ABOUT THE SPEAKER
Robin Hanson - Futurist, social scientist
Does humanity have a future as uploaded minds? In his work, Robin Hanson asks this and other extra-large questions.

Why you should listen

In his book, The Age of Em: Work, Love and Life When Robots Rule the Earth, Robin Hanson re-imagines humanity's role as our tech becomes smarter. A pioneer in prediction markets, also known as information markets and idea futures, Hanson has been known since the 1980s for taking the very very long view on topics as varied as (a selected list) spatial product competition, health incentive contracts, group insurance, product bans, evolutionary psychology and bioethics of health care, voter information incentives, incentives to fake expertise, Bayesian classification, agreeing to disagree, self-deception in disagreement, probability elicitation, wiretaps, image reconstruction, the history of science prizes, reversible computation, the origin of life, the survival of humanity, very long term economic growth, growth given machine intelligence and interstellar colonization.

Meanwhile, he has developed new technologies for conditional, combinatorial and intermediated trading, and he studied insider trading, manipulation and other foul play. Hanson is associate professor of economics at George Mason University and a research associate at the Future of Humanity Institute of Oxford University. His next book is The Elephant in the Brain, co-authored with Kevin Simler, due in 2018.

More profile about the speaker
Robin Hanson | Speaker | TED.com
TED2017

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

羅賓韓森: 如果我們把我們的大腦上傳到電腦上,會如何?

Filmed:
1,430,160 views

認識一下「仿真腦」──模仿人類大腦的機器,它們可以像它們複製的大腦一樣思考、感覺和運作。未來學家和社會科學家羅賓韓森描述了一個可能的未來,仿真腦接管了全球經濟,在超快速電腦上運作,並進行自我複製完成多任務處理,只留給人類一個選擇:永久退休。來聽聽韓森所描述的那不可思議的未來,如果機器人統治地球時將會發生的事情。
- Futurist, social scientist
Does humanity have a future as uploaded minds? In his work, Robin Hanson asks this and other extra-large questions. Full bio

Double-click the English transcript below to play the video.

