Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,000 --> 00:00:02,960
(trumpet jazz music)
2
00:00:04,000 --> 00:00:10,074
Advertise your product or brand here
contact www.OpenSubtitles.org today
3
00:00:12,320 --> 00:00:14,176
- The flaming trumpet.
4
00:00:14,200 --> 00:00:16,496
Once when I was in junior high school,
5
00:00:16,520 --> 00:00:19,456
preparing to perform in
the school talent show,
6
00:00:19,480 --> 00:00:22,096
my father asked me if I thought
it would be a good effect
7
00:00:22,120 --> 00:00:25,336
to have fire shooting out
of the trumpet as I played.
8
00:00:25,360 --> 00:00:27,896
(trumpet jazz music)
9
00:00:27,920 --> 00:00:28,897
Of course, I had to admit
10
00:00:28,921 --> 00:00:31,600
that that would indeed be very impressive.
11
00:00:32,480 --> 00:00:35,536
He then went into his workshop
and began tinkering around.
12
00:00:35,560 --> 00:00:39,176
(trumpet jazz music)
13
00:00:39,200 --> 00:00:42,656
And by the very next day,
I found myself holding
14
00:00:42,680 --> 00:00:42,697
in my hands a flaming trumpet.
And by the very next day,
I found myself holding
15
00:00:42,721 --> 00:00:46,096
in my hands a flaming trumpet.
16
00:00:46,120 --> 00:00:49,240
(cheerful piano music)
17
00:01:40,480 --> 00:01:41,257
- Who's Claude Shannon?
18
00:01:41,281 --> 00:01:42,416
- Who's Claude Shannon?
19
00:01:42,440 --> 00:01:43,217
- Claude Shannon?
20
00:01:43,241 --> 00:01:44,496
- You're asking me this question?
21
00:01:44,520 --> 00:01:47,736
- I often go back to this...
22
00:01:47,760 --> 00:01:47,777
Well, you've probably heard the word bit.
- I often go back to this...
23
00:01:47,801 --> 00:01:49,856
Well, you've probably heard the word bit.
24
00:01:49,880 --> 00:01:49,897
- We live in the information age
Well, you've probably heard the word bit.
25
00:01:49,921 --> 00:01:51,696
- We live in the information age
26
00:01:51,720 --> 00:01:53,776
and he's the father of
the information theory.
27
00:01:53,800 --> 00:01:56,736
- Doesn't everybody know
who Claude Shannon is?
28
00:01:56,760 --> 00:01:56,777
- If you value the fact
that your iPhone can contact
- Doesn't everybody know
who Claude Shannon is?
29
00:01:56,801 --> 00:02:01,256
- If you value the fact
that your iPhone can contact
30
00:02:01,280 --> 00:02:05,176
and call for you an Uber,
and then place a phone call,
31
00:02:05,200 --> 00:02:08,536
and then text a friend,
and then check your email,
32
00:02:08,560 --> 00:02:12,776
all of those things you owe to
Claude Shannon and his work.
33
00:02:12,800 --> 00:02:14,536
- I truly believe he's at the same level
34
00:02:14,560 --> 00:02:14,577
as Newton or Einstein.
- I truly believe he's at the same level
35
00:02:14,601 --> 00:02:16,136
as Newton or Einstein.
36
00:02:16,160 --> 00:02:17,696
- Why don't people know about him?
37
00:02:17,720 --> 00:02:19,576
They know about Alexander Graham Bell,
38
00:02:19,600 --> 00:02:19,617
they know about Henry Ford,
but so few people have heard
They know about Alexander Graham Bell,
39
00:02:19,641 --> 00:02:22,696
they know about Henry Ford,
but so few people have heard
40
00:02:22,720 --> 00:02:26,536
of Claude Shannon, and his
impact is bigger than all
41
00:02:26,560 --> 00:02:28,376
of the household names.
42
00:02:28,400 --> 00:02:29,856
- I didn't know that that was one person,
43
00:02:29,880 --> 00:02:31,136
and whoever that one person was
44
00:02:31,160 --> 00:02:33,096
is probably the most
important person in our time.
45
00:02:33,120 --> 00:02:34,456
- What would the world look like
46
00:02:34,480 --> 00:02:36,560
if Claude Shannon hadn't done what he did?
47
00:02:39,960 --> 00:02:41,896
(piano key boings)
48
00:02:41,920 --> 00:02:42,976
(gentle instrumental music)
49
00:02:43,000 --> 00:02:43,017
- By the 1980s, the revolution
(gentle instrumental music)
50
00:02:43,041 --> 00:02:44,976
- By the 1980s, the revolution
51
00:02:45,000 --> 00:02:45,017
that he started is in full swing,
- By the 1980s, the revolution
52
00:02:45,041 --> 00:02:48,216
that he started is in full swing,
53
00:02:48,240 --> 00:02:50,680
but he is nowhere to be found.
54
00:02:51,880 --> 00:02:53,296
He's a founding father,
55
00:02:53,320 --> 00:02:55,776
but he's a distant historical figure,
56
00:02:55,800 --> 00:02:57,480
sort of like George Washington.
57
00:02:58,840 --> 00:03:00,800
And nobody really knows what he's doing.
58
00:03:01,600 --> 00:03:05,120
(gentle instrumental music)
59
00:03:39,560 --> 00:03:40,337
(Claude clears throat)
60
00:03:40,361 --> 00:03:41,496
- Is this where
you want me to sit?
61
00:03:41,520 --> 00:03:42,760
- Yes, please.
62
00:03:44,480 --> 00:03:45,560
Let's see...
63
00:03:49,920 --> 00:03:51,296
- Here we go.
64
00:03:51,320 --> 00:03:52,656
- Thank you again for agreeing
65
00:03:52,680 --> 00:03:52,697
to see me, Dr. Shannon.
- Thank you again for agreeing
66
00:03:52,721 --> 00:03:54,160
to see me, Dr. Shannon.
67
00:03:55,160 --> 00:03:57,016
This is an amazing room.
68
00:03:57,040 --> 00:04:01,840
- I guess it must seem
pretty funny a grown man
69
00:04:01,920 --> 00:04:04,776
has a room full of toys. (chuckles)
70
00:04:04,800 --> 00:04:04,817
My life's work. (laughs)
has a room full of toys. (chuckles)
71
00:04:04,841 --> 00:04:06,496
My life's work. (laughs)
72
00:04:06,520 --> 00:04:08,056
- I don't
think you need to worry
73
00:04:08,080 --> 00:04:09,816
about just being remembered for your toys.
74
00:04:09,840 --> 00:04:13,216
- Claude doesn't overly worry
about how he'll be remembered.
75
00:04:13,240 --> 00:04:13,255
- No.
- Claude doesn't overly worry
about how he'll be remembered.
76
00:04:13,279 --> 00:04:14,057
- No.
77
00:04:14,081 --> 00:04:16,616
How'd you uncover me again, Michelle?
78
00:04:16,640 --> 00:04:18,616
- As I said
on the phone, I'm working
79
00:04:18,640 --> 00:04:22,016
on a project about the
founders of the modern age,
80
00:04:22,040 --> 00:04:25,136
and I kept seeing over and over again,
81
00:04:25,160 --> 00:04:28,776
"Claude Shannon, the father
of Information Theory."
82
00:04:28,800 --> 00:04:30,256
(Claude chuckles)
83
00:04:30,280 --> 00:04:33,040
But nobody seemed to know
what you're doing now.
84
00:04:34,280 --> 00:04:37,456
Your 1948 paper was a blueprint
85
00:04:37,480 --> 00:04:40,096
for what people are calling
The Information Age.
86
00:04:40,120 --> 00:04:40,137
- Well, I wrote that nearly 40 years ago.
for what people are calling
The Information Age.
87
00:04:40,161 --> 00:04:42,496
- Well, I wrote that nearly 40 years ago.
88
00:04:42,520 --> 00:04:43,496
- Yes.
89
00:04:43,520 --> 00:04:43,537
In a sense you saw the future.
- Yes.
90
00:04:43,561 --> 00:04:45,736
In a sense you saw the future.
91
00:04:45,760 --> 00:04:46,760
How did you do that?
92
00:04:47,560 --> 00:04:48,936
(Betty chuckles)
93
00:04:48,960 --> 00:04:48,977
(Claude chuckles)
(Betty chuckles)
94
00:04:49,001 --> 00:04:50,336
(Claude chuckles)
95
00:04:50,360 --> 00:04:52,096
- How long do you plan on being here?
96
00:04:52,120 --> 00:04:52,135
(Claude and Betty laugh)
- How long do you plan on being here?
97
00:04:52,159 --> 00:04:52,897
(Claude and Betty laugh)
98
00:04:52,921 --> 00:04:54,456
- Maybe I should get some coffee.
99
00:04:54,480 --> 00:04:55,257
- Oh, good.
100
00:04:55,281 --> 00:04:57,816
(Claude and Betty laugh)
101
00:04:57,840 --> 00:05:02,600
The fact is, Michelle,
that my mind wanders around
102
00:05:07,120 --> 00:05:09,680
and conceives of different
things day and night.
103
00:05:11,800 --> 00:05:13,960
Well, like a science
fiction writer, I think,
104
00:05:15,240 --> 00:05:16,680
"What if it were like this?"
105
00:05:19,160 --> 00:05:20,416
And I never particularly cared
106
00:05:20,440 --> 00:05:25,240
whether anybody was interested
in it or wanted to work on it
107
00:05:27,240 --> 00:05:28,536
or thought I was right.
108
00:05:28,560 --> 00:05:28,577
(dramatic music)
or thought I was right.
109
00:05:28,601 --> 00:05:29,936
(dramatic music)
110
00:05:29,960 --> 00:05:33,896
I was always interested in
communication things, whether
111
00:05:33,920 --> 00:05:33,937
or not people thought that's
what I should be working on.
I was always interested in
communication things, whether
112
00:05:33,961 --> 00:05:36,976
or not people thought that's
what I should be working on.
113
00:05:37,000 --> 00:05:37,017
(dramatic music)
or not people thought that's
what I should be working on.
114
00:05:37,041 --> 00:05:39,680
(dramatic music)
115
00:05:42,400 --> 00:05:44,976
(man shouts)
116
00:05:45,000 --> 00:05:46,496
- Shannon
thought about the problem
117
00:05:46,520 --> 00:05:49,896
of communication in a way that nobody else
118
00:05:49,920 --> 00:05:50,960
had ever thought of.
119
00:05:54,160 --> 00:05:56,616
The basic problem is
how do you get a message
120
00:05:56,640 --> 00:05:56,657
from one place to another?
The basic problem is
how do you get a message
121
00:05:56,681 --> 00:05:58,240
from one place to another?
122
00:06:00,760 --> 00:06:03,816
The challenge was always how to go faster,
123
00:06:03,840 --> 00:06:06,696
how to go farther, and
how to be more accurate.
124
00:06:06,720 --> 00:06:08,656
(horse neighs)
125
00:06:08,680 --> 00:06:11,376
The huge advance was the discovery
126
00:06:11,400 --> 00:06:14,200
of how to use electricity to communicate.
127
00:06:15,960 --> 00:06:17,376
(phone ringing)
128
00:06:17,400 --> 00:06:18,616
- Hello?
129
00:06:18,640 --> 00:06:18,655
Hello? - Hello?
130
00:06:18,679 --> 00:06:19,416
Hello?
131
00:06:19,440 --> 00:06:19,457
- But initially it was very unreliable,
Hello?
132
00:06:19,481 --> 00:06:21,880
- But initially it was very unreliable,
133
00:06:24,280 --> 00:06:27,080
and efforts to improve
things were just haphazard.
134
00:06:28,560 --> 00:06:31,816
And then, Shannon came
along and asked a question
135
00:06:31,840 --> 00:06:33,800
that hadn't even occurred to anyone,
136
00:06:34,920 --> 00:06:38,696
"Could you construct a universal
theory of communication
137
00:06:38,720 --> 00:06:41,400
"that could allow perfect communication?"
138
00:06:43,520 --> 00:06:45,360
He drew a very simple diagram.
139
00:06:47,680 --> 00:06:49,640
You have some information you wanna send
140
00:06:50,720 --> 00:06:53,896
and you have to convert it
into something like sound waves
141
00:06:53,920 --> 00:06:56,896
or electrical pulses or
light that can be sent
142
00:06:56,920 --> 00:07:00,480
over a channel or medium,
like air or wires.
143
00:07:01,680 --> 00:07:04,136
There's things that can disrupt
the signal on that medium,
144
00:07:04,160 --> 00:07:04,177
like static or smoke or noise
There's things that can disrupt
the signal on that medium,
145
00:07:04,201 --> 00:07:07,096
like static or smoke or noise
146
00:07:07,120 --> 00:07:07,137
before it gets to the receiver.
Like static or smoke or noise
147
00:07:07,161 --> 00:07:09,456
before it gets to the receiver.
148
00:07:09,480 --> 00:07:11,416
And then the receiver
converts the signal back
149
00:07:11,440 --> 00:07:11,457
into something that's
hopefully really close
And then the receiver
converts the signal back
150
00:07:11,481 --> 00:07:14,896
into something that's
hopefully really close
151
00:07:14,920 --> 00:07:16,360
to the original message.
