What I loved most about Empire of AI

There’s a lot to like about Karen Hao’s Empire of AI, our book of the month for June. It’s a detailed and argumentative title that explores what is, in many ways, the story of now. Empire of AI goes over the origins of OpenAI and its CEO Sam Altman (including the financial, political and ideological battles fought inside and outside the organization) and it details the shape of artificial intelligence more generally—how it works, what it demands, and how it changes the world around us.

But there are particular areas of the book that I thought made it stand out, so I wanted to get specific about a few of the things I enjoyed most.

I love that it’s thorough. I fully expected Karen to go into lots of detail in her reporting—I had the privilege of being her colleague for a couple of years when we were both at MIT Technology Review, which is when she first started writing about OpenAI (I edited her a few times, but not on those stories.) So I wasn’t surprised that she brings her trademark level of high intensity across nearly 500 pages. Still, it takes effort to be that thorough, and the book would have probably been just as commercially successful if she skipped lightly over a few areas. The result is that this isn’t a quick, easy read: it takes time and effort to get through. But the meticulousness means that you feel as if you’ve taken a masterclass by the end. 

⌘

I love that it’s transparent. Karen has a lot of opinions about OpenAI and its people and the technologies it represents. But those are never hidden away. They are formed from different experiences, different conversations, different sources—and those all get air time. She regularly credits other reporters or journalists who have broken stories about OpenAI and its dramatic little world; she tells you about the people who help inform her opinion, whether they’re activists in Silicon Valley, informed insiders, or people living elsewhere in the world. You are welcome to agree or disagree with what she presents (and I think it leans a little heavily on a few people for perspective) but you know where it’s coming from. This is fairness versus the “view from nowhere”.

⌘

I love that it’s expansive. Right now there are lots of books about AI and the people who shape it out—and I want to talk about some of them in a future newsletter. But for an industry that is changing the way we work and live, this book takes a much wider and more global view of things. I loved that Karen goes to report from Chile and Kenya and Colombia and elsewhere. There’s a scene where she is walking around a park in Chile that’s been put together by Google in return for its deal to build a huge data center that strips the region of important natural resources. She writes:

“Past a graveyard dedicated to deceased pets, we pull up to what looks like an abandoned plot of grassland with tufts of shrubs and a scattered handful of nutrient-starved trees jutting out of the soil. Most days the land is so parched it looks like parts of the Atacama Desert; today the rain is turning everything into mud. In the middle of the plot, a purple sign announces in Spanish, “Welcome to the Quilicura Urban Forest,” a project, it explains, that Google began in 2019 to give back to the community for hosting its data center.”

Google brags, she says, about the urban forest and its benefits for locals, about how the forest mitigates the costs of the technology, about how residents use this green space. She looks at reality.

“There are no residents. The place is too far from any bus line, and there are no homes in the surrounding area to speak of. Outside the modest plot, too small to fit Google’s data center itself, a dozen stray dogs meander around, barking and rummaging through the trash. The spokesperson said the forest is being “updated to “evolve the experience for the community.””

While it sounds kind of obvious to actually go and witness something with your own eyes, you actually don’t get that in every piece of writing about the technology industry, and it gives the arguments in the book new dimension.

But maybe above all, I love that it suggests there are different ways for this to go. The old saying is that history is written by the victors, and in Silicon Valley this leads to a sort of determinism—the idea that a certain approach is technologically inevitable. Once a change has happened, it feels like it couldn’t have gone any other way. Of course computers do this; of course phones and apps and private data and digitization would lead us here; of course AI was going to go in this direction.

You can see a lot of this inevitablistic thinking in some of the ideologies that underpin AI development, like rationalism and effective altruism, which often seem like mathematical boardgames. 

It is, of course, in the best interests of OpenAI and Sam Altman to make their model look inevitable: that their capital-hungry, huge compute, immense scale approach to technology and artificial intelligence is the only way it could have happened. 

But throughout the book Karen details the alternative paths that weren’t taken. Did OpenAI need to start an LLM arms race? Did it need to train its models on the sanity of workers in the global south, or the resources of people living in arid, underprivileged areas? And, towards the book’s end, she looks at what alternative models really look like if we don’t accept that “this happened” is the same as “this had to happen.”

What if we concentrate on less resource-intensive approaches, like those we’re seeing emerge from China and elsewhere? What if communities get to really own the way technologies digest and use their cultural work, like the people using AI to revitalize the Maori language?

⌘

Don’t forget, you’ll be able to ask Karen questions about the book when she joins us for a live Q&A on Thursday June 26 at 8pm Eastern/5pm Pacific. If you have any thoughts or ideas, please email me and I can wrap them into our interview. If you’re reading along, I’d be excited to know what you loved—or hated—most about Empire of AI.

Onward

Bobbie