00:13
Someday日後, we may可能 have robots機器人
as smart聰明 as people,
0
1521
4444
總有一天,機器人
會像人類一樣聰明,
00:17
artificial人造 intelligence情報, AIAI.
1
5989
2474
人工智慧,AI。
00:20
How could that happen發生?
2
8487
1476
那要如何做到呢?
00:22
One route路線 is that we'll just keep
accumulating積累 better software軟件,
3
10812
3184
其中一個方式就是,
我們只要繼續累積更好的軟體,
00:26
like we've我們已經 been doing for 70 years年份.
4
14020
1992
就像我們過去七十年所做的。
00:28
At past過去 rates利率 of progress進展,
that may可能 take centuries百年.
5
16543
2936
但以過去的進步速率推斷,
這樣要花上數百年。
00:32
Some say it'll它會 happen發生 a lot faster更快
6
20138
2350
有人說,它會提早實現,
00:34
as we discover發現 grand盛大 new
powerful強大 theories理論 of intelligence情報.
7
22512
3491
因為我們發現了
全新的強大智慧理論。
00:38
I'm skeptical懷疑的.
8
26407
1199
我持懷疑態度。
00:40
But a third第三 scenario腳本
9
28608
3061
但有第三種情況,
00:43
is what I'm going to talk about today今天.
10
31693
1827
也是我今天要講的。
00:45
The idea理念 is to port港口 the software軟件
from the human人的 brain.
11
33544
2738
這個想法是從人腦中移植出軟體。
00:48
To do this, we're going to need
three technologies技術 to be good enough足夠,
12
36972
3301
要做到這一點,我們需要
讓三種技術的發展足夠成熟,
00:52
and none沒有 of them are there yet然而.
13
40297
1555
但目前我們這些都還沒達到。
00:54
First, we're going to need
lots of cheap低廉, fast快速, parallel平行 computers電腦.
14
42417
4714
第一,我們需要許多
廉價、快速、平行的電腦。
01:00
Second第二, we're going to need
to scan掃描 individual個人 human人的 brains大腦
15
48855
4207
第二,我們需要在適當的
空間與正確的化學成分下
01:05
in fine spatial空間的 and chemical化學 detail詳情,
16
53086
2238
掃瞄我們的大腦。
01:07
to see exactly究竟 what cells細胞 are where,
connected連接的 to what, of what type類型.
17
55348
4028
這樣才能看出細胞的
精確位置、連結、類型。
01:11
And third第三, we're going to need
computer電腦 models楷模
18
59400
4356
第三,我們需要一種電腦模型,
01:15
of how each kind of brain cell細胞 works作品 --
19
63780
2222
關於各類腦細胞如何運作的模型──
01:19
taking服用 input輸入 signals信號,
changing改變 interval間隔 state
20
67002
2580
可以讓我們輸入訊號,
改變區間的狀態,
01:21
and sending發出 output產量 signals信號.
21
69606
1294
並再發送出訊號的模型。
01:22
If we have good enough足夠 models楷模
of all the kinds of brain cells細胞
22
70924
3818
如果我們有夠好的各種腦細胞模型、
01:26
and a good enough足夠 model模型 of the brain,
23
74766
1763
夠好的大腦模型,
01:28
we can put it together一起 to make
a good enough足夠 model模型 of an entire整個 brain,
24
76553
3434
我們可以把它們結合起來,
做出一個完整的大腦模型,
01:32
and that model模型 would have the same相同
input-output輸入輸出 behavior行為 as the original原版的.
25
80011
4003
那個模型與原本大腦所想要的
輸入和輸出行為一致。
所以,如果你跟它說話,
它可能會回話。
01:36
So if you talk to it, it might威力 talk back.
26
84038
2485
01:38
If you ask it to do things,
it might威力 do them.
27
86547
2097
如果你要它做事,它可能會去做。
01:40
And if we could do that,
everything would change更改.
28
88668
2822
如果我們能做到這樣,
一切都會改變。
01:43
People have been talking
about this idea理念 for decades幾十年,
29
91514
2568
幾十年來人們以「上傳」的方式
01:46
under the name名稱 of "uploads上傳."
30
94106
1789
談論這個想法。
01:47
I'm going to call them "emsEMS."
31
95919
1484
我打算稱它為「仿真腦(ems)」。
01:50
When they talk about it, they say,
32
98355
2604
當他們在談論這個想法時,他們說:
01:52
"Is this even possible可能?
33
100983
1544
「這是可能的嗎?」
01:54
If you made製作 one, would it be conscious意識?
Or is it just an empty machine?
34
102838
3351
如果能做出來,它會是有意識的嗎?
還是只是個空機器?
01:58
If you made製作 one of me,
is that me or someone有人 else其他?"
35
106213
2452
如果你仿了我的大腦,
它會是我還是另一個人?
02:01
These are all fascinating迷人 questions問題
that I'm going to ignore忽視 ...
36
109036
3102
我將忽略這些令人著迷的問題…...
02:04
(Laughter笑聲)
37
112162
2365
(笑聲)
02:07
because I see a neglected被忽視的 question:
38
115041
2089
因為我看到了一個被忽視的問題:
02:09
What would actually其實 happen發生?
39
117154
1610
究竟會發生什麽事?
02:12
I became成為 obsessed痴迷 with this question.
40
120440
1904
我對這個問題很著迷。
02:15
I spent花費 four years年份 trying to analyze分析 it,
41
123993
2953
我花了四年時間來分析它,
02:18
using運用 standard標準 academic學術的 tools工具,
to guess猜測 what would happen發生,
42
126970
3547
用標準的學術工具,
來猜測可能會發生什麼事,
02:22
and I'm here to tell you what I found發現.
43
130541
1849
我來這裡是要告訴各位我的發現。
02:24
But be warned警告 --
44
132928
1382
但,請注意──
02:26
I'm not offering inspiration靈感,
I'm offering analysis分析.
45
134722
3407
我不是要提供靈感,
我是要提供分析。
02:30
I see my job工作 as telling告訴 you
what's most likely容易 to happen發生