152
00:07:18,720 --> 00:07:22,856
The diagram looked trivial,
but it allowed Shannon
153
00:07:22,880 --> 00:07:25,376
to break the problem of communication down
154
00:07:25,400 --> 00:07:29,096
into its essential parts and
create a universal theory
155
00:07:29,120 --> 00:07:31,560
that you can apply to anything you want.
156
00:07:32,720 --> 00:07:36,096
- This is a calculator I
built to do calculations
157
00:07:36,120 --> 00:07:37,216
with Roman numerals.
158
00:07:37,240 --> 00:07:38,736
I call it THROBAC:
159
00:07:38,760 --> 00:07:38,777
Thrifty Roman Numeral
Backward Looking Computer.
I call it THROBAC:
160
00:07:38,801 --> 00:07:41,720
Thrifty Roman Numeral
Backward Looking Computer.
161
00:07:45,600 --> 00:07:48,296
- Did you make that here at MIT?
162
00:07:48,320 --> 00:07:48,337
- No, I made it when I
was at Bell Laboratories
- Did you
make that here at MIT?
163
00:07:48,361 --> 00:07:51,776
- No, I made it when I
was at Bell Laboratories
164
00:07:51,800 --> 00:07:54,336
when I should've been working
on communication things.
165
00:07:54,360 --> 00:07:56,456
- It didn't have a big commercial market.
166
00:07:56,480 --> 00:07:59,696
(Claude chuckles)
167
00:07:59,720 --> 00:08:01,136
- There are some important things
168
00:08:01,160 --> 00:08:03,016
that made Claude Shannon the thinker
169
00:08:03,040 --> 00:08:03,057
and the mind that he was.
That made Claude Shannon the thinker
170
00:08:03,081 --> 00:08:04,736
and the mind that he was.
171
00:08:04,760 --> 00:08:06,256
I believe it was important that he grew up
172
00:08:06,280 --> 00:08:08,576
in Gaylord, Michigan, I
think it was important
173
00:08:08,600 --> 00:08:10,296
that he came of age in the Midwest,
174
00:08:10,320 --> 00:08:10,337
and I think it was equally
important that he came of age
that he came of age in the Midwest,
175
00:08:10,361 --> 00:08:12,896
and I think it was equally
important that he came of age
176
00:08:12,920 --> 00:08:16,856
in a town that built things,
where a boy like him could,
177
00:08:16,880 --> 00:08:19,656
with his best friend, go
to his friend's backyard
178
00:08:19,680 --> 00:08:21,480
and build an elevator in a barn.
179
00:08:24,040 --> 00:08:25,456
- My mother was a principal
180
00:08:25,480 --> 00:08:25,497
at the high school.
- My mother was a principal
181
00:08:25,521 --> 00:08:28,136
at the high school.
182
00:08:28,160 --> 00:08:30,376
My father had a bunch of jobs.
183
00:08:30,400 --> 00:08:31,600
He had a furniture store.
184
00:08:33,080 --> 00:08:34,480
He was a probate judge.
185
00:08:36,360 --> 00:08:38,480
And he was also an undertaker.
186
00:08:42,600 --> 00:08:43,880
He got me an erector set.
187
00:08:44,920 --> 00:08:46,800
It's still one of my favorite toys.
188
00:08:49,880 --> 00:08:52,696
My grandfather was a very
interesting character,
189
00:08:52,720 --> 00:08:54,200
he was always inventing things.
190
00:08:55,760 --> 00:08:57,976
He was intent on determining if the turn
191
00:08:58,000 --> 00:09:01,160
of the century was 1900 or 1901.
192
00:09:02,720 --> 00:09:04,896
I was very impressed he spent so much time
193
00:09:04,920 --> 00:09:07,480
on something most people
thought was totally useless.
194
00:09:12,440 --> 00:09:15,760
(cheerful pianola music)
195
00:09:43,280 --> 00:09:45,216
- So, would you
say your first experiment
196
00:09:45,240 --> 00:09:47,976
in communication was a failure?
197
00:09:48,000 --> 00:09:50,336
- No, but
there was a lot of noise
198
00:09:50,360 --> 00:09:50,377
on the line.
- No, but
there was a lot of noise
199
00:09:50,401 --> 00:09:51,936
on the line.
200
00:09:51,960 --> 00:09:54,256
- Motivation to
come up with something better?
201
00:09:54,280 --> 00:09:54,297
- A little error could turn
- Motivation to
come up with something better?
202
00:09:54,321 --> 00:09:55,976
- A little error could turn
203
00:09:56,000 --> 00:09:59,736
"Do you like me?" into "Do you hear me?"
204
00:09:59,760 --> 00:09:59,777
And which she was answering
was very important to me.
"Do you like me?" into "Do you hear me?"
205
00:09:59,801 --> 00:10:03,536
And which she was answering
was very important to me.
206
00:10:03,560 --> 00:10:05,456
- So, you used Morse code?
207
00:10:05,480 --> 00:10:06,257
- Oh, you bet.
208
00:10:06,281 --> 00:10:09,056
Morse already invented the binary code,
209
00:10:09,080 --> 00:10:09,097
with dots and dashes, and he
had the great idea of going
Morse already invented the binary code,
210
00:10:09,121 --> 00:10:12,176
with dots and dashes, and he
had the great idea of going
211
00:10:12,200 --> 00:10:12,217
to a print shop to find out
which letters we use the most.
with dots and dashes, and he
had the great idea of going
212
00:10:12,241 --> 00:10:16,376
to a print shop to find out
which letters we use the most.
213
00:10:16,400 --> 00:10:16,417
E, the most common letter,
gets the shortest code, dot.
to a print shop to find out
which letters we use the most.
214
00:10:16,441 --> 00:10:19,800
E, the most common letter,
gets the shortest code, dot.
215
00:10:21,000 --> 00:10:25,800
T, the second most common,
gets the dash, and so on.
216
00:10:26,480 --> 00:10:28,096
Exactly the sort of things I was thinking
217
00:10:28,120 --> 00:10:29,280
about 100 years later.
218
00:10:30,320 --> 00:10:31,536
- He had what we would probably call now
219
00:10:31,560 --> 00:10:32,416
a free range childhood,
220
00:10:32,440 --> 00:10:32,455
even though there wasn't
really a term for it back
a free range childhood,
221
00:10:32,479 --> 00:10:34,216
even though there wasn't
really a term for it back
222
00:10:34,240 --> 00:10:36,656
in the 1910s and '20s;
but these were things
223
00:10:36,680 --> 00:10:39,656
like rigging up a barbed wire telegraph
224
00:10:39,680 --> 00:10:39,695
to connect his house
to his neighbors house
like rigging up a barbed wire telegraph
225
00:10:39,719 --> 00:10:41,136
to connect his house
to his neighbors house
226
00:10:41,160 --> 00:10:43,856
along the fences, things like
learning wigwag signaling,
227
00:10:43,880 --> 00:10:46,016
which is the idea of sending
Morse code with flags
228
00:10:46,040 --> 00:10:46,057
so that your body kind of becomes a sort
which is the idea of sending
Morse code with flags
229
00:10:46,081 --> 00:10:48,216
so that your body kind of becomes a sort
230
00:10:48,240 --> 00:10:49,520
of Morse code telegraph,
231
00:10:51,080 --> 00:10:52,616
things like solving math puzzles.
232
00:10:52,640 --> 00:10:53,896
Shannon's favorite story growing up
233
00:10:53,920 --> 00:10:55,856
was Edgar Allan Poe's "The Goldbug."
234
00:10:55,880 --> 00:10:59,240
(gentle clavichord music)
235
00:11:00,280 --> 00:11:04,280
- Poe loved codes and
cryptography, just like me.
236
00:11:05,080 --> 00:11:06,496
He even invited his readers
237
00:11:06,520 --> 00:11:06,537
to send him their own
cryptograms that he'd solve.
He even invited his readers
238
00:11:06,561 --> 00:11:09,800
to send him their own
cryptograms that he'd solve.
239
00:11:14,360 --> 00:11:17,816
In "The Goldbug," a man
finds a scrap of paper
240
00:11:17,840 --> 00:11:21,096
with a coded message that
tells where the pirate,
241
00:11:21,120 --> 00:11:23,616
Captain Kidd, buried his treasure.
242
00:11:23,640 --> 00:11:27,160
(gentle instrumental music)
243
00:11:32,280 --> 00:11:34,816
The message is written
in invisible ink though
244
00:11:34,840 --> 00:11:34,857
and only appears when
it's held up to a flame.
The message is written
in invisible ink though
245
00:11:34,881 --> 00:11:37,720
and only appears when
it's held up to a flame.
246
00:11:42,080 --> 00:11:45,656
Poe shows how you can break the code
247
00:11:45,680 --> 00:11:48,616
by connecting the most used symbols
248
00:11:48,640 --> 00:11:50,880
to the most used letters in English.
249
00:11:51,760 --> 00:11:55,280
(gentle instrumental music)
250
00:11:59,640 --> 00:12:02,280
It's just like Morse did, in reverse.
251
00:12:04,640 --> 00:12:07,696
Just like Alan Turing
did, too, during the war,
252
00:12:07,720 --> 00:12:07,737
looking for patterns to
break the German codes.
Just like Alan Turing
did, too, during the war,
253
00:12:07,761 --> 00:12:10,640
looking for patterns to
break the German codes.
254
00:12:12,360 --> 00:12:15,720
Finding patterns, that's the key.
255
00:12:21,480 --> 00:12:25,080
- The foremost characteristic
of my father was playfulness.
256
00:12:26,200 --> 00:12:27,680
Play infused everything he did.
257
00:12:29,720 --> 00:12:32,696
And you can see that in
all the machines he made,
258
00:12:32,720 --> 00:12:34,400
like, who makes a flaming trumpet?
259
00:12:35,600 --> 00:12:38,016
Making a chair lift to go
down to lake and come back,
260
00:12:38,040 --> 00:12:40,720
that was a playful thing that
he did with Marvin Minsky.
261
00:12:42,840 --> 00:12:44,040
Nothing was sacred.
262
00:12:46,560 --> 00:12:49,296
My brother, Andrew, had an idea
for a rocket-powered Frisbee
263
00:12:49,320 --> 00:12:49,337
and my father loved that.
My brother, Andrew, had an idea
for a rocket-powered Frisbee
264
00:12:49,361 --> 00:12:51,016
and my father loved that.
265
00:12:51,040 --> 00:12:53,760
(Frisbee buzzing)
266
00:12:55,240 --> 00:12:59,176
If we had a tray of coffee cake, somebody,
267
00:12:59,200 --> 00:13:00,336
if they were cutting a piece,
268
00:13:00,360 --> 00:13:02,016
they usually wouldn't
cut a piece off the end,
269
00:13:02,040 --> 00:13:04,040
they would cut a piece out of the middle.
270
00:13:05,360 --> 00:13:07,456
And they might cut it out
in an interesting shape,
271
00:13:07,480 --> 00:13:08,736
it wouldn't just be a square.
272
00:13:08,760 --> 00:13:08,777
It might be an octagon,
it might be a pentagon,
it wouldn't just be a square.
273
00:13:08,801 --> 00:13:10,896
It might be an octagon,
it might be a pentagon,
274
00:13:10,920 --> 00:13:11,920
something like that.
275
00:13:14,680 --> 00:13:16,696
He tried to fool with
everything a little bit
276
00:13:16,720 --> 00:13:16,737
and see if he could make it more fun.
He tried to fool with
everything a little bit
277
00:13:16,761 --> 00:13:18,616
and see if he could make it more fun.
278
00:13:18,640 --> 00:13:21,776
- I was a big fan of "Alice in Wonderland"
279
00:13:21,800 --> 00:13:23,856
and "Through the Looking-Glass."
280
00:13:23,880 --> 00:13:26,240
You know, Lewis Carroll
was a mathematician, too.
281
00:13:27,800 --> 00:13:28,577
(rabbit sneezes)
282
00:13:28,601 --> 00:13:33,336
Wonderland is full of word games and codes
283
00:13:33,360 --> 00:13:35,400
and mathematical ideas.
284
00:13:38,000 --> 00:13:41,776
"Through the Looking" is
almost like a big chess game
285
00:13:41,800 --> 00:13:44,336
with crazy rules of logic.
286
00:13:44,360 --> 00:13:46,696
- Stop, stop, stop,
stop, stop, stop, stop.
287
00:13:46,720 --> 00:13:47,497
- Carroll liked
288
00:13:47,521 --> 00:13:49,256
to invent funny word puzzles
289
00:13:49,280 --> 00:13:52,000
that you could solve with mathematics.
290
00:13:54,640 --> 00:13:57,056
He used something called Boolean algebra
291
00:13:57,080 --> 00:14:00,096
to convert the words into equations.
292
00:14:00,120 --> 00:14:02,696
The only numbers in
Boolean algebra are zero
293
00:14:02,720 --> 00:14:06,360
and one, nothing and everything.
294
00:14:07,760 --> 00:14:10,056
And you could reach a conclusion by adding
295
00:14:10,080 --> 00:14:10,097
and multiplying the equations.
And you could reach a conclusion by adding
296
00:14:10,121 --> 00:14:11,920
and multiplying the equations.
297
00:14:13,920 --> 00:14:16,256
- So, what's the answer?
298
00:14:16,280 --> 00:14:18,720
- Babies cannot manage crocodiles.
299
00:14:20,160 --> 00:14:23,136
- He couldn't do mental
arithmetic any better
300
00:14:23,160 --> 00:14:23,177
than the average person.
