What would happen if we upload our brains to computers? | Robin Hanson

譯者: Lilian Chiu 審譯者: 易帆 余

Someday, we may have robots as smart as people,

總有一天，機器人 會像人類一樣聰明，

artificial intelligence, AI.

人工智慧，AI。

How could that happen?

那要如何做到呢？

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

其中一個方式就是， 我們只要繼續累積更好的軟體，

like we've been doing for 70 years.

就像我們過去七十年所做的。

At past rates of progress, that may take centuries.

但以過去的進步速率推斷， 這樣要花上數百年。

Some say it'll happen a lot faster

有人說，它會提早實現，

as we discover grand new powerful theories of intelligence.

因為我們發現了 全新的強大智慧理論。

I'm skeptical.

我持懷疑態度。

But a third scenario

但有第三種情況，

is what I'm going to talk about today.

也是我今天要講的。

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

這個想法是從人腦中移植出軟體。

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

要做到這一點，我們需要 讓三種技術的發展足夠成熟，

and none of them are there yet.

但目前我們這些都還沒達到。

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

第一，我們需要許多 廉價、快速、平行的電腦。

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

第二，我們需要在適當的 空間與正確的化學成分下

in fine spatial and chemical detail,

掃瞄我們的大腦。

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

這樣才能看出細胞的 精確位置、連結、類型。

And third, we're going to need computer models

第三，我們需要一種電腦模型，

of how each kind of brain cell works --

關於各類腦細胞如何運作的模型──

taking input signals, changing interval state

可以讓我們輸入訊號， 改變區間的狀態，

and sending output signals.

並再發送出訊號的模型。

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

如果我們有夠好的各種腦細胞模型、

and a good enough model of the brain,

夠好的大腦模型，

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

我們可以把它們結合起來， 做出一個完整的大腦模型，

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

那個模型與原本大腦所想要的 輸入和輸出行為一致。

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

所以，如果你跟它說話， 它可能會回話。

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

如果你要它做事，它可能會去做。

And if we could do that, everything would change.

如果我們能做到這樣， 一切都會改變。

People have been talking about this idea for decades,

幾十年來人們以「上傳」的方式

under the name of "uploads."

談論這個想法。

I'm going to call them "ems."

我打算稱它為「仿真腦（ems）」。

When they talk about it, they say,

當他們在談論這個想法時，他們說：

"Is this even possible?

「這是可能的嗎？」

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

如果能做出來，它會是有意識的嗎？ 還是只是個空機器？

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

如果你仿了我的大腦， 它會是我還是另一個人？

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

我將忽略這些令人著迷的問題…...

(Laughter)

（笑聲）

because I see a neglected question:

因為我看到了一個被忽視的問題：

What would actually happen?

究竟會發生什麽事？

I became obsessed with this question.

我對這個問題很著迷。

I spent four years trying to analyze it,

我花了四年時間來分析它，

using standard academic tools, to guess what would happen,

用標準的學術工具， 來猜測可能會發生什麼事，

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

我來這裡是要告訴各位我的發現。

But be warned --

但，請注意──

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

我不是要提供靈感， 我是要提供分析。

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

我認為我的工作是 告訴你們最有可能發生的情況，

if we did the least to avoid it.

如果我們不花心思去預防，

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

如果你對於我今天告訴你的事 並不感到擔心，

you're just not paying attention.

那表示你沒在注意聽。

(Laughter)

（笑聲）

OK, the first thing I can tell you

好，我能告訴各位的第一件事是，

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

仿真腦的大部分時間 都是在虛擬現實生活中度過的。

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

如果你使用虛擬實境， 這就是你的樣子。

And this is what you might see:

這也是你可能會看到的：

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

陽光在水面上閃閃發光， 可能還會聽到海鷗從上面飛過，

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

你甚至可以用先進的硬體

with advanced hardware.

感受到風吹撫過臉頰或 聞到海水的味道。

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

如果你要花很多時間在這裡，

you might want a dashboard

你可能會想要一個儀表板，

where you could do things like make a phone call,

讓你可以做一些事， 比如打電話、

move to a new virtual world,

移動到一個新的虛擬世界、

check your bank account.

查看銀行帳戶。

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

這是你在虛擬實境中 看起來的樣子，

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

這是仿真腦在虛擬實境中 看起來的樣子。

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

它是電腦硬體， 就放在某個伺服器的機架上。

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

但它仍然能夠 看見和體驗同樣的東西。

But some things are different for ems.

