by the artist Michael Najjar,

這是張相片

and it's real,

由藝術家Michael Najjar 拍攝的

in the sense that he went there to Argentina

這張相片是真的

to take the photo.

也就是說,他親自到阿根廷,那座山的所在處

But it's also a fiction. There's a lot of work that went into it after that.

拍攝這張照片。

And what he's done

但也可以說,這是張虛構的相片。這張相片的完作花了很多功夫。

is he's actually reshaped, digitally,

他對相片動了些手腳:

all of the contours of the mountains

數位化重整

to follow the vicissitudes of the Dow Jones index.

整片山脈的形體輪廓,

So what you see,

使其隨道瓊指數曲線變化。

that precipice, that high precipice with the valley,

你所看到的

is the 2008 financial crisis.

那個峭壁, 那個有處凹陷的高聳峭壁

The photo was made

代表2008年的金融危機。

when we were deep in the valley over there.

拍攝這張相片時

I don't know where we are now.

我們的金融情勢正處於低谷,

This is the Hang Seng index

不曉得我們現在處於何種形勢。

for Hong Kong.

這是恆生指數,

And similar topography.

香港股市價格的重要指標。

I wonder why.

(兩張相片)地形相似,

And this is art. This is metaphor.

我想知道為什麼

But I think the point is

這是藝術;這是種象徵。

that this is metaphor with teeth,

但我認為重點是

and it's with those teeth that I want to propose today

這個象徵有“牙齒”。

that we rethink a little bit

就是因為這些“牙齒”,我今天提議

about the role of contemporary math --

我們稍微重新思考

not just financial math, but math in general.

當代數學的角色;

That its transition

不只金融數學,還有普通數學。

from being something that we extract and derive from the world

「它」的演變:

to something that actually starts to shape it --

從我們鑽研這個世界,抽絲撥繭而取得的發現

the world around us and the world inside us.

到實際開始形成「它」的重要發現,

And it's specifically algorithms,

這含括我們的外在世界和我們內在的世界。

which are basically the math

說明確些,它是演算法,

that computers use to decide stuff.

基本上,是種數學─

They acquire the sensibility of truth

─電腦用來測定東西的數學。

because they repeat over and over again,

演算法掌握高度精確的計量,

and they ossify and calcify,

因為它們一而再,再而三的重覆著;

and they become real.

然後漸漸成型，發展出基本架構

And I was thinking about this, of all places,

然後它們變得實際且可靠。

on a transatlantic flight a couple of years ago,

我當時正在思考這點, 真是太湊巧了!

because I happened to be seated

就在幾年前,橫越大西洋的班機上,

next to a Hungarian physicist about my age

因為我的座位碰巧

and we were talking

在一位年紀與我相仿的匈牙利物理學家隔壁

about what life was like during the Cold War

我們談論關於

for physicists in Hungary.

匈牙利冷戰期間

And I said, "So what were you doing?"

物理學家的生活情況。

And he said, "Well we were mostly breaking stealth."

我說:「你那時在做什麼?」

And I said, "That's a good job. That's interesting.

他說:「嗯,我們大多在打擊祕密行動。」

How does that work?"

「那是個好工作,有趣吧,

And to understand that,

那是怎麼運作的?」

you have to understand a little bit about how stealth works.

要了解那之前

And so -- this is an over-simplification --

你必須稍稍了解祕密行動的運作。

but basically, it's not like

這是個超簡單化的例子,

you can just pass a radar signal

基本上,它不像是

right through 156 tons of steel in the sky.

你可以藉由156噸在天空飛的鋼鐵

It's not just going to disappear.

傳送雷達信號。

But if you can take this big, massive thing,

飛機不會消失不見。

and you could turn it into

但若你能將這個龐大、具規模的東西

a million little things --

變成

something like a flock of birds --

百萬個小玩意

well then the radar that's looking for that

─像鳥群一樣的東西─

has to be able to see

那麼雷達偵測到那一群群的小東西

every flock of birds in the sky.

必定會看到

And if you're a radar, that's a really bad job.

在空中有“一群群的鳥”

And he said, "Yeah." He said, "But that's if you're a radar.

若你是一個雷達,事情可就糟了。

So we didn't use a radar;

他說:「對,但那是,如果你是個雷達。

we built a black box that was looking for electrical signals,

所以我們不用雷達,

electronic communication.

