ABOUT THE SPEAKERS
Sebastian Thrun - Educator, entrepreneur
Sebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us.

Why you should listen

Sebastian Thrun is an educator, entrepreneur and troublemaker. After a long life as a professor at Stanford University, Thrun resigned from tenure to join Google. At Google, he founded Google X, home to self-driving cars and many other moonshot technologies. Thrun also founded Udacity, an online university with worldwide reach, and Kitty Hawk, a "flying car" company. He has authored 11 books, 400 papers, holds 3 doctorates and has won numerous awards.

More profile about the speaker
Sebastian Thrun | Speaker | TED.com
Chris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading.

Why you should listen

Chris Anderson is the Curator of TED, a nonprofit devoted to sharing valuable ideas, primarily through the medium of 'TED Talks' -- short talks that are offered free online to a global audience.

Chris was born in a remote village in Pakistan in 1957. He spent his early years in India, Pakistan and Afghanistan, where his parents worked as medical missionaries, and he attended an American school in the Himalayas for his early education. After boarding school in Bath, England, he went on to Oxford University, graduating in 1978 with a degree in philosophy, politics and economics.

Chris then trained as a journalist, working in newspapers and radio, including two years producing a world news service in the Seychelles Islands.

Back in the UK in 1984, Chris was captivated by the personal computer revolution and became an editor at one of the UK's early computer magazines. A year later he founded Future Publishing with a $25,000 bank loan. The new company initially focused on specialist computer publications but eventually expanded into other areas such as cycling, music, video games, technology and design, doubling in size every year for seven years. In 1994, Chris moved to the United States where he built Imagine Media, publisher of Business 2.0 magazine and creator of the popular video game users website IGN. Chris eventually merged Imagine and Future, taking the combined entity public in London in 1999, under the Future name. At its peak, it published 150 magazines and websites and employed 2,000 people.

This success allowed Chris to create a private nonprofit organization, the Sapling Foundation, with the hope of finding new ways to tackle tough global issues through media, technology, entrepreneurship and, most of all, ideas. In 2001, the foundation acquired the TED Conference, then an annual meeting of luminaries in the fields of Technology, Entertainment and Design held in Monterey, California, and Chris left Future to work full time on TED.

He expanded the conference's remit to cover all topics, including science, business and key global issues, while adding a Fellows program, which now has some 300 alumni, and the TED Prize, which grants its recipients "one wish to change the world." The TED stage has become a place for thinkers and doers from all fields to share their ideas and their work, capturing imaginations, sparking conversation and encouraging discovery along the way.

In 2006, TED experimented with posting some of its talks on the Internet. Their viral success encouraged Chris to begin positioning the organization as a global media initiative devoted to 'ideas worth spreading,' part of a new era of information dissemination using the power of online video. In June 2015, the organization posted its 2,000th talk online. The talks are free to view, and they have been translated into more than 100 languages with the help of volunteers from around the world. Viewership has grown to approximately one billion views per year.

Continuing a strategy of 'radical openness,' in 2009 Chris introduced the TEDx initiative, allowing free licenses to local organizers who wished to organize their own TED-like events. More than 8,000 such events have been held, generating an archive of 60,000 TEDx talks. And three years later, the TED-Ed program was launched, offering free educational videos and tools to students and teachers.

More profile about the speaker
Chris Anderson | Speaker | TED.com
TED2017

Sebastian Thrun and Chris Anderson: What AI is -- and isn't

賽巴斯汀索朗及克里斯安德森: 新世代的電腦已為自己寫程式了

Filmed:
1,575,780 views

教育家和企業家賽巴斯汀索朗鼓吹使用人工智慧來讓人類從重覆性工作中解脫、釋放創意。在與 TED 策展人克里斯安德森的這場啟發靈感、充滿資訊的談話中,索朗討論深度學習的進步、為什麼我們用不著害怕人工智慧失控,以及用機器來幫忙做無聊乏味的工作會使得社會更好。索朗說:「我們只發明了 1% 的有趣東西。…… 因為我相信所有的人都極富創意 …… 所以人工智慧能賦予我們將創意轉化為行動的能力。」
- Educator, entrepreneur
Sebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us. Full bio - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading. Full bio