但對仿真腦來說， 有些東西是不同的。

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

第一，你可能會留意到 虛擬實境並不全是真實的，

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

但仿真腦感受到的真實度， 就像你現在對這空間

or as anything ever feels.

或對任何其它東西 感受到的真實度一樣。

And ems also have some more action possibilities.

仿真腦還有更多的行為可能性。

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

比如，你的大腦 總是以同樣速度運作，

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

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，

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

因此，如果它們覺得周圍的世界 發展地似乎太快了，

they can just speed up their mind,

它們可以加快思考的速度，

and the world around them would seem to slow down.

這樣它們周遭的世界 似乎就會慢了下來。

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

此外，仿真腦可以在 那個時候複製它自己。

This copy would remember everything the same,

這個複製品會記住所有的東西，

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

如果它用同樣的速度開始運作， 以同樣的速度觀察世界，

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

它們甚至可能需要被告知： 「你只是個複製品。」

And em could make archive copies,

仿真腦可以做歸檔副本，

and with enough archives,

一旦有足夠的歸檔副本，

an em can be immortal --

仿真腦就可以永生——

in principle, though not usually in practice.

那是理論上，雖然實務上不常見。

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

仿真腦可以移動它的腦， 也就是代表它的腦的那台電腦，

from one physical location to another.

從一個實體地點移到另一個地點。

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

仿真腦可以用光速在全世界移動，

and by moving to a new location,

藉由移動到一個新地點，

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

它可以更快速地與新地點 附近的仿真腦做互動。

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

目前，我說的都是仿真腦能做什麼。

What do ems choose to do?

那仿真腦會選擇做什麼？

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

要了解這一點，我們得要 先了解三項關鍵事實。

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

第一，定義上， 仿真腦在相同情境下，

會模仿出人類會做的事。

in the same situation.

So their lives and behavior are very human.

所以它們的生活和行為 都非常像人類。

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

它們與人主要的差異在於 它們活在不同的世界裡。

Second, ems need real resources to survive.

第二，仿真腦需要真實資源來存活。

You need food and shelter or you'll die.

你需要食物和住所，不然就會死亡。

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

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。

For every subjective minute that an em experiences,

仿真腦所經歷的主觀每一分鐘，

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

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。

Third, ems are poor.

第三，仿真腦很窮。

(Laughter)

（笑聲）

The em population can grow quicker than the em economy,

仿真腦數量的成長會快過 仿真腦的經濟成長，

so that means wages fall down to em subsistence levels.

那意味著薪資水平會落到 仿真腦能維持生計的水平。

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

那意味著仿真腦 大部份的時間得要工作。

So that means this is what ems usually see:

那意味著仿真腦通常看見的是這個：

beautiful and luxurious, but desks --

美麗又奢華的，辦公桌──

they're working most of the time.

它們大部份的時間都在工作。

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

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，

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

但這是人類歷史上常見的情況，

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

這基本是所有野生動物的生活方式，

so we know what humans do in this situation.

所以我們知道人類 在這種情況下會做什麼。

Humans basically do what it takes to survive,

人類基本上會為了生存而行動，

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

這就是讓我說了這麽多 關於仿真腦世界的原因。

When creatures are rich, like you,

當生物富有時，就像你們一樣，

you have to know a lot about what they want

你得先要知道他們想要什麼，

to figure out what they do.

才能猜出他們會做什麼。

When creatures are poor,

當生物很貧窮時，

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

你知道的，他們會做的 就只是生存下來。

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

我們已經從仿真腦的觀點 談了仿真腦的世界。

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

現在我們退後一步， 來看看它們的整個世界。

First, the em world grows much faster than ours,

第一，仿真腦世界的成長速度 比我們的世界快很多，

roughly a hundred times faster.

大約快上一百倍。

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

我們經歷了一、兩個世紀的改變，

they would experience in a year or two.

它們在一、兩年內就能完成。

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

我其實不想預測更遙遠的未來，

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

因為，理論上到那時就會有 其它事發生，我們無法知道。

Second, the typical emulation runs even faster,

第二，典型的仿真腦運作得更快，

roughly a thousand times human speed.

大約是人類的一千倍。

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

所以，對它們而言，它們會在 這一、兩年就經歷我們的數千年，

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

相對於你對你的周圍世界的感覺，

than your world seems to change for you.