我們建造一個黑箱子,用它搜尋電波,

And whenever we saw a flock of birds that had electronic communication,

電子通訊。

we thought, 'Probably has something to do with the Americans.'"

任何時候,我們發現帶有電子通訊的鳥群,

And I said, "Yeah.

我們會認為這很可能跟美國人有關。」

That's good.

我接著說:「是啊,

So you've effectively negated

真行,

60 years of aeronautic research.

你們成功地消磨了

What's your act two?

60年的航空學研究心血。

What do you do when you grow up?"

你接著要做什麼?

And he said,

當你長大成人以後,你從事什麼工作?」

"Well, financial services."

他回答:

And I said, "Oh."

「嗯,金融服務業。」

Because those had been in the news lately.

我驚呼:「喔!」

And I said, "How does that work?"

那一陣子相關報導一直在新聞出現。

And he said, "Well there's 2,000 physicists on Wall Street now,

我問:「進行的如何?」

and I'm one of them."

他說:「現有2,000名物理學家在華爾街(美國金融中心),

And I said, "What's the black box for Wall Street?"

我是他們其中一人。」

And he said, "It's funny you ask that,

我接著問:「華爾街用的『黑箱』是什麼?」

because it's actually called black box trading.

他回說:「你這樣問很好笑,

And it's also sometimes called algo trading,

事實上,人們會稱它為『黑箱交易』

algorithmic trading."

有時也稱為

And algorithmic trading evolved in part

「演算法交易」(algorithmic trading 或algo trading）

because institutional traders have the same problems

「演算法交易」的演進發展,有部分是因為

that the United States Air Force had,

某些機構交易人遇到相同的問題;

which is that they're moving these positions --

而那問題美國空軍也同樣遭遇到,

whether it's Proctor & Gamble or Accenture, whatever --

他們都在「移動這些位置」──

they're moving a million shares of something

不論是寶僑(Proctor&Gamble)或埃森哲(Accenture:管理顧問、技術服務公司)

through the market.

他們都在移動一百萬股的東西,

And if they do that all at once,

透過市場交易而進行。

it's like playing poker and going all in right away.

如果他們一次就挪動全部,

You just tip your hand.

就像玩撲克牌,把剩下的所有籌碼一次全部壓上,

And so they have to find a way --

你只會過早洩露底餡;

and they use algorithms to do this --

所以他們必須找到方法

to break up that big thing

─他們用演算法,有系統的操作─

into a million little transactions.

將龐然大數化整為零

And the magic and the horror of that

成為百萬個小交易。

is that the same math

恐怖的是這個魔術正是

that you use to break up the big thing

「相同的數學」

into a million little things

─用來瓦解龐然巨物

can be used to find a million little things

變成百萬個小東西─

and sew them back together

可以用來計算出百萬個零星單位

and figure out what's actually happening in the market.

又將他們統整在一起

So if you need to have some image

並推算出實際在市場上發生的事情。

of what's happening in the stock market right now,

如果你立即需要

what you can picture is a bunch of algorithms

一些股市交易的樣貌,

that are basically programmed to hide,

你可以構想到的是,成串的運算法

and a bunch of algorithms that are programmed to go find them and act.

基本上被設計為隱藏不顯示

And all of that's great, and it's fine.

和成串的運算法被設計為可搜尋並執行。

And that's 70 percent

整個設計的真是太棒了,又精確。

of the United States stock market,

那是百分之七十的

70 percent of the operating system

美國股票市場,

formerly known as your pension,

這個百分之七十的營運系統

your mortgage.

之前堪稱為某些人的“退休金”

And what could go wrong?

某人的“抵押借款”。

What could go wrong

會有什麼錯呢?

is that a year ago,

事情出了差池:

nine percent of the entire market just disappears in five minutes,

一年前

and they called it the Flash Crash of 2:45.

整體股市的百分之九突然消失了五分鐘,

All of a sudden, nine percent just goes away,

人們稱之為『瞬間當機2:45』

and nobody to this day

突然, 百分之九就這樣不見了,

can even agree on what happened

直到今天,仍沒有人

because nobody ordered it, nobody asked for it.

對發生的事取得一致的意見,

Nobody had any control over what was actually happening.

