Great Games of Thought
A bearded man in old Chinese dress seated on a riverside rock with a river and rocky mountains behind him; at the far right, the scene runs into the machines and desk of an old laboratory

Great Games of Thought · Game 2  ·   ·  13 min 44 s

If AI Does All My Work, What Should I Do Next?

Zhuangzi×Turing

On the left, a man in old Chinese dress seated on a riverside rock; on the right, a man in a tweed jacket seated at a desk before a blackboard; the two seen side by side

Watch on YouTube (13:44) · Audio in Korean · English subtitles

The Two

Zhuangzi and Turing

Zhuangzi — an imagined portraitAn imagined portrait

Zhuangzi

About 2,300 years ago

Once I dreamt I was a butterfly, fluttering happily, not knowing I was Zhuang Zhou.
Reconstruction; Zhuangzi 'On Seeing Things as Equal' (Inner Chapters)
Turing — an imagined portraitAn imagined portrait

Can machines think? I propose to consider this question.
Reconstruction · Record, Turing 1950 paper first sentence

History · Philosophy · Science

Looking back

History

Qin, bound tight by law, fell just 15 years after unifying China.

Philosophy

from the outside, we can't fully know what thinks or what's alive. So don't judge only by usefulness — learn to play freely.

Science

A healthy heart doesn't beat evenly, and evolution begins with small copying mistakes.

Read

Script and sources

The video's English subtitles, line by line. The dialogue in this video is a dramatization based on the writings the two men left.

Script

99 lines · 12 chapters

The codebreaker in the opening scene is a fictional character. [Record] confirmed in documents · [Tradition] a story handed down (whether it really happened has not been confirmed) · [Reconstruction] lines written on the basis of what the two left · [Interpretation] scholars' interpretation · [Debated] not yet settled

0:00

0:06

0:05Codebreaker

Weather again.

Reconstruction

0:08Narrator

In the Atlantic, German submarines were sinking ships bound for Britain almost every day. They sent orders with a cipher machine called Enigma, and believed it was too complex for anyone to break.

Record

0:19Narrator

But German operators sent weather reports at the same time every day, in almost the same format. Captured codebooks plus this predictable habit, and the U-boat code began to be broken in earnest.

Record

0:32Host

Even the best lock opens if it's predictable.

0:36Host

Ever felt a little uneasy when YouTube knows exactly which video you'll watch next? It means you're predictable.

0:46Host

If AI does all my work, what should I do next? Today we look for the answer in Zhuangzi, 2,300 years ago.

0:54Host

Stay to the end and you'll take away three things. From history: why the state that first unified China collapsed in just 15 years. From philosophy: the question a butterfly dream asks of AI.

1:07Host

From science: the surprising fact that a healthy heart doesn't beat evenly like a machine.

1:22

1:21Narrator

About 2,300 years ago, China's seven major states were at war with one another — the Warring States period.

Record

1:27Narrator

The western state of Qin would later unify China under the First Emperor. Led by a reformer named Shang Yang, it bound the whole country tightly with law.

Record

1:36Narrator

Households were grouped in fives to watch one another, and high rank went only to those who earned merit in war.

Record

1:42Host

The whole country began to run like a precise machine.

1:46Narrator

In the same era, in the state of Song, there was a minor official who tended a lacquer-tree grove. His name was Zhuang Zhou — the man we call Zhuangzi.

Record, Shiji etc.

1:55Narrator

He was poor. The story goes that when he went to borrow grain, the lender said he'd lend him a fortune later, once his taxes came in. Zhuangzi shot back:

Tradition, Zhuangzi 'External Things' (Miscellaneous Chapters; by Zhuangzi's school)

2:04Zhuangzi

On my way here, a carp in a puddle in a wheel rut called out to me, asking for just a dipper of water.

Reconstruction · Tradition, Zhuangzi 'External Things' (Miscellaneous Chapters; by Zhuangzi's school)

2:10Zhuangzi

So I said: soon I'll go to the southern kingdoms and bring you river water. The carp got angry.

Reconstruction · Tradition, Zhuangzi 'External Things' (Miscellaneous Chapters; by Zhuangzi's school)

2:16Zhuangzi

'Then look for me at the dried-fish shop.'