46
138153
2794
我認為我的工作是
告訴你們最有可能發生的情況,
02:32
if we did the least最小 to avoid避免 it.
47
140971
1777
如果我們不花心思去預防,
02:36
If you aren't at least最小 a bit disturbed不安
by something I tell you here,
48
144132
3302
如果你對於我今天告訴你的事
並不感到擔心,
02:39
you're just not paying付款 attention注意.
49
147458
1594
那表示你沒在注意聽。
02:41
(Laughter笑聲)
50
149076
1351
(笑聲)
02:42
OK, the first thing I can tell you
51
150451
2596
好,我能告訴各位的第一件事是,
02:45
is that emsEMS spend most
of their life in virtual虛擬 reality現實.
52
153071
3133
仿真腦的大部分時間
都是在虛擬現實生活中度過的。
02:48
This is what you might威力 look like
if you were using運用 virtual虛擬 reality現實.
53
156819
4484
如果你使用虛擬實境,
這就是你的樣子。
02:54
And this is what you might威力 see:
54
162074
2150
這也是你可能會看到的:
02:57
sunlight陽光 glinting閃閃發光 off of water,
you might威力 hear gulls海鷗 flying飛行 above以上,
55
165244
3596
陽光在水面上閃閃發光,
可能還會聽到海鷗從上面飛過,
03:00
you might威力 even feel the wind
on your cheeks臉頰 or smell seawater海水,
56
168864
3031
你甚至可以用先進的硬體
03:03
with advanced高級 hardware硬件.
57
171919
1325
感受到風吹撫過臉頰或
聞到海水的味道。
03:06
Now, if you were to spend
a lot of time here,
58
174082
2097
如果你要花很多時間在這裡,
03:08
you might威力 want a dashboard儀表板
59
176203
1266
你可能會想要一個儀表板,
03:09
where you could do things like
make a phone電話 call,
60
177493
2494
讓你可以做一些事,
比如打電話、
03:12
move移動 to a new virtual虛擬 world世界,
61
180011
1550
移動到一個新的虛擬世界、
03:13
check your bank銀行 account帳戶.
62
181585
1300
查看銀行帳戶。
03:15
Now, while this is what
you would look like in virtual虛擬 reality現實,
63
183960
3657
這是你在虛擬實境中
看起來的樣子,
03:19
this is what an emEM
would look like in virtual虛擬 reality現實.
64
187641
3128
這是仿真腦在虛擬實境中
看起來的樣子。
03:22
It's computer電腦 hardware硬件
sitting坐在 in a server服務器 rack somewhere某處.
65
190793
2756
它是電腦硬體,
就放在某個伺服器的機架上。
03:25
But still, it could see
and experience經驗 the same相同 thing.
66
193573
3158
但它仍然能夠
看見和體驗同樣的東西。
03:29
But some things are different不同 for emsEMS.
67
197282
1830
但對仿真腦來說,
有些東西是不同的。
03:31
First, while you'll你會 probably大概 always notice注意
that virtual虛擬 reality現實 isn't entirely完全 real真實,
68
199690
4786
第一,你可能會留意到
虛擬實境並不全是真實的,
03:36
to an emEM, it can feel as real真實 to them
as this room房間 feels感覺 to you now
69
204500
3254
但仿真腦感受到的真實度,
就像你現在對這空間
03:39
or as anything ever feels感覺.
70
207778
1638
或對任何其它東西
感受到的真實度一樣。
03:41
And emsEMS also have
some more action行動 possibilities可能性.
71
209440
2412
仿真腦還有更多的行為可能性。
03:44
For example, your mind心神 just always
runs運行 at the same相同 speed速度,
72
212389
2716
比如,你的大腦
總是以同樣速度運作,
03:47
but an emEM can add more or less
computer電腦 hardware硬件 to run faster更快 or slower比較慢,
73
215129
3841
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,
03:50
and therefore因此, if the world世界 around them
seems似乎 to be going too fast快速,
74
218994
3688
因此,如果它們覺得周圍的世界
發展地似乎太快了,
它們可以加快思考的速度,
03:54
they can just speed速度 up their mind心神,
75
222706
1619
03:56
and the world世界 around them
would seem似乎 to slow down.
76
224349
2350
這樣它們周遭的世界
似乎就會慢了下來。
03:58
In addition加成, an emEM can make
a copy複製 of itself本身 at that moment時刻.
77
226723
4841
此外,仿真腦可以在
那個時候複製它自己。
04:04
This copy複製 would remember記得
everything the same相同,
78
232120
2097
這個複製品會記住所有的東西,
04:06
and if it starts啟動 out with the same相同 speed速度,
looking at the same相同 speed速度,
79
234241
3316
如果它用同樣的速度開始運作,
以同樣的速度觀察世界,
04:09
it might威力 even need to be told,
"You are the copy複製."
80
237581
2651
它們甚至可能需要被告知:
「你只是個複製品。」
04:12
And emEM could make archive檔案 copies副本,
81
240779
1573
仿真腦可以做歸檔副本,
04:14
and with enough足夠 archives檔案,
82
242376
1824
一旦有足夠的歸檔副本,
04:16
an emEM can be immortal不朽 --
83
244224
2072
仿真腦就可以永生——
04:18
in principle原理, though雖然 not
usually平時 in practice實踐.
84
246320
2190
那是理論上,雖然實務上不常見。
04:21
And an emEM can move移動 its brain,
the computer電腦 that represents代表 its brain,
85
249411
3532
仿真腦可以移動它的腦,
也就是代表它的腦的那台電腦,
04:24
from one physical物理 location位置 to another另一個.
86
252967
2087
從一個實體地點移到另一個地點。
04:28
Ems埃姆斯 can actually其實 move移動 around the world世界
at the speed速度 of light,
87
256037
3318
仿真腦可以用光速在全世界移動,
04:31
and by moving移動 to a new location位置,
88
259379
1532
藉由移動到一個新地點,
04:32
they can interact相互作用 more quickly很快