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

00:12
Chris克里斯 Anderson安德森: Help us understand理解
what machine learning學習 is,
0
904
2886
克里斯安德森:幫我們
了解一下機器學習是什麼,
00:15
because that seems似乎 to be the key driver司機
1
3814
2054
因為機器學習似乎是
00:17
of so much of the excitement激動
and also of the concern關心
2
5892
2737
推動人工智慧
一些令人興奮及重要議題的關鍵動因,
00:20
around artificial人造 intelligence情報.
3
8653
1494
00:22
How does machine learning學習 work?
4
10171
1643
機器學習如何運作?
00:23
Sebastian塞巴斯蒂安 Thrun史朗: So, artificial人造
intelligence情報 and machine learning學習
5
11838
3896
賽巴斯汀索朗:人工智慧和機器學習
大約有六十年歷史,
00:27
is about 60 years年份 old
6
15758
2002
00:29
and has not had a great day
in its past過去 until直到 recently最近.
7
17784
4269
一直到近期才有輝煌的日子可言。
00:34
And the reason原因 is that today今天,
8
22077
2924
原因是因為現今
00:37
we have reached到達 a scale規模
of computing計算 and datasets數據集
9
25025
3973
我們的計算能力和
資料集規模已經達到
讓機器變聰明所必要的條件。
00:41
that was necessary必要 to make machines smart聰明.
10
29022
2637
00:43
So here's這裡的 how it works作品.
11
31683
1751
它的運作方式是這樣的。
00:45
If you program程序 a computer電腦 today今天,
say, your phone電話,
12
33458
3497
如果現在你要為一台電腦
寫程式,比如你的手機,
00:48
then you hire聘請 software軟件 engineers工程師
13
36979
2335
你會僱用軟體工程師,
00:51
that write a very,
very long kitchen廚房 recipe食譜,
14
39338
3854
他們會寫一份
非常非常長的廚房食譜,
00:55
like, "If the water is too hot,
turn down the temperature溫度.
15
43216
3132
比如「如果水太熱,就把溫度調低。
00:58
If it's too cold, turn up
the temperature溫度."
16
46372
2279
如果水太冷,把溫度調高。」
01:00
The recipes食譜 are not just 10 lines long.
17
48675
2849
食譜長度並不是只有十行。
01:03
They are millions百萬 of lines long.
18
51548
2603
它們長達數百萬行。
01:06
A modern現代 cell細胞 phone電話
has 12 million百萬 lines of code.
19
54175
4084
一台現代手機有
1200 萬行的程式碼。
01:10
A browser瀏覽器 has five million百萬 lines of code.
20
58283
2646
一個瀏覽器有五百萬行的程式碼。
01:12
And each bug竊聽器 in this recipe食譜
can cause原因 your computer電腦 to crash緊急.
21
60953
4969
食譜中的每一個錯誤,
都會造成你的電腦當機。
01:17
That's why a software軟件 engineer工程師
makes品牌 so much money.
22
65946
3075
那就是為什麼軟體工程師
能賺那麼多錢。
01:21
The new thing now is that computers電腦
can find their own擁有 rules規則.
23
69953
3660
現在的新發展是,電腦能
找到它們自己的規則。
01:25
So instead代替 of an expert專家
deciphering破譯, step by step,
24
73637
3606
所以不再需要找一個專家,
來針對每個情況的規則
01:29
a rule規則 for every一切 contingency偶然性,
25
77267
2148
一步一步地做理解辨識,
01:31
what you do now is you give
the computer電腦 examples例子
26
79439
3074
現在你的做法是,給電腦一些範例,
01:34
and have it infer推斷 its own擁有 rules規則.
27
82537
1581
讓它推導出它自己的規則。
01:36
A really good example is AlphaGoAlphaGo,
which哪一個 recently最近 was won韓元 by Google谷歌.
28
84142
4306
最近 Google 的阿爾法圍棋贏得比賽,
就是一個很好的例子。
01:40
Normally一般, in game遊戲 playing播放,
you would really write down all the rules規則,
29
88472
3687
通常,在玩遊戲時,
你得要寫下所有的規則,
01:44
but in AlphaGo'sAlphaGo的 case案件,
30
92183
1785
但在阿爾法圍棋的這個例子,
01:45
the system系統 looked看著 over a million百萬 games遊戲
31
93992
2066
系統是去看了一百萬場比賽,
01:48
and was able能夠 to infer推斷 its own擁有 rules規則
32
96082
2192
能推導出它自己的規則,
01:50
and then beat擊敗 the world's世界
residing居住 Go champion冠軍.
33
98298
2738
然後打敗世界現在的棋王。
01:53
That is exciting扣人心弦, because it relieves可以減壓
the software軟件 engineer工程師
34
101853
3509
這讓人很興奮,
因為軟體工程師能鬆口氣了,
01:57
of the need of being存在 super smart聰明,
35
105386
1819
他們不需要超聰明,
01:59
and pushes the burden負擔 towards the data數據.
36
107229
2325
這個重任已轉到資料上。
02:01
As I said, the inflection拐點 point
where this has become成為 really possible可能 --
37
109578
4534
如我所言,這件事的反轉點在於──
02:06
very embarrassing尷尬, my thesis論文
was about machine learning學習.
38
114136
2746
很慚愧,我的論文主題是機器學習,
02:08
It was completely全然
insignificant微不足道, don't read it,
39
116906
2205
它完全不重要,請別去讀它,
02:11
because it was 20 years年份 ago
40
119135
1350
因為那是二十年前寫的,
02:12
and back then, the computers電腦
were as big as a cockroach蟑螂 brain.
41
120509
2907
那時,電腦和蟑螂大腦一樣大。
02:15
Now they are powerful強大 enough足夠
to really emulate仿真
42
123440
2331
現在,電腦強大到能夠真正地模擬
02:17
kind of specialized專門 human人的 thinking思維.
43
125795
2076
人類的特定思想。
02:19
And then the computers電腦
take advantage優點 of the fact事實
44
127895
2313
接著,電腦也因為
02:22
that they can look at
much more data數據 than people can.
45
130232
2500
可以比人類看更多的資料
進而取得優勢,
02:24
So I'd say AlphaGoAlphaGo looked看著 at
more than a million百萬 games遊戲.
46
132756
3080
阿爾法圍棋已經研究過
一百多萬場比賽。
02:27
No human人的 expert專家 can ever
study研究 a million百萬 games遊戲.
47
135860
2839
沒有任何人類專家能夠
研究到一百萬場比賽。
02:30
Google谷歌 has looked看著 at over
a hundred billion十億 web捲筒紙 pages網頁.
48
138723
3182
Google 已看過了一千億個網頁。
02:33
No person can ever study研究
a hundred billion十億 web捲筒紙 pages網頁.
49
141929
2650
從來沒有人有能力研究
一千億個網頁。
02:36
So as a result結果,
the computer電腦 can find rules規則
50
144603
2714
因此,電腦能找出一些
02:39
that even people can't find.
51
147341
1755
人類找不出來的規則。
02:41
CACA: So instead代替 of looking ahead
to, "If he does that, I will do that,"
52
149120
4312
克:換句話說,不太像是:
「如果他那樣下,我就這樣下。」
02:45
it's more saying, "Here is what
looks容貌 like a winning勝利 pattern模式,
53
153456
3072
比較像是在說:
「下這裡像是獲勝的模式,
02:48
here is what looks容貌 like
a winning勝利 pattern模式."
54
156552
2079
下那裡像是獲勝的模式。」
02:50
STST: Yeah. I mean, think about
how you raise提高 children孩子.
55
158655
2517
賽:是的,想想看
你如何養育你的孩子。
你並不會花前十八年的時間,
對每種狀況給孩子一條規則,
02:53
You don't spend the first 18 years年份
giving kids孩子 a rule規則 for every一切 contingency偶然性
56
161196
3644
然後放他們自由,
他們就會做出這個大程式。
02:56
and set them free自由
and they have this big program程序.
57
164864
2347
他們會摔跤,會爬起來,
他們會被賞巴掌或打屁股,
02:59
They stumble絆倒, fall秋季, get up,
they get slapped耳光 or spanked打屁股,
58
167235
2719
03:01
and they have a positive experience經驗,
a good grade年級 in school學校,
59
169978
2884
他們會有正向的經驗,
在學校有好成績,
03:04
and they figure數字 it out on their own擁有.
60
172886
1834
他們會靠自己去了解這些。
03:06
That's happening事件 with computers電腦 now,
61
174744
1737
現在電腦也是這樣,
突然間讓電腦寫程式就變簡單了。
03:08
which哪一個 makes品牌 computer電腦 programming程序設計
so much easier更輕鬆 all of a sudden突然.
62
176505
3029
03:11
Now we don't have to think anymore.
We just give them lots of data數據.
63
179558
3175
我們不用再花腦筋思考了。
只要給它們大量資料即可。
克:所以這是自動駕駛車的能力
03:14
CACA: And so, this has been key
to the spectacular壯觀 improvement起色
64
182757
3422
03:18
in power功率 of self-driving自駕車 cars汽車.
65
186203
3064
能夠有重大改善的關鍵。
03:21
I think you gave me an example.
66
189291
1739
我想你給了我一個例子。
03:23
Can you explain說明 what's happening事件 here?
67
191054
2685
你能否解釋一下這裡發生了什麼事?
03:25
STST: This is a drive駕駛 of a self-driving自駕車 car汽車
68
193763
3564
賽:這是自動駕駛車的行車,
03:29
that we happened發生 to have at UdacityUdacity
69
197351
1957
我們優達學城(Udacity)碰巧有,
03:31
and recently最近 made製作
into a spin-off分拆 called Voyage航程.
70
199332
2398
最近變成稱為 Voyage 的副產品。
03:33
We have used this thing
called deep learning學習
71
201754
2574
我們用所謂的「深度學習」
03:36
to train培養 a car汽車 to drive駕駛 itself本身,
72
204352
1623
來訓練汽車自動駕駛,
03:37
and this is driving主動
from Mountain View視圖, California加州,
73
205999
2387
這趟行程從加州的山景城出發
03:40
to San Francisco弗朗西斯科
74
208410
1168
前往舊金山,
03:41
on El薩爾瓦多 Camino卡米諾 Real真實 on a rainy多雨的 day,
75
209602
2259
在雨天行駛 El Camino Real 路名,
03:43
with bicyclists騎自行車 and pedestrians行人
and 133 traffic交通 lights燈火.
76
211885
3524
路上有腳踏車騎士及行人,
途中經過 133 個交通燈號。
03:47
And the novel小說 thing here is,
77
215433
2636
新奇的是,
許多個月前,我成立了
Google 自動駕駛汽車團隊,
03:50
many許多, many許多 moons月亮 ago, I started開始
the Google谷歌 self-driving自駕車 car汽車 team球隊.
78
218093
3120
03:53
And back in the day, I hired僱用
the world's世界 best最好 software軟件 engineers工程師
79
221237
3181
那時,我僱用了世界上
最好的軟體工程師,
03:56
to find the world's世界 best最好 rules規則.
80
224442
1607
來找出世界上最好的規則。
03:58
This is just trained熟練.
81
226073
1754
這只是訓練出來的。
03:59
We drive駕駛 this road 20 times,
82
227851
3336
這條路我們開了二十次,
04:03
we put all this data數據
into the computer電腦 brain,
83
231211
2447
我們把所有資料放到電腦的大腦中,
04:05
and after a few少數 hours小時 of processing處理,
84
233682
2082
經過幾小時的處理之後,
04:07
it comes up with behavior行為
that often經常 surpasses超過 human人的 agility敏捷.
85
235788
3926
它所找出的行為,
通常都能勝過人類的機敏。
04:11
So it's become成為 really easy簡單 to program程序 it.
86
239738
2017
所以變得很容易為它寫程式。
04:13
