artificial intelligence, AI.

總有一天，機器人 會像人類一樣聰明，

How could that happen?

人工智慧，AI。

One route is that we'll just keep accumulating better software,

那要如何做到呢？

like we've been doing for 70 years.

其中一個方式就是， 我們只要繼續累積更好的軟體，

At past rates of progress, that may take centuries.

就像我們過去七十年所做的。

Some say it'll happen a lot faster

但以過去的進步速率推斷， 這樣要花上數百年。

as we discover grand new powerful theories of intelligence.

有人說，它會提早實現，

I'm skeptical.

因為我們發現了 全新的強大智慧理論。

But a third scenario

我持懷疑態度。

is what I'm going to talk about today.

但有第三種情況，

The idea is to port the software from the human brain.

也是我今天要講的。

To do this, we're going to need three technologies to be good enough,

這個想法是從人腦中移植出軟體。

and none of them are there yet.

要做到這一點，我們需要 讓三種技術的發展足夠成熟，

First, we're going to need lots of cheap, fast, parallel computers.

但目前我們這些都還沒達到。

Second, we're going to need to scan individual human brains

第一，我們需要許多 廉價、快速、平行的電腦。

in fine spatial and chemical detail,

第二，我們需要在適當的 空間與正確的化學成分下

to see exactly what cells are where, connected to what, of what type.

掃瞄我們的大腦。

And third, we're going to need computer models

這樣才能看出細胞的 精確位置、連結、類型。

of how each kind of brain cell works --

第三，我們需要一種電腦模型，

taking input signals, changing interval state

關於各類腦細胞如何運作的模型──

and sending output signals.

可以讓我們輸入訊號， 改變區間的狀態，

If we have good enough models of all the kinds of brain cells

並再發送出訊號的模型。

and a good enough model of the brain,

如果我們有夠好的各種腦細胞模型、

we can put it together to make a good enough model of an entire brain,

夠好的大腦模型，

and that model would have the same input-output behavior as the original.

我們可以把它們結合起來， 做出一個完整的大腦模型，

So if you talk to it, it might talk back.

那個模型與原本大腦所想要的 輸入和輸出行為一致。

If you ask it to do things, it might do them.

所以，如果你跟它說話， 它可能會回話。

And if we could do that, everything would change.

如果你要它做事，它可能會去做。

People have been talking about this idea for decades,

如果我們能做到這樣， 一切都會改變。

under the name of "uploads."

幾十年來人們以「上傳」的方式

I'm going to call them "ems."

談論這個想法。

When they talk about it, they say,

我打算稱它為「仿真腦（ems）」。

"Is this even possible?

當他們在談論這個想法時，他們說：

If you made one, would it be conscious? Or is it just an empty machine?

「這是可能的嗎？」

If you made one of me, is that me or someone else?"

如果能做出來，它會是有意識的嗎？ 還是只是個空機器？

These are all fascinating questions that I'm going to ignore ...

如果你仿了我的大腦， 它會是我還是另一個人？

(Laughter)

我將忽略這些令人著迷的問題…...

because I see a neglected question:

（笑聲）

What would actually happen?

因為我看到了一個被忽視的問題：

I became obsessed with this question.

究竟會發生什麽事？

I spent four years trying to analyze it,

我對這個問題很著迷。

using standard academic tools, to guess what would happen,

我花了四年時間來分析它，

and I'm here to tell you what I found.

用標準的學術工具， 來猜測可能會發生什麼事，

But be warned --

我來這裡是要告訴各位我的發現。

I'm not offering inspiration, I'm offering analysis.

但，請注意──

I see my job as telling you what's most likely to happen

我不是要提供靈感， 我是要提供分析。

if we did the least to avoid it.

我認為我的工作是 告訴你們最有可能發生的情況，

If you aren't at least a bit disturbed by something I tell you here,

如果我們不花心思去預防，

you're just not paying attention.

如果你對於我今天告訴你的事 並不感到擔心，

(Laughter)

那表示你沒在注意聽。

OK, the first thing I can tell you

（笑聲）

is that ems spend most of their life in virtual reality.

好，我能告訴各位的第一件事是，

This is what you might look like if you were using virtual reality.

仿真腦的大部分時間 都是在虛擬現實生活中度過的。

And this is what you might see:

如果你使用虛擬實境， 這就是你的樣子。

sunlight glinting off of water, you might hear gulls flying above,

這也是你可能會看到的：

you might even feel the wind on your cheeks or smell seawater,

陽光在水面上閃閃發光， 可能還會聽到海鷗從上面飛過，

with advanced hardware.

你甚至可以用先進的硬體

Now, if you were to spend a lot of time here,

感受到風吹撫過臉頰或 聞到海水的味道。

you might want a dashboard

如果你要花很多時間在這裡，

where you could do things like make a phone call,

你可能會想要一個儀表板，

move to a new virtual world,

讓你可以做一些事， 比如打電話、

check your bank account.

