The Stronger AI Gets, the More I Need Philosophy

Today I was reading Hassabis: The Brain Behind Google's AI (《哈萨比斯：谷歌AI之脑》). It's the first biography I've read of anyone from the AI world. Before this, I'd never heard his name (or I'd forgotten it), but I'd been using a product his team created for a long time: Gemini.

Besides Gemini, I'd heard about him, or about things closely tied to him, at least twice. The most recent was the 2024 Nobel Prize in Medicine. I remember the news said the prize went to a scientist who works on AI, not a medical scientist.

Before that, you have to go back to the first half of 2017. To be precise, I was about to take my high school entrance exams. All I remember is that a machine called AlphaGo beat a Go player. That's it.

Now I open Hassabis: The Brain Behind Google's AI, and the preface puts it like this:

"In 2016, DeepMind's AlphaGo defeated the South Korean player Lee Sedol. The next year, it beat Ke Jie, the Chinese prodigy who was then ranked number one in the world."

Two sentences, neither long nor short, but they caught my interest.

I know who Ke Jie is. I know he plays Go. But I didn't know he was the one who lost to AlphaGo, and I didn't know how high he ranked or how big a name he was in the Go world. I know him because he plays Teamfight Tactics online, and his style online is on the funny side.

So, out of boredom, I looked up Ke Jie's record. Once I did, his story left me deep in thought.

Ke Jie was born in 1997. In 2014, at 17, he won the 2nd Bailing Cup world championship.¹

He won two more titles in 2015 and 2016, making him a three-time champion at 19. You can't call achievements like that anything but brilliant. With that much success at that age, how could anyone not be on top of the world? Many people say Ke Jie was cocky back then, but I think if I'd achieved that much at that age, it would have been hard not to be cocky.

In March 2016, AlphaGo beat Lee Sedol, a top South Korean player. Then, between December 29 that year and January 4 the next, it played online Go against players across Asia and won sixty games in a row.

A new prodigy was born. But it wasn't a person. It was something people created.

Back then Ke Jie posted on Weibo: "Lee Sedol can't beat AlphaGo, but I can." So bold, so cool.

In May 2017, at the Future of Go Summit in Wuzhen, AlphaGo beat Ke Jie 3 to 0.

The father of AlphaGo is Demis Hassabis, the subject of this book. Besides being a physicist and neuroscientist, he's also a Go enthusiast. AlphaGo went from amateur level to world number one in just two years.

Two things about this stunned me.

The first is the development of artificial intelligence. Human Go has a history of thousands of years. The sheer number of stones, moves, and game records is enormous, and the game keeps moving forward. Today's players keep cracking the games left by players before them, and the game keeps evolving.

That slow but healthy development seems to have been broken that day in 2017, in Wuzhen.

The passage I quoted above is understated, but what it really means is a transformation of the entire Go world.

That's the first thing.

The second is Ke Jie himself. In 2017, Ke Jie was China's top player, unbeaten by anyone. Back then, nobody on Earth knew how strong artificial intelligence could get, or what effect it could have on human civilization.

I'm sure Ke Jie didn't know either. Who could guess they were standing at a turning point of the age, and that their own name would be written into it? He didn't know. He was young and proud. He claimed he could beat AlphaGo, carrying all his past results and confidence on his back.

What was Ke Jie back then? Invincible.

The novel Sword Snow Stride has a famous line: "Below the immortals, I'm invincible. Above them, I'll trade my life for theirs."

Ke Jie had the strength and nerve to match that line. And China's materialist education has never believed in gods or buddhas. Even if AlphaGo really were an immortal, he could still trade one for one.

How does that saying go? When the train of the times rolls past, there's always someone under the wheels adding friction.

Did losing make Ke Jie one of the people adding friction? I don't think so. I think Ke Jie played a different role.

The biggest difference between him and ordinary people was that he stood at the very top of the Go world, looking down on everyone. If an ordinary player had lost to AI, humanity could still say: what's the point of beating an amateur? Why don't you go play Ke Jie?

Ke Jie went. He lost. But then what?

In Ke Jie, I see a special kind of tension. It's precisely because he was so strong that his match against AI makes you think so hard: if the human mind has limits, then one day, when AI's mind completely surpasses ours, even far exceeds ours, how should humans live with ourselves?

You can tell Ke Jie really wanted to win back then. Today, now that AI Go players are everywhere, he's said on his livestream: when you play Go against AI, at first you still think about beating it, but after a game or two you think, why am I even fighting this thing?

But back then, Ke Jie didn't yet have such a concrete understanding of AI players. I don't understand Go, but in the material I found, the second game between Ke Jie and AlphaGo was given an extremely high rating by AlphaGo itself.² And online, I saw many people say that second game was the most complete calculation in history, the peak of human Go strength at the time.