This is 100 percent百分 autonomous自主性,
about 33 miles英里, an hour小時 and a half.
87
241779
3803
這是 100% 自主的,
大約 33 英哩,一小時半。
04:17
CACA: So, explain說明 it -- on the big part部分
of this program程序 on the left,
88
245606
3630
克:解釋一下這程式左半邊的大部分,
04:21
you're seeing眼看 basically基本上 what
the computer電腦 sees看到 as trucks卡車 and cars汽車
89
249260
3257
我們可以看到電腦
所看到的卡車與汽車,
04:24
and those dots overtaking超車 it and so forth向前.
90
252541
2886
還有那些超過它的點。
04:27
STST: On the right side, you see the camera相機
image圖片, which哪一個 is the main主要 input輸入 here,
91
255451
3762
賽:右側的是攝影機的影像,
也就是主要的輸入,
04:31
and it's used to find lanes車道,
other cars汽車, traffic交通 lights燈火.
92
259237
2676
用來找車道、其它車輛、交通號誌。
04:33
The vehicle車輛 has a radar雷達
to do distance距離 estimation估計.
93
261937
2489
這車用個雷達來估算距離。
04:36
This is very commonly常用 used
in these kind of systems系統.
94
264450
2621
這是這類系統常用的方式。
04:39
On the left side you see a laser激光 diagram,
95
267095
1992
左邊的是雷射圖,
04:41
where you see obstacles障礙 like trees樹木
and so on depicted描繪 by the laser激光.
96
269111
3200
可以看到雷射槍描繪出來的障礙,
如樹木等等。
04:44
But almost幾乎 all the interesting有趣 work
is centering定心 on the camera相機 image圖片 now.
97
272335
3436
但幾乎所有有趣的部份
都以攝影機影像為中心。
我們其實在從精準的感測器,
像是雷達和雷射,
04:47
We're really shifting over from precision精確
sensors傳感器 like radars雷達 and lasers激光器
98
275795
3476
轉換到極便宜的一般感測器。
04:51
into very cheap低廉, commoditized商品化 sensors傳感器.
99
279295
1842
04:53
A camera相機 costs成本 less than eight dollars美元.
100
281161
1987
一台攝影機的成本不到 $8。
04:55
CACA: And that green綠色 dot
on the left thing, what is that?
101
283172
2793
克:左邊的綠點是什麼?
04:57
Is that anything meaningful富有意義的?
102
285989
1371
是有意義的嗎?
04:59
STST: This is a look-ahead展望未來 point
for your adaptive自適應 cruise巡航 control控制,
103
287384
3668
賽:這是「向前看」的點,
供自動調整航程控制用,
05:03
so it helps幫助 us understand理解
how to regulate調節 velocity速度
104
291076
2477
它會根據前車的距離
05:05
based基於 on how far
the cars汽車 in front面前 of you are.
105
293577
2634
幫助我們調整速度。
05:08
CACA: And so, you've also
got an example, I think,
106
296235
2716
克:我想,你應該也可以舉個例說明
05:10
of how the actual實際
learning學習 part部分 takes place地點.
107
298975
2381
學習的部份實際上如何進行。
05:13
Maybe we can see that. Talk about this.
108
301380
2458
也許我們可以
邊看那畫面,邊談這個。
05:15
STST: This is an example where we posed構成
a challenge挑戰 to UdacityUdacity students學生們
109
303862
3643
賽:這是我們挑戰
Udacity 學生的一個例子,
05:19
to take what we call
a self-driving自駕車 car汽車 NanodegreeNanodegree.
110
307529
3131
是取得「自駕車奈米學位」的挑戰。
05:22
We gave them this dataset數據集
111
310684
1495
我們給他們這個資料集,
05:24
and said "Hey, can you guys figure數字 out
how to steer駕駛 this car汽車?"
112
312203
3054
說:「你們能不能想出
要如何駕駛這台車?」
05:27
And if you look at the images圖片,
113
315281
1624
如果從影像來看,
05:28
it's, even for humans人類, quite相當 impossible不可能
to get the steering操舵 right.
114
316929
4073
即使是人類操縱也很難駕駛好。
05:33
And we ran a competition競爭 and said,
"It's a deep learning學習 competition競爭,
115
321026
3591
我們進行了一項競賽,並說:
「這是場深度學習競賽,
05:36
AIAI competition競爭,"
116
324641
1173
人工智慧競賽。」
05:37
and we gave the students學生們 48 hours小時.
117
325838
1887
我們給學生 48 小時。
05:39
So if you are a software軟件 house
like Google谷歌 or FacebookFacebook的,
118
327749
4172
如果你是間軟體公司,
如 Google 或臉書,
05:43
something like this costs成本 you
at least最小 six months個月 of work.
119
331945
2717
像這樣的東西會花你
至少六個月的功夫。
05:46
So we figured想通 48 hours小時 is great.
120
334686
2202
所以我們認為 48 小時是很棒的。
05:48
And within 48 hours小時, we got about
100 submissions提交 from students學生們,
121
336912
3467
在 48 小時內,我們得到了
約一百件學生提交的結果,
05:52
and the top最佳 four got it perfectly完美 right.
122
340403
3370
前四名完全無誤。
05:55
It drives驅動器 better than I could
drive駕駛 on this imagery意象,
123
343797
2640
它駕駛得比我能在
這影像上駕駛得還要好,
05:58
using運用 deep learning學習.
124
346461
1189
用的就是深度學習。
05:59
And again, it's the same相同 methodology方法.
125
347674
1799
同樣的方法,
06:01
It's this magical神奇 thing.
126
349497
1164
很神奇,
06:02
When you give enough足夠 data數據
to a computer電腦 now,
127
350685
2085
當你提供電腦足夠的資料,
06:04
and give enough足夠 time
to comprehend理解 the data數據,
128
352794
2140
並給它足夠時間來理解那些資料,
06:06
it finds認定 its own擁有 rules規則.
129
354958
1445
它就會自己找到規則。
06:09
CACA: And so that has led to the development發展
of powerful強大 applications應用
130
357339
4845
克:所以那就導致了
強大應用程式的發展,
06:14
in all sorts排序 of areas.
131
362208
1525
在各領域都有。
06:15
You were talking to me
the other day about cancer癌症.
132
363757
2668
之前你有和我談過癌症的事。
06:18
Can I show顯示 this video視頻?
133
366449
1189
我能播那段影片嗎?
06:19
STST: Yeah, absolutely絕對, please.
CACA: This is cool.
134
367662
2354
賽:當然,請放。
克:這很酷。
06:22
STST: This is kind of an insight眼光
into what's happening事件
135
370040
3534
賽:這有點像是對完全不同的領域
06:25
in a completely全然 different不同 domain.
136
373598
2429
洞察所發生的事。
06:28
This is augmenting增廣, or competing競爭 --
137
376051
3752
在旁觀者眼裡,
這是擴增,或者可說是
06:31
it's in the eye of the beholder旁觀者 --
138
379827
1749
06:33
with people who are being存在 paid支付
400,000 dollars美元 a year,
139
381600
3454
與年賺 $40 萬美元的人競爭:
06:37
dermatologists皮膚科醫生,
140
385078
1237
皮膚科醫生,
06:38
highly高度 trained熟練 specialists專家.
141
386339
1983
他們是受過高度訓練的專家,
06:40
It takes more than a decade of training訓練
to be a good dermatologist皮膚科醫生.
142
388346
3561
要受十年以上的訓練才可能
成為好的皮膚科醫生。
06:43
What you see here is
the machine learning學習 version of it.
143
391931
3196
這裡所看到的是它的機器學習版本。
06:47
It's called a neural神經 network網絡.
144
395151
1841
稱為「神經網路」,
06:49
"Neural神經 networks網絡" is the technical技術 term術語
for these machine learning學習 algorithms算法.
145
397016
3742
神經網路是機器學習
演算法的專有名詞,
大約出現於 1980 年代。
06:52
They've他們已經 been around since以來 the 1980s.
146
400782
1789
06:54
This one was invented發明 in 1988
by a FacebookFacebook的 Fellow同伴 called Yann LeCunLeCun,
147
402595
4640
這個是在 1988 年由臉書的
研究專員揚勒丘恩所發明,
06:59
and it propagates傳播 data數據 stages階段
148
407259
3558
它傳播數據的階段
07:02
through通過 what you could think of
as the human人的 brain.
149
410841
2578
透過一種你可視為是人腦的方式。
07:05
It's not quite相當 the same相同 thing,
but it emulates模擬 the same相同 thing.
150
413443
2966
它不是人腦,但它模仿人腦。
一個階段接著一個階段,
07:08
It goes stage階段 after stage階段.
151
416433
1302
07:09
In the very first stage階段, it takes
the visual視覺 input輸入 and extracts提取物 edges邊緣
152
417759
3637
在第一個階段取得視覺輸入,
粹取出邊緣、細竿,和點。
07:13
and rods and dots.
153
421420
2612
07:16
And the next下一個 one becomes
more complicated複雜 edges邊緣
154
424056
3037
下個階段就變成更複雜的邊緣
07:19
and shapes形狀 like little half-moons半月.
155
427117
3191
以及形狀,像是半月。
07:22
And eventually終於, it's able能夠 to build建立
really complicated複雜 concepts概念.
156
430332
4443
最終,它能建立出非常複雜的概念。
07:26
Andrew安德魯 Ng has been able能夠 to show顯示
157
434799
2048
吳恩達就展示過,
07:28
that it's able能夠 to find
cat faces面孔 and dog faces面孔
158
436871
3480
它能夠在非常大量的影像中找出
貓和狗的臉。
07:32
in vast廣大 amounts of images圖片.
159
440375
1661
07:34
What my student學生 team球隊
at Stanford斯坦福 has shown顯示 is that
160
442060
2724
我在史丹佛的學生團隊也展示過,
07:36
if you train培養 it on 129,000 images圖片
of skin皮膚 conditions條件,
161
444808
6073
如果你用十二萬九千張
皮膚症狀的影像來訓練它,
07:42
including包含 melanoma黑色素瘤 and carcinomas,
162
450905
2565
包括黑色素瘤和癌,
07:45
you can do as good a job工作
163
453494
3301
你就能和最好的人類皮膚科醫生
07:48
as the best最好 human人的 dermatologists皮膚科醫生.
164
456819
2197
做得一樣好。
07:51
And to convince說服 ourselves我們自己
that this is the case案件,
165
459040
2549
為了說服我們自己確實是如此,
07:53
we captured捕獲 an independent獨立 dataset數據集
that we presented呈現 to our network網絡
166
461613
3990
我們取得了一個獨立的資料集,
拿給我們的網路看,
07:57
and to 25 board-certified認證資格
Stanford-level斯坦福級 dermatologists皮膚科醫生,
167
465627
4342
也拿給 25 位認證過的
史丹佛水準的皮膚科醫生看,
08:01
and compared相比 those.
168
469993
1672
來做比較。
08:03
And in most cases,
169
471689
1504
在大部份狀況,
08:05
they were either on par平價 or above以上
the performance性能 classification分類 accuracy準確性
170
473217
3875
在分類正確性上,
網路的表現都和人類皮膚科醫生
08:09
of human人的 dermatologists皮膚科醫生.
171
477116
1467
並駕齊驅或更好。
08:10
CACA: You were telling告訴 me an anecdote軼事.
172
478607
1746
克:你跟我說過一則軼事。
08:12
I think about this image圖片 right here.
173
480377
1957
上面的這張影像。
08:14
What happened發生 here?
174
482358