with emsEMS near that new location位置.
89
260935
3264
它可以更快速地與新地點
附近的仿真腦做互動。
04:36
So far, I've been talking about
what emsEMS can do.
90
264223
4277
目前,我說的都是仿真腦能做什麼。
04:41
What do emsEMS choose選擇 to do?
91
269227
1900
那仿真腦會選擇做什麼?
04:44
To understand理解 that, we'll need
to understand理解 three key facts事實.
92
272405
2977
要了解這一點,我們得要
先了解三項關鍵事實。
04:48
First, emsEMS by definition定義 do what
the human人的 they emulate仿真 would do
93
276096
6151
第一,定義上,
仿真腦在相同情境下,
會模仿出人類會做的事。
04:54
in the same相同 situation情況.
94
282271
1301
04:56
So their lives生活 and behavior行為
are very human人的.
95
284305
2643
所以它們的生活和行為
都非常像人類。
04:58
They're mainly主要 different不同 because
they're living活的 in a different不同 world世界.
96
286972
3297
它們與人主要的差異在於
它們活在不同的世界裡。
05:02
Second第二, emsEMS need
real真實 resources資源 to survive生存.
97
290293
3298
第二,仿真腦需要真實資源來存活。
05:05
You need food餐飲 and shelter庇護 or you'll你會 die.
98
293615
3663
你需要食物和住所,不然就會死亡。
05:09
Also, emsEMS need computer電腦 hardware硬件,
energy能源, cooling冷卻, or they can't exist存在.
99
297302
4310
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。
05:13
For every一切 subjective主觀 minute分鐘
that an emEM experiences經驗,
100
301636
2575
仿真腦所經歷的主觀每一分鐘,
05:16
someone有人, usually平時 that emEM,
had to work to pay工資 for it.
101
304235
3274
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。
05:20
Third第三, emsEMS are poor較差的.
102
308136
1690
第三,仿真腦很窮。
05:21
(Laughter笑聲)
103
309850
1150
(笑聲)
05:23
The emEM population人口 can grow增長
quicker更快 than the emEM economy經濟,
104
311715
2635
仿真腦數量的成長會快過
仿真腦的經濟成長,
05:26
so that means手段 wages工資 fall秋季 down
to emEM subsistence生活 levels水平.
105
314374
3335
那意味著薪資水平會落到
仿真腦能維持生計的水平。
05:29
That means手段 emsEMS have to be working加工
most of the time.
106
317733
2604
那意味著仿真腦
大部份的時間得要工作。
05:32
So that means手段
this is what emsEMS usually平時 see:
107
320361
2928
那意味著仿真腦通常看見的是這個:
05:35
beautiful美麗 and luxurious豪華, but desks書桌 --
108
323684
2548
美麗又奢華的,辦公桌──
05:38
they're working加工 most of the time.
109
326256
1724
它們大部份的時間都在工作。
05:40
Now, a subsistence生活 wage工資 scenario腳本,
you might威力 think, is exotic異國情調 and strange奇怪,
110
328283
3969
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,
05:44
but it's actually其實 the usual通常 case案件
in human人的 history歷史,
111
332276
2341
但這是人類歷史上常見的情況,
05:46
and it's how pretty漂亮 much
all wild野生 animals動物 have ever lived生活,
112
334641
2714
這基本是所有野生動物的生活方式,
05:49
so we know what humans人類 do
in this situation情況.
113
337379
2276
所以我們知道人類
在這種情況下會做什麼。
05:51
Humans人類 basically基本上 do
what it takes to survive生存,
114
339679
2930
人類基本上會為了生存而行動,
05:54
and this is what lets讓我們 me say
so much about the emEM world世界.
115
342633
3309
這就是讓我說了這麽多
關於仿真腦世界的原因。
05:58
When creatures生物 are rich豐富, like you,
116
346752
2389
當生物富有時,就像你們一樣,
06:01
you have to know a lot
about what they want
117
349165
2023
你得先要知道他們想要什麼,
06:03
to figure數字 out what they do.
118
351212
1341
才能猜出他們會做什麼。
06:04
When creatures生物 are poor較差的,
119
352577
1170
當生物很貧窮時,
06:05
you know that they mostly大多 do
what it takes to survive生存.
120
353771
2657
你知道的,他們會做的
就只是生存下來。
06:08
So we've我們已經 been talking about the emEM world世界
from the point of view視圖 of the emsEMS --
121
356776
3624
我們已經從仿真腦的觀點
談了仿真腦的世界。
06:12
now, let's step back
and look at their whole整個 world世界.
122
360424
2515
現在我們退後一步,
來看看它們的整個世界。
第一,仿真腦世界的成長速度
比我們的世界快很多,
06:14
First, the emEM world世界 grows成長
much faster更快 than ours我們的,
123
362963
3080
06:18
roughly大致 a hundred times faster更快.
124
366067
2464
大約快上一百倍。
06:20
So the amount of change更改
we would experience經驗 in a century世紀 or two,
125
368555
3163
我們經歷了一、兩個世紀的改變,
06:23
they would experience經驗 in a year or two.
126
371742
1858
它們在一、兩年內就能完成。
06:25
And I'm not really willing願意 to project項目
this age年齡 much beyond that,
127
373624
3038
我其實不想預測更遙遠的未來,
因為,理論上到那時就會有
其它事發生,我們無法知道。
06:28
because plausibly振振有詞 by then something else其他
will happen發生, I don't know what.
128
376686
3437
第二,典型的仿真腦運作得更快,
06:32
Second第二, the typical典型 emulation仿真
runs運行 even faster更快,
129
380147
2358
06:34
roughly大致 a thousand times human人的 speed速度.