Reconstruction · Tradition, Zhuangzi 'External Things' (Miscellaneous Chapters; by Zhuangzi's school)

2:19Host

Don't cover today's urgent need with a big promise for later. He lived 2,300 years ago, yet he talks just like someone today.

2:29Narrator

It is said that the king of Chu sent envoys offering him what today would be the prime minister's post. Zhuangzi, still holding his fishing rod, replied:

Tradition

2:39Zhuangzi

Rather than a dead tortoise kept in the ancestral temple, I'd be a living one dragging its tail in the mud.

Reconstruction · Tradition, Zhuangzi 'Autumn Floods' (Outer Chapters; by Zhuangzi's school)

2:47Narrator

Scholars see Zhuangzi, with Laozi, as the thinker of living naturally without forcing things — what's called wu wei, 'effortless action.'

Interpretation

2:56Narrator

Zhuangzi's sparring partner in argument was his friend Huizi. After Huizi died, passing his grave, Zhuangzi is said to have remarked that he no longer had anyone to talk with.

Tradition, Zhuangzi 'Xu Wugui' (Miscellaneous Chapters; by Zhuangzi's school)

3:06

3:06Narrator

Alan Turing. Born in London in 1912.

Record

3:09Narrator

In 1936, at 24, Turing sketched an imaginary machine on paper — one that reads and writes symbols by following a few simple rules. The root of every computer today: the Turing machine.

Record

Note — Turing's paper was received by the London Mathematical Society on 28 May 1936, when he was 23, and read before it in November, when he was 24. "Root" is an interpretation: the foundation of the theory of computation.

3:22Narrator

When war broke out, at Bletchley Park he designed the Bombe, a huge machine for breaking Enigma.

Record

3:29Narrator

He was an eccentric. He reportedly cycled to work in a gas mask because of hay fever and chained his favorite mug to a radiator so no one would take it. He ran a marathon in 2 hours 46 minutes.

Tradition, Hodges biography (colleagues' recollections); Record (marathon, 1947)

3:41Host

A genius — and a bit adorable.

3:44Narrator

And in 1950 he began a paper with this first sentence:

Record

3:49Turing

Can machines think? I propose to consider this question.

Reconstruction · Record, Turing 1950 paper first sentence

3:54Narrator

Scholars call Turing the father of computer science.

Interpretation

3:57Narrator

His proposed test, the Turing test, is still used as a yardstick for how human-like AI is — though many still ask whether good imitation really means thinking.

Interpretation, debated

4:08Host

One, it is said, turned down high office and went fishing; the other tried to end a war with a machine run by rules.

4:17

4:16Host

Three words are all you need to follow today's story.

4:20Host

To operate. Put a coin in a vending machine and the same can drops out every time. Moving by rules — that's operating.

Note · Operate = move by rules

4:28Host

To think: not following what's set, but weighing things for yourself.

Note · Think = weigh things for yourself

4:33Host

And to be alive. On which step of this staircase does AI stand today?

Note · Life = to be alive

4:41

4:41Zhuangzi

See how freely the minnows swim. That is the joy of fish.

Reconstruction · Tradition, Zhuangzi 'Autumn Floods' (Outer Chapters; by Zhuangzi's school)

4:46Huishi

You're not a fish — how do you know fish are happy?

Reconstruction · Tradition, Zhuangzi 'Autumn Floods' (Outer Chapters; by Zhuangzi's school)

4:50Zhuangzi

You're not me — how do you know I don't know the joy of fish?

Reconstruction · Tradition, Zhuangzi 'Autumn Floods' (Outer Chapters; by Zhuangzi's school)

5:02Host

The question asked on that bridge 2,300 years ago is one we now ask AI every day.

5:11

5:10Host

The two never met. What follows is an imagined conversation built from what they wrote.

5:20Turing

Break down the process of human calculation into very small rules.

Reconstruction; Turing 1936 paper §9

5:25Turing

If it can be broken down, a machine can do it too.

Reconstruction; Turing 1936 paper §9

5:32Zhuangzi

Let me tell you about Ding, a man who butchered oxen. At first, all he could see was the ox.