1484
這裡發生了什麼事?
08:15
STST: This was last Thursday星期四.
That's a moving移動 piece.
175
483866
4008
賽:時間是上星期四,
是個進行中的故事。
08:19
What we've我們已經 shown顯示 before and we published發表
in "Nature性質" earlier this year
176
487898
3600
我們之前展示過,且今年稍早
也刊在「Nature」期刊中,
08:23
was this idea理念 that we show顯示
dermatologists皮膚科醫生 images圖片
177
491522
2484
想法是,我們讓皮膚科醫生看影像,
08:26
and our computer電腦 program程序 images圖片,
178
494030
1539
也讓我們的電腦程式看,
08:27
and count計數 how often經常 they're right.
179
495593
1627
計算它們多常判斷正確。
08:29
But all these images圖片 are past過去 images圖片.
180
497244
1778
但所有影像都是過去的影像。
08:31
They've他們已經 all been biopsied活檢 to make sure
we had the correct正確 classification分類.
181
499046
3460
都已經過切片檢查,確保分類正確。
08:34
This one wasn't.
182
502530
1172
這一張卻不是。
08:35
This one was actually其實 doneDONE at Stanford斯坦福
by one of our collaborators合作者.
183
503726
3179
這張其實是在史丹佛
由我們的合作者之一做的。
08:38
The story故事 goes that our collaborator合作者,
184
506929
2314
故事是,我們的合作者
08:41
who is a world-famous世界知名 dermatologist皮膚科醫生,
one of the three best最好, apparently顯然地,
185
509267
3391
是世界知名的皮膚科醫生,
很顯然是最好的三位之一,
08:44
looked看著 at this mole and said,
"This is not skin皮膚 cancer癌症."
186
512682
2935
他看著這個痣,說:
「這不是皮膚癌。」
08:47
And then he had
a second第二 moment時刻, where he said,
187
515641
2476
他想了一下,接著又說:
08:50
"Well, let me just check with the app應用."
188
518141
1866
「讓我用應用程式確認一下。」
08:52
So he took out his iPhone蘋果手機
and ran our piece of software軟件,
189
520031
2699
他拿出他的 iPhone,
執行我們的軟體,
可說是我們的「口袋皮膚科醫生」,
08:54
our "pocket口袋 dermatologist皮膚科醫生," so to speak說話,
190
522754
2121
08:56
and the iPhone蘋果手機 said: cancer癌症.
191
524899
2994
而 iPhone 說:癌症。
08:59
It said melanoma黑色素瘤.
192
527917
1306
它說是黑色素瘤。
09:01
And then he was confused困惑.
193
529849
1233
他很困惑。
09:03
And he decided決定, "OK, maybe I trust相信
the iPhone蘋果手機 a little bit more than myself,"
194
531106
4551
他決定:「好,也許我應該相信
iPhone 比相信我自己多一點點。」
09:07
and he sent發送 it out to the lab實驗室
to get it biopsied活檢.
195
535681
2735
他把它送去實驗室做切片檢查。
09:10
And it came來了 up as an aggressive侵略性 melanoma黑色素瘤.
196
538440
2469
結果是惡性黑色素瘤。
09:13
So I think this might威力 be the first time
that we actually其實 found發現,
197
541545
3067
我想,這可能是我們第一次
真正在深度學習的實做中遇到,
09:16
in the practice實踐 of using運用 deep learning學習,
198
544636
2487
09:19
an actual實際 person whose誰的 melanoma黑色素瘤
would have gone走了 unclassified機密,
199
547147
3372
如果沒有深度學習的話,
這個人的黑色素瘤就不會被發現。
09:22
had it not been for deep learning學習.
200
550543
2115
09:24
CACA: I mean, that's incredible難以置信.
201
552682
1560
克:那很了不起。
09:26
(Applause掌聲)
202
554266
1769
(掌聲)
09:28
It feels感覺 like there'd這紅色 be an instant瞬間 demand需求
for an app應用 like this right now,
203
556059
3600
感覺現在對於像這樣的應用程式,
有很迫切的需求,
09:31
that you might威力 freak怪物 out a lot of people.
204
559683
1966
你可能會嚇壞很多人。
09:33
Are you thinking思維 of doing this,
making製造 an app應用 that allows允許 self-checking自我檢查?
205
561673
3527
你有想過要這麼做嗎?
做個自我檢測的應用程式?
賽:我的收件匣被關於癌症
應用程式的信件給淹滿了,
09:37
STST: So my in-box箱內 is flooded
about cancer癌症 apps應用,
206
565224
4973
09:42
with heartbreaking令人心碎 stories故事 of people.
207
570221
2303
信上都是人們的心碎故事。
09:44
I mean, some people have had
10, 15, 20 melanomas黑色素 removed去除,
208
572548
3204
有些人已經移除了
10、15、20 個黑色素瘤,
09:47
and are scared害怕 that one
might威力 be overlooked忽視, like this one,
209
575776
3952
很害怕會漏掉任何一個,就像這個,
09:51
and also, about, I don't know,
210
579752
1741
還有些內容是,我不知道,
09:53
flying飛行 cars汽車 and speaker揚聲器 inquiries查詢
these days, I guess猜測.
211
581517
2732
飛天車、演說邀請,我猜是吧。
09:56
My take is, we need more testing測試.
212
584273
2738
我的反應是,我們需要更多測試。
09:59
I want to be very careful小心.
213
587449
1778
我想要非常小心。
10:01
It's very easy簡單 to give a flashy華而不實 result結果
and impress a TEDTED audience聽眾.
214
589251
3666
很容易就可以丟出亮眼的結果
來讓 TED 觀眾印象深刻。
10:04
It's much harder更難 to put
something out that's ethical合乎道德的.
215
592941
2627
要端出合乎道德的東西就難很多。
10:07
And if people were to use the app應用
216
595592
2394
如果人們要用這個應用程式,
10:10
and choose選擇 not to consult請教
the assistance幫助 of a doctor醫生
217
598010
2797
且選擇不去諮詢醫生的協助,
10:12
because we get it wrong錯誤,
218
600831
1583
而我們弄錯的話,
10:14
I would feel really bad about it.
219
602438
1653
我就會感覺非常糟。
10:16
So we're currently目前 doing clinical臨床 tests測試,
220
604115
1925
所以我們目前在做臨床實驗,
10:18
and if these clinical臨床 tests測試 commence開始
and our data數據 holds持有 up,
221
606064
2798
如果這些實驗開始之後,
我們的資料站得住腳,
10:20
we might威力 be able能夠 at some point
to take this kind of technology技術
222
608886
2990
在某個時點,我們或許可以把這技術
10:23
and take it out of the Stanford斯坦福 clinic診所
223
611900
1892
拿到史丹佛臨床課之外,
10:25
and bring帶來 it to the entire整個 world世界,
224
613816
1658
把它帶給全世界,
帶到史丹佛的醫生
從來沒有去過的地方。
10:27
places地方 where Stanford斯坦福
doctors醫生 never, ever set foot腳丫子.
225
615498
2456
10:30
CACA: And do I hear this right,
226
618617
2580
克:我有沒有聽正確,
10:33
that it seemed似乎 like what you were saying,
227
621221
1966
聽起來像是你在說
10:35
because you are working加工
with this army軍隊 of UdacityUdacity students學生們,
228
623211
4254
因為你在和這支
Udacity 學生大軍合作,
10:39
that in a way, you're applying應用
a different不同 form形成 of machine learning學習
229
627489
3221
以某種方式,你們在應用
一種不同形式的機器學習,
可能會發生在公司中的形式,
10:42
than might威力 take place地點 in a company公司,
230
630734
1735
10:44
which哪一個 is you're combining結合 machine learning學習
with a form形成 of crowd人群 wisdom智慧.
231
632493
3484
也就是你們將機器學習
與一種群眾智慧結合。
10:48
Are you saying that sometimes有時 you think
that could actually其實 outperform跑贏大市
232
636001
3384
你是不是在說,
有時你認為那能夠勝過
公司所能做到的,甚至大型公司?
10:51
what a company公司 can do,
even a vast廣大 company公司?
233
639409
2050
賽:我相信現在有一些
讓我很興奮的例子,
10:53
STST: I believe there's now
instances實例 that blow打擊 my mind心神,
234
641483
2940
10:56
and I'm still trying to understand理解.
235
644447
1758
我還在試著了解。
10:58
What Chris克里斯 is referring to
is these competitions比賽 that we run.
236
646229
3937
克里斯指的是
我們的競賽才進行了大約
四十八小時就打開來用;
11:02
We turn them around in 48 hours小時,
237
650190
2268
11:04
and we've我們已經 been able能夠 to build建立
a self-driving自駕車 car汽車
238
652482
2252
而我們建的自駕車
11:06
that can drive駕駛 from Mountain View視圖
to San Francisco弗朗西斯科 on surface表面 streets街道.
239
654758
3387
能從山景城開上馬路去到舊金山;
11:10
It's not quite相當 on par平價 with Google谷歌
after seven years年份 of Google谷歌 work,
240
658169
3584
它尚未趕上 Google
投入七年心血的成果,
11:13
but it's getting得到 there.
241
661777
2528
但是就快追上了。
11:16
And it took us only two engineers工程師
and three months個月 to do this.
242
664329
3084
我們的研發只花了兩個工程師
用了三個月就做到這樣,
11:19
And the reason原因 is, we have
an army軍隊 of students學生們
243
667437
2856
原因是,我們有一支學生大軍,
11:22
who participate參加 in competitions比賽.
244
670317
1850
參與競賽的那些學生。
11:24
We're not the only ones那些
who use crowdsourcing眾包.
245
672191
2220
我們並非唯一使用「群眾外包」的人,
11:26
Uber尤伯杯 and Didi迪迪 use crowdsource眾包 for driving主動.
246
674435
2223
Uber 和 Didi 用群眾外包做駕駛,
11:28
Airbnb製作的Airbnb uses使用 crowdsourcing眾包 for hotels酒店.
247
676682
2759
Airbnb 用群眾外包做飯店。
11:31
There's now many許多 examples例子
where people do bug-findingbug 查找 crowdsourcing眾包
248
679465
4007
現在有許多例子是
群眾外包除錯工作
11:35
or protein蛋白 folding摺頁, of all things,
in crowdsourcing眾包.
249
683496
2804
或群眾外包蛋白質摺疊等。
11:38
But we've我們已經 been able能夠 to build建立
this car汽車 in three months個月,
250
686324
2915
但我們能在三個月內建造這台車,
11:41
so I am actually其實 rethinking重新思考
251
689263
3655
所以我其實在重新思考,
11:44
how we organize組織 corporations公司.
252
692942
2238
我們要如何組織企業。
11:47
We have a staff員工 of 9,000 people
who are never hired僱用,
253
695204
4696
我們有從未被僱用的九千名員工,
我也從未開除他們,
11:51
that I never fire.
254
699924
1308
11:53
They show顯示 up to work
and I don't even know.
255
701256
2362
我不知道他們什麼時候工作。
11:55
Then they submit提交 to me
maybe 9,000 answers答案.
256
703642
3058
後來他們提交大約九千份答案給我。
11:58
I'm not obliged有義務的 to use any of those.
257
706724
2176
我沒有義務要用任何一個答案。
12:00
I end結束 up -- I pay工資 only the winners獲獎者,
258
708924
1991
最後我只付錢給贏家,
12:02
so I'm actually其實 very cheapskate小氣鬼 here,
which哪一個 is maybe not the best最好 thing to do.
259
710939
3718
所以在這裡我算是個小氣鬼,
這不見得是最好的做法。
12:06
But they consider考慮 it part部分