130
382529
2189
大約是人類的一千倍。
06:37
So for them, they experience經驗
thousands數千 of years年份 in this year or two,
131
385316
5340
所以,對它們而言,它們會在
這一、兩年就經歷我們的數千年,
06:42
and for them, the world世界 around them
is actually其實 changing改變 more slowly慢慢地
132
390680
3192
相對於你對你的周圍世界的感覺,
06:45
than your world世界 seems似乎 to change更改 for you.
133
393896
1942
它們會覺得它們周圍的
世界改變很緩慢。
06:48
Third第三, emsEMS are crammed臨時抱佛腳 together一起
in a small number of very dense稠密 cities城市.
134
396277
4484
第三,仿真腦會擠在
少數幾個高密度城市中。
06:52
This is not only how they see
themselves他們自己 in virtual虛擬 reality現實,
135
400785
3466
這不只是它們在虛擬實境中
看到的自身狀況,
06:56
it's also how they actually其實 are
physically物理 crammed臨時抱佛腳 together一起.
136
404275
2896
它們的實體也確實是
擠在同一個地方的。
06:59
So at emEM speeds速度, physical物理 travel旅行
feels感覺 really painfully痛苦 slow,
137
407474
3886
以仿真腦速度來看,
實體旅行是要命的緩慢,
07:03
so most emEM cities城市 are self-sufficient自給,
138
411384
2322
所以大部份的仿真腦城市
是自給自足的,
07:05
most war戰爭 is cyber網絡 war戰爭,
139
413730
1750
大部份的戰爭是網路戰爭,
07:07
and most of the rest休息 of the earth地球
away from the emEM cities城市
140
415504
2691
而地球上遠離
仿真腦城市的其它地方,
07:10
is left to the humans人類, because the emsEMS
really aren't that interested有興趣 in it.
141
418219
3674
就留給人類,因為仿真腦
對這些地方真的不感興趣。
07:14
Speaking請講 of humans人類,
142
422252
1508
說到人類,
07:15
you were wanting希望 to hear about that.
143
423784
1784
你會想要聽到這個。
07:17
Humans人類 must必須 retire退休, at once一旦, for good.
144
425592
4466
人類必須要退休,
一旦發生了,你就好好退休吧。
07:22
They just can't compete競爭.
145
430887
1523
人類無法和仿真腦競爭。
07:24
Now, humans人類 start開始 out owning擁有
all of the capital首都 in this world世界.
146
432918
3072
一開始,人類擁有
這個世界上所有的資本。
07:28
The economy經濟 grows成長 very fast快速,
their wealth財富 grows成長 very fast快速.
147
436014
2869
經濟成長非常快,
他們的財富也成長非常快。
07:30
Humans人類 get rich豐富, collectively.
148
438907
2384
整體而言,人類變富有了。
07:34
As you may可能 know, most humans人類 today今天
don't actually其實 own擁有 that much
149
442228
3239
如你所知,現今大部份人
除了擁有工作能力外,
07:37
besides除了 their ability能力 to work,
150
445491
1841
其它其實擁有的並不多。
07:39
so between之間 now and then,
they need to acquire獲得 sufficient足夠 assets資產,
151
447356
3548
所以從現在起到那時後,
他們需要獲得足夠的資產、
07:42
insurance保險 or sharing分享 arrangements安排,
152
450928
1833
準備保險或股權的配置,
07:44
or they may可能 starve餓死.
153
452785
1461
不然他們就要挨餓了。
07:46
I highly高度 recommend推薦 avoiding避免 this outcome結果.
154
454270
2015
我強烈建議要避免這個結果。
07:48
(Laughter笑聲)
155
456309
1285
(笑聲)
07:50
Now, you might威力 wonder奇蹟,
156
458465
1223
你們可能會好奇,
07:51
why would emsEMS let humans人類 exist存在?
157
459712
1478
為什麼仿真腦要讓人類存在?
07:53
Why not kill them, take their stuff東東?
158
461214
1715
為什麼不殺了人類,
拿走他們的東西?
07:55
But notice注意 we have many許多
unproductive非生產性 retirees退休人員 around us today今天,
159
463701
3077
別忘了,現今我們身邊就有很多
沒有生產力的退休者,
07:58
and we don't kill them
and take their stuff東東.
160
466802
2105
我們不會殺了他們,拿他們的東西。
08:00
(Laughter笑聲)
161
468931
1634
(笑聲)
08:02
In part部分, that's because it would disrupt破壞
the institutions機構 we share分享 with them.
162
470589
4420
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。
08:07
Other groups would wonder奇蹟 who's誰是 next下一個,
163
475524
2116
其他族群就會想,誰會下一個遭殃?
08:09
so plausibly振振有詞, emsEMS may可能 well let humans人類
retire退休 in peace和平 during the age年齡 of emEM.
164
477664
4726
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。
08:14
You should worry擔心 more that
the age年齡 of emEM only lasts持續 a year or two
165
482414
3049
你們比較該擔心的是,
仿真腦的時代可能只有一、兩年,
而你卻不知道接下來會發生什麼事。
08:17
and you don't know what happens發生 next下一個.
166
485487
1828
08:21
Ems埃姆斯 are very much like humans人類,
167
489207
2366
仿真腦和人類很像,
08:23
but they are not like the typical典型 human人的.
168
491597
2309
但它們不像典型的人類。
08:26
The typical典型 emEM is a copy複製
of the few少數 hundred most productive生產的 humans人類.
169
494438
5983
典型的仿真腦是
幾百個最有生產力的人的翻版。
08:33
So in fact事實, they are as elite原種,
compared相比 to the typical典型 human人的,
170
501329
3580
所以事實上,
與一般人相比它們是菁英,
08:36
as the typical典型 billionaire億萬富翁,
Nobel諾貝爾 Prize winner優勝者,
171
504933
2262
它們是典型的億萬富翁、
諾貝爾獎得主、
08:39
Olympic奧林匹克 gold medalist得主, head of state.
172