它們會覺得它們周圍的 世界改變很緩慢。

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

第三，仿真腦會擠在 少數幾個高密度城市中。

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

這不只是它們在虛擬實境中 看到的自身狀況，

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

它們的實體也確實是 擠在同一個地方的。

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

以仿真腦速度來看， 實體旅行是要命的緩慢，

so most em cities are self-sufficient,

所以大部份的仿真腦城市 是自給自足的，

most war is cyber war,

大部份的戰爭是網路戰爭，

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

而地球上遠離 仿真腦城市的其它地方，

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

就留給人類，因為仿真腦 對這些地方真的不感興趣。

Speaking of humans,

說到人類，

you were wanting to hear about that.

你會想要聽到這個。

Humans must retire, at once, for good.

人類必須要退休， 一旦發生了，你就好好退休吧。

They just can't compete.

人類無法和仿真腦競爭。

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

一開始，人類擁有 這個世界上所有的資本。

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

經濟成長非常快， 他們的財富也成長非常快。

Humans get rich, collectively.

整體而言，人類變富有了。

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

如你所知，現今大部份人 除了擁有工作能力外，

besides their ability to work,

其它其實擁有的並不多。

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

所以從現在起到那時後， 他們需要獲得足夠的資產、

insurance or sharing arrangements,

準備保險或股權的配置，

or they may starve.

不然他們就要挨餓了。

I highly recommend avoiding this outcome.

我強烈建議要避免這個結果。

(Laughter)

（笑聲）

Now, you might wonder,

你們可能會好奇，

why would ems let humans exist?

為什麼仿真腦要讓人類存在？

Why not kill them, take their stuff?

為什麼不殺了人類， 拿走他們的東西？

But notice we have many unproductive retirees around us today,

別忘了，現今我們身邊就有很多 沒有生產力的退休者，

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

我們不會殺了他們，拿他們的東西。

(Laughter)

（笑聲）

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

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。

Other groups would wonder who's next,

其他族群就會想，誰會下一個遭殃？

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

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。

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

你們比較該擔心的是， 仿真腦的時代可能只有一、兩年，

and you don't know what happens next.

而你卻不知道接下來會發生什麼事。

Ems are very much like humans,

仿真腦和人類很像，

but they are not like the typical human.

但它們不像典型的人類。

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

典型的仿真腦是 幾百個最有生產力的人的翻版。

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

所以事實上， 與一般人相比它們是菁英，

as the typical billionaire, Nobel Prize winner,

它們是典型的億萬富翁、 諾貝爾獎得主、

Olympic gold medalist, head of state.

奧運金牌得主、國家元首。

Ems look on humans perhaps with nostalgia and gratitude,

仿真腦在看人類時，可能是帶著 懷舊之情和感激之情的，

but not so much respect,

但不會帶有很多尊重，

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

就像你仔細想想， 你是怎麽看待你的祖先的。

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

我們知道許多關於 不同個體在生產力方面的不同之處。

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

我們可以用那些知識 來預測仿真腦的特徵──

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

比如，它們往往是 聰明的、認真的、努力的、

married, religious, middle-aged.

已婚的、有宗教信仰的、中年的。

These are features of ems.

這些是仿真腦的特點。

Em world also contains enormous variety.

仿真腦世界也有很高的多樣性。

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

它不只延續了人類的各種多樣性，

including variety of industry and profession,

包括產業及職業的多樣性，

they also have many new kinds of variety,

它們也在很多新型特點上 體現了多樣性，

and one of the most important is mind speed.

最重要的之一就是思維速度。

Ems can plausibly go from human speed

仿真腦的速度， 慢可以慢到與人類一樣的速度

up to a million times faster than human speed,

快可以快到比人類快 幾百萬倍的速度，

and down to a billion times slower than human speed.

往下還可以比人類慢 十億倍的速度。

Faster ems tend to have markers of high status.

更快的仿真腦往往具有較高的地位。

They embody more wealth. They win arguments.

它們有比較多的財富。 它們能在爭論中勝出。

They sit at premium locations.

它們位於金字塔頂端。

Slower ems are mostly retirees,

比較慢的仿真腦多半是已退休的，

and they are like the ghosts of our literature.

它們就像是我們小說中看到的鬼魂。

If you recall, ghosts are all around us --

如果你有注意到， 書中的鬼魂都在我們周圍，

you can interact with them if you pay the price.

如果你願意付出代價， 就可以與它們互動。

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

但它們知道的不多， 它們的影響不大，

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

且它們沉陷在過去中， 但那有什麼意義？

(Laughter)

（笑聲）

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

仿真腦的人生結構 也有較高的多樣性。

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

這是你的人生： 開始、結束就這麼簡單。

This is the life of an em,

這是仿真腦的人生，

who every day splits off some short-term copies

每天都會分裂出一些短暫的複製品，

to do short-term tasks and then end.