of their education教育, too, which哪一個 is nice不錯.
260
714681
3185
但他們認為這是他們
教育的一部份,這樣想很好。
12:09
But these students學生們 have been able能夠
to produce生產 amazing驚人 deep learning學習 results結果.
261
717890
4201
但這些學生能夠產出非常
了不起的深度學習結果。
12:14
So yeah, the synthesis合成 of great people
and great machine learning學習 is amazing驚人.
262
722115
3861
所以,厲害的人結合
偉大的機器學習是很驚人的。
克:加里卡斯帕洛夫
在(TED 2017)第一天說,
12:18
CACA: I mean, Gary加里 Kasparov卡斯帕羅夫 said on
the first day [of TEDTED2017]
263
726000
2814
12:20
that the winners獲獎者 of chess, surprisingly出奇,
turned轉身 out to be two amateur業餘 chess players玩家
264
728848
5412
很意外的,棋賽的贏家
是兩位業餘的棋手,
12:26
with three mediocre-ish平庸的,
mediocre-to-good平庸到好, computer電腦 programs程式,
265
734284
5371
用三個平庸、中上的電腦程式
12:31
that could outperform跑贏大市 one grand盛大 master
with one great chess player播放機,
266
739679
3163
就勝過了一個大師
和一個很棒的棋手,
12:34
like it was all part部分 of the process處理.
267
742866
1743
就像這過程的一部份,
12:36
And it almost幾乎 seems似乎 like
you're talking about a much richer更豐富 version
268
744633
3335
幾乎和你談的想法同樣,
12:39
of that same相同 idea理念.
269
747992
1200
而是更豐富的版本。
12:41
STST: Yeah, I mean, as you followed其次
the fantastic奇妙 panels面板 yesterday昨天 morning早上,
270
749216
3857
賽:是的,昨天早上的小組討論很棒,
12:45
two sessions會議 about AIAI,
271
753097
1994
兩場關於人工智慧的討論,
12:47
robotic機器人 overlords霸主 and the human人的 response響應,
272
755115
2167
機器超載和人類回應,
12:49
many許多, many許多 great things were said.
273
757306
1982
說到很多很棒的內容。
12:51
But one of the concerns關注 is
that we sometimes有時 confuse迷惑
274
759312
2687
但是讓人擔心的事情之一
是有時我們混淆了
12:54
what's actually其實 been doneDONE with AIAI
with this kind of overlord霸王 threat威脅,
275
762023
4062
人工智慧實際做的事
和機器超載的威脅,
也就是人工智慧發展出意識,對吧?
12:58
where your AIAI develops發展
consciousness意識, right?
276
766109
3424
13:01
The last thing I want
is for my AIAI to have consciousness意識.
277
769557
2971
我最不想要人工智慧有意識。
13:04
I don't want to come into my kitchen廚房
278
772552
1716
我可不想進到廚房,
13:06
and have the refrigerator冰箱 fall秋季 in love
with the dishwasher洗碗機
279
774292
4193
發現冰箱愛上了洗碗機,
13:10
and tell me, because I wasn't nice不錯 enough足夠,
280
778509
2124
然後告訴我,因為我不夠好,
13:12
my food餐飲 is now warm.
281
780657
1837
我的食物現在溫的。
13:14
I wouldn't不會 buy購買 these products製品,
and I don't want them.
282
782518
2891
我不會買這些產品,
我也不想要它們。
13:17
But the truth真相 is, for me,
283
785825
1802
但,事實是,對我來說,
13:19
AIAI has always been
an augmentation增強 of people.
284
787651
2720
人工智慧一直都是人的擴增。
13:22
It's been an augmentation增強 of us,
285
790893
1676
它一直是我們的擴增,
13:24
to make us stronger.
286
792593
1457
讓我們更強大。
13:26
And I think Kasparov卡斯帕羅夫 was exactly究竟 correct正確.
287
794074
2831
我認為卡斯帕洛夫完全正確。
13:28
It's been the combination組合
of human人的 smarts智慧 and machine smarts智慧
288
796929
3849
一直都是人類的聰明
結合機器的聰明,
13:32
that make us stronger.
289
800802
1464
才讓我們更強。
13:34
The theme主題 of machines making製造 us stronger
is as old as machines are.
290
802290
4587
機器讓我們更強的主題,
就像機器本身一樣老。
13:39
The agricultural農業的 revolution革命 took
place地點 because it made製作 steam蒸汽 engines引擎
291
807567
3758
發生農業革命是因為
做出了蒸汽引擎以及耕作設備,
13:43
and farming農業 equipment設備
that couldn't不能 farm農場 by itself本身,
292
811349
2666
它們不會自己耕作或取代我們,
13:46
that never replaced更換 us;
it made製作 us stronger.
293
814039
2122
而是會讓我們更強。
13:48
And I believe this new wave of AIAI
will make us much, much stronger
294
816185
3738
而我相信,這波新的人工智慧風潮
會讓我們人類更強大許多。
13:51
as a human人的 race種族.
295
819947
1183
13:53
CACA: We'll come on to that a bit more,
296
821765
1813
克:我們等等會再談那個話題,
13:55
but just to continue繼續 with the scary害怕 part部分
of this for some people,
297
823602
3671
但先繼續聊這個
對一些人來說很駭人的部份,
13:59
like, what feels感覺 like it gets得到
scary害怕 for people is when you have
298
827297
3558
對人們來說,會覺得害怕的是
14:02
a computer電腦 that can, one,
rewrite改寫 its own擁有 code,
299
830879
4618
你讓電腦能重寫它自己的程式,
14:07
so, it can create創建
multiple copies副本 of itself本身,
300
835521
3584
它就能複製多個自己,
14:11
try a bunch of different不同 code versions版本,
301
839129
1897
嘗試各種不同版本的程式,
14:13
possibly或者 even at random隨機,
302
841050
1775
甚至可能是隨機嘗試,
14:14
and then check them out and see
if a goal目標 is achieved實現 and improved改善.
303
842849
3632
然後再確認看看
目標是否有達成或改善。
14:18
So, say the goal目標 is to do better
on an intelligence情報 test測試.
304
846505
3641
所以,比如,目標是要在一項
智力測驗中得到更好的成績。
14:22
You know, a computer電腦
that's moderately適當地 good at that,
305
850170
3894
一台電腦只要還算擅長,
14:26
you could try a million百萬 versions版本 of that.
306
854088
2509
就能嘗試一百萬個版本,
14:28
You might威力 find one that was better,
307
856621
2090
可能會找到一版比較理想,
14:30
and then, you know, repeat重複.
308
858735
2004
重覆做下去。
14:32
And so the concern關心 is that you get
some sort分類 of runaway逃跑 effect影響
309
860763
3040
擔心的是,你會有某種失控效應,
14:35
where everything is fine
on Thursday星期四 evening晚間,
310
863827
3008
在星期四晚上一切都很好,
14:38
and you come back into the lab實驗室
on Friday星期五 morning早上,
311
866859
2336
你星期五早上回到實驗室,
14:41
and because of the speed速度
of computers電腦 and so forth向前,
312
869219
2449
因為電腦的速度等等,
一切就天翻地覆,突然間──
14:43
things have gone走了 crazy, and suddenly突然 --
313
871692
1903
14:45
STST: I would say this is a possibility可能性,
314
873619
2020
賽:我會說,這有可能,
14:47
but it's a very remote遠程 possibility可能性.
315
875663
1916
卻是非常遙遠的可能。
14:49
So let me just translate翻譯
what I heard聽說 you say.
316
877603
3337
所以讓我翻譯一下我剛聽你說的。
14:52
In the AlphaGoAlphaGo case案件,
we had exactly究竟 this thing:
317
880964
2704
阿爾法圍棋的例子就有這樣的狀況:
14:55
the computer電腦 would play
the game遊戲 against反對 itself本身
318
883692
2315
電腦會自己對抗自己來下棋,
14:58
and then learn學習 new rules規則.
319
886031
1250
接著學習新規則。
14:59
And what machine learning學習 is
is a rewriting重寫 of the rules規則.
320
887305
3235
機器學習就是重寫規則。
15:02
It's the rewriting重寫 of code.
321
890564
1769
就是重寫程式。
15:04
But I think there was
absolutely絕對 no concern關心
322
892357
2845
但我認為完全不用擔心
15:07
that AlphaGoAlphaGo would take over the world世界.
323
895226
2426
阿爾法圍棋會佔領世界。
15:09
It can't even play chess.
324
897676
1464
它不會下西洋棋。
15:11
CACA: No, no, no, but now,
these are all very single-domain單域 things.
325
899164
5147
克:不,不,現在這些
都還是非常單一領域的東西。
15:16
But it's possible可能 to imagine想像.
326
904335
2879
但有可能去想像,
15:19
I mean, we just saw a computer電腦
that seemed似乎 nearly幾乎 capable
327
907238
3089
我是指,我們剛看到幾乎有能力
15:22
of passing通過 a university大學 entrance入口 test測試,
328
910351
2655
通過大學入學測驗的電腦,
15:25
that can kind of -- it can't read
and understand理解 in the sense that we can,
329
913030
3688
就像它無法用
我們的方式去閱讀及了解,
15:28
but it can certainly當然 absorb吸收 all the text文本
330
916742
1987
但它絕對可以吸收所有的文字,
15:30
and maybe see increased增加
patterns模式 of meaning含義.
331
918753
2899
也許能看到越來越多有意義的模式。
15:33
Isn't there a chance機會 that,
as this broadens變寬 out,
332
921676
3694
有沒有可能,當拓展更廣時,
15:37
there could be a different不同
kind of runaway逃跑 effect影響?
333
925394
2466
會是不同種類的失控效應?
15:39
STST: That's where
I draw the line, honestly老老實實.
334
927884
2078
賽:老實說,我會把底線設在那裡。
15:41
And the chance機會 exists存在 --
I don't want to downplay淡化 it --
335
929986
2643
存在這可能性,我不想低估它,
15:44
but I think it's remote遠程, and it's not
the thing that's on my mind心神 these days,
336
932653
3672
但我認為它很遙遠,
現在我腦中不會去想這個,
因為我認為大革命是另一回事。
15:48
because I think the big revolution革命
is something else其他.
337
936349
2512
目前為止,人工智慧所有的成功,
15:50
Everything successful成功 in AIAI
to the present當下 date日期
338
938885
2922
15:53
has been extremely非常 specialized專門,
339
941831
2214
都是極度專門化的,
15:56
and it's been thriving on a single idea理念,
340
944069
2489
一直以來,它能興盛全靠一個辦法:
15:58
which哪一個 is massive大規模的 amounts of data數據.
341
946582
2739
大量的資料。
16:01
The reason原因 AlphaGoAlphaGo works作品 so well
is because of massive大規模的 numbers數字 of Go plays播放,
342
949345
4147
阿爾法圍棋能如此成功
是因為下過大量的圍棋棋譜,
16:05
and AlphaGoAlphaGo can't drive駕駛 a car汽車
or fly a plane平面.
343
953516
3255
阿爾法圍棋無法開車或開飛機。
16:08
The Google谷歌 self-driving自駕車 car汽車
or the UdacityUdacity self-driving自駕車 car汽車
344
956795
3031
Google 的自動駕駛汽車或
Udacity 的自動駕駛汽車
16:11
thrives蓬勃發展 on massive大規模的 amounts of data數據,
and it can't do anything else其他.
345
959850
3240
能成功是因為有大量的資料,
