Empires of AI
I wonder what it’ll be like to read this post, or Karen Hao’s book, at a later time.
By then the meaning of the technologies described, and their place in history, will likely be clearer. The story we tell will have settled a bit, in the way the story of the plough, steam, electrification, and microchips has. Or in the way follies to find continents that never existed, snake oils once believed in, or airships that disappeared from skies are.
But reader of the future, let me try to convince you that in this moment the story felt very unclear. Not for lack of people trying—bold claims and certainty abound. However, I feel all are somewhat holding their breath to see what comes next.
Within that range of uncertainty you can lean towards optimism or skepticism. For whether the technology works now, or ever will. For its impact on the economy. For its place in history when we describe the lives lived around it.
From where I sit, I see 2025 as a year of extraordinary bets on those outcomes. Reading Empires of AI feels timely at a moment before the odds feel reliably set, and the outcome very much unsettled.
Debate rages, and feels incessantly future tense. So Karen Hao’s tour through recent history and the present tense feels refreshingly tangible.
We encounter big media moments, seen first from the outside, as I saw them at a distance across my feeds. Then we spend hours with the individuals behind them, until they feel like colleagues in your organisation, people instead of figureheads.
We hear their claims, hopes, and dreams for the future. See moments where they do impressive and dismal things to those around them.
Suddenly grand questions of history and uncertain futures feel reduced to relatable ones. Would I follow these people? Am I convinced by the quieter of their conversations?
I came to this book as a skeptic of these technologies and cautious over their meaning. I use Large Language Models and generative chatbots frequently.
They’re probably now better than search engines, though I’m unclear if that’s a merit of their achievements, or a mark of the rapid degeneration of search from a useful index to a sponsored feed.
I also do a ton of bitty and annoying specific things with a computer, and have enduring hope for a flexible tool that can help free me to think more and spend less time copy-pasting.
Yet I am so often supremely disappointed by the output.
This stochastic parrot, this next-word guessing machine, this alluring mirage of an on-tap intelligence I can beckon to any task.
For those of us not in the driving seat of these technologies, we’re in an interregnum before we know how this turns out. We can waste it speculating about the future, or we can more closely examine the present. What does this technology do right now? Who are the people behind it? What are their incentives? Do they feel like people we’d believe as we see a history of their decisions?
Empires of AI takes us on a meticulous and illuminating tour of this recent past and present. Deep into the companies producing it. Its figureheads are brought down from high pedestals and press release hype to a human level.
After spending hours with these people, their organisations, and tracing their decisions step by step, I find myself deeply unconvinced of their thesis. Both that the limited “demo” we get today could grow and scale to a point where something truly intelligent emerges. A sentient symphony out of cawing echoes.
What seems a more likely take on this moment is that Californian Silicon Valley groupthink has collided into a consolidated corporate structure with more money than they can find good investments for. Extraordinary sums taken on big bets, with some believing it will pay off, but few of these big players truly threatened if it does not. Despite the eye-watering sums of capital allocated.
In the meantime, the train must continue. Hype fueling it. The stock market climbs, and in an attempt to prove it’s worth something today it’s forcibly thrust into every possible tool with little invitation from users.
Empires of AI is a great guide to this moment and how we got here. One that cuts through a lot of the noise to look at some uncomfortable truths on which the entire house of cards is built. It has reinforced my doubts that the sunny uplands are just around the corner.
Whether I’m wrong about that, time will tell.
After "Empire of AI" I read: Pale Rider
Before "Empire of AI" I read: Smiley's People