- He couldn't do mental
arithmetic any better
301
00:14:23,201 --> 00:14:24,896
than the average person.
302
00:14:24,920 --> 00:14:27,816
And if he had two numbers
to multiply, 26 times 42,
303
00:14:27,840 --> 00:14:30,336
he'd go on the blackboard, 26, 42,
304
00:14:30,360 --> 00:14:34,256
and do long multiplication, which was...
305
00:14:34,280 --> 00:14:37,816
You'd think he'd just be,
you know, a numbers genius.
306
00:14:37,840 --> 00:14:40,576
No, he was much bigger than that.
307
00:14:40,600 --> 00:14:43,376
- I got two Bachelor's
degrees at Michigan.
308
00:14:43,400 --> 00:14:46,760
One in electrical engineering,
the other in mathematics.
309
00:14:48,440 --> 00:14:51,496
I really wasn't sure what to do at all,
310
00:14:51,520 --> 00:14:54,056
and I saw a little postcard on the wall
311
00:14:54,080 --> 00:14:56,456
that said MIT was looking
for a research assistant
312
00:14:56,480 --> 00:14:56,497
to run a machine that
could solve equations.
that said MIT was looking
for a research assistant
313
00:14:56,521 --> 00:14:59,696
to run a machine that
could solve equations.
314
00:14:59,720 --> 00:15:02,576
They called it The Thinking Machine.
315
00:15:02,600 --> 00:15:04,176
That sounded interesting.
316
00:15:04,200 --> 00:15:06,800
Vannevar Bush (chuckles)
317
00:15:08,640 --> 00:15:12,096
right over there, he invented it.
318
00:15:12,120 --> 00:15:15,376
He was pretty famous in the
world of engineering by then,
319
00:15:15,400 --> 00:15:20,200
so I applied and spent the
next four years at MIT.
320
00:15:21,920 --> 00:15:26,136
(intense instrumental music)
321
00:15:26,160 --> 00:15:26,177
- The electromechanical
Differential Analyzer
(intense instrumental music)
322
00:15:26,201 --> 00:15:28,696
- The electromechanical
Differential Analyzer
323
00:15:28,720 --> 00:15:28,737
performs scientific and industrial marvels
- The electromechanical
Differential Analyzer
324
00:15:28,761 --> 00:15:31,256
performs scientific and industrial marvels
325
00:15:31,280 --> 00:15:34,056
at the Massachusetts
Institute of Technology.
326
00:15:34,080 --> 00:15:37,256
Perforated tape introduces the
data to the 100-ton marvel,
327
00:15:37,280 --> 00:15:37,297
which solves problems
Perforated tape introduces the
data to the 100-ton marvel,
328
00:15:37,321 --> 00:15:38,496
which solves problems
329
00:15:38,520 --> 00:15:41,256
beyond the ordinary
methods of calculation.
330
00:15:41,280 --> 00:15:42,216
- The Thinking Machine,
331
00:15:42,240 --> 00:15:42,257
which officially was called
the Differential Analyzer,
- The Thinking Machine,
332
00:15:42,281 --> 00:15:46,816
which officially was called
the Differential Analyzer,
333
00:15:46,840 --> 00:15:51,616
was the smartest calculating
machine of the time.
334
00:15:51,640 --> 00:15:55,896
All sorts of gears and rotating
shafts and wheels and so on,
335
00:15:55,920 --> 00:15:57,720
it must've weighed 100 tons.
336
00:15:58,600 --> 00:16:02,176
My job was to set the different wheels
337
00:16:02,200 --> 00:16:04,440
and things to solve equations.
338
00:16:06,080 --> 00:16:10,096
The heart of the machine
though was electric switches,
339
00:16:10,120 --> 00:16:10,137
just like the switches you find
The heart of the machine
though was electric switches,
340
00:16:10,161 --> 00:16:11,376
just like the switches you find
341
00:16:11,400 --> 00:16:13,960
in a telephone network to route calls.
342
00:16:15,400 --> 00:16:16,736
It occurred to me,
343
00:16:16,760 --> 00:16:16,777
because I'd taken the Boolean
algebra course in Michigan,
It occurred to me,
344
00:16:16,801 --> 00:16:21,560
because I'd taken the Boolean
algebra course in Michigan,
345
00:16:24,640 --> 00:16:26,136
that there was a beautiful parallel
346
00:16:26,160 --> 00:16:30,960
between electric switches,
which are just on or off,
347
00:16:35,200 --> 00:16:39,016
and Boolean algebra
which reduced everything
348
00:16:39,040 --> 00:16:40,136
to true and false.
349
00:16:40,160 --> 00:16:42,800
I used one for true and zero for false.
350
00:16:44,240 --> 00:16:45,696
I worked out the mathematics,
351
00:16:45,720 --> 00:16:47,856
which had some funny things like,
352
00:16:47,880 --> 00:16:47,897
if a true statement plus a true statement
which had some funny things like,
353
00:16:47,921 --> 00:16:50,896
if a true statement plus a true statement
354
00:16:50,920 --> 00:16:55,680
equals true statement, that
means one plus one equals one.
355
00:16:58,480 --> 00:17:01,016
I wrote it all up for my Master's thesis
356
00:17:01,040 --> 00:17:05,840
and that was the beginning
of my great career. (laughs)
357
00:17:09,200 --> 00:17:11,376
- So, his Master's thesis is considered
358
00:17:11,400 --> 00:17:14,976
the most important Master's
thesis of all time.
359
00:17:15,000 --> 00:17:15,017
And he writes this at the age of 21.
the most important Master's
thesis of all time.
360
00:17:15,041 --> 00:17:18,376
And he writes this at the age of 21.
361
00:17:18,400 --> 00:17:20,056
He even wins a prize.
362
00:17:20,080 --> 00:17:21,616
It's the Noble Prize.
363
00:17:21,640 --> 00:17:25,256
Now, that's not Noble,
it's N-O-B-L-E is the prize
364
00:17:25,280 --> 00:17:26,097
that he wins.
365
00:17:26,121 --> 00:17:29,136
It's very prestigious, it's
given to the top engineer.
366
00:17:29,160 --> 00:17:32,736
He's the youngest recipient
ever, and he gets prize money,
367
00:17:32,760 --> 00:17:32,777
a certificate, he's
invited to give the paper
He's the youngest recipient
ever, and he gets prize money,
368
00:17:32,801 --> 00:17:35,776
a certificate, he's
invited to give the paper
369
00:17:35,800 --> 00:17:38,136
to an audience of engineers
in Washington D.C.
370
00:17:38,160 --> 00:17:41,440
and, best part, he gets written
up in the New York Times.
371
00:17:43,000 --> 00:17:45,656
- Shannon's Master's thesis
372
00:17:45,680 --> 00:17:45,697
is his first fundamental leap of insight.
- Shannon's Master's thesis
373
00:17:45,721 --> 00:17:50,176
is his first fundamental leap of insight.
374
00:17:50,200 --> 00:17:53,200
George Boole came up with laws of thought,
375
00:17:54,360 --> 00:17:55,896
and they were nice and beautiful,
376
00:17:55,920 --> 00:18:00,120
but Shannon was looking at
electromechanical switches,
377
00:18:01,440 --> 00:18:03,696
and what is the connection between laws
378
00:18:03,720 --> 00:18:06,160
of thought and switches?
379
00:18:07,800 --> 00:18:09,736
(clinking)
380
00:18:09,760 --> 00:18:12,480
Well, Shannon realized
they are the same thing.
381
00:18:13,400 --> 00:18:15,560
(clinking)
382
00:18:20,440 --> 00:18:22,016
And that was just mind-boggling,
383
00:18:22,040 --> 00:18:25,296
because if the laws of
thought could be implemented
384
00:18:25,320 --> 00:18:25,337
in switches then it's not outlandish
because if the laws of
thought could be implemented
385
00:18:25,361 --> 00:18:30,120
in switches then it's not outlandish
386
00:18:31,240 --> 00:18:33,880
to surmise that at some
point machines could think.
387
00:18:34,960 --> 00:18:38,480
(gentle instrumental music)
388
00:18:40,200 --> 00:18:41,400
- No, no, not that way.
389
00:18:42,840 --> 00:18:45,496
Back up, back up a little bit.
390
00:18:45,520 --> 00:18:45,537
Theseus would've figured this out already.
Back up, back up a little bit.
391
00:18:45,561 --> 00:18:47,856
Theseus would've figured this out already.
392
00:18:47,880 --> 00:18:50,256
- I've heard people say
that your Master's thesis
393
00:18:50,280 --> 00:18:54,176
was the most important
Master's thesis of the century.
394
00:18:54,200 --> 00:18:54,217
Did you have a eureka moment?
was the most important
Master's thesis of the century.
395
00:18:54,241 --> 00:18:56,976
Did you have a eureka moment?
396
00:18:57,000 --> 00:19:00,400
- (chuckles) I might have,
if I'd known how to spell it.
397
00:19:01,840 --> 00:19:04,720
I had a lot of fun
figuring that thing out.
398
00:19:06,840 --> 00:19:07,880
I had a great time.
399
00:19:09,200 --> 00:19:11,696
(gentle instrumental music)
400
00:19:11,720 --> 00:19:14,816
- I was going to Radcliffe,
the girls' part of Harvard,
401
00:19:14,840 --> 00:19:19,000
and I had two friends who
said, "Oh, Norma, come with us.
402
00:19:20,720 --> 00:19:23,600
"We know some nice guys over at MIT."
403
00:19:24,560 --> 00:19:27,720
I said, "Okay," and we went.
404
00:19:30,680 --> 00:19:35,216
And it was a living room
with three bedrooms off it.
405
00:19:35,240 --> 00:19:39,176
Now, two of the guys
who lived there came out
406
00:19:39,200 --> 00:19:39,217
and talked to my friends,
and the door was open
Now, two of the guys
who lived there came out
407
00:19:39,241 --> 00:19:44,040
and talked to my friends,
and the door was open
408
00:19:44,440 --> 00:19:47,096
to the third bedroom.
409
00:19:47,120 --> 00:19:50,160
(cheerful jazz music)
410
00:19:54,560 --> 00:19:59,056
And I could see someone
in it not coming forth,
411
00:19:59,080 --> 00:20:03,080
but what came forth from
there was Bix Beiderbecke.
412
00:20:04,800 --> 00:20:08,416
(cheerful jazz music)
413
00:20:08,440 --> 00:20:10,520
So, I walked over that way
414
00:20:16,240 --> 00:20:21,040
and said, "Why don't you come out?"
415
00:20:21,280 --> 00:20:24,056
And he came just sort of halfway out.
416
00:20:24,080 --> 00:20:27,120
(cheerful jazz music)
417
00:20:43,000 --> 00:20:45,016
And he said, "Why don't you come in?"
418
00:20:45,040 --> 00:20:45,057
(cheerful jazz music)
And he said, "Why don't you come in?"
419
00:20:45,081 --> 00:20:48,120
(cheerful jazz music)
420
00:21:05,200 --> 00:21:09,440
He was so funny and charming,
and delicious and cute,
421
00:21:11,840 --> 00:21:14,616
and saw things, I mean, his comments
422
00:21:14,640 --> 00:21:18,400
about things were like an artist.
423
00:21:20,120 --> 00:21:23,520
He didn't sound like a
scientist, not at all.
424
00:21:25,760 --> 00:21:28,520
He liked poetry, and
he liked E.E. Cummings.
425
00:21:29,680 --> 00:21:34,040
He wrote a poem to my
bosom and I tore it up.
426
00:21:35,920 --> 00:21:40,720
Yes, I'm ashamed and unhappy
that I did, but there you are.
427
00:21:42,360 --> 00:21:44,096
(gentle guitar music)
428
00:21:44,120 --> 00:21:46,280
He wanted to learn to fly, and he did.
429
00:21:48,000 --> 00:21:52,040
He had just done his
one plus one equals one,
430
00:21:53,080 --> 00:21:57,880
it was his Master's thesis,
and he had to prepare French
431
00:21:58,080 --> 00:22:00,416
to get his PhD.
432
00:22:00,440 --> 00:22:03,960
(gentle instrumental music)
433
00:22:06,400 --> 00:22:10,536
And I had to teach him enough
French to pass the doctorate.
434
00:22:10,560 --> 00:22:14,080
(gentle instrumental music)
435
00:22:26,880 --> 00:22:29,336
I love it that he wasn't
that good at that.
436
00:22:29,360 --> 00:22:29,377
(big band music)
I love it that he wasn't
that good at that.
437
00:22:29,401 --> 00:22:32,040
(big band music)
438
00:22:35,840 --> 00:22:39,376
- "Dear Dr. Bush, I did not,
439
00:22:39,400 --> 00:22:44,056
"as you may have anticipated,
marry a lady scientist,
440
00:22:44,080 --> 00:22:45,896
"but rather a writer.
441
00:22:45,920 --> 00:22:47,536
"She was helping me with my French
442
00:22:47,560 --> 00:22:51,296
"and apparently it ripened into
something more than French."
443
00:22:51,320 --> 00:22:53,960
(big band music)
444
00:22:59,560 --> 00:23:02,376
- So, you turned
to biology for your PhD?
445
00:23:02,400 --> 00:23:03,656
(Claude laughs)
446
00:23:03,680 --> 00:23:04,680
- Vann Bush.