它們無法做其他事,
甚至無法開摩托車。
16:15
It can't even control控制 a motorcycle摩托車.
346
963114
1727
它是非常明確、專門領域的功能,
16:16
It's a very specific具體,
domain-specific特定領域 function功能,
347
964865
2762
16:19
and the same相同 is true真正 for our cancer癌症 app應用.
348
967651
1907
我們的癌症應用程式也是如此。
16:21
There has been almost幾乎 no progress進展
on this thing called "general一般 AIAI,"
349
969582
3236
所謂的「一般性人工智慧」幾無進展,
16:24
where you go to an AIAI and say,
"Hey, invent發明 for me special特別 relativity相對論
350
972842
4000
就是你可以叫它:
「嘿,為我發明
狹義相對論或弦理論」的那種
16:28
or string theory理論."
351
976866
1666
16:30
It's totally完全 in the infancy嬰兒期.
352
978556
1931
完全還在嬰兒期。
16:32
The reason原因 I want to emphasize注重 this,
353
980511
2127
我想要強調這點的理由
16:34
I see the concerns關注,
and I want to acknowledge確認 them.
354
982662
3838
是我知道人們擔心,我聽見了。
16:38
But if I were to think about one thing,
355
986524
2886
但如果要我思考一件事,我會自問:
16:41
I would ask myself the question,
"What if we can take anything repetitive重複
356
989434
5563
「如果我們能夠把任何重覆事物的
16:47
and make ourselves我們自己
100 times as efficient高效?"
357
995021
3473
效率提高一百倍,會如何?」
16:51
It so turns out, 300 years年份 ago,
we all worked工作 in agriculture農業
358
999170
4249
事實證明,三百年前我們都從事農業,
16:55
and did farming農業 and did repetitive重複 things.
359
1003443
2051
耕種,做重覆性的事。
16:57
Today今天, 75 percent百分 of us work in offices辦事處
360
1005518
2556
現今,我們有 75% 的人
在辦公室工作,
17:00
and do repetitive重複 things.
361
1008098
2124
做重覆性的事。
17:02
We've我們已經 become成為 spreadsheet電子表格 monkeys猴子.
362
1010246
2183
我們已變成了試算表猴子。
17:04
And not just low-end低端 labor勞動.
363
1012453
2054
不只是低階勞工,
17:06
We've我們已經 become成為 dermatologists皮膚科醫生
doing repetitive重複 things,
364
1014531
2754
我們的皮膚科醫生
已經開始做重覆性工作,
17:09
lawyers律師 doing repetitive重複 things.
365
1017309
1749
律師也做重覆性工作。
17:11
I think we are at the brink邊緣
of being存在 able能夠 to take an AIAI,
366
1019082
3823
我認為我們正處於
能夠採用 AI 的邊緣,
17:14
look over our shoulders肩膀,
367
1022929
1718
保持警覺,
17:16
and they make us maybe 10 or 50 times
as effective有效 in these repetitive重複 things.
368
1024671
4058
可以提高我們執行
重複性工作的效率十或五十倍。
17:20
That's what is on my mind心神.
369
1028753
1275
我在想的是這個。
17:22
CACA: That sounds聲音 super exciting扣人心弦.
370
1030052
2450
克:那聽起來非常讓人興奮。
17:24
The process處理 of getting得到 there seems似乎
a little terrifying可怕的 to some people,
371
1032526
3530
對於一些人來說,要達成
那樣的過程似乎有點嚇人,
17:28
because once一旦 a computer電腦
can do this repetitive重複 thing
372
1036080
3180
因為一旦電腦能做重覆性的事,
17:31
much better than the dermatologist皮膚科醫生
373
1039284
3434
且做得比皮膚科醫生好,
17:34
or than the driver司機, especially特別,
is the thing that's talked about
374
1042742
3230
尤其做得比司機好,
這是現在熱門的話題,
17:37
so much now,
375
1045996
1290
17:39
suddenly突然 millions百萬 of jobs工作 go,
376
1047310
1958
突然間,幾百萬個工作就沒了,
17:41
and, you know, the country's in revolution革命
377
1049292
2695
你知道的,這個國家正在革命之中,
17:44
before we ever get to the more
glorious輝煌 aspects方面 of what's possible可能.
378
1052011
4329
我們都還來不及去做到
可能達成的輝煌面。
17:48
STST: Yeah, and that's an issue問題,
and it's a big issue問題,
379
1056364
2517
賽:是啊,那是個課題,大課題,
17:50
and it was pointed out yesterday昨天 morning早上
by several一些 guest客人 speakers音箱.
380
1058905
4196
昨天早上有幾位嘉賓指出這一點。
17:55
Now, prior to me showing展示 up onstage在舞台上,
381
1063125
2754
在我上台之前,
17:57
I confessed供認不諱 I'm a positive,
optimistic樂觀 person,
382
1065903
3739
我坦白說我是個正面、樂觀的人,
18:01
so let me give you an optimistic樂觀 pitch瀝青,
383
1069666
2389
讓我為各位定個樂觀的調,
18:04
which哪一個 is, think of yourself你自己
back 300 years年份 ago.
384
1072079
4795
就是,試想你回到三百年前,
18:08
Europe歐洲 just survived倖存 140 years年份
of continuous連續 war戰爭,
385
1076898
3996
歐洲剛結束了持續 140 年的戰爭,
18:12
none沒有 of you could read or write,
386
1080918
1711
你們都不會讀或寫,
18:14
there were no jobs工作 that you hold保持 today今天,
387
1082653
2945
沒有你們現在做的工作,
18:17
like investment投資 banker銀行家
or software軟件 engineer工程師 or TV電視 anchor.
388
1085622
4096
比如投資銀行家、
軟體工程師、電視台主播,
18:21
We would all be in the fields領域 and farming農業.
389
1089742
2414
我們都在田野裡耕種。
18:24
Now here comes little Sebastian塞巴斯蒂安
with a little steam蒸汽 engine發動機 in his pocket口袋,
390
1092180
3573
現在,來了一個小賽巴斯汀,
口袋中有個小蒸氣引擎,
說:「嘿,各位,看看這個。
18:27
saying, "Hey guys, look at this.
391
1095777
1548
18:29
It's going to make you 100 times
as strong強大, so you can do something else其他."
392
1097349
3595
它會讓你強大一百倍,
這樣你們就可以做其它事了。」
在那個年代,沒有真正的舞台,
18:32
And then back in the day,
there was no real真實 stage階段,
393
1100968
2470
18:35
but Chris克里斯 and I hang out
with the cows奶牛 in the stable穩定,
394
1103462
2526
但克里斯和我在畜舍中
和乳牛在一起,
18:38
and he says, "I'm really
concerned關心 about it,
395
1106012
2100
他說:「我真的很擔心這事,
18:40
because I milk牛奶 my cow every一切 day,
and what if the machine does this for me?"
396
1108136
3652
我每天給乳牛擠奶,
如果讓機器來為我擠,會如何?
我提到這一點的原因是,
18:43
The reason原因 why I mention提到 this is,
397
1111812
1702
18:46
we're always good in acknowledging承認
past過去 progress進展 and the benefit效益 of it,
398
1114360
3603
我們向來都很擅長認可
過去的進展和它帶來的益處,
18:49
like our iPhonesiPhone手機 or our planes飛機
or electricity電力 or medical supply供應.
399
1117987
3354
就像我們的 iPhone、
飛機、電力、醫材。
18:53
We all love to live生活 to 80,
which哪一個 was impossible不可能 300 years年份 ago.
400
1121365
4245
我們都喜歡活到八十歲,
這在三百年前是不可能的。
18:57
But we kind of don't apply應用
the same相同 rules規則 to the future未來.
401
1125634
4156
但我們似乎不太會用
同樣的規則看未來。
19:02
So if I look at my own擁有 job工作 as a CEOCEO,
402
1130621
3207
如果我看我自己的工作,執行長,
19:05
I would say 90 percent百分
of my work is repetitive重複,
403
1133852
3140
我會說,我 90% 的工作
是重覆性的,
19:09
I don't enjoy請享用 it,
404
1137016
1351
我並不享受做那些,
19:10
I spend about four hours小時 per day
on stupid, repetitive重複 email電子郵件.
405
1138391
3978
我每天要花大約四小時在
愚蠢、重覆性的電子郵件上。
19:14
And I'm burning燃燒 to have something
that helps幫助 me get rid擺脫 of this.
406
1142393
3641
我極渴望有什麼能協助我擺脫這些。
19:18
Why?
407
1146058
1158
為什麼?
19:19
Because I believe all of us
are insanely瘋狂地 creative創作的;
408
1147240
3003
因為我相信我們所有人
都極度有創意;
19:22
I think the TEDTED community社區
more than anybody任何人 else其他.
409
1150731
3194
我認為,比起其他人,
TED 社區更是如此。
19:25
But even blue-collar藍領 workers工人;
I think you can go to your hotel旅館 maid女傭
410
1153949
3559
但即使藍領階級勞工;我認為
你可以去找你的飯店服務員,
19:29
and have a drink with him or her,
411
1157532
2402
和他或她喝杯飲料,
19:31
and an hour小時 later後來,
you find a creative創作的 idea理念.
412
1159958
2717
一小時後,你會找到一個創意想法。
19:34
What this will empower授權
is to turn this creativity創造力 into action行動.
413
1162699
4140
人工智慧能賦予人能力,
將創意轉化為行動。
19:39
Like, what if you could
build建立 Google谷歌 in a day?
414
1167265
3442
比如,如果你能在一天就
建造出 Google,會如何?
19:43
What if you could sit over beer啤酒
and invent發明 the next下一個 SnapchatSnapchat,
415
1171221
3316
如果你能坐著喝啤酒,就發明出
下一個 Snapchat,會如何?
19:46
whatever隨你 it is,
416
1174561
1165
不論發明的是什麼,
19:47
and tomorrow明天 morning早上 it's up and running賽跑?
417
1175750
2187
明早它就可以開始運作,會如何?
19:49
And that is not science科學 fiction小說.
418
1177961
1773
那不是科幻小說。
19:51
What's going to happen發生 is,
419
1179758
1254
將會發生的事是,
19:53
we are already已經 in history歷史.
420
1181036
1867
我們已經在歷史中。
19:54
We've我們已經 unleashed如虎添翼 this amazing驚人 creativity創造力
421
1182927
3228
我們已經釋放出了這了不起的創意,
19:58
by de-slaving取消奴隸 us from farming農業
422
1186179
1611
讓我們脫離耕種的奴役,
19:59
and later後來, of course課程, from factory work
423
1187814
3363
當然,之後又脫離了
工廠工作的奴役,
20:03
and have invented發明 so many許多 things.
424
1191201
3162
且發明出了這麼多東西。
20:06
It's going to be even better,
in my opinion意見.
425
1194387
2178
依我所見,將來還會更好。
20:08
And there's going to be
great side effects效果.
426
1196589
2072
將來會有很大的副作用。
20:10
One of the side effects效果 will be
427
1198685
1489
其中一項副作用會是,
20:12
that things like food餐飲 and medical supply供應
and education教育 and shelter庇護
428
1200198
4795
很多東西,比如食物、
醫材、教育、庇護所
20:17
and transportation運輸
429
1205017
1177
以及交通,
20:18
will all become成為 much more
affordable實惠 to all of us,
430
1206218
2441
都會變為大家更負擔得起,
20:20
not just the rich豐富 people.
431
1208683
1322
不只是有錢人的專利。
20:22
CACA: Hmm.
432
1210029
1182
克:嗯。
20:23
So when Martin馬丁 Ford argued爭論, you know,
that this time it's different不同
433
1211235