因為沒人“下令”當機;沒人自找麻煩。

All they had

大家對實際正在發生的事情束手無策

was just a monitor in front of them

他們只有

that had the numbers on it

盯著面前的電腦螢幕,

and just a red button

電腦螢幕上的數字,

that said, "Stop."

和一顆紅色按紐

And that's the thing,

上面寫著: 『停止』

is that we're writing things,

事情就是這樣,

we're writing these things that we can no longer read.

我們正在編寫的「東西」,

And we've rendered something

我們正在編寫這些連自己都看不懂的東西。

illegible,

我們已經對「某種東西」投降了,

and we've lost the sense

某種「難以辨識」的東西。

of what's actually happening

而且我們失去了

in this world that we've made.

對實際正發生之事的判別力

And we're starting to make our way.

就在我們自己創造的這個世界中,

There's a company in Boston called Nanex,

況且我們正開始邁向成功。

and they use math and magic

在波士頓有間公司叫Nanex(該公司開發市場數據供給系統),

and I don't know what,

他們用數學和魔法

and they reach into all the market data

和我不知道的什麼來的

and they find, actually sometimes, some of these algorithms.

他們深入研究市場數據資料

And when they find them they pull them out

他們確實發現值得重視的東西:某些演算法

and they pin them to the wall like butterflies.

當他們發現這些演算程序,便把它們擷取出來

And they do what we've always done

並將它們像蝴蝶一樣釘在牆上。

when confronted with huge amounts of data that we don't understand --

他們做大家總是會做的事情,

which is that they give them a name

當面臨龐大又不懂的數據資料時,

and a story.

為其命名

So this is one that they found,

和揑造故事。

they called the Knife,

這是他們的發現:

the Carnival,

他們稱為『刀』

the Boston Shuffler,

『嘉年華會』(Carnival)

Twilight.

『波士頓通勤者』(Boston Shuffler )

And the gag is

『暮光』

that, of course, these aren't just running through the market.

好玩的是

You can find these kinds of things wherever you look,

當然,這些不光是存在於金融市場;

once you learn how to look for them.

你能在任何你看得到的地方,發現這些東西,

You can find it here: this book about flies

一旦你明白如何找尋到它們(演算法)。

that you may have been looking at on Amazon.

從這兒你可以發現:這是本關於蒼蠅的書,

You may have noticed it

你可能已在亞馬遜看到這本書;

when its price started at 1.7 million dollars.

你可能已經注意到

It's out of print -- still ...

它的價格從一百七十萬元起價時,

(Laughter)

這本書是絶版的......仍然絶版中。

If you had bought it at 1.7, it would have been a bargain.

(笑笑)

A few hours later, it had gone up

如果能以一百七十萬的價格買下它是很划算的

to 23.6 million dollars,

稍後幾小時,它飆漲至

plus shipping and handling.

兩千三百六十萬元,

And the question is:

包含運費和手續費。

Nobody was buying or selling anything; what was happening?

問題是:

And you see this behavior on Amazon

這並無產生任何買賣行為；發生了什麼事?

as surely as you see it on Wall Street.

你在亞馬遜見到這樣的行為,

And when you see this kind of behavior,

確實跟你在華爾街看到的一般。

what you see is the evidence

當你見到這種行為:

of algorithms in conflict,

你所看到的顯然正是

algorithms locked in loops with each other,

矛盾的演算程序,

without any human oversight,

演算程序被彼此套住,卡在電腦程式回路中;

without any adult supervision

沒有任何“人類監管”

to say, "Actually, 1.7 million is plenty."

沒有任何“成人監護”

(Laughter)

來告訴你,“其實,一百七十萬已經夠多了!”

And as with Amazon, so it is with Netflix.

(笑笑)

And so Netflix has gone through

如同亞馬遜,Netflix(美國公司,經營線上串流影片)也一樣。

several different algorithms over the years.

多年來, Netflix採用過

They started with Cinematch, and they've tried a bunch of others --

好幾個不同的演算程序。

there's Dinosaur Planet; there's Gravity.

他們從Cinematch(推薦系統軟體)開始,也試了一連串其他的軟體。

They're using Pragmatic Chaos now.