From AlphaGo's rating and the discussion online, you can see that Ke Jie didn't just give it some thought. He met it the way you meet a powerful enemy, defending his honor as humanity's top player.

But in the end, he lost.

I find it hard to put myself in Ke Jie's shoes and imagine how he felt. Being beaten by AlphaGo at the peak of his life is as dramatic as a novel.

It's timing, it's fate. That's the second thing that stunned me.

Could this human-versus-machine battle in the Go world be a small rehearsal for the relationship between humans and AI? Or, put another way, now that top AIs officially play Go better than humans, how is human Go doing?

Before getting to that, I want to mention one more thing. After losing to AlphaGo, Ke Jie went on a 22-game winning streak, and over the following years up to 2020, he reached eight career titles.

From then on, AI's influence on Go gradually showed itself. Plenty of so-called "AI moves" appeared, moves that humans would never have played in the past. Doesn't that count as AI's influence on Go?

Scale this up to the lives of all humanity. If human civilization is a torch, will AI make its flame burn brighter or smaller?

Judging by Go, it seems brighter. Imagine it: Ke Jie first plays what people call a perfect game in the human-versus-machine battle, then goes on to the peak of his career after losing. It easily reminds me of the heroes in novels who break through in the middle of a great battle, and of the Divine Move in the TV series Hikaru no Go.

Whether it's a breakthrough at the edge of life and death or the Divine Move, they have one thing in common: you can only reach them by crossing swords with a master as strong as you are.

Break through, and you emerge like a butterfly from its cocoon. Fail, and you die and your path ends.

That's how it is for the strong. What about the rest of us ordinary people?

It reminded me of A Record of a Mortal's Journey to Immortality. One idea runs through the whole novel: "immortals and mortals are worlds apart." From the moment Han Li sets out on the path of cultivation, he understands this. He devotes himself completely to the Way and ruthlessly lets go of anything in life that might tie him down.

People who cultivate immortality never get mixed up in mortal affairs. Whenever immortals and mortals cross paths, bloodshed is sure to follow.

Why am I bringing this up?

I think the two have something in common. Anyone who has used AI and played seriously with vibe coding knows how powerful AI is now.

If I crudely split people into three groups, I think the relationship goes something like this:

The first group is people already using AI deeply. They're like the cultivators in A Record of a Mortal's Journey to Immortality, and fairly advanced ones. They know how powerful AI is, put it to work in production, and have their own understanding of it and their own ways of using it.

The second group has used AI and knows how powerful it is, but their use stays on the surface. For reasons like time and money, they never quite step fully into this world. They're like the people who want to set out on the path of cultivation but never make it, because of talent, competition, and other problems.

The third group has only used chat AI, or never used AI at all. They're like the mortals in A Record of a Mortal's Journey to Immortality.

As I said, this is a crude split. It's purely a simple classification, with no higher or lower, noble or base.

What I really want to say is that of these three groups, the most anxious is actually the second. These people long to make more progress, but for all kinds of reasons they keep hesitating, and they feel torn and lost because of it.

Mortals, on the other hand, may not have the power to move heaven and earth, but they get peace of mind. They can have their own happiness and joy in the human world without chasing the so-called Great Way of eternal life.

The real world isn't a novel. You can't cultivate immortality. Whether you get into AI in this era is entirely up to you. I make a few thousand yuan a month, and I can still use Opus 4.8 and Codex 5.5. So it all comes down to a single choice.

Once Han Li chose the path of cultivation, he was as careful as he could be all the way, devoted completely to the Way.

What I want to say is that in the real world, ordinary people have their own Great Way too. What matters isn't how great the Way is. It's that the Way is always under your own feet.

Today, and in the era to come, philosophy will matter more than ever.

The arrival of AI will hit the human mind in a way it never has before. When machines can easily write, calculate, and answer the things once thought to make humans human, young minds will be more prone than ever to losing their way: does anything I do still mean anything? Nihilism grows quietly in exactly that crack.

And philosophy is the anchor handed to people in all that rocking.

The wind and waves won't stop, but you won't drift away with them.

The same goes for civilization. Whether it grows grand or small has never been up to AI. What really decides it is whether civilization's own anchor is still there.

In the past, that anchor was gods and faith. Later, it was reason and science.

As for the future, I can't yet say what it will be.

Footnotes

1. https://zh.wikipedia.org/wiki/%E6%9F%AF%E6%B4%81 ↩

2. http://www.eweiqi.com/index.php?m=content&c=index&a=show&catid=131&id=28301 ↩