移動到一個新的虛擬世界、

Now, while this is what you would look like in virtual reality,

查看銀行帳戶。

this is what an em would look like in virtual reality.

這是你在虛擬實境中 看起來的樣子，

It's computer hardware sitting in a server rack somewhere.

這是仿真腦在虛擬實境中 看起來的樣子。

But still, it could see and experience the same thing.

它是電腦硬體， 就放在某個伺服器的機架上。

But some things are different for ems.

但它仍然能夠 看見和體驗同樣的東西。

First, while you'll probably always notice that virtual reality isn't entirely real,

但對仿真腦來說， 有些東西是不同的。

to an em, it can feel as real to them as this room feels to you now

第一，你可能會留意到 虛擬實境並不全是真實的，

or as anything ever feels.

但仿真腦感受到的真實度， 就像你現在對這空間

And ems also have some more action possibilities.

或對任何其它東西 感受到的真實度一樣。

For example, your mind just always runs at the same speed,

仿真腦還有更多的行為可能性。

but an em can add more or less computer hardware to run faster or slower,

比如，你的大腦 總是以同樣速度運作，

and therefore, if the world around them seems to be going too fast,

但仿真腦可以透過增加或減少一些 電腦硬體來運作地更快或更慢，

they can just speed up their mind,

因此，如果它們覺得周圍的世界 發展地似乎太快了，

and the world around them would seem to slow down.

它們可以加快思考的速度，

In addition, an em can make a copy of itself at that moment.

這樣它們周遭的世界 似乎就會慢了下來。

This copy would remember everything the same,

此外，仿真腦可以在 那個時候複製它自己。

and if it starts out with the same speed, looking at the same speed,

這個複製品會記住所有的東西，

it might even need to be told, "You are the copy."

如果它用同樣的速度開始運作， 以同樣的速度觀察世界，

And em could make archive copies,

它們甚至可能需要被告知： 「你只是個複製品。」

and with enough archives,

仿真腦可以做歸檔副本，

an em can be immortal --

一旦有足夠的歸檔副本，

in principle, though not usually in practice.

仿真腦就可以永生——

And an em can move its brain, the computer that represents its brain,

那是理論上，雖然實務上不常見。

from one physical location to another.

仿真腦可以移動它的腦， 也就是代表它的腦的那台電腦，

Ems can actually move around the world at the speed of light,

從一個實體地點移到另一個地點。

and by moving to a new location,

仿真腦可以用光速在全世界移動，

they can interact more quickly with ems near that new location.

藉由移動到一個新地點，

So far, I've been talking about what ems can do.

它可以更快速地與新地點 附近的仿真腦做互動。

What do ems choose to do?

目前，我說的都是仿真腦能做什麼。

To understand that, we'll need to understand three key facts.

那仿真腦會選擇做什麼？

First, ems by definition do what the human they emulate would do

要了解這一點，我們得要 先了解三項關鍵事實。

in the same situation.

第一，定義上， 仿真腦在相同情境下，

So their lives and behavior are very human.

會模仿出人類會做的事。

They're mainly different because they're living in a different world.

所以它們的生活和行為 都非常像人類。

Second, ems need real resources to survive.

它們與人主要的差異在於 它們活在不同的世界裡。

You need food and shelter or you'll die.

第二，仿真腦需要真實資源來存活。

Also, ems need computer hardware, energy, cooling, or they can't exist.

你需要食物和住所，不然就會死亡。

For every subjective minute that an em experiences,

同樣的，仿真腦需要電腦硬體、 能源、冷卻，不然它們將不能生存。

someone, usually that em, had to work to pay for it.

仿真腦所經歷的主觀每一分鐘，

Third, ems are poor.

都是得有人付出才能得到的， 而付出的那個人通常是仿真腦本身。

(Laughter)

第三，仿真腦很窮。

The em population can grow quicker than the em economy,

（笑聲）

so that means wages fall down to em subsistence levels.

仿真腦數量的成長會快過 仿真腦的經濟成長，

That means ems have to be working most of the time.

那意味著薪資水平會落到 仿真腦能維持生計的水平。

So that means this is what ems usually see:

那意味著仿真腦 大部份的時間得要工作。

beautiful and luxurious, but desks --

那意味著仿真腦通常看見的是這個：

they're working most of the time.

美麗又奢華的，辦公桌──

Now, a subsistence wage scenario, you might think, is exotic and strange,

它們大部份的時間都在工作。

but it's actually the usual case in human history,

你可能會認為，這種勉強糊口的 工資的情景是奇異並奇怪的，

and it's how pretty much all wild animals have ever lived,

但這是人類歷史上常見的情況，

so we know what humans do in this situation.