447
00:23:06,160 --> 00:23:10,960
Vann liked how I used my
queer algebra so much,
448
00:23:11,520 --> 00:23:14,336
that's what he called it,
that he suggested I try
449
00:23:14,360 --> 00:23:15,720
to apply it to genetics.
450
00:23:16,760 --> 00:23:18,920
I didn't know anything about genetics.
451
00:23:20,240 --> 00:23:22,336
This was in 1940, before Watson
452
00:23:22,360 --> 00:23:24,760
and Crick had even
figured out DNA structure.
453
00:23:27,560 --> 00:23:30,016
But, you know, DNA might be
454
00:23:30,040 --> 00:23:33,040
the ultimate information storage device.
455
00:23:36,080 --> 00:23:40,136
So, I read all the
introductory genetic text books
456
00:23:40,160 --> 00:23:43,096
that I could find, and
what I discovered was
457
00:23:43,120 --> 00:23:46,696
that you could represent
different genetic possibilities
458
00:23:46,720 --> 00:23:48,960
mathematically in matrices.
459
00:23:51,320 --> 00:23:53,976
And then you could just use matrix algebra
460
00:23:54,000 --> 00:23:56,616
to make predictions of
how traits were passed on
461
00:23:56,640 --> 00:23:56,657
through generations.
to make predictions of
how traits were passed on
462
00:23:56,681 --> 00:23:57,800
through generations.
463
00:24:00,520 --> 00:24:04,120
But I don't think the biologists
knew what to make of it.
464
00:24:05,320 --> 00:24:07,320
Maybe it was too far out for them.
465
00:24:09,160 --> 00:24:13,216
- He'd got a national research
fellowship in mathematics
466
00:24:13,240 --> 00:24:17,536
to the Institute for
Advanced Study in June 1940,
467
00:24:17,560 --> 00:24:19,656
and that's when we moved to Princeton.
468
00:24:19,680 --> 00:24:23,200
(gentle instrumental music)
469
00:24:25,040 --> 00:24:27,336
Claude liked my taste.
470
00:24:27,360 --> 00:24:32,160
What he liked particularly
was a painting that I loved,
471
00:24:33,280 --> 00:24:38,040
the Seurat, Dimanche a la Grande
Jatte, Sunday in the Park.
472
00:24:41,560 --> 00:24:43,760
He liked the timelessness of it.
473
00:24:44,720 --> 00:24:48,320
(intense instrumental music)
474
00:24:57,520 --> 00:24:58,856
I looked at that a million times
475
00:24:58,880 --> 00:25:03,040
and I didn't think it was a
painting about time, but he did.
476
00:25:04,600 --> 00:25:08,000
- I met Einstein, Gödel, von Neumann,
477
00:25:09,080 --> 00:25:11,416
but you've got to get the
picture right, Michelle,
478
00:25:11,440 --> 00:25:15,256
I said I met them but they
probably wouldn't say they met me
479
00:25:15,280 --> 00:25:16,576
because they wouldn't remember.
480
00:25:16,600 --> 00:25:19,536
But I remember being very impressed
481
00:25:19,560 --> 00:25:21,976
when I met Einstein, for example.
482
00:25:22,000 --> 00:25:22,017
He always walked to work,
when I met Einstein, for example.
483
00:25:22,041 --> 00:25:24,016
He always walked to work,
484
00:25:24,040 --> 00:25:26,160
and when I drove by I'd wave at him.
485
00:25:27,000 --> 00:25:29,920
I don't think he had any idea who I was,
486
00:25:31,240 --> 00:25:33,896
but he probably thought I
was some kind of weirdo.
487
00:25:33,920 --> 00:25:36,720
- Did you try to
talk to them about your ideas,
488
00:25:38,800 --> 00:25:40,840
about information and...
489
00:25:42,640 --> 00:25:43,640
Dr. Shannon?
490
00:25:46,680 --> 00:25:49,256
- Oh, I...
491
00:25:49,280 --> 00:25:53,576
I had a couple of good
conversations with Hermann Weyl,
492
00:25:53,600 --> 00:25:53,617
who was my supervisor, about
the revolutions taking place
I had a couple of good
conversations with Hermann Weyl,
493
00:25:53,641 --> 00:25:58,440
who was my supervisor, about
the revolutions taking place
494
00:25:58,800 --> 00:26:02,016
in science: quantum mechanics,
495
00:26:02,040 --> 00:26:04,296
the Heisenberg uncertainty principle
496
00:26:04,320 --> 00:26:06,776
saying things couldn't
be measured precisely,
497
00:26:06,800 --> 00:26:10,920
Kurt Gödel showing the
limitations in mathematics.
498
00:26:12,520 --> 00:26:14,456
But the people at the institute
499
00:26:14,480 --> 00:26:17,640
were very pure mathematics-oriented
500
00:26:19,040 --> 00:26:21,320
and not so interested in
more practical things.
501
00:26:23,400 --> 00:26:26,600
(gentle clarinet music)
502
00:26:33,680 --> 00:26:36,496
I didn't run across anyone
who was much interested
503
00:26:36,520 --> 00:26:38,760
in the things I was trying to do.
504
00:26:41,040 --> 00:26:44,296
- The last six months in Princeton,
505
00:26:44,320 --> 00:26:46,360
something was going wrong.
506
00:26:49,400 --> 00:26:52,056
When he came home from
there, he was unhappy
507
00:26:52,080 --> 00:26:53,920
and he didn't even wanna talk about it.
508
00:26:54,960 --> 00:26:56,936
And it wasn't going well.
509
00:26:56,960 --> 00:27:00,160
(gentle clarinet music)
510
00:27:09,760 --> 00:27:12,880
He stayed home, he didn't wanna go out,
511
00:27:14,280 --> 00:27:16,656
he talked less and less,
512
00:27:16,680 --> 00:27:16,697
he played the clarinet more and more.
He talked less and less,
513
00:27:16,721 --> 00:27:18,800
he played the clarinet more and more.
514
00:27:23,160 --> 00:27:25,056
- Was it possibly
related to what was going on
515
00:27:25,080 --> 00:27:26,336
in the world?
516
00:27:26,360 --> 00:27:27,816
The build up to the war?
517
00:27:27,840 --> 00:27:29,016
- Not at all.
518
00:27:29,040 --> 00:27:31,400
He didn't wanna know
about anything political.
519
00:27:33,520 --> 00:27:35,536
I thought he was just unhappy
520
00:27:35,560 --> 00:27:37,400
with his work at the institute.
521
00:27:39,200 --> 00:27:40,456
- Is that why
he left the institute
522
00:27:40,480 --> 00:27:42,016
after two months?
523
00:27:42,040 --> 00:27:42,055
- No. after two months?
524
00:27:42,079 --> 00:27:43,016
- No.
525
00:27:43,040 --> 00:27:43,057
We got there in September - No.
526
00:27:43,081 --> 00:27:45,296
We got there in September
527
00:27:45,320 --> 00:27:45,337
and I was living in
Princeton through July.
We got there in September
528
00:27:45,361 --> 00:27:49,320
and I was living in
Princeton through July.
529
00:27:50,360 --> 00:27:51,137
- But he was only working
530
00:27:51,161 --> 00:27:52,760
at the institute for two months.
531
00:27:55,200 --> 00:27:56,936
- Are you telling me that
when I thought he was
532
00:27:56,960 --> 00:27:59,776
at the Institute for Advanced
Study he wasn't there?
533
00:27:59,800 --> 00:28:01,416
When I was with him?
534
00:28:01,440 --> 00:28:04,296
You mean he resigned
and he didn't tell me?
535
00:28:04,320 --> 00:28:06,880
Was he going somewhere else?
536
00:28:07,920 --> 00:28:09,256
What was he doing?
537
00:28:09,280 --> 00:28:09,295
- He was working
What was he doing?
538
00:28:09,319 --> 00:28:10,057
- He was working
539
00:28:10,081 --> 00:28:11,496
for the National Defense
Research Committee
540
00:28:11,520 --> 00:28:11,537
through Princeton University.
for the National Defense
Research Committee
541
00:28:11,561 --> 00:28:12,920
through Princeton University.
542
00:28:14,400 --> 00:28:15,800
- Well, that's interesting.
543
00:28:17,080 --> 00:28:20,056
Listen, something funny was going on
544
00:28:20,080 --> 00:28:23,776
because he looked different,
545
00:28:23,800 --> 00:28:25,720
he was a completely different guy.
546
00:28:26,800 --> 00:28:28,840
Something peculiar was happening.
547
00:28:32,240 --> 00:28:33,920
It was hard living with him.
548
00:28:37,200 --> 00:28:39,016
I said, "You know, I can't live this way,
549
00:28:39,040 --> 00:28:41,496
"and I've got to get a divorce."
550
00:28:41,520 --> 00:28:44,720
(gentle clarinet music)
551
00:28:55,120 --> 00:28:56,120
- Uh.
552
00:28:57,360 --> 00:28:59,600
About halfway through
the year at Princeton,
553
00:29:01,040 --> 00:29:02,120
I could smell the...
554
00:29:04,400 --> 00:29:06,080
I could smell the war coming along.
555
00:29:07,560 --> 00:29:11,736
It occurred to me that
I would be more useful
556
00:29:11,760 --> 00:29:16,560
and safer working full-time
for the war effort.
557
00:29:16,640 --> 00:29:21,440
Safer against the draft,
which I didn't exactly fancy.
558
00:29:24,000 --> 00:29:28,800
So, what I did was I resigned
my fellowship halfway through,
559
00:29:30,920 --> 00:29:35,720
and I went down and asked if I
could get a job at Bell Labs.
560
00:29:36,560 --> 00:29:38,960
(blues music)
561
00:29:44,440 --> 00:29:46,696
♪ I went to the restaurant monday ♪
562
00:29:46,720 --> 00:29:49,040
- He played the blues day and night.
563
00:29:50,200 --> 00:29:52,536
Teddy Grace was the singer that he adored.
564
00:29:52,560 --> 00:29:54,096
Oh, I loved her, too.
565
00:29:54,120 --> 00:29:54,137
I still love her.
Oh, I loved her, too.
566
00:29:54,161 --> 00:29:55,161
I still love her.
567
00:29:56,000 --> 00:29:59,136
Very sad blues that he
played, but beautiful ones.
568
00:29:59,160 --> 00:30:01,856
He had a very big record player.
569
00:30:01,880 --> 00:30:01,897
And you could hear it through the ceiling.
He had a very big record player.
570
00:30:01,921 --> 00:30:03,920
And you could hear it through the ceiling.
571
00:30:06,160 --> 00:30:08,496
He played it day and night,
that's how I met him.
572
00:30:08,520 --> 00:30:12,616
I said, you know, "Even
though I'm up all night,
573
00:30:12,640 --> 00:30:14,936
"a good part of the
night, but you keep me up.
574
00:30:14,960 --> 00:30:16,456
"I can't even study.
575
00:30:16,480 --> 00:30:17,816
"It's too noisy."
576
00:30:17,840 --> 00:30:19,280
He said, "Well, I'll move it."
577
00:30:21,200 --> 00:30:23,360
And then he asked me
would I go out with him.
578
00:30:26,160 --> 00:30:28,896
He was very good-looking, really striking.
579
00:30:28,920 --> 00:30:32,176
He had this black thing, you know, a curl
580
00:30:32,200 --> 00:30:32,217
that came over his forehead always,
He had this black thing, you know, a curl
581
00:30:32,241 --> 00:30:34,016
that came over his forehead always,
582
00:30:34,040 --> 00:30:35,400
that he had to shake off.
583
00:30:36,200 --> 00:30:37,416
He loved jazz.
584
00:30:37,440 --> 00:30:40,160
I mean, it was a passion
with him, a real passion.
585
00:30:41,120 --> 00:30:43,496
He said the improvisation,
the unpredictability
586
00:30:43,520 --> 00:30:45,800
of it is what he loved about it, he said.
587
00:30:49,560 --> 00:30:51,416
When I asked him, "What are you doing?
588
00:30:51,440 --> 00:30:53,416
"How are you making a living?"
589
00:30:53,440 --> 00:30:55,960
He said, "I work with numbers."
590
00:30:56,800 --> 00:30:58,880
And I thought, "Oh, he's an accountant."
591
00:31:00,920 --> 00:31:02,960
You know, I think it was a suffering time.
592
00:31:03,960 --> 00:31:08,536
I had German measles, and
Claude was trying to get them.
593
00:31:08,560 --> 00:31:12,576
He said if he got it then he
wouldn't have to go to work,
594
00:31:12,600 --> 00:31:14,600
because he hated the work he was doing.
595
00:31:17,160 --> 00:31:21,496
- Initially they had me working
on missile firing systems,
596
00:31:21,520 --> 00:31:25,056
but there was another group
at Bell working on systems
597
00:31:25,080 --> 00:31:27,496
to scramble the communications
598
00:31:27,520 --> 00:31:27,537
between Churchill and Roosevelt.
To scramble the communications
599
00:31:27,561 --> 00:31:30,416
between Churchill and Roosevelt.
600
00:31:30,440 --> 00:31:31,417
That was the sort of thing
601
00:31:31,441 --> 00:31:34,416
I was interested in, communications.
602
00:31:34,440 --> 00:31:37,056
So, I asked if I could go work with them.
603
00:31:37,080 --> 00:31:39,576
Most of the cryptography group were trying
604
00:31:39,600 --> 00:31:43,336
to figure out how to
crack specific messages,
605
00:31:43,360 --> 00:31:46,256
very practical stuff,
code breaking things,
606
00:31:46,280 --> 00:31:48,496
like Sherlock Holmes and so forth.