4341
所以,當馬丁福特主張,
這次會有所不同,
20:27
because the intelligence情報
that we've我們已經 used in the past過去
434
1215600
3453
因為我們在過去用來
20:31
to find new ways方法 to be
435
1219077
2483
找出新方式的智慧,
20:33
will be matched匹配 at the same相同 pace步伐
436
1221584
2279
將會以同樣的速度
20:35
by computers電腦 taking服用 over those things,
437
1223887
2291
被接手那些事的電腦給比過,
20:38
what I hear you saying
is that, not completely全然,
438
1226202
3078
我聽到你說並不完全如此,
20:41
because of human人的 creativity創造力.
439
1229304
2951
因為人類有創意。
20:44
Do you think that that's fundamentally從根本上
different不同 from the kind of creativity創造力
440
1232279
3785
你認為那和電腦能做的那種創意
20:48
that computers電腦 can do?
441
1236088
2696
在根本上是不同的嗎?
20:50
STST: So, that's my firm公司
belief信仰 as an AIAI person --
442
1238808
4434
賽:身為人工智慧人,我堅定相信
20:55
that I haven't沒有 seen看到
any real真實 progress進展 on creativity創造力
443
1243266
3803
我尚未看到任何真正創意上的進展,
20:59
and out-of-the-box盒子外面 thinking思維.
444
1247949
1407
也沒有創造性思維。
21:01
What I see right now -- and this is
really important重要 for people to realize實現,
445
1249380
3623
我現在看到的──
人們很需要了解這一點,
21:05
because the word "artificial人造
intelligence情報" is so threatening危險的,
446
1253027
2903
因為「人工智慧」這詞
深具威脅性的,
史帝芬史匹柏拍了一部電影,
21:07
and then we have Steve史蒂夫 Spielberg斯皮爾伯格
tossing折騰 a movie電影 in,
447
1255954
2523
電影中電腦突然成了我們的主人,
21:10
where all of a sudden突然
the computer電腦 is our overlord霸王,
448
1258501
2413
21:12
but it's really a technology技術.
449
1260938
1452
但它其實只是一項技術,
21:14
It's a technology技術 that helps幫助 us
do repetitive重複 things.
450
1262414
2982
協助我們做重覆工作的技術,
21:17
And the progress進展 has been
entirely完全 on the repetitive重複 end結束.
451
1265420
2913
而進展完全在重覆性方面:
21:20
It's been in legal法律 document文件 discovery發現.
452
1268357
2228
在法律文件探索上有進展,
21:22
It's been contract合同 drafting製圖.
453
1270609
1680
在合約起草上有進展,
21:24
It's been screening篩查 X-raysX射線 of your chest胸部.
454
1272313
4223
在判讀胸腔 X 光上有進展。
21:28
And these things are so specialized專門,
455
1276560
1773
這些工作都很專門化,
21:30
I don't see the big threat威脅 of humanity人性.
456
1278357
2391
我看不出對人類有什麼大威脅。
21:32
In fact事實, we as people --
457
1280772
1794
事實上,我們人類──
21:34
I mean, let's face面對 it:
we've我們已經 become成為 superhuman超人.
458
1282590
2385
我們得承認,我們已經變成超人。
21:36
We've我們已經 made製作 us superhuman超人.
459
1284999
1764
我們已經把自己變成超人。
21:38
We can swim游泳 across橫過
the Atlantic大西洋 in 11 hours小時.
460
1286787
2632
我們可以在 11 小時泳渡大西洋。
21:41
We can take a device設備 out of our pocket口袋
461
1289443
2074
我們能從口袋中拿出一個裝置
21:43
and shout all the way to Australia澳大利亞,
462
1291541
2147
然後對著遙遠的澳洲大吼,
21:45
and in real真實 time, have that person
shouting叫喊 back to us.
463
1293712
2600
而且對方還會即時吼回來。
21:48
That's physically物理 not possible可能.
We're breaking破壞 the rules規則 of physics物理.
464
1296336
3624
在物理上是不可能的,
我們打破了物理的規則。
21:51
When this is said and doneDONE,
we're going to remember記得 everything
465
1299984
2943
說到底,
我們會記得曾經說過和看過的一切,
21:54
we've我們已經 ever said and seen看到,
466
1302951
1213
21:56
you'll你會 remember記得 every一切 person,
467
1304188
1496
你們將會記得每個人,
21:57
which哪一個 is good for me
in my early stages階段 of Alzheimer's老年癡呆症.
468
1305708
2626
對在阿滋海默前期的我是件好事。
抱歉,我剛在說什麼?我忘了。
22:00
Sorry, what was I saying? I forgot忘記.
469
1308358
1677
22:02
CACA: (Laughs)
470
1310059
1578
克:(笑聲)
賽:我們將來可能會有
超過 1,000 的智商。
22:03
STST: We will probably大概 have
an IQ智商 of 1,000 or more.
471
1311661
3077
22:06
There will be no more
spelling拼字 classes for our kids孩子,
472
1314762
3425
我們的孩子將不用再學習拼字,
22:10
because there's no spelling拼字 issue問題 anymore.
473
1318211
2086
因為將不再有拼字問題。
22:12
There's no math數學 issue問題 anymore.
474
1320321
1832
將不再有數學問題。
22:14
And I think what really will happen發生
is that we can be super creative創作的.
475
1322177
3510
我認為會發生的是,
我們會超級有創意。
22:17
And we are. We are creative創作的.
476
1325711
1857
我們是有創意的,
22:19
That's our secret秘密 weapon武器.
477
1327592
1552
那是我們的秘密武器。
22:21
CACA: So the jobs工作 that are getting得到 lost丟失,
478
1329168
2153
克:所以正在消失中的工作,
22:23
in a way, even though雖然
it's going to be painful痛苦,
479
1331345
2494
在某個層面上,即使會很痛苦,
22:25
humans人類 are capable
of more than those jobs工作.
480
1333863
2047
人類的能力是超過這些工作的。
22:27
This is the dream夢想.
481
1335934
1218
這是個夢想。
22:29
The dream夢想 is that humans人類 can rise上升
to just a new level水平 of empowerment權力
482
1337176
4247
夢想是人類可以崛起
爬升到賦能與探索的新層級。
22:33
and discovery發現.
483
1341447
1657
22:35
That's the dream夢想.
484
1343128
1452
這是個夢想。
賽:想想這一點:
22:36
STST: And think about this:
485
1344604
1643
22:38
if you look at the history歷史 of humanity人性,
486
1346271
2021
如果你去看人類的歷史,
22:40
that might威力 be whatever隨你 --
60-100,000 years年份 old, give or take --
487
1348316
3328
也許 6~10 萬年前左右,
22:43
almost幾乎 everything that you cherish珍視
in terms條款 of invention發明,
488
1351668
3726
幾乎你所珍惜的一切,
發明、科技、我們建造的東西,
22:47
of technology技術, of things we've我們已經 built內置,
489
1355418
2151
22:49
has been invented發明 in the last 150 years年份.
490
1357593
3099
都是在最近的 150 年間發明的。
22:53
If you toss折騰 in the book and the wheel,
it's a little bit older舊的.
491
1361756
3048
如果你把書和輪子放進來,
那就古老一些。
22:56
Or the axe斧頭.
492
1364828
1169
或是斧頭。
22:58
But your phone電話, your sneakers球鞋,
493
1366021
2790
但你的手機、你的運動鞋、
23:00
these chairs椅子, modern現代
manufacturing製造業, penicillin青黴素 --
494
1368835
3551
這些椅子、現代工業、盤尼西林──
23:04
the things we cherish珍視.
495
1372410
1714
我們珍視的東西。
23:06
Now, that to me means手段
496
1374148
3658
對我來說,那意味著,
23:09
the next下一個 150 years年份 will find more things.
497
1377830
3041
接下來的 150 年會發現更多東西。
23:12
In fact事實, the pace步伐 of invention發明
has gone走了 up, not gone走了 down, in my opinion意見.
498
1380895
4154
事實上,依我所見,發明的速度
已經變快了,不是變慢。
23:17
I believe only one percent百分 of interesting有趣
things have been invented發明 yet然而. Right?
499
1385073
4905
我相信,我們才只發明了 1%
有趣的東西出來。對吧?
23:22
We haven't沒有 cured治愈 cancer癌症.
500
1390002
1988
我們還沒有治癒癌症。
23:24
We don't have flying飛行 cars汽車 -- yet然而.
Hopefully希望, I'll change更改 this.
501
1392014
3718
我們沒有飛天車,還沒有。
希望我能改變這一點。
23:27
That used to be an example
people laughed笑了 about. (Laughs)
502
1395756
3257
以前那是個會讓人發笑的例子。
(笑聲)
23:31
It's funny滑稽, isn't it?
Working加工 secretly偷偷 on flying飛行 cars汽車.
503
1399037
2992
很有趣,是吧?
暗地裡致力發明飛天車。
23:34
We don't live生活 twice兩次 as long yet然而. OK?
504
1402053
2683
我們的壽命還沒到兩倍長。是嗎?
23:36
We don't have this magic魔法
implant注入 in our brain
505
1404760
2785
我們在大腦中還沒有神奇的植入物
23:39
that gives us the information信息 we want.
506
1407569
1832
能給予我們想要的資訊。
23:41
And you might威力 be appalled by it,
507
1409425
1526
你可能會覺得它很駭人,
23:42
but I promise諾言 you,
once一旦 you have it, you'll你會 love it.
508
1410975
2444
但我保證,一旦你有了它
就會愛上它。
23:45
I hope希望 you will.
509
1413443
1166
我希望你會。
23:46
It's a bit scary害怕, I know.
510
1414633
1909
它有點可怕,我知道。
還有好多我認為我們應該
發明的東西還沒被發明出來。
23:48
There are so many許多 things
we haven't沒有 invented發明 yet然而
511
1416566
2254
23:50
that I think we'll invent發明.
512
1418844
1268
我們沒有重力保護罩。
23:52
We have no gravity重力 shields盾牌.
513
1420136
1306
我們無法把自己從一地
用光束傳送到另一地。
23:53
We can't beam光束 ourselves我們自己
from one location位置 to another另一個.
514
1421466
2553
那聽起來很荒謬,
23:56
That sounds聲音 ridiculous荒謬,
515
1424043
1151
但大約 200 年前,
23:57
but about 200 years年份 ago,
516
1425218
1288
23:58
experts專家 were of the opinion意見
that flight飛行 wouldn't不會 exist存在,
517
1426530
2667
專家認為飛機不會存在,
24:01
even 120 years年份 ago,
518
1429221
1324
甚至 120 年前。
24:02
and if you moved移動 faster更快
than you could run,
519
1430569
2582
還有認為如果你移動速度
比你的跑步速度快,
24:05
you would instantly即刻 die.
520
1433175
1520
你就會馬上死掉。
24:06
So who says we are correct正確 today今天
that you can't beam光束 a person
521
1434719
3569
所以現在誰敢肯定說
我們不能把一個人用光束
24:10
from here to Mars火星?
522
1438312
2249
從這裡傳送到火星?
24:12
CACA: Sebastian塞巴斯蒂安, thank you so much
523
1440585
1569
克:賽巴斯汀,非常謝謝你
24:14
for your incredibly令人難以置信 inspiring鼓舞人心 vision視力
and your brilliance.
524
1442178
2682
分享啟發靈感的遠景和你的智慧。
24:16
Thank you, Sebastian塞巴斯蒂安 Thrun史朗.
525
1444884
1323
謝謝你,賽巴斯汀索朗。
24:18
STST: That was fantastic奇妙. (Applause掌聲)
526
1446231
1895
賽:很棒的經驗。(掌聲)
Translated by Lilian Chiu
Reviewed by Yang Jasson