Reconstruction; Zhuangzi 'Nourishing Life' (Inner Chapters)

5:38Zhuangzi

After some years, he met the ox not with his eyes but with his mind.

Reconstruction; Zhuangzi 'Nourishing Life' (Inner Chapters)

5:43Zhuangzi

His knife slid by itself through the gaps between bone and flesh, and after nineteen years its edge was still like new.

Reconstruction; Zhuangzi 'Nourishing Life' (Inner Chapters)

5:51Turing

Can machines think? The question is too vague to argue about.

Reconstruction; Turing 1950 paper §1;§6

5:56Turing

So let's turn it into a game.

Reconstruction; Turing 1950 paper §1

5:58Turing

Talk only in writing with someone in another room, and guess whether it's a person or a machine.

Reconstruction; Turing 1950 paper §1;§2

6:02Turing

The imitation game.

Reconstruction; Turing 1950 paper §1

6:07Zhuangzi

My friend Huizi asked the same: 'You're not a fish — how do you know?'

Reconstruction; Zhuangzi 'Autumn Floods' (Outer Chapters; by Zhuangzi's school)

6:13Turing

Push that view to its extreme, and the only way to know another person thinks is to be that person.

Reconstruction; Turing 1950 paper §6(4)

6:17Turing

So we all live by the polite convention that everyone thinks.

Reconstruction; Turing 1950 paper §6(4)

6:25Host

One was written by a man 2,300 years ago; the other by AI for this video. Which one is human?

6:36Host

A is the human — Zhuangzi. B was written by AI. If you hesitated, Turing's question has already reached you.

6:46Zhuangzi

Once I dreamt I was a butterfly, fluttering happily, not knowing I was Zhuang Zhou.

Reconstruction; Zhuangzi 'On Seeing Things as Equal' (Inner Chapters)

6:52Zhuangzi

Suddenly I woke — and there I was, Zhuang Zhou.

Reconstruction; Zhuangzi 'On Seeing Things as Equal' (Inner Chapters)

6:57Zhuangzi

Did Zhuang Zhou dream he was a butterfly, or is a butterfly dreaming it is Zhuang Zhou?

Reconstruction; Zhuangzi 'On Seeing Things as Equal' (Inner Chapters)

7:04Turing

I do not wish to give the impression that there is no mystery about consciousness.

Reconstruction; Turing 1950 paper §6(4)

7:10Host

Hold on — today even scientists seriously ask whether our world might be a virtual one, like a game inside a computer. The butterfly's dream is back.

7:22Turing

Lady Lovelace said a machine can only do what we tell it to do — it can't originate anything new.

Reconstruction; Turing 1950 paper §6(6)

7:27Turing

Yet machines take me by surprise with great frequency.

Reconstruction; Turing 1950 paper §6(6)

7:31Turing

If we build a learning machine, it's probably wise to include a little randomness.

Reconstruction; Turing 1950 paper §7

7:40Zhuangzi

Huizi grumbled: he planted gourd seeds from the king, and the gourd grew so big —

Reconstruction; Zhuangzi 'Free and Easy Wandering' (Inner Chapters)

7:46Zhuangzi

too heavy to hold water, too big to use as a ladle — that he just smashed it.

Reconstruction; Zhuangzi 'Free and Easy Wandering' (Inner Chapters)

7:51Zhuangzi

Why didn't he think of making it into a big boat and floating on the river?

Reconstruction; Zhuangzi 'Free and Easy Wandering' (Inner Chapters)

8:00

7:59Narrator

Qin, the state of rules, unified China in 221 BC — and collapsed in just 15 years.

Record

8:07Host

Many historians blame its harsh laws.

Interpretation

8:11Narrator

Enigma had a strange weakness: no letter could ever be encrypted as itself. Turing and his colleagues exploited this gap — and the operators' predictable habits.

Record

8:23Narrator

All this stayed secret until the 1970s, so Turing could never tell the world what he had done. In 1952, under British law of the time that treated homosexuality as a crime, he was convicted.

Record

8:37Narrator

He died in June 1954, at 41. The official verdict was suicide, though some researchers point to the possibility of an accident.