這基本是所有野生動物的生活方式，

Humans basically do what it takes to survive,

所以我們知道人類 在這種情況下會做什麼。

and this is what lets me say so much about the em world.

人類基本上會為了生存而行動，

When creatures are rich, like you,

這就是讓我說了這麽多 關於仿真腦世界的原因。

you have to know a lot about what they want

當生物富有時，就像你們一樣，

to figure out what they do.

你得先要知道他們想要什麼，

When creatures are poor,

才能猜出他們會做什麼。

you know that they mostly do what it takes to survive.

當生物很貧窮時，

So we've been talking about the em world from the point of view of the ems --

你知道的，他們會做的 就只是生存下來。

now, let's step back and look at their whole world.

我們已經從仿真腦的觀點 談了仿真腦的世界。

First, the em world grows much faster than ours,

現在我們退後一步， 來看看它們的整個世界。

roughly a hundred times faster.

第一，仿真腦世界的成長速度 比我們的世界快很多，

So the amount of change we would experience in a century or two,

大約快上一百倍。

they would experience in a year or two.

我們經歷了一、兩個世紀的改變，

And I'm not really willing to project this age much beyond that,

它們在一、兩年內就能完成。

because plausibly by then something else will happen, I don't know what.

我其實不想預測更遙遠的未來，

Second, the typical emulation runs even faster,

因為，理論上到那時就會有 其它事發生，我們無法知道。

roughly a thousand times human speed.

第二，典型的仿真腦運作得更快，

So for them, they experience thousands of years in this year or two,

大約是人類的一千倍。

and for them, the world around them is actually changing more slowly

所以，對它們而言，它們會在 這一、兩年就經歷我們的數千年，

than your world seems to change for you.

相對於你對你的周圍世界的感覺，

Third, ems are crammed together in a small number of very dense cities.

它們會覺得它們周圍的 世界改變很緩慢。

This is not only how they see themselves in virtual reality,

第三，仿真腦會擠在 少數幾個高密度城市中。

it's also how they actually are physically crammed together.

這不只是它們在虛擬實境中 看到的自身狀況，

So at em speeds, physical travel feels really painfully slow,

它們的實體也確實是 擠在同一個地方的。

so most em cities are self-sufficient,

以仿真腦速度來看， 實體旅行是要命的緩慢，

most war is cyber war,

所以大部份的仿真腦城市 是自給自足的，

and most of the rest of the earth away from the em cities

大部份的戰爭是網路戰爭，

is left to the humans, because the ems really aren't that interested in it.

而地球上遠離 仿真腦城市的其它地方，

Speaking of humans,

就留給人類，因為仿真腦 對這些地方真的不感興趣。

you were wanting to hear about that.

說到人類，

Humans must retire, at once, for good.

你會想要聽到這個。

They just can't compete.

人類必須要退休， 一旦發生了，你就好好退休吧。

Now, humans start out owning all of the capital in this world.

人類無法和仿真腦競爭。

The economy grows very fast, their wealth grows very fast.

一開始，人類擁有 這個世界上所有的資本。

Humans get rich, collectively.

經濟成長非常快， 他們的財富也成長非常快。

As you may know, most humans today don't actually own that much

整體而言，人類變富有了。

besides their ability to work,

如你所知，現今大部份人 除了擁有工作能力外，

so between now and then, they need to acquire sufficient assets,

其它其實擁有的並不多。

insurance or sharing arrangements,

所以從現在起到那時後， 他們需要獲得足夠的資產、

or they may starve.

準備保險或股權的配置，

I highly recommend avoiding this outcome.

不然他們就要挨餓了。

(Laughter)

我強烈建議要避免這個結果。

Now, you might wonder,

（笑聲）

why would ems let humans exist?

你們可能會好奇，

Why not kill them, take their stuff?

為什麼仿真腦要讓人類存在？

But notice we have many unproductive retirees around us today,

為什麼不殺了人類， 拿走他們的東西？

and we don't kill them and take their stuff.

別忘了，現今我們身邊就有很多 沒有生產力的退休者，

(Laughter)

我們不會殺了他們，拿他們的東西。

In part, that's because it would disrupt the institutions we share with them.

（笑聲）

Other groups would wonder who's next,

部份原因是因為這樣會使得 我們與他們共同的制度被瓦解。

so plausibly, ems may well let humans retire in peace during the age of em.

其他族群就會想，誰會下一個遭殃？

You should worry more that the age of em only lasts a year or two

所以仿真腦比較合理的做法，是讓 人類在仿真腦的時代能平靜地退休。

and you don't know what happens next.

你們比較該擔心的是， 仿真腦的時代可能只有一、兩年，

Ems are very much like humans,

而你卻不知道接下來會發生什麼事。

but they are not like the typical human.