507219
2349
奧運金牌得主、國家元首。
08:42
Ems埃姆斯 look on humans人類
perhaps也許 with nostalgia懷舊之情 and gratitude感謝,
173
510527
4302
仿真腦在看人類時,可能是帶著
懷舊之情和感激之情的,
08:46
but not so much respect尊重,
174
514853
1508
但不會帶有很多尊重,
08:48
which哪一個 is, if you think about it,
how you think about your ancestors祖先.
175
516385
3195
就像你仔細想想,
你是怎麽看待你的祖先的。
08:51
We know many許多 things about how humans人類
differ不同 in terms條款 of productivity生產率.
176
519604
3267
我們知道許多關於
不同個體在生產力方面的不同之處。
08:54
We can just use those
to predict預測 features特徵 of emsEMS --
177
522895
2419
我們可以用那些知識
來預測仿真腦的特徵──
08:57
for example, they tend趨向 to be smart聰明,
conscientious有良心, hard-working用功,
178
525338
3096
比如,它們往往是
聰明的、認真的、努力的、
09:00
married已婚, religious宗教, middle-aged中年.
179
528458
1685
已婚的、有宗教信仰的、中年的。
09:02
These are features特徵 of emsEMS.
180
530461
1612
這些是仿真腦的特點。
09:04
EmEM world世界 also contains包含 enormous巨大 variety品種.
181
532097
2228
仿真腦世界也有很高的多樣性。
09:06
Not only does it continue繼續 on with most
of the kinds of variety品種 that humans人類 do,
182
534710
3668
它不只延續了人類的各種多樣性,
09:10
including包含 variety品種 of industry行業
and profession職業,
183
538402
2395
包括產業及職業的多樣性,
09:12
they also have many許多 new kinds of variety品種,
184
540821
2162
它們也在很多新型特點上
體現了多樣性,
09:15
and one of the most important重要
is mind心神 speed速度.
185
543007
2345
最重要的之一就是思維速度。
09:17
Ems埃姆斯 can plausibly振振有詞 go from human人的 speed速度
186
545712
3834
仿真腦的速度,
慢可以慢到與人類一樣的速度
09:21
up to a million百萬 times
faster更快 than human人的 speed速度,
187
549570
2880
快可以快到比人類快
幾百萬倍的速度,
09:24
and down to a billion十億 times
slower比較慢 than human人的 speed速度.
188
552474
2777
往下還可以比人類慢
十億倍的速度。
09:28
Faster更快 emsEMS tend趨向 to have
markers標記 of high status狀態.
189
556226
2766
更快的仿真腦往往具有較高的地位。
09:31
They embody體現 more wealth財富.
They win贏得 arguments參數.
190
559016
2160
它們有比較多的財富。
它們能在爭論中勝出。
09:33
They sit at premium額外費用 locations地點.
191
561200
1663
它們位於金字塔頂端。
09:34
Slower比較慢 emsEMS are mostly大多 retirees退休人員,
192
562887
1820
比較慢的仿真腦多半是已退休的,
09:36
and they are like the ghosts
of our literature文學.
193
564731
2250
它們就像是我們小說中看到的鬼魂。
09:39
If you recall召回, ghosts are all around us --
194
567005
2041
如果你有注意到,
書中的鬼魂都在我們周圍,
09:41
you can interact相互作用 with them
if you pay工資 the price價錢.
195
569070
2270
如果你願意付出代價,
就可以與它們互動。
09:43
But they don't know much,
they can't influence影響 much,
196
571364
2429
但它們知道的不多,
它們的影響不大,
09:45
and they're obsessed痴迷 with the past過去,
so what's the point?
197
573817
2698
且它們沉陷在過去中,
但那有什麼意義?
09:48
(Laughter笑聲)
198
576539
1151
(笑聲)
09:49
Ems埃姆斯 also have more variety品種
in the structure結構體 of their lives生活.
199
577714
2957
仿真腦的人生結構
也有較高的多樣性。
09:52
This is your life: you start開始
and you end結束, really simple簡單.
200
580695
2682
這是你的人生:
開始、結束就這麼簡單。
09:55
This is the life of an emEM,
201
583401
1613
這是仿真腦的人生,
09:57
who every一切 day splits拆分 off
some short-term短期 copies副本
202
585038
2293
每天都會分裂出一些短暫的複製品,
09:59
to do short-term短期 tasks任務 and then end結束.
203
587355
1958
用來完成短期任務,然後結束。
10:01
We'll talk more about
those short term術語 versions版本 in a moment時刻,
204
589337
4127
我們稍後再討論這些短期版本,
10:05
but they are much more efficient高效
205
593488
1610
但是它們的效率更高,
10:07
because they don't have to rest休息
for the next下一個 day.
206
595122
2365
因為它們不需要在第二天休息。
這個仿真腦是比較機會主義者的。
10:09
This emEM is more opportunistic機會主義的.
207
597511
1508
10:11
They make more copies副本 of themselves他們自己
when there's more demand需求 for that.
208
599043
3287
當有需要時,它們就會
把自己複製更多份。
10:14
They don't know which哪一個 way
the future's期貨 going.
209
602354
2097
它們不知道未來的走向。
10:16
This is an emEM designer設計師,
210
604475
1299
這是一個仿真腦設計師,
10:17
who conceives of a large system系統
211
605798
1602
它構想了一個很大的系統,
10:19
and then breaks休息 recursively遞歸 into copies副本
who elaborate闡述 that,
212
607424
3485
然後再遞歸地分裂成複製品
來詳細說明它的構想,
10:22
so emsEMS can implement實行
larger, more coherent相干 designs設計.
213
610933
3412
因此,仿真腦可以實現更大、
更連貫的設計。
10:26
This an emulation仿真 plumber水管工人
214
614369
1358
這是個仿真腦世界裡的水管工,
10:27
who remembers記得 that every一切 day,
for the last 20 years年份,
215
615751
2445
它記得過去二十年的每一天,
10:30
they only ever worked工作
two hours小時 a day, a life of leisure閒暇.