607
00:31:48,520 --> 00:31:48,537
But nobody seemed to be very interested
like Sherlock Holmes and so forth.
608
00:31:48,561 --> 00:31:50,416
But nobody seemed to be very interested
609
00:31:50,440 --> 00:31:50,457
in theoretical questions.
But nobody seemed to be very interested
610
00:31:50,481 --> 00:31:53,400
in theoretical questions.
611
00:31:55,200 --> 00:31:58,536
Like, how many messages
you needed to intercept
612
00:31:58,560 --> 00:32:00,120
before you could get a solution?
613
00:32:00,920 --> 00:32:04,280
Or, could you create a
completely unbreakable code?
614
00:32:08,400 --> 00:32:11,976
That was the sort of thing
I was more interested
615
00:32:12,000 --> 00:32:13,000
in trying to do.
616
00:32:14,160 --> 00:32:16,456
- Norma had told
him to get in analysis,
617
00:32:16,480 --> 00:32:18,456
because he must've been
very depressed already
618
00:32:18,480 --> 00:32:19,680
when he was with her.
619
00:32:20,520 --> 00:32:23,216
- I didn't get as much response
620
00:32:23,240 --> 00:32:23,257
to my cryptography work as I expected.
- I didn't get as much response
621
00:32:23,281 --> 00:32:26,056
to my cryptography work as I expected.
622
00:32:26,080 --> 00:32:27,976
I guess it wasn't very useful,
623
00:32:28,000 --> 00:32:30,656
practically speaking, to cryptographers.
624
00:32:30,680 --> 00:32:30,697
But after the war though, I
was called in as a consultant
practically speaking, to cryptographers.
625
00:32:30,721 --> 00:32:33,776
But after the war though, I
was called in as a consultant
626
00:32:33,800 --> 00:32:33,817
to the National Security
Agency in Washington.
But after the war though, I
was called in as a consultant
627
00:32:33,841 --> 00:32:36,296
to the National Security
Agency in Washington.
628
00:32:36,320 --> 00:32:37,560
So, I guess someone...
629
00:32:38,480 --> 00:32:39,800
I probably better...
630
00:32:40,840 --> 00:32:42,296
Even though this was a long time ago,
631
00:32:42,320 --> 00:32:45,760
I guess I shouldn't tell you about that.
632
00:32:47,160 --> 00:32:49,760
- I asked him, "How is
it going in analysis?"
633
00:32:52,040 --> 00:32:53,856
He said, "The guy's
finally learned something
634
00:32:53,880 --> 00:32:53,897
"about mathematics and
how to think properly."
He said, "The guy's
finally learned something
635
00:32:53,921 --> 00:32:56,760
"about mathematics and
how to think properly."
636
00:32:59,440 --> 00:33:02,656
He said that what had
hurt him so in life was
637
00:33:02,680 --> 00:33:06,240
that there was certainty
until a certain time,
638
00:33:07,160 --> 00:33:10,856
that once he knew a thing to
be true it would remain true,
639
00:33:10,880 --> 00:33:12,280
and everything had changed.
640
00:33:13,560 --> 00:33:15,680
Everything became sort
of relative, he said.
641
00:33:17,240 --> 00:33:21,240
He had believed in black and
white truth, absolute truth.
642
00:33:22,800 --> 00:33:24,656
And those had gone for him.
643
00:33:24,680 --> 00:33:26,536
He had to adjust to that.
644
00:33:26,560 --> 00:33:30,256
(gentle instrumental music)
645
00:33:30,280 --> 00:33:30,297
"With love and kisses
and curses and hisses.
(gentle instrumental music)
646
00:33:30,321 --> 00:33:34,296
"With love and kisses
and curses and hisses.
647
00:33:34,320 --> 00:33:36,976
"Claude, 1944."
648
00:33:37,000 --> 00:33:38,576
That was a long time ago.
649
00:33:38,600 --> 00:33:42,120
(gentle instrumental music)
650
00:33:44,440 --> 00:33:47,080
It doesn't seem to be
that long ago, really not.
651
00:33:49,080 --> 00:33:51,480
I didn't know how much
I cared for him, really.
652
00:33:53,080 --> 00:33:54,080
I should've known.
653
00:33:55,360 --> 00:33:58,280
That secret life, I think,
really got between us.
654
00:33:59,520 --> 00:34:04,320
- I found this 1945
cryptography report you wrote.
655
00:34:05,600 --> 00:34:09,816
And it uses the words
"information theory".
656
00:34:09,840 --> 00:34:09,857
And I think that that's the
first time that phrase appeared,
And it uses the words
"information theory".
657
00:34:09,881 --> 00:34:12,336
And I think that that's the
first time that phrase appeared,
658
00:34:12,360 --> 00:34:15,560
those two words together
in print, anywhere.
659
00:34:16,440 --> 00:34:17,536
- Is that so?
660
00:34:17,560 --> 00:34:18,696
- Yes.
661
00:34:18,720 --> 00:34:19,720
See?
662
00:34:20,840 --> 00:34:24,456
And I'm wondering, your monumental paper
663
00:34:24,480 --> 00:34:27,656
on information theory
didn't come out until 1948.
664
00:34:27,680 --> 00:34:29,096
- Hmm.
665
00:34:29,120 --> 00:34:30,376
- So is it accurate to say
666
00:34:30,400 --> 00:34:33,656
that information theory came
out of your cryptography work?
667
00:34:33,680 --> 00:34:35,536
- No.
668
00:34:35,560 --> 00:34:35,577
Look, Michelle, you're
looking for a simple...
- No.
669
00:34:35,601 --> 00:34:37,920
Look, Michelle, you're
looking for a simple...
670
00:34:39,360 --> 00:34:41,696
It would be more accurate
to say they were going on
671
00:34:41,720 --> 00:34:41,737
at the same time.
It would be more accurate
to say they were going on
672
00:34:41,761 --> 00:34:43,256
at the same time.
673
00:34:43,280 --> 00:34:46,680
In my mind, they're connected.
674
00:34:47,520 --> 00:34:49,896
How you make something unintelligible
675
00:34:49,920 --> 00:34:52,656
or how you make it
intelligible are two sides
676
00:34:52,680 --> 00:34:54,096
of the same coin.
677
00:34:54,120 --> 00:34:55,696
- But officially
you were just working
678
00:34:55,720 --> 00:34:57,336
on cryptography, right?
679
00:34:57,360 --> 00:34:57,377
- Are you looking to work
on cryptography, right?
680
00:34:57,401 --> 00:34:58,696
- Are you looking to work
681
00:34:58,720 --> 00:35:00,656
for the district attorney's
office? (chuckles)
682
00:35:00,680 --> 00:35:00,695
- No.
for the district attorney's
office? (chuckles)
683
00:35:00,719 --> 00:35:01,456
- No.
684
00:35:01,480 --> 00:35:02,336
No, I'm just...
685
00:35:02,360 --> 00:35:02,377
- The cryptography work was just a way
No, I'm just...
686
00:35:02,401 --> 00:35:04,696
- The cryptography work was just a way
687
00:35:04,720 --> 00:35:07,776
of legitimizing the other work I was doing
688
00:35:07,800 --> 00:35:11,520
at home outside office hours, so to speak,
689
00:35:13,200 --> 00:35:16,160
to make it look like I
was working on decent,
690
00:35:17,080 --> 00:35:20,176
worthwhile things, if
you know what I mean.
691
00:35:20,200 --> 00:35:20,215
- Did you feel like you had
worthwhile things, if
you know what I mean.
692
00:35:20,239 --> 00:35:21,097
- Did you feel like you had
693
00:35:21,121 --> 00:35:22,976
to hide the other work?
694
00:35:23,000 --> 00:35:23,017
- Not hide exactly, but I didn't know
to hide the other work?
695
00:35:23,041 --> 00:35:27,296
- Not hide exactly, but I didn't know
696
00:35:27,320 --> 00:35:28,880
if it would all work out.
697
00:35:32,200 --> 00:35:34,176
- Shannon works on his universal theory
698
00:35:34,200 --> 00:35:37,400
of communication for 10 years.
699
00:35:38,840 --> 00:35:41,256
He works on it through a failed marriage
700
00:35:41,280 --> 00:35:44,296
and through this war work
that he really dislikes.
701
00:35:44,320 --> 00:35:47,416
He works at nights, he
works on the weekends,
702
00:35:47,440 --> 00:35:50,216
he collects snatches
of time here and there,
703
00:35:50,240 --> 00:35:50,257
and he works alone.
he collects snatches
of time here and there,
704
00:35:50,281 --> 00:35:51,976
and he works alone.
705
00:35:52,000 --> 00:35:52,017
He doesn't tell anyone
what he's working on.
and he works alone.
706
00:35:52,041 --> 00:35:54,320
He doesn't tell anyone
what he's working on.
707
00:35:57,680 --> 00:35:59,616
What he realizes is that in order
708
00:35:59,640 --> 00:36:03,056
to understand communication,
he needs a new understanding
709
00:36:03,080 --> 00:36:04,840
of the thing you wanna communicate.
710
00:36:05,920 --> 00:36:09,840
He needs a radical new
definition of information.
711
00:36:13,160 --> 00:36:16,936
- Information, which they
called intelligence at the time,
712
00:36:16,960 --> 00:36:19,136
was a slippery concept.
713
00:36:19,160 --> 00:36:19,177
I wanted to try to find a
way to treat information
was a slippery concept.
714
00:36:19,201 --> 00:36:22,256
I wanted to try to find a
way to treat information
715
00:36:22,280 --> 00:36:26,176
like a physical thing, like
energy that you could measure.
716
00:36:26,200 --> 00:36:29,720
(gentle instrumental music)
717
00:36:34,400 --> 00:36:37,320
What I realized was,
to measure information,
718
00:36:38,160 --> 00:36:41,480
you have to look at it
without regard to meaning.
719
00:36:42,320 --> 00:36:44,376
That upsets a lot of people.
720
00:36:44,400 --> 00:36:44,417
That content is irrelevant.
That upsets a lot of people.
721
00:36:44,441 --> 00:36:47,560
That content is irrelevant.
722
00:36:48,680 --> 00:36:51,176
(intense instrumental music)
723
00:36:51,200 --> 00:36:54,456
The point of sending a message
is to remove uncertainty,
724
00:36:54,480 --> 00:36:57,776
something you didn't
know but now you know.
725
00:36:57,800 --> 00:36:59,296
Do you have a coin, Michelle?
726
00:36:59,320 --> 00:37:00,856
- A coin?
- Yes, a coin.
727
00:37:00,880 --> 00:37:02,536
I'd get up to get one myself,
728
00:37:02,560 --> 00:37:05,496
but you'll just tell me to
sit back down. (chuckles)
729
00:37:05,520 --> 00:37:06,896
- Um...
730
00:37:06,920 --> 00:37:08,496
Yes.
731
00:37:08,520 --> 00:37:09,936
- There.
732
00:37:09,960 --> 00:37:12,216
(Claude clears throat)
733
00:37:12,240 --> 00:37:13,056
Heads or tails?
734
00:37:13,080 --> 00:37:14,056
- Tails.
735
00:37:14,080 --> 00:37:14,857
- Heads.
736
00:37:14,881 --> 00:37:17,976
A coin toss is an example
of the simplest form
737
00:37:18,000 --> 00:37:18,017
of information to
communicate, a binary choice.
A coin toss is an example
of the simplest form
738
00:37:18,041 --> 00:37:22,576
of information to
communicate, a binary choice.
739
00:37:22,600 --> 00:37:27,400
True or false, heads or
tails, a one or a zero.
740
00:37:28,880 --> 00:37:29,777
A bit.
741
00:37:29,801 --> 00:37:33,320
(gentle instrumental music)
742
00:37:35,080 --> 00:37:37,336
- Shannon's big idea number one:
743
00:37:37,360 --> 00:37:40,696
convert all information to bits.
744
00:37:40,720 --> 00:37:42,176
(coin clinking)
745
00:37:42,200 --> 00:37:45,720
(gentle instrumental music)
746
00:37:49,680 --> 00:37:51,456
(glass crashing)
747
00:37:51,480 --> 00:37:55,000
(gentle instrumental music)
748
00:37:57,560 --> 00:38:01,120
All information can be
represented as zeros and ones.
749
00:38:04,120 --> 00:38:08,536
Pictures, music, (upbeat music)
750
00:38:08,560 --> 00:38:13,336
video, all information can
be treated the same way.
751
00:38:13,360 --> 00:38:16,880
(gentle instrumental music)
752
00:38:19,400 --> 00:38:22,816
Which means the problem
of communication reduces
753
00:38:22,840 --> 00:38:27,096
to just two things:
how do I convert things
754
00:38:27,120 --> 00:38:28,120
to zeros and ones,
755
00:38:31,720 --> 00:38:34,280
and how do I make sure that those zeros
756
00:38:35,080 --> 00:38:36,640
and ones get through accurately?
757
00:38:40,040 --> 00:38:41,576
- Is your coin legal?
758
00:38:41,600 --> 00:38:41,617
- Sorry?
- Is your coin legal?
759
00:38:41,641 --> 00:38:42,896
- Sorry?
760
00:38:42,920 --> 00:38:44,736
- Your coin's not counterfeit, is it?
