Altman, at the peak of his generative-AI-powered superstardom, is unceremoniously fired by his own board. Cue the five-day corporate civil war that reads like Game of Thrones meets The Social Network, shot through the lens of Joseph Weizenbaum’s warning: “Once its inner workings are explained… its magic crumbles away.”

But this is not just gossip from the OpenAI war room. Hao uses the palace intrigue as a launchpad into something deeper. A forensic investigation of how a startup with utopian dreams of democratizing AGI turned into a gilded vessel for empire-building, riding waves of compute, capitalism, and control.

Building God With Venture Capital

Empire of AI traces the arc of OpenAI from its founding ideals (openness, safety, and shared benefits) to its slow metamorphosis into a commercial juggernaut. It’s the story of how a nonprofit designed to stop Google from monopolizing AGI ended up entering the same race, weaponizing Microsoft’s cloud billions and gorging itself on data, energy, and global labor. Through over 300 interviews, Hao reconstructs not just timelines and Slack threads, but the very pulse of a movement hijacked by ambition. She blends investigative journalism with cultural anthropology, mapping OpenAI’s rise onto a much older template: empire.

And not just any empire — a colonial one.

What Makes This Book Sing

Style, Scope, and Emotional Voltage

There are good tech books. And then there are cultural reckoners. Hao’s style is cinematic but surgical. She doesn’t merely report. She anatomizes. Each chapter flows like a scene. A boardroom revolt, a data-labeled sweatshop in Kenya, a proxy war with DeepMind, a heartbroken employee watching their dream vaporize under fluorescent lights and billion-dollar valuations. What’s most extraordinary, though, is how Hao expands the aperture. This is not just the story of Altman, Brockman, or Sutskever. It’s the story of an entire geopolitical apparatus. A clash between techno-utopianism and the real-world scaffolding of exploitation, labor, and planetary cost.

“Over the years, I’ve found only one metaphor that encapsulates the nature of what these AI power players are: empires.”

In one line, Hao distills the entire moral architecture of the book. This isn’t about whether AGI will save or doom us. It’s about who gets to decide that future. And how they already are.

Standout Moments & Quotable Power

Two quotes merit framing:

“Altman had long touted the board’s ability to fire him as its most important governance mechanism.”

“AI is not a product. It is a projection of power.”

What Works. And What Falters

What works:

Exhaustive Reporting

Hao’s research is rigorous, triangulated, and transparent. She distinguishes between confirmed quotes, paraphrased accounts, and unverified claims with journalistic precision.

Moral Clarity Without Sanctimony

Hao doesn’t preach. She observes and indicts with facts.

Global Perspective

Most Silicon Valley exposés are Western loops. This one loops in Kenyan data laborers, South American data centers, and the sweat behind the silicon.

Where it could push further:

Philosophical Depth

While rich in ethical framing, the book sometimes leaves the metaphysics of AGI shallow. The big, juicy ontological questions: What is Intelligence? Consciousness? Soul? is hinted at but not wrestled with.

More Human Voices

The political drama is gripping, but the emotional weight might’ve hit harder with more unfiltered profiles of workers, annotators, and skeptics who aren’t in boardrooms.

Final Verdict

Who Should Read This Book?

If you’re:

• A technologist who believes AI can save the world
• A policymaker trying to regulate a tsunami
• A student wondering how the future is actually getting made
• Or just someone who thinks power always wears a hoodie and a smile…

Then Empire of AI is your must-read. Hao has crafted a masterpiece, part exposé, part epic, part call to arms.

This book is not just about OpenAI. It’s about the stories we tell ourselves about progress, and who gets to tell them. It’s about what happens when the people who build gods start believing they are them.

🔥 Empire of AI is The Big Short for artificial intelligence. A taut, thrilling, revelatory autopsy of the empire being built in our names. Read it now, before the future is no longer up for vote.