▲Back to top

ABOUT THE SPEAKERS
Sebastian Thrun - Educator, entrepreneur
Sebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us.

Why you should listen

Sebastian Thrun is an educator, entrepreneur and troublemaker. After a long life as a professor at Stanford University, Thrun resigned from tenure to join Google. At Google, he founded Google X, home to self-driving cars and many other moonshot technologies. Thrun also founded Udacity, an online university with worldwide reach, and Kitty Hawk, a "flying car" company. He has authored 11 books, 400 papers, holds 3 doctorates and has won numerous awards.

More profile about the speaker
Sebastian Thrun | Speaker | TED.com
Chris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading.

Why you should listen

Chris Anderson is the Curator of TED, a nonprofit devoted to sharing valuable ideas, primarily through the medium of 'TED Talks' -- short talks that are offered free online to a global audience.

Chris was born in a remote village in Pakistan in 1957. He spent his early years in India, Pakistan and Afghanistan, where his parents worked as medical missionaries, and he attended an American school in the Himalayas for his early education. After boarding school in Bath, England, he went on to Oxford University, graduating in 1978 with a degree in philosophy, politics and economics.

Chris then trained as a journalist, working in newspapers and radio, including two years producing a world news service in the Seychelles Islands.

Back in the UK in 1984, Chris was captivated by the personal computer revolution and became an editor at one of the UK's early computer magazines. A year later he founded Future Publishing with a $25,000 bank loan. The new company initially focused on specialist computer publications but eventually expanded into other areas such as cycling, music, video games, technology and design, doubling in size every year for seven years. In 1994, Chris moved to the United States where he built Imagine Media, publisher of Business 2.0 magazine and creator of the popular video game users website IGN. Chris eventually merged Imagine and Future, taking the combined entity public in London in 1999, under the Future name. At its peak, it published 150 magazines and websites and employed 2,000 people.

This success allowed Chris to create a private nonprofit organization, the Sapling Foundation, with the hope of finding new ways to tackle tough global issues through media, technology, entrepreneurship and, most of all, ideas. In 2001, the foundation acquired the TED Conference, then an annual meeting of luminaries in the fields of Technology, Entertainment and Design held in Monterey, California, and Chris left Future to work full time on TED.

He expanded the conference's remit to cover all topics, including science, business and key global issues, while adding a Fellows program, which now has some 300 alumni, and the TED Prize, which grants its recipients "one wish to change the world." The TED stage has become a place for thinkers and doers from all fields to share their ideas and their work, capturing imaginations, sparking conversation and encouraging discovery along the way.

In 2006, TED experimented with posting some of its talks on the Internet. Their viral success encouraged Chris to begin positioning the organization as a global media initiative devoted to 'ideas worth spreading,' part of a new era of information dissemination using the power of online video. In June 2015, the organization posted its 2,000th talk online. The talks are free to view, and they have been translated into more than 100 languages with the help of volunteers from around the world. Viewership has grown to approximately one billion views per year.

Continuing a strategy of 'radical openness,' in 2009 Chris introduced the TEDx initiative, allowing free licenses to local organizers who wished to organize their own TED-like events. More than 8,000 such events have been held, generating an archive of 60,000 TEDx talks. And three years later, the TED-Ed program was launched, offering free educational videos and tools to students and teachers.

More profile about the speaker
Chris Anderson | Speaker | TED.com