用來完成短期任務，然後結束。

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

我們稍後再討論這些短期版本，

but they are much more efficient

但是它們的效率更高，

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

因為它們不需要在第二天休息。

This em is more opportunistic.

這個仿真腦是比較機會主義者的。

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

當有需要時，它們就會 把自己複製更多份。

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

它們不知道未來的走向。

This is an em designer,

這是一個仿真腦設計師，

who conceives of a large system

它構想了一個很大的系統，

and then breaks recursively into copies who elaborate that,

然後再遞歸地分裂成複製品 來詳細說明它的構想，

so ems can implement larger, more coherent designs.

因此，仿真腦可以實現更大、 更連貫的設計。

This an emulation plumber

這是個仿真腦世界裡的水管工，

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

它記得過去二十年的每一天，

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

它們一天只工作兩小時， 很悠閒的人生。

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

但實情是，它們每天都得 複製出數千個複製品，

each of whom did a two-hour plumbing job,

每個複製品得做兩個小時的工作，

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

它們之中只有一個 第二天會接著工作。

客觀上，它們 99% 的時間 都在努力工作。

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

主觀上，它們還記得 甚麼是悠閑的生活。

Subjectively, they remember a life of leisure.

(Laughter)

（笑聲）

This, again, is you. You start and you end.

同樣的，這還是你。 你開始，你結束。

This could be you if at the start of party,

這可能是你，如果在派對開始時，

you took a drug that meant you would not remember that party

你吃了顆藥， 它會讓你在那天之後再也

ever after that day.

想不起那場派對。

Some people do this, I'm told.

據我所知，有些人會這樣做。

Toward the end of the party,

到派對尾聲時，

will you say to yourself, "I'm about to die, this is terrible.

我要問你的是，你是否會告訴自己： 「我快要死了，這太可怕了。

That person tomorrow isn't me, because they won't remember what I do."

明天的那個我其實不是我， 因為他不會記得我做過的事。」

Or you could say, "I will go on tomorrow. I just won't remember what I did."

或者說：「明天，我還是我， 只是不記得我過去做了什麽。」

This is an em who splits off a short-term copy

這是分裂出短期複製品的仿真腦，

to do a short-term task and then end.

用來做短期工作，然後就結束。

They have the same two attitude possibilities.

它們同樣也有這兩種可能的態度。

They can say, "I'm a new short-term creature with a short life. I hate this."

它們會說：「我是新的短期生物， 生命很短。我討厭這樣。」

或「我是更大生物的一部份， 只是記不得是哪一部份。」

Or "I'm a part of a larger creature who won't remember this part."

I predict they'll have that second attitude,

我預測它們會採用第二種態度，

不是因為哲學正確性，而是因為 那種態度能幫助它們相處。

not because it's philosophically correct, but because it helps them get along.

Today, if the president says we must invade Iraq,

如果今天總統說 我們必須要侵略伊拉克，

and you say, "Why?"

你說：「為什麼？」

and they say, "State secret,"

他們說：「國家機密。」

you're not sure if you can trust them,

你不確定是否可以信任他們，

但對仿真腦而言，總統的複製品和 你的複製品可以進入一個保險箱，

but for ems, a copy of the president and a copy of you can go inside a safe,

explain all their secret reasons,

解釋它們所有的秘密原因，

and then one bit comes out from your copy to yourself,

然後有一個位元訊號 從你的複製品回來告知你，

telling you if you were convinced.

讓你知道你是否有被說服。

So now you can know there is a good reason.

這樣你就會知道有個好理由存在。

I know you guys are all eager to evaluate this world.

我知道你們都很渴望評估那個世界。

You're eager to decide if you love it or hate it.

你想盡快決定 你是否喜歡它或討厭它。

But think: your ancestors from thousands of years ago

但想想：你數千年前的祖先

would have loved or hated your world

會根據他們聽聞的最初的幾件事，

based on the first few things they heard about it,

來選擇愛或恨你現在這個世界嗎？

因為你的世界聽起來是很奇怪的。

because your world is really just weird.

So before judging a strange future world, you should really learn a lot about it,

所以在評斷一個奇怪的 未來世界之前，

你應該好好學習它，

maybe read a whole book about it,

也許讀一本關於它的書，

and then, if you don't like it, work to change it.

然後，如果你不喜歡它， 就努力去改變它。

Thank you.

謝謝。

(Applause)

（掌聲）