仿真腦和人類很像，

The typical em is a copy of the few hundred most productive humans.

但它們不像典型的人類。

So in fact, they are as elite, compared to the typical human,

典型的仿真腦是 幾百個最有生產力的人的翻版。

as the typical billionaire, Nobel Prize winner,

所以事實上， 與一般人相比它們是菁英，

Olympic gold medalist, head of state.

它們是典型的億萬富翁、 諾貝爾獎得主、

Ems look on humans perhaps with nostalgia and gratitude,

奧運金牌得主、國家元首。

but not so much respect,

仿真腦在看人類時，可能是帶著 懷舊之情和感激之情的，

which is, if you think about it, how you think about your ancestors.

但不會帶有很多尊重，

We know many things about how humans differ in terms of productivity.

就像你仔細想想， 你是怎麽看待你的祖先的。

We can just use those to predict features of ems --

我們知道許多關於 不同個體在生產力方面的不同之處。

for example, they tend to be smart, conscientious, hard-working,

我們可以用那些知識 來預測仿真腦的特徵──

married, religious, middle-aged.

比如，它們往往是 聰明的、認真的、努力的、

These are features of ems.

已婚的、有宗教信仰的、中年的。

Em world also contains enormous variety.

這些是仿真腦的特點。

Not only does it continue on with most of the kinds of variety that humans do,

仿真腦世界也有很高的多樣性。

including variety of industry and profession,

它不只延續了人類的各種多樣性，

they also have many new kinds of variety,

包括產業及職業的多樣性，

and one of the most important is mind speed.

它們也在很多新型特點上 體現了多樣性，

Ems can plausibly go from human speed

最重要的之一就是思維速度。

up to a million times faster than human speed,

仿真腦的速度， 慢可以慢到與人類一樣的速度

and down to a billion times slower than human speed.

快可以快到比人類快 幾百萬倍的速度，

Faster ems tend to have markers of high status.

往下還可以比人類慢 十億倍的速度。

They embody more wealth. They win arguments.

更快的仿真腦往往具有較高的地位。

They sit at premium locations.

它們有比較多的財富。 它們能在爭論中勝出。

Slower ems are mostly retirees,

它們位於金字塔頂端。

and they are like the ghosts of our literature.

比較慢的仿真腦多半是已退休的，

If you recall, ghosts are all around us --

它們就像是我們小說中看到的鬼魂。

you can interact with them if you pay the price.

如果你有注意到， 書中的鬼魂都在我們周圍，

But they don't know much, they can't influence much,

如果你願意付出代價， 就可以與它們互動。

and they're obsessed with the past, so what's the point?

但它們知道的不多， 它們的影響不大，

(Laughter)

且它們沉陷在過去中， 但那有什麼意義？

Ems also have more variety in the structure of their lives.

（笑聲）

This is your life: you start and you end, really simple.

仿真腦的人生結構 也有較高的多樣性。

This is the life of an em,

這是你的人生： 開始、結束就這麼簡單。

who every day splits off some short-term copies

這是仿真腦的人生，

to do short-term tasks and then end.

每天都會分裂出一些短暫的複製品，

We'll talk more about those short term versions in a moment,

用來完成短期任務，然後結束。

but they are much more efficient

我們稍後再討論這些短期版本，

because they don't have to rest for the next day.

但是它們的效率更高，

This em is more opportunistic.

因為它們不需要在第二天休息。

They make more copies of themselves when there's more demand for that.

這個仿真腦是比較機會主義者的。

They don't know which way the future's going.

當有需要時，它們就會 把自己複製更多份。

This is an em designer,

它們不知道未來的走向。

who conceives of a large system

這是一個仿真腦設計師，

and then breaks recursively into copies who elaborate that,

它構想了一個很大的系統，

so ems can implement larger, more coherent designs.

然後再遞歸地分裂成複製品 來詳細說明它的構想，

This an emulation plumber

因此，仿真腦可以實現更大、 更連貫的設計。

who remembers that every day, for the last 20 years,

這是個仿真腦世界裡的水管工，

they only ever worked two hours a day, a life of leisure.

它記得過去二十年的每一天，

But what really happened is, every day they had a thousand copies,

它們一天只工作兩小時， 很悠閒的人生。

each of whom did a two-hour plumbing job,

但實情是，它們每天都得 複製出數千個複製品，

and only one of them went on to the next day.

每個複製品得做兩個小時的工作，

Objectively, they're working well over 99 percent of the time.

它們之中只有一個 第二天會接著工作。

Subjectively, they remember a life of leisure.

客觀上，它們 99% 的時間 都在努力工作。

(Laughter)

主觀上，它們還記得 甚麼是悠閑的生活。