761
00:38:44,760 --> 00:38:46,816
- Not that I know of.
762
00:38:46,840 --> 00:38:46,857
- So, if I toss it 100 times,
- Not that I know of.
763
00:38:46,881 --> 00:38:50,096
- So, if I toss it 100 times,
764
00:38:50,120 --> 00:38:52,736
each time I toss it there's
an equal probability
765
00:38:52,760 --> 00:38:55,576
that it's a head or tail.
766
00:38:55,600 --> 00:38:59,816
I'd need 100 bits of
information, one for each toss,
767
00:38:59,840 --> 00:39:01,536
to communicate the results.
768
00:39:01,560 --> 00:39:01,577
Now, if it turns out you're
some sort of a con woman
to communicate the results.
769
00:39:01,601 --> 00:39:06,336
Now, if it turns out you're
some sort of a con woman
770
00:39:06,360 --> 00:39:09,576
and you thought it might
be fun to pull the wool
771
00:39:09,600 --> 00:39:12,760
over some gullible old man's eyes...
772
00:39:14,400 --> 00:39:15,400
Heads, right?
773
00:39:16,200 --> 00:39:17,536
- It's heads on both sides.
774
00:39:17,560 --> 00:39:20,056
- Yeah, I didn't need
any information from you
775
00:39:20,080 --> 00:39:21,736
to know the result.
776
00:39:21,760 --> 00:39:24,256
If something's completely predictable,
777
00:39:24,280 --> 00:39:24,297
there's zero information.
If something's completely predictable,
778
00:39:24,321 --> 00:39:26,256
there's zero information.
779
00:39:26,280 --> 00:39:29,016
But, on the other hand, if
something's completely random,
780
00:39:29,040 --> 00:39:31,456
you need the information about everything.
781
00:39:31,480 --> 00:39:32,576
It's counterintuitive,
782
00:39:32,600 --> 00:39:32,617
but gibberish contains more information
It's counterintuitive,
783
00:39:32,641 --> 00:39:36,336
but gibberish contains more information
784
00:39:36,360 --> 00:39:37,736
than great literature.
785
00:39:37,760 --> 00:39:39,376
- Not if it's "Finnegans Wake."
786
00:39:39,400 --> 00:39:39,415
(Claude laughs)
- Not if it's "Finnegans Wake."
787
00:39:39,439 --> 00:39:40,337
(Claude laughs)
788
00:39:40,361 --> 00:39:43,696
- No, James Joyce is
unpredictable. (chuckles)
789
00:39:43,720 --> 00:39:45,416
- Or the Jabberwocky poem,
790
00:39:45,440 --> 00:39:47,416
"Twas brillig, and the slithy toves.
791
00:39:47,440 --> 00:39:49,536
"Did gyre and gimble in the wabe."
792
00:39:49,560 --> 00:39:50,457
(Claude laughs)
793
00:39:50,481 --> 00:39:53,176
- You could barely guess
the next letter of that.
794
00:39:53,200 --> 00:39:58,000
But normal English though, I calculated,
795
00:39:58,640 --> 00:40:00,920
is over 50% redundant.
796
00:40:03,400 --> 00:40:05,976
You know if certain
words follow each other,
797
00:40:06,000 --> 00:40:06,017
there's grammar rules.
You know if certain
words follow each other,
798
00:40:06,041 --> 00:40:07,160
there's grammar rules.
799
00:40:08,400 --> 00:40:09,536
When you learn a language,
800
00:40:09,560 --> 00:40:11,896
you inherently know the statistics,
801
00:40:11,920 --> 00:40:14,336
that's why you can drop
letters and even words
802
00:40:14,360 --> 00:40:16,080
and still understand the message.
803
00:40:17,680 --> 00:40:18,816
- Looking at the redundancy
804
00:40:18,840 --> 00:40:22,336
in English gives Shannon
his big thought number two:
805
00:40:22,360 --> 00:40:24,936
compress your information.
806
00:40:24,960 --> 00:40:28,480
(gentle instrumental music)
807
00:40:33,920 --> 00:40:35,240
Eliminate the redundancy.
808
00:40:37,240 --> 00:40:39,216
Just send what you can't predict.
809
00:40:39,240 --> 00:40:42,760
(gentle instrumental music)
810
00:40:44,120 --> 00:40:45,736
And then he asked the question
811
00:40:45,760 --> 00:40:48,456
that nobody had really thought of before,
812
00:40:48,480 --> 00:40:51,896
is there a minimum size,
a minimum number of bits
813
00:40:51,920 --> 00:40:54,256
that I can shrink my information to
814
00:40:54,280 --> 00:40:56,256
and not lose anything essential?
815
00:40:56,280 --> 00:40:58,840
(playful music)
816
00:41:01,320 --> 00:41:03,800
He discovers there is a minimum
817
00:41:04,760 --> 00:41:07,040
and he shows how to calculate it.
818
00:41:08,160 --> 00:41:11,840
His formula is based on the
probabilities in the message.
819
00:41:13,280 --> 00:41:17,896
It has a very intriguing
form, it's almost identical
820
00:41:17,920 --> 00:41:21,176
to a fundamental quantity
in physics called entropy,
821
00:41:21,200 --> 00:41:24,216
which is related to
randomness and disorder.
822
00:41:24,240 --> 00:41:28,416
(intense instrumental music)
823
00:41:28,440 --> 00:41:30,520
Shannon calls his quantity entropy, too.
824
00:41:31,640 --> 00:41:34,616
And what he proves is that the entropy
825
00:41:34,640 --> 00:41:34,657
is actually the answer to his question,
And what he proves is that the entropy
826
00:41:34,681 --> 00:41:37,776
is actually the answer to his question,
827
00:41:37,800 --> 00:41:39,976
"What's the best I can do?"
828
00:41:40,000 --> 00:41:43,856
The entropy is the absolute minimum size
829
00:41:43,880 --> 00:41:47,616
you can compress your
message without any loss.
830
00:41:47,640 --> 00:41:51,080
(intense orchestral music)
831
00:41:58,760 --> 00:42:02,736
Shannon's ideas that you
can compress a huge amount
832
00:42:02,760 --> 00:42:05,816
of information into something very small
833
00:42:05,840 --> 00:42:05,857
is what makes The
Information Age possible.
of information into something very small
834
00:42:05,881 --> 00:42:08,520
is what makes The
Information Age possible.
835
00:42:12,960 --> 00:42:15,056
- Do you know what the first
message was that was sent
836
00:42:15,080 --> 00:42:15,097
on the Washington-Moscow hotline?
- Do you know what the first
message was that was sent
837
00:42:15,121 --> 00:42:18,496
on the Washington-Moscow hotline?
838
00:42:18,520 --> 00:42:18,537
- Uh...
on the Washington-Moscow hotline?
839
00:42:18,561 --> 00:42:19,576
- Uh...
840
00:42:19,600 --> 00:42:21,016
"Do you hear me?"
841
00:42:21,040 --> 00:42:21,055
(Claude chuckles)
"Do you hear me?"
842
00:42:21,079 --> 00:42:21,937
(Claude chuckles)
843
00:42:21,961 --> 00:42:25,376
- "The quick brown fox jumped
over the lazy dog's back."
844
00:42:25,400 --> 00:42:28,560
One, two, three, four, five,
six, seven, eight, nine, zero.
845
00:42:29,680 --> 00:42:32,440
It uses every letter of the alphabet,
846
00:42:34,240 --> 00:42:36,096
but it isn't the most efficient
847
00:42:36,120 --> 00:42:36,137
because some of the letters are repeated.
But it isn't the most efficient
848
00:42:36,161 --> 00:42:38,976
because some of the letters are repeated.
849
00:42:39,000 --> 00:42:40,936
Betty and I had a grand
time trying to come up
850
00:42:40,960 --> 00:42:40,977
with the shortest sentence that
uses every letter only once.
Betty and I had a grand
time trying to come up
851
00:42:41,001 --> 00:42:45,616
with the shortest sentence that
uses every letter only once.
852
00:42:45,640 --> 00:42:48,256
- We spent a lot of time
looking in the dictionary
853
00:42:48,280 --> 00:42:51,696
for words heavy on consonants
and light on vowels.
854
00:42:51,720 --> 00:42:53,856
- The hard part is that
there's only a few vowels
855
00:42:53,880 --> 00:42:55,936
and every word needs to have one.
856
00:42:55,960 --> 00:42:59,016
Squdgy Fez Blank Jimp Crwth Vox.
857
00:42:59,040 --> 00:43:02,480
(intense orchestral music)
858
00:43:06,880 --> 00:43:10,496
- Shannon's ideas that you
should first change everything
859
00:43:10,520 --> 00:43:14,776
to zeros and ones and then
compress it is a huge step
860
00:43:14,800 --> 00:43:17,616
towards the universal
theory of communication,
861
00:43:17,640 --> 00:43:20,736
but it's not the end of the story,
862
00:43:20,760 --> 00:43:23,496
because you still have to get those zeros
863
00:43:23,520 --> 00:43:23,537
and ones across your channel,
because you still have to get those zeros
864
00:43:23,561 --> 00:43:25,160
and ones across your channel,
865
00:43:26,000 --> 00:43:28,696
and the channel has physical limitations,
866
00:43:28,720 --> 00:43:28,737
whether it's a wire or a pipe
with water flowing through it.
and the channel has physical limitations,
867
00:43:28,761 --> 00:43:33,200
whether it's a wire or a pipe
with water flowing through it.
868
00:43:34,800 --> 00:43:37,736
And then, there's always noise, (static)
869
00:43:37,760 --> 00:43:39,656
things like static or interference,
870
00:43:39,680 --> 00:43:43,640
that can mix up or flip the
zeros and ones and wipe you out.
871
00:43:48,400 --> 00:43:52,016
Shannon wonders, "Can I
calculate for any channel
872
00:43:52,040 --> 00:43:54,776
"a speed limit beyond which you know
873
00:43:54,800 --> 00:43:54,817
"things are gonna fall apart?"
"a speed limit beyond which you know
874
00:43:54,841 --> 00:43:56,360
"things are gonna fall apart?"
875
00:43:57,400 --> 00:43:59,456
He comes up with another formula,
876
00:43:59,480 --> 00:44:01,840
it's a theoretical maximum speed,
877
00:44:02,760 --> 00:44:06,000
an information speed limit,
like the speed of light.
878
00:44:07,160 --> 00:44:09,160
We call it the Shannon limit.
879
00:44:10,200 --> 00:44:12,816
You can't go faster
than the Shannon limit.
880
00:44:12,840 --> 00:44:15,960
If you try, you go off a cliff.
881
00:44:17,440 --> 00:44:19,616
- The idea that there was a speed limit
882
00:44:19,640 --> 00:44:22,376
to information was so out of the blue.
883
00:44:22,400 --> 00:44:25,176
I mean, even just
conceptually this was a shock.
884
00:44:25,200 --> 00:44:26,296
It was like when Einstein said
885
00:44:26,320 --> 00:44:27,936
that you couldn't make things go faster
886
00:44:27,960 --> 00:44:29,536
than the speed of light.
887
00:44:29,560 --> 00:44:31,696
But what Shannon showed is if you tried
888
00:44:31,720 --> 00:44:34,896
to get past this limit,
above the Shannon limit,
889
00:44:34,920 --> 00:44:37,896
you cannot recover any useful information.
890
00:44:37,920 --> 00:44:40,456
(static)
891
00:44:40,480 --> 00:44:43,856
- Shannon's third big idea, huge idea:
892
00:44:43,880 --> 00:44:47,056
use mathematics to eliminate noise.
893
00:44:47,080 --> 00:44:47,097
This seemed crazy, what
does that even mean?
use mathematics to eliminate noise.
894
00:44:47,121 --> 00:44:50,320
This seemed crazy, what
does that even mean?
895
00:44:51,120 --> 00:44:54,976
He proved that mathematics
could eliminate errors
896
00:44:55,000 --> 00:44:57,536
from static and interference.
897
00:44:57,560 --> 00:45:01,200
You could use mathematical
codes to overcome noise.
898
00:45:03,560 --> 00:45:06,216
And then, as the icing on the cake,
899
00:45:06,240 --> 00:45:09,376
Shannon claimed this could work right up
900
00:45:09,400 --> 00:45:10,560
to the Shannon limit.
901
00:45:11,360 --> 00:45:13,376
If you could find the right codes,
902
00:45:13,400 --> 00:45:16,656
you could have perfect
communication with as little errors
903
00:45:16,680 --> 00:45:20,416
as you wanted, and that
was the revolution.
904
00:45:20,440 --> 00:45:21,656
(alarm-clock ringing)
905
00:45:21,680 --> 00:45:23,096
(wrist-watch bleeping)
906
00:45:23,120 --> 00:45:23,137
(telephone ringing)
(wrist-watch bleeping)
907
00:45:23,161 --> 00:45:24,376
(telephone ringing)
908
00:45:24,400 --> 00:45:26,136
(cell phone ringing)
909
00:45:26,160 --> 00:45:29,600
(intense orchestral music)
910
00:45:38,680 --> 00:45:42,176
(satellite bleeping)
911
00:45:42,200 --> 00:45:44,456
- Suppose you wanted
to use Shannon's ideas
912
00:45:44,480 --> 00:45:48,336
to help with one of the most
difficult known problems
913
00:45:48,360 --> 00:45:51,616
in communication:
communicating with a spouse.