216
618220
2781
它們一天只工作兩小時,
很悠閒的人生。
10:33
But what really happened發生 is,
every一切 day they had a thousand copies副本,
217
621025
3097
但實情是,它們每天都得
複製出數千個複製品,
10:36
each of whom did a two-hour兩小時 plumbing水暖 job工作,
218
624146
1953
每個複製品得做兩個小時的工作,
10:38
and only one of them
went on to the next下一個 day.
219
626123
2165
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。
10:40
Objectively客觀, they're working加工
well over 99 percent百分 of the time.
220
628312
2999
主觀上,它們還記得
甚麼是悠閑的生活。
10:43
Subjectively主觀, they remember記得
a life of leisure閒暇.
221
631335
2190
10:45
(Laughter笑聲)
222
633549
1008
(笑聲)
10:46
This, again, is you.
You start開始 and you end結束.
223
634581
2088
同樣的,這還是你。
你開始,你結束。
10:48
This could be you
if at the start開始 of party派對,
224
636693
2112
這可能是你,如果在派對開始時,
10:50
you took a drug藥物 that meant意味著
you would not remember記得 that party派對
225
638829
2811
你吃了顆藥,
它會讓你在那天之後再也
10:53
ever after that day.
226
641664
1618
想不起那場派對。
10:55
Some people do this, I'm told.
227
643306
1929
據我所知,有些人會這樣做。
10:57
Toward the end結束 of the party派對,
228
645670
1635
到派對尾聲時,
10:59
will you say to yourself你自己,
"I'm about to die, this is terrible可怕.
229
647329
2926
我要問你的是,你是否會告訴自己:
「我快要死了,這太可怕了。
11:02
That person tomorrow明天 isn't me,
because they won't慣於 remember記得 what I do."
230
650279
3303
明天的那個我其實不是我,
因為他不會記得我做過的事。」
11:05
Or you could say, "I will go on tomorrow明天.
I just won't慣於 remember記得 what I did."
231
653606
3961
或者說:「明天,我還是我,
只是不記得我過去做了什麽。」
11:09
This is an emEM who splits拆分 off
a short-term短期 copy複製
232
657591
2429
這是分裂出短期複製品的仿真腦,
11:12
to do a short-term短期 task任務 and then end結束.
233
660044
1864
用來做短期工作,然後就結束。
11:13
They have the same相同 two
attitude態度 possibilities可能性.
234
661932
2206
它們同樣也有這兩種可能的態度。
11:16
They can say, "I'm a new short-term短期
creature生物 with a short life. I hate討厭 this."
235
664162
3679
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。」
11:19
Or "I'm a part部分 of a larger creature生物
who won't慣於 remember記得 this part部分."
236
667865
3134
我預測它們會採用第二種態度,
11:23
I predict預測 they'll他們會 have
that second第二 attitude態度,
237
671023
2118
不是因為哲學正確性,而是因為
那種態度能幫助它們相處。
11:25
not because it's philosophically哲學 correct正確,
but because it helps幫助 them get along沿.
238
673165
3676
11:28
Today今天, if the president主席 says
we must必須 invade入侵 Iraq伊拉克,
239
676865
2328
如果今天總統說
我們必須要侵略伊拉克,
11:31
and you say, "Why?"
240
679217
1151
你說:「為什麼?」
11:32
and they say, "State secret秘密,"
241
680392
1382
他們說:「國家機密。」
11:33
you're not sure if you can trust相信 them,
242
681798
1850
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,
11:35
but for emsEMS, a copy複製 of the president主席
and a copy複製 of you can go inside a safe安全,
243
683672
3616
11:39
explain說明 all their secret秘密 reasons原因,
244
687312
1654
解釋它們所有的秘密原因,
11:40
and then one bit comes out
from your copy複製 to yourself你自己,
245
688990
2531
然後有一個位元訊號
從你的複製品回來告知你,
11:43
telling告訴 you if you were convinced相信.
246
691545
1856
讓你知道你是否有被說服。
11:45
So now you can know
there is a good reason原因.
247
693425
2223
這樣你就會知道有個好理由存在。
11:47
I know you guys are all eager急於
to evaluate評估 this world世界.
248
695976
2545
我知道你們都很渴望評估那個世界。
11:50
You're eager急於 to decide決定
if you love it or hate討厭 it.
249
698545
2334
你想盡快決定
你是否喜歡它或討厭它。
11:52
But think: your ancestors祖先
from thousands數千 of years年份 ago
250
700903
2543
但想想:你數千年前的祖先
11:55
would have loved喜愛 or hated your world世界
251
703470
1753
會根據他們聽聞的最初的幾件事,
11:57
based基於 on the first few少數 things
they heard聽說 about it,
252
705247
2421
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。
11:59
because your world世界
is really just weird奇怪的.
253
707692
1853
12:01
So before judging判斷 a strange奇怪 future未來 world世界,
you should really learn學習 a lot about it,
254
709579
3818
所以在評斷一個奇怪的
未來世界之前,
你應該好好學習它,
也許讀一本關於它的書,
12:05
maybe read a whole整個 book about it,
255
713417
1570
12:07
and then, if you don't like it,
work to change更改 it.
256
715031
2344
然後,如果你不喜歡它,
就努力去改變它。
12:09
Thank you.
257
717409
1019
謝謝。
12:10
(Applause掌聲)
258
718452
4833
(掌聲)
Translated by Lilian Chiu
Reviewed by Yi-Fan Yu