Record; debated

8:45Narrator

In 2013 the Queen pardoned him, and in 2021 his face appeared on the £50 note.

Record

8:53Narrator

Then, March 2016, Seoul: Google's Go AI, AlphaGo, versus Lee Sedol, 9-dan. In game two, AlphaGo's move 37 was one it estimated a human would play only once in ten thousand times.

Record

9:08Narrator

Game four: Lee Sedol's move 78. With this move AlphaGo failed to predict, a human won a single game.

Record

9:16Host

A move the machine couldn't predict — and a move the human couldn't predict.

9:23

9:23Host

Before I share my view, I'd like to hear yours.

9:31Host

There's no right answer. So far, those were the facts and scholars' readings. From here on, it's my own view.

9:38

9:38Host

At work, AI drafts our reports; at home, algorithms pick what we watch tonight.

9:43Host

AI is imperfect too, so it keeps learning — and it's becoming perfect very fast. We race too, trying to work faster and more accurately than everyone else. But once it's all done, what comes next?

10:00Host

Here's my answer. What comes after perfect is not an end, but a new evolution.

10:07Host

Life did exactly that. Genes copy themselves almost perfectly, but once in a while they make a mistake. Over long ages, nature sifts those small mistakes and they add up into new life. That's evolution.

10:21Host

Qin, on the other hand, bound tight by law, vanished just 15 years after unifying China. Living things are never finished. Within the time we're given, we keep searching for ourselves and learning.

10:36Host

The body is the same. A healthy heart doesn't beat evenly like a machine. The beats wobble slightly — and when that wobble disappears, doctors see it as a sign of illness.

Record (research)

10:46Host

An age when AI does more and more of the useful work. Rather than straining to be more useful, life is a play — so enjoy it, don't stress. Wu wei, naturally.

Note · Wu wei, naturally = without forcing things

10:59Host

Zhuangzi already knew, 2,300 years ago: to roam freely without measuring everything by its use. That's the story of the gourd too big to be useful — make it a boat and float down the river.

Note · 'Free and Easy Wandering', the first chapter of the Zhuangzi = roaming freely

11:11Zhuangzi

Why didn't he think of making it into a big boat and floating on the river?

Reconstruction, Zhuangzi 'Free and Easy Wandering' (Inner Chapters)

11:15Host

But what about food on the table? We all have to do work we don't want, for money. Zhuangzi himself was a minor official guarding a lacquer grove, and is said to have been so poor he went to borrow grain.

11:27Host

Philosophy won't pay your salary. The work you must do is the stage. On it, once a day, you add a small wobble of your own choosing. Cook Ding, the butcher, roamed freely like that within his own work.

Note · 'Nourishing Life', the chapter with Cook Ding's story = how to nourish life

11:43Host

I think AI will take the predictable, repetitive work first. What's left, then, is probably helping people live like human beings. I dream of a world where that work pays.

11:56Host

So here's my suggestion: why not try studying history, philosophy, science — and literature too?

12:02Host

Make your own reason to live — your own will — with your own hands. Making new mistakes, and learning again. Searching for what AI can't give us.

12:14Host

Great Games of Thought will study it with you — easily, and with fun. Not staying asleep, but waking each other's minds.

12:21Host

And from today, look after your body, take a little less stress, and try walking a road you've never taken.

12:27Host

Rather than living only for tomorrow's big promises, live seeing that you are alive today. What the carp needed, too, was this one ladle of water — now.

12:38Host

Each today you gather becomes the story that makes you.

12:44

12:44Narrator

History. Qin, bound tight by law, fell just 15 years after unifying China.

12:49Narrator

Philosophy: from the outside, we can't fully know what thinks or what's alive. So don't judge only by usefulness — learn to play freely.

13:00Narrator

Science. A healthy heart doesn't beat evenly, and evolution begins with small copying mistakes.

13:12

13:11Host

One last question. On the day AI does all your work, what will you be doing?

Sources and research

Sources are listed in their original Korean.