914
00:45:51,640 --> 00:45:54,136
- One communication
scenario I can think of
915
00:45:54,160 --> 00:45:54,177
applying Shannon's ideas to is
communication in a marriage,
- One communication
scenario I can think of
916
00:45:54,201 --> 00:45:58,576
applying Shannon's ideas to is
communication in a marriage,
917
00:45:58,600 --> 00:45:58,617
which, as we all know,
can be really difficult.
applying Shannon's ideas to is
communication in a marriage,
918
00:45:58,641 --> 00:46:01,616
which, as we all know,
can be really difficult.
919
00:46:01,640 --> 00:46:01,657
- This communication is often
prone to serious errors,
which, as we all know,
can be really difficult.
920
00:46:01,681 --> 00:46:05,056
- This communication is often
prone to serious errors,
921
00:46:05,080 --> 00:46:07,536
which sometimes lead to disasters.
922
00:46:07,560 --> 00:46:07,577
- There's all this noise,
there's interference,
which sometimes lead to disasters.
923
00:46:07,601 --> 00:46:10,256
- There's all this noise,
there's interference,
924
00:46:10,280 --> 00:46:12,136
maybe they had a bad day or a good day,
925
00:46:12,160 --> 00:46:13,936
you know, maybe the kids are
shouting in the background
926
00:46:13,960 --> 00:46:14,896
or the dog's barking.
927
00:46:14,920 --> 00:46:14,937
And so, how do you convey
exactly what you wanna say
or the dog's barking.
928
00:46:14,961 --> 00:46:18,376
And so, how do you convey
exactly what you wanna say
929
00:46:18,400 --> 00:46:20,416
to your spouse in a way
that they understand?
930
00:46:20,440 --> 00:46:23,816
- One way of communicating is
repeating your message over
931
00:46:23,840 --> 00:46:23,855
and over again.
- One way of communicating is
repeating your message over
932
00:46:23,879 --> 00:46:24,777
and over again.
933
00:46:24,801 --> 00:46:28,176
- And that probably will
eventually get through,
934
00:46:28,200 --> 00:46:32,096
but your partner may get annoyed with you.
935
00:46:32,120 --> 00:46:34,936
- And another way is to shout louder.
936
00:46:34,960 --> 00:46:38,616
- But that could also be misconstrued.
937
00:46:38,640 --> 00:46:38,657
- So, what Shannon really said is
- But that could also be misconstrued.
938
00:46:38,681 --> 00:46:40,136
- So, what Shannon really said is
939
00:46:40,160 --> 00:46:42,776
that nagging, you know,
doesn't really work,
940
00:46:42,800 --> 00:46:45,376
or speaking louder doesn't work either.
941
00:46:45,400 --> 00:46:45,417
- It's how you pick your words.
Or speaking louder doesn't work either.
942
00:46:45,441 --> 00:46:47,896
- It's how you pick your words.
943
00:46:47,920 --> 00:46:50,256
- So, we should say
things perhaps many times
944
00:46:50,280 --> 00:46:53,136
but in different fashions,
in different presentations.
945
00:46:53,160 --> 00:46:53,177
We should code them well,
but in different fashions,
in different presentations.
946
00:46:53,201 --> 00:46:54,736
We should code them well,
947
00:46:54,760 --> 00:46:54,777
and at the end of our communication
We should code them well,
948
00:46:54,801 --> 00:46:57,136
and at the end of our communication
949
00:46:57,160 --> 00:46:59,736
we should always add an
error-correcting code,
950
00:46:59,760 --> 00:47:01,736
which in this case
could be something like,
951
00:47:01,760 --> 00:47:02,960
"But I love you."
952
00:47:05,400 --> 00:47:07,096
(bomb explodes)
953
00:47:07,120 --> 00:47:07,137
- When Shannon's paper came
out, it was like a bombshell.
(bomb explodes)
954
00:47:07,161 --> 00:47:11,056
- When Shannon's paper came
out, it was like a bombshell.
955
00:47:11,080 --> 00:47:14,936
To be able to communicate with
as few errors as you wanted,
956
00:47:14,960 --> 00:47:16,280
people couldn't believe it.
957
00:47:17,960 --> 00:47:20,120
Some people couldn't understand the math.
958
00:47:22,160 --> 00:47:24,896
Engineers thought it was so far removed
959
00:47:24,920 --> 00:47:24,937
from anything they could build,
Engineers thought it was so far removed
960
00:47:24,961 --> 00:47:26,800
from anything they could build,
961
00:47:28,640 --> 00:47:31,080
they just basically thought
it was science fiction.
962
00:47:32,840 --> 00:47:35,136
- But there were some people
who became big champions
963
00:47:35,160 --> 00:47:35,177
and saw the potential in
what Shannon had done.
- But there were some people
who became big champions
964
00:47:35,201 --> 00:47:37,976
and saw the potential in
what Shannon had done.
965
00:47:38,000 --> 00:47:41,656
They became totally enamored
of this new field, information.
966
00:47:41,680 --> 00:47:44,376
So, he's placed on a list
of the leading scientists
967
00:47:44,400 --> 00:47:44,417
of the day, alongside Richard
Feynman and Watson and Crick.
So, he's placed on a list
of the leading scientists
968
00:47:44,441 --> 00:47:48,656
of the day, alongside Richard
Feynman and Watson and Crick.
969
00:47:48,680 --> 00:47:52,496
He's given write ups in Time
Magazine and Fortune Magazine
970
00:47:52,520 --> 00:47:53,936
and Scientific American.
971
00:47:53,960 --> 00:47:53,977
And he's even given a
spread in Vogue Magazine
and Scientific American.
972
00:47:54,001 --> 00:47:56,896
And he's even given a
spread in Vogue Magazine
973
00:47:56,920 --> 00:47:59,376
with a photo shoot from
Henri Cartier-Bresson,
974
00:47:59,400 --> 00:48:00,896
the renowned photographer,
975
00:48:00,920 --> 00:48:04,496
whose other photo shoots
included Mahatma Gandhi,
976
00:48:04,520 --> 00:48:08,056
Chairman Mao's funeral,
Queen Elizabeth's coronation,
977
00:48:08,080 --> 00:48:11,200
and Claude Shannon at his
home in Massachusetts.
978
00:48:12,480 --> 00:48:14,696
- When Shannon's paper came out in 1948,
979
00:48:14,720 --> 00:48:18,616
people embraced it far beyond
what Shannon ever envisioned
980
00:48:18,640 --> 00:48:19,696
it would be embraced as,
981
00:48:19,720 --> 00:48:22,656
which was a mathematical
theory of information.
982
00:48:22,680 --> 00:48:24,856
So, it started being applied to philosophy
983
00:48:24,880 --> 00:48:24,897
and pretty much almost
every field embraced it,
So, it started being applied to philosophy
984
00:48:24,921 --> 00:48:27,256
and pretty much almost
every field embraced it,
985
00:48:27,280 --> 00:48:28,976
"Oh, we're gonna use
this mathematical theory
986
00:48:29,000 --> 00:48:30,896
"of information in what we do."
987
00:48:30,920 --> 00:48:33,456
And it got a little bit out of control,
988
00:48:33,480 --> 00:48:36,920
and it was being applied in
ways that didn't make any sense.
989
00:48:40,440 --> 00:48:41,480
- Just a bit more.
990
00:48:42,680 --> 00:48:45,600
- I found this wonderful T-shirt
991
00:48:46,440 --> 00:48:47,456
(Claude laughs)
992
00:48:47,480 --> 00:48:49,016
in the London airport.
993
00:48:49,040 --> 00:48:49,817
(Michelle chuckles)
994
00:48:49,841 --> 00:48:51,616
- Were you famous
when your paper came out?
995
00:48:51,640 --> 00:48:52,497
- Not too much.
996
00:48:52,521 --> 00:48:53,776
- He wrote a paper telling them
997
00:48:53,800 --> 00:48:55,376
to get off the bandwagon.
998
00:48:55,400 --> 00:48:56,976
- I was worried that there was still a lot
999
00:48:57,000 --> 00:49:00,816
of hard slugging left to do
in our own professional house.
1000
00:49:00,840 --> 00:49:03,976
- I read that a
major American mathematician
1001
00:49:04,000 --> 00:49:05,056
criticized your paper
1002
00:49:05,080 --> 00:49:08,216
for not being mathematically
rigorous enough.
1003
00:49:08,240 --> 00:49:11,336
- Maybe it needed people a
little brighter to understand it.
1004
00:49:11,360 --> 00:49:14,136
- Russian mathematicians
got it right away.
1005
00:49:14,160 --> 00:49:15,496
- Yes, they did.
1006
00:49:15,520 --> 00:49:20,000
- But the codes that
you said had to exist,
1007
00:49:20,840 --> 00:49:23,056
you didn't show how to get them.
1008
00:49:23,080 --> 00:49:27,880
And now it's 35 or so years
later and all your ideas
1009
00:49:28,120 --> 00:49:32,096
about converting everything
to digital, compressing audio,
1010
00:49:32,120 --> 00:49:33,536
those things are all happening,
1011
00:49:33,560 --> 00:49:38,360
but as far as your fundamental
theorem, nobody's been able
1012
00:49:38,800 --> 00:49:41,440
to find codes that get anywhere
near the Shannon limit.
1013
00:49:42,600 --> 00:49:44,856
Do you think that...
1014
00:49:44,880 --> 00:49:48,160
Are you worried that
you'll be proven right?
1015
00:49:50,440 --> 00:49:55,240
- I knew what I was doing,
and it came out exactly right.
1016
00:49:57,200 --> 00:50:00,496
Some people question
whether the technology
1017
00:50:00,520 --> 00:50:04,640
will ever be powerful
enough to implement it, but
1018
00:50:08,080 --> 00:50:12,896
there's a lot of bright
young people trying
1019
00:50:12,920 --> 00:50:14,496
to find those codes.
1020
00:50:14,520 --> 00:50:17,400
(roulette rattling)
1021
00:50:21,480 --> 00:50:23,336
- The first few times I met him,
1022
00:50:23,360 --> 00:50:26,376
he was giving a lecture or something.
1023
00:50:26,400 --> 00:50:30,296
I was very young then,
very shy, very immature,
1024
00:50:30,320 --> 00:50:32,040
afraid to go up and talk to him.
1025
00:50:32,840 --> 00:50:34,096
I was working on a problem
1026
00:50:34,120 --> 00:50:35,896
that I thought was very important.
1027
00:50:35,920 --> 00:50:35,937
It was an information theoretic problem.
That I thought was very important.
1028
00:50:35,961 --> 00:50:37,776
It was an information theoretic problem.
1029
00:50:37,800 --> 00:50:40,296
I knew he would be interested in it also.
1030
00:50:40,320 --> 00:50:42,896
I started to talk to him about it.
1031
00:50:42,920 --> 00:50:47,416
And suddenly, the problem
seemed too complicated for him,
1032
00:50:47,440 --> 00:50:48,976
which surprised me, but he said,
1033
00:50:49,000 --> 00:50:50,976
"Let's simplify it a little bit."
1034
00:50:51,000 --> 00:50:54,336
So we took one piece out, then
we took another piece out,
1035
00:50:54,360 --> 00:50:55,696
and another piece out,
1036
00:50:55,720 --> 00:50:57,640
suddenly it was a silly little problem
1037
00:50:58,720 --> 00:51:01,176
and the answer became obvious.
1038
00:51:01,200 --> 00:51:04,400
And then we slowly put the pieces back in
1039
00:51:06,280 --> 00:51:07,600
and the problem was solved.
1040
00:51:08,560 --> 00:51:09,537
How do you do that?
1041
00:51:09,561 --> 00:51:12,736
How do you take pieces
out in the right way
1042
00:51:12,760 --> 00:51:15,496
so that when you get done taking them out
1043
00:51:15,520 --> 00:51:17,000
you can solve the problem?
1044
00:51:17,840 --> 00:51:20,136
He would build these little experiments
1045
00:51:20,160 --> 00:51:23,256
and his little experiments
would give him some insights
1046
00:51:23,280 --> 00:51:26,000
about whether the general
problem could be solved.
1047
00:51:28,120 --> 00:51:28,897
- Hello.
1048
00:51:28,921 --> 00:51:30,816
I'm Claude Shannon, a mathematician here
1049
00:51:30,840 --> 00:51:32,520
at the Bell Telephone Laboratories.
1050
00:51:34,320 --> 00:51:35,096
This is Theseus.
1051
00:51:35,120 --> 00:51:35,137
Theseus is an
electrically-controlled mouse.
This is Theseus.
1052
00:51:35,161 --> 00:51:38,096
Theseus is an
electrically-controlled mouse.
1053
00:51:38,120 --> 00:51:40,496
He has the ability to solve
a certain class of problems
1054
00:51:40,520 --> 00:51:43,616
by trial and error means, and
then remember the solution.
1055
00:51:43,640 --> 00:51:45,840
In other words, he can
learn from experience.
1056
00:51:48,680 --> 00:51:51,256
Like his classical namesake,
Theseus has a problem
1057
00:51:51,280 --> 00:51:53,000
with finding his way through a maze.
1058
00:51:55,720 --> 00:52:00,520
The mouse moves by this electromagnet.
1059
00:52:03,400 --> 00:52:06,376
And its brain are these switches,
1060
00:52:06,400 --> 00:52:07,656
just like the switches you'd find
1061
00:52:07,680 --> 00:52:09,576
in a telephone switch board.