▲Back to top

ABOUT THE SPEAKER
Robin Hanson - Futurist, social scientist
Does humanity have a future as uploaded minds? In his work, Robin Hanson asks this and other extra-large questions.

Why you should listen

In his book, The Age of Em: Work, Love and Life When Robots Rule the Earth, Robin Hanson re-imagines humanity's role as our tech becomes smarter. A pioneer in prediction markets, also known as information markets and idea futures, Hanson has been known since the 1980s for taking the very very long view on topics as varied as (a selected list) spatial product competition, health incentive contracts, group insurance, product bans, evolutionary psychology and bioethics of health care, voter information incentives, incentives to fake expertise, Bayesian classification, agreeing to disagree, self-deception in disagreement, probability elicitation, wiretaps, image reconstruction, the history of science prizes, reversible computation, the origin of life, the survival of humanity, very long term economic growth, growth given machine intelligence and interstellar colonization.

Meanwhile, he has developed new technologies for conditional, combinatorial and intermediated trading, and he studied insider trading, manipulation and other foul play. Hanson is associate professor of economics at George Mason University and a research associate at the Future of Humanity Institute of Oxford University. His next book is The Elephant in the Brain, co-authored with Kevin Simler, due in 2018.

More profile about the speaker
Robin Hanson | Speaker | TED.com

Data provided by TED.

This site was created in May 2015 and the last update was on January 12, 2020. It will no longer be updated.

We are currently creating a new site called "eng.lish.video" and would be grateful if you could access it.

If you have any questions or suggestions, please feel free to write comments in your language on the contact form.

Privacy Policy

Developer's Blog

Buy Me A Coffee