有Dinosaur Planet團隊、Gravity團隊各別研發的推薦系統。

Pragmatic Chaos is, like all of Netflix algorithms,

他們現在使用 Pragmatic Chaos研發的系統。

trying to do the same thing.

像所有Netflix的運算系統,

It's trying to get a grasp on you,

Pragmatic Chaos研發的推薦系統,試圖做相同的事。

on the firmware inside the human skull,

它試著去掌控你們,

so that it can recommend what movie

控制人類頭顱內的思考邏輯,

you might want to watch next --

以便它能推薦你

which is a very, very difficult problem.

下次你也許想看的電影─

But the difficulty of the problem

─這是非常高難度的難題。

and the fact that we don't really quite have it down,

但問題和事實的艱難度

it doesn't take away

─我們不是真的掌握問題的事實─

from the effects Pragmatic Chaos has.

並沒減損

Pragmatic Chaos, like all Netflix algorithms,

Pragmatic Chaos的影嚮。

determines, in the end,

Pragmatic Chaos,如同所有Netflix運算系統,

60 percent

至終裁定

of what movies end up being rented.

百分之六十的

So one piece of code

哪些電影最後會被租借。

with one idea about you

所以一片程式編碼

is responsible for 60 percent of those movies.

─紀錄著你們看片的喜好─

But what if you could rate those movies

得為百分之六十的電影負責。

before they get made?

但倘若你能評估這些電影,

Wouldn't that be handy?

在電影製作前作預測呢?

Well, a few data scientists from the U.K. are in Hollywood,

那不就簡便多了?

and they have "story algorithms" --

嗯,在好萊塢,一些來自英國的數據科學家

a company called Epagogix.

擁有故事情節演算程式系統──

And you can run your script through there,

一間公司叫Epagogix(英國一家預測劇本未來票房好壞的公司)

and they can tell you, quantifiably,

你可以拿劇本請這間公司幫你預測;

that that's a 30 million dollar movie

他們會提供你數據:

or a 200 million dollar movie.

那是一部可賣三千萬的電影

And the thing is, is that this isn't Google.

或是一部兩億的賣座電影。

This isn't information.

事情是......這不是Google;

These aren't financial stats; this is culture.

這不是情報資料;

And what you see here,

這些不是金融統計;這是文化。

or what you don't really see normally,

你們在這裡見到的,

is that these are the physics of culture.

或者說,實際上,你通常不會察覺的

And if these algorithms,

是物理文化

like the algorithms on Wall Street,

而且若這些演算系統

just crashed one day and went awry,

像華爾街的演算系統

how would we know?

某天突然當機,出岔子了

What would it look like?

我們如何會知道.....

And they're in your house. They're in your house.

那會如何?

These are two algorithms competing for your living room.

再者,它們就在你的房子內,它們就在你的房子內

These are two different cleaning robots

兩個演算系統在競爭你的客廳。

that have very different ideas about what clean means.

兩個不同的清潔機器人

And you can see it

對乾淨的定義有不同的概念。

if you slow it down and attach lights to them,

而且你能從中看到演算程序,

and they're sort of like secret architects in your bedroom.

如果讓它慢下來,為它們裝上LCD燈的話,你們就能見識到。

And the idea that architecture itself

而且他們有點像在你卧房內的袐密建築師。

is somehow subject to algorithmic optimization

況且建築學本身的概念

is not far-fetched.

從某種角度而言,是基於演算法的最佳化

It's super-real and it's happening around you.

一點也不牽強喔,

You feel it most

超真實而且就在存在你週遭。

when you're in a sealed metal box,

你感受最深的時刻是,

a new-style elevator;

當你在一個密閉的金屬箱子內

they're called destination-control elevators.

─一臺新型的電梯─

These are the ones where you have to press what floor you're going to go to

他們被稱為「終點控制電梯」。

before you get in the elevator.

這些是電梯,你可以按鈕到你要去的樓層

And it uses what's called a bin-packing algorithm.

在你“進電梯前”按鈕。

So none of this mishegas

它使用所謂的「裝著演算法的盒子」。

of letting everybody go into whatever car they want.

也就是說,這一點也不異常或瘋狂,

Everybody who wants to go to the 10th floor goes into car two,

讓每個人選擇進入任何一台電梯。

and everybody who wants to go to the third floor goes into car five.