1062
00:52:09,600 --> 00:52:12,096
Betty did most of the wiring, by the way.
1063
00:52:12,120 --> 00:52:12,137
I took it over to Bell
Labs one day to show them.
Betty did most of the wiring, by the way.
1064
00:52:12,161 --> 00:52:15,736
I took it over to Bell
Labs one day to show them.
1065
00:52:15,760 --> 00:52:17,776
They liked it very much,
they were very impressed.
1066
00:52:17,800 --> 00:52:19,800
They even wanted to use it in a film.
1067
00:52:20,680 --> 00:52:22,616
- When did you build this?
1068
00:52:22,640 --> 00:52:23,457
It must've been one
1069
00:52:23,481 --> 00:52:26,296
of the earliest artificial
intelligence devices.
1070
00:52:26,320 --> 00:52:26,335
- Um...
of the earliest artificial
intelligence devices.
1071
00:52:26,359 --> 00:52:27,359
- Um...
1072
00:52:27,960 --> 00:52:28,960
Uh...
1073
00:52:29,880 --> 00:52:30,880
Betty!
1074
00:52:34,720 --> 00:52:36,336
When did we make Theseus?
1075
00:52:36,360 --> 00:52:36,377
- 1950.
- Right.
When did we make Theseus?
1076
00:52:36,401 --> 00:52:38,416
- 1950.
- Right.
1077
00:52:38,440 --> 00:52:41,696
- He made a little turtle
first that would walk
1078
00:52:41,720 --> 00:52:45,416
around the house and bump
into things, and back off.
1079
00:52:45,440 --> 00:52:45,457
And then the mouse.
around the house and bump
into things, and back off.
1080
00:52:45,481 --> 00:52:47,376
And then the mouse.
1081
00:52:47,400 --> 00:52:47,417
- So, you said that you did most
And then the mouse.
1082
00:52:47,441 --> 00:52:49,296
- So, you said that you did most
1083
00:52:49,320 --> 00:52:53,416
of your information theory
work between 1943 and 1945,
1084
00:52:53,440 --> 00:52:58,176
so why didn't you publish
your paper until 1948?
1085
00:52:58,200 --> 00:53:01,096
- Laziness, I guess.
1086
00:53:01,120 --> 00:53:01,137
I have a number of papers
that are unpublished,
- Laziness, I guess.
1087
00:53:01,161 --> 00:53:03,256
I have a number of papers
that are unpublished,
1088
00:53:03,280 --> 00:53:04,177
they're even more tardy than that.
1089
00:53:04,201 --> 00:53:06,256
- Oh, we've got boxes and boxes
1090
00:53:06,280 --> 00:53:08,656
of unfinished papers upstairs.
1091
00:53:08,680 --> 00:53:09,960
Who knows what's in them?
1092
00:53:10,880 --> 00:53:13,216
Maybe we should punish you
and make you go through them.
1093
00:53:13,240 --> 00:53:14,376
(Claude laughs)
1094
00:53:14,400 --> 00:53:18,576
- Well, I never liked writing very much.
1095
00:53:18,600 --> 00:53:22,936
I liked ideas, but I was
lazy about putting them down
1096
00:53:22,960 --> 00:53:24,616
on paper, which is how
you get the acclaim.
1097
00:53:24,640 --> 00:53:25,577
- Hmm.
1098
00:53:25,601 --> 00:53:28,616
After we were married, he
would dictate his papers to me
1099
00:53:28,640 --> 00:53:30,496
and I would write them up.
1100
00:53:30,520 --> 00:53:31,297
- You were a mathematician
1101
00:53:31,321 --> 00:53:32,456
at Bell Labs, too, right?
1102
00:53:32,480 --> 00:53:33,680
- I married a computer.
1103
00:53:35,520 --> 00:53:36,856
- My mother was much, much better
1104
00:53:36,880 --> 00:53:39,136
at arithmetic than my father was.
1105
00:53:39,160 --> 00:53:40,856
She went to Douglas College, which is part
1106
00:53:40,880 --> 00:53:42,456
of Rutgers University,
1107
00:53:42,480 --> 00:53:45,976
which was a wonderful opportunity for her.
1108
00:53:46,000 --> 00:53:50,456
She was a math major, did
very well, and graduated just
1109
00:53:50,480 --> 00:53:52,040
at the beginning of World War II
1110
00:53:52,920 --> 00:53:56,776
when everybody was
looking for math majors.
1111
00:53:56,800 --> 00:53:59,816
And she had several job offers,
which was exciting for her.
1112
00:53:59,840 --> 00:54:02,216
She accepted the one at Bell Labs to be,
1113
00:54:02,240 --> 00:54:02,257
what they called, a computer.
She accepted the one at Bell Labs to be,
1114
00:54:02,281 --> 00:54:03,896
what they called, a computer.
1115
00:54:03,920 --> 00:54:07,296
Her term is, it was a
secretarial pool for math majors.
1116
00:54:07,320 --> 00:54:10,376
So, it was a room full of
women who had been math majors
1117
00:54:10,400 --> 00:54:12,536
who would then do the calculating
1118
00:54:12,560 --> 00:54:15,736
for the male mathematicians and engineers.
1119
00:54:15,760 --> 00:54:17,856
After Bell, my mother took up weaving,
1120
00:54:17,880 --> 00:54:17,897
which is very mathematical.
After Bell, my mother took up weaving,
1121
00:54:17,921 --> 00:54:19,256
which is very mathematical.
1122
00:54:19,280 --> 00:54:19,297
She became an accomplished weaver
which is very mathematical.
1123
00:54:19,321 --> 00:54:20,816
She became an accomplished weaver
1124
00:54:20,840 --> 00:54:23,536
and a pioneer in using
computers in hand-weaving.
1125
00:54:23,560 --> 00:54:25,896
My father, of course,
had to build her a loom.
1126
00:54:25,920 --> 00:54:25,937
My parents were a great partnership.
My father, of course,
had to build her a loom.
1127
00:54:25,961 --> 00:54:27,696
My parents were a great partnership.
1128
00:54:27,720 --> 00:54:30,216
They both appreciated
each other's intelligence,
1129
00:54:30,240 --> 00:54:32,920
and together they were a great team.
1130
00:54:34,600 --> 00:54:38,856
- But I didn't meet Claude
until after the 1948 paper,
1131
00:54:38,880 --> 00:54:42,296
so I couldn't make him
write that one quicker.
1132
00:54:42,320 --> 00:54:44,496
(Claude laughs)
1133
00:54:44,520 --> 00:54:45,536
- You have to understand, too,
1134
00:54:45,560 --> 00:54:49,080
that during the war this
wasn't a priority, this work.
1135
00:54:50,360 --> 00:54:53,576
And I didn't think in the early stages
1136
00:54:53,600 --> 00:54:53,617
it would have such an impact.
And I didn't think in the early stages
1137
00:54:53,641 --> 00:54:55,000
it would have such an impact.
1138
00:54:55,800 --> 00:54:59,576
(intense instrumental music)
1139
00:54:59,600 --> 00:54:59,617
- One of the most beautiful
things about Shannon's paper is
(intense instrumental music)
1140
00:54:59,641 --> 00:55:02,496
- One of the most beautiful
things about Shannon's paper is
1141
00:55:02,520 --> 00:55:02,537
that he reduced the general
problem of communication
- One of the most beautiful
things about Shannon's paper is
1142
00:55:02,561 --> 00:55:06,176
that he reduced the general
problem of communication
1143
00:55:06,200 --> 00:55:08,896
to just two numbers: the minimum size
1144
00:55:08,920 --> 00:55:08,937
you could compress a
message, it's entropy,
to just two numbers: the minimum size
1145
00:55:08,961 --> 00:55:12,256
you could compress a
message, it's entropy,
1146
00:55:12,280 --> 00:55:15,656
and the maximum rate you
could send it without errors,
1147
00:55:15,680 --> 00:55:16,680
the Shannon limit.
1148
00:55:17,480 --> 00:55:19,616
And it just started a race to find ways
1149
00:55:19,640 --> 00:55:20,880
to reach those limits,
1150
00:55:22,440 --> 00:55:25,080
and Bell Labs was at
the center of that race.
1151
00:55:27,320 --> 00:55:29,256
It was the top research institute
1152
00:55:29,280 --> 00:55:31,816
in the world for communications.
1153
00:55:31,840 --> 00:55:31,857
They just invented the
transistor right down the hall
in the world for communications.
1154
00:55:31,881 --> 00:55:34,136
They just invented the
transistor right down the hall
1155
00:55:34,160 --> 00:55:36,856
from Shannon's office in 1947,
1156
00:55:36,880 --> 00:55:36,897
the year before his paper came out.
From Shannon's office in 1947,
1157
00:55:36,921 --> 00:55:38,936
the year before his paper came out.
1158
00:55:38,960 --> 00:55:42,096
They were geniuses across many fields,
1159
00:55:42,120 --> 00:55:44,616
but even there Shannon stood out.
1160
00:55:44,640 --> 00:55:48,160
(gentle instrumental music)
1161
00:55:49,520 --> 00:55:51,736
- I would call it the grand puzzle,
1162
00:55:51,760 --> 00:55:54,240
that's what I call it, the Shannon puzzle.
1163
00:55:55,600 --> 00:55:58,176
All of this great stuff where he showed
1164
00:55:58,200 --> 00:56:01,720
that communicating
information is possible,
1165
00:56:02,600 --> 00:56:04,856
he proved it mathematically,
1166
00:56:04,880 --> 00:56:09,256
but he never told anyone
how to actually do it.
1167
00:56:09,280 --> 00:56:10,776
So, that is the puzzle.
1168
00:56:10,800 --> 00:56:12,040
It's tantalizing.
1169
00:56:12,960 --> 00:56:16,256
You know that those codes exist,
1170
00:56:16,280 --> 00:56:16,297
but Shannon gave really
almost no clue how to go
You know that those codes exist,
1171
00:56:16,321 --> 00:56:21,080
but Shannon gave really
almost no clue how to go
1172
00:56:22,120 --> 00:56:23,320
about finding them.
1173
00:56:26,920 --> 00:56:27,696
- Of course, that only means
1174
00:56:27,720 --> 00:56:27,737
that Theseus can solve any
problem that can be solved.
- Of course, that only means
1175
00:56:27,761 --> 00:56:31,456
that Theseus can solve any
problem that can be solved.
1176
00:56:31,480 --> 00:56:34,256
Like the rest of us, he
occasionally finds himself
1177
00:56:34,280 --> 00:56:36,256
in a situation something like this.
1178
00:56:36,280 --> 00:56:39,880
(intense instrumental music)
1179
00:56:48,560 --> 00:56:50,896
- Why did you
decide to leave Bell Labs
1180
00:56:50,920 --> 00:56:53,096
and come to MIT?
1181
00:56:53,120 --> 00:56:54,057
(Claude sighs)
1182
00:56:54,081 --> 00:56:56,616
- They gave me a lot of freedom at Bell,
1183
00:56:56,640 --> 00:56:56,657
but I started to feel a little stale.
- They gave me a lot of freedom at Bell,
1184
00:56:56,681 --> 00:57:01,360
but I started to feel a little stale.
1185
00:57:02,280 --> 00:57:05,416
I also always felt that the freedom
1186
00:57:05,440 --> 00:57:09,880
I took there was something
of a special favor.
1187
00:57:12,040 --> 00:57:13,840
I thought at MIT...
1188
00:57:14,920 --> 00:57:19,720
Well, the freedom in the
work hours that I took
1189
00:57:21,680 --> 00:57:24,160
might be a little less unusual.
1190
00:57:25,040 --> 00:57:29,376
(intense instrumental music)
1191
00:57:29,400 --> 00:57:32,016
- In the mid '50s, MIT is a powerhouse
1192
00:57:32,040 --> 00:57:33,576
in technology in general,
1193
00:57:33,600 --> 00:57:35,640
and especially in information theory.
1194
00:57:36,920 --> 00:57:41,496
They make a huge effort to
recruit Shannon and they get him.
1195
00:57:41,520 --> 00:57:44,096
And people expect he's gonna be the leader
1196
00:57:44,120 --> 00:57:47,040
in delivering the promise
of his 1948 paper.
1197
00:57:48,480 --> 00:57:49,856
He's not.
1198
00:57:49,880 --> 00:57:53,216
He writes some important
extensions of his work,
1199
00:57:53,240 --> 00:57:55,536
but he starts coming to
campus less and less.
1200
00:57:55,560 --> 00:57:58,096
And, in fact, at a certain point
1201
00:57:58,120 --> 00:57:59,616
he's coming in so infrequently
1202
00:57:59,640 --> 00:58:02,240
they actually put two graduate
students in his office.
1203
00:58:03,480 --> 00:58:06,976
And people are having trouble
finding the codes he said had
1204
00:58:07,000 --> 00:58:10,976
to exist, so there begins
to be some skepticism
1205
00:58:11,000 --> 00:58:12,360
about an information theory.
1206
00:58:13,320 --> 00:58:15,960
And everybody's wondering,
"Where's Claude Shannon?"
1207
00:58:20,560 --> 00:58:23,680
(cheerful piano music)
1208
00:58:28,320 --> 00:58:30,256
- When I was quite
young, he would dress up
1209
00:58:30,280 --> 00:58:32,720
and go to work the way