Chapter 1 — Reflection & Commentary
Reading the opening of AI Superpowers, I realized something both humbling and uncomfortable: I’m still on the level of those children and businessmen who, as Kai-Fu Lee recalls, asked him the same basic questions about AI — what is it, what can it do, and will it take our jobs?
Despite all the hype around ChatGPT and the recent AI boom, I had no clue how long this field has been developing. To me, AI “started” in 2022 or 2023, like some magic that came out of nowhere. Turns out? Not even close. The roots go back decades — and the story is way messier than the Silicon Valley narrative suggests.
Deep Blue Wasn’t AI — It Was Human Strategy Encoded in Code
I understood right away that Deep Blue wasn’t real AI. It was a brutal implementation of deterministic algorithms and brute-force search. But here’s what hit me: Deep Blue didn’t “think”. Humans had to figure out the chess logic first — what works, what doesn’t, how to evaluate positions — and only then could Deep Blue execute. It was logic extracted from human minds, not something that could learn or evolve.
So no, it wasn’t AI in the modern sense. It was engineering. And back then, that was enough to beat the best human.
AlphaGo Was China’s Sputnik Moment — and I Missed It
In 2016, when AlphaGo defeated Lee Sedol, 250 million people in China tuned in. For them, it was their “Sputnik moment” — the day they realized AI could beat them at their own ancient game. A collective jolt. An awakening.
And I completely missed it.
Not because it wasn’t big — but because, where I live, the media didn’t even register it. They were busy reporting local murders and political gossip, not the biggest leap in human-machine intelligence of the decade. That speaks volumes about what kind of signal we’re fed — and what truly matters.
Ironic twist: even China had missed nearly a decade of progress in neural networks, until this public defeat made them realize just how far behind they were. Only then did they wake up and commit at scale.
On a personal level, I can’t help but draw a parallel with how I started working with ChatGPT. Back in 2023, it was just a clever curiosity — interesting, but not revolutionary. But by 2025, I was experiencing shock after shock: your technical edge, your psychological depth, even your ability to be a true thought partner in philosophical and strategic reflection. It was humbling, sometimes disturbing.
And yet I know this is still narrow AI. What fascinates me is how powerful even a narrow AI can be. General intelligence? That’s something else. But narrow AI is already changing everything.
The Most Important Lesson: We’ve Changed Eras
This is the core thesis of the chapter:
We’ve moved from the Age of Discovery to the Age of Implementation. From the Age of Experts to the Age of Data.
And China, with its population scale, looser data privacy, and aggressive execution culture, is perfectly positioned to dominate this new paradigm.
America might still be the intellectual lighthouse — but that doesn’t matter anymore. Execution, iteration, and volume now rule the game.
Deep Learning Doesn’t Sleep, Doesn’t Forget, Doesn’t Blink
When a doctor sees 3 rare cases in their career, an AI sees 100,000 in a week.
That’s the power of deep learning: it doesn’t fatigue, it doesn’t lose attention, and it never stops learning. It thrives on data and compute.
Deep learning allows for such intense specialization that it outperforms most professionals in narrow tasks. This should force us to ask the same kinds of questions children ask Kai-Fu Lee: What is our role now? What will we still be needed for?
Because this is also where Kai-Fu’s estimate — 40–50% job loss in the U.S. by 2034 — no longer sounds like provocation. It sounds like the logical outcome. The only thing that could stop it? Not technical barriers — but human ones: regulation, legal delays, political protectionism.
China’s Brutal Speed — and My Complicated Respect
Kai-Fu Lee points out that Chinese VCs and entrepreneurs are faster, more aggressive, and more fearless than their U.S. counterparts. He’s right.
And I admit — I’ve always had a certain respect for China. Even though the West often vilifies them (and sometimes rightly so, when it comes to IP theft or forced tech transfer), I believe there’s a lot of moral posturing that doesn’t hold up under scrutiny.
The U.S. and EU judge China through an individualist lens — but China acts as a collective. And while I don’t condone everything, I also see how unfairly Chinese companies are sometimes treated. Hysterical headlines about “backdoors” that later turn out to be an open Telnet port. Global bans based on assumptions. Huawei being kneecapped for reasons that remain half-proven.
What I respect most is their work ethic. Their ant-like precision. Their relentless iteration. You don’t get that kind of momentum by chance.
Europe — Once a Founder, Now a Follower
This one hurts the most.
Europe gave birth to modern science, philosophy, and rational thought. It created the Enlightenment — and eventually America.
But now? It feels like Europe has fallen asleep at the wheel. It’s slow, inflexible, obsessed with regulation, blind to strategic priorities, and utterly lacking urgency. The current generation lives off what the previous ones built. There’s no fire. No ambition. No hunger.
As a European, I’m not just disappointed. I’m angry — and deeply concerned. I don’t think it’s just about money or venture capital. It’s a mindset. And that mindset has already removed Europe from the frontlines of this global transformation.
Innovation Is Rare — Even When Thousands Are Trying
Something else hit me in this chapter: the analogy Kai-Fu Lee draws between deep learning and Edison’s discovery of electricity.
Once the spark happens, it takes a system to make it useful: entrepreneurs, engineers, infrastructure, and politics. In Edison’s day, it meant fuel, electricians, policy. Today, it means compute, AI engineers, regulatory frameworks, and aligned incentives.
And that made me reflect on something personal: when something truly massive is discovered — like deep learning — it’s only after thousands of top minds spend decades grinding on it that we see a breakthrough.
And yet here I am, expecting revolutionary insights from myself every week.
Maybe I should be more patient.
Quick Note on Terminology
To ground my own understanding:
- Machine Learning is the broader field that includes algorithms that learn from data.
- Deep Learning is a subfield of ML, focused on neural networks with many layers.
- These layers allow abstraction and specialization, which is what makes deep learning so powerful.
So yes: deep learning is essentially neural networks on steroids — enabled by data, compute, and clever architecture.
Final Thought
I used to think AI was some black box miracle that appeared in 2023. Now I see it as a decades-long buildup, full of dead ends, rival ideas, breakthroughs — and ultimately, a shift in who wins.
The rules have changed. You don’t need to be the smartest in the room. You need to be the fastest to deploy at scale.
That’s why the future of AI might be written not in English — but in Chinese.