  1. Record사마천, 『사기』 권63 — 장자(이름 장주), 몽 땅의 칠원리(옻나무 밭을 맡은 관리). 몽을 송나라 땅으로 보는 것은 유향 『별록』 등 · Wikisource
  2. Record『장자』 내편 — 소요유(너무 큰 박) · 양생주(포정) · 제물론(나비 꿈) · Wikisource
  3. Tradition『장자』 외편 — 추수(다리 위 물고기 문답 · 초나라의 재상 제안과 거북, 재상 제안은 『사기』 권63에도) / 잡편 — 외물(곡식을 꾸러 간 장자와 수레바퀴 자국 속 붕어) · 서무귀(혜시의 무덤) — 외편·잡편은 장자 학파의 글로 보는 견해가 많음 · Wikisource
  4. Record『사기』 권68 「상군열전」 — 상앙의 변법(다섯 집 연좌, 군공에 따른 작위) · 권6 「진시황본기」 — 진나라의 통일(기원전 221)과 멸망(기원전 206) · [해석] 멸망 원인을 가혹한 법에서 찾는 견해(가의 「과진론」 이래) · Wikisource
  5. Record앨런 튜링, 「계산 가능한 수에 관하여」(1936, 9절) · 「계산 기계와 지능」(Computing Machinery and Intelligence, Mind, 1950 — 1절 흉내 게임, 6절 (4) 예의 바른 약속·의식의 신비, 6절 (6) 러브레이스의 반론과 '기계는 나를 아주 자주 놀라게 한다', 7절 배우는 기계와 우연) · DOI
  6. Record에이다 러브레이스(1815~1852), 배비지의 해석 기관에 관한 주석(1843): '해석 기관은 무엇을 새로 만들어 낸다고 내세우지 않는다. 우리가 시킬 줄 아는 일은 무엇이든 할 수 있다.'(튜링 1950 인용) · Internet Archive
  7. Record에니그마를 처음 푼 것은 1932년 폴란드 암호국(마리안 레예프스키). 1941년 U보트 암호는 영국 해군이 붙잡은 암호표(기상선 뮌헨호·U-110·기상선 라우엔부르크호)에, 정해진 시각·형식으로 나가던 날씨 보고 같은 전문이 더해져 본격적으로 풀림. 봄브는 튜링이 설계(1939~1940)하고 웰치먼이 개량 · TNMOC · ethw.org · bletchleypark.org.uk
  8. Tradition방독면 쓰고 자전거 출근, 머그잔 쇠사슬 일화 — 동료 회고(앤드루 호지스, 『앨런 튜링의 이미테이션 게임』, 김희주·한지원 옮김, 동아시아, 2015) · [기록] 마라톤 2시간 46분(1947) · Aladin · turing.org.uk
  9. Record1946년 대영제국 훈장(OBE)을 받았으나 전시 공적의 내용은 1970년대까지 비밀 · 판결(1952)과 죽음(1954, 공식 판정 자살, 사고설은 [논쟁 중]) · 사후 사면(2013) · 50파운드 지폐(2021) · turing.org.uk · gov.uk · bankofengland.co.uk
  10. Record2016년 3월 서울, 알파고와 이세돌 9단의 다섯 경기(두 번째 경기 알파고의 37번째 수, 네 번째 경기 이세돌 9단의 78번째 수 — 딥마인드 발표) · Google DeepMind
  11. Record · Research돌연변이(유전자 복사 실수)는 진화의 원재료이고 자연선택 등이 함께 작용 · 심박 변이 감소와 건강(유럽심장학회·북미심박학회 태스크포스, 1996) · DOI · evolution.berkeley.edu
  12. Debated시뮬레이션 가설(닉 보스트롬, 2003) · DOI
  13. Note영상 속 'B' 문장은 이 영상을 위해 AI가 쓴 글입니다 · 'A' 문장은 『장자』 제물론(내편) 나비 꿈 구절(昔者莊周夢為胡蝶…不知周也)의 우리말 옮김입니다 · Wikisource

If you find a mistake, please tell us in the video's comments. We will check it and correct it. How We Check Facts

A man in old-style clothes crouching on a dirt road with wheel ruts, looking down at a carp in the water pooled in a rut

One last question. On the day AI does all your work, what will you be doing?

Share your thoughts in the YouTube comments

Keep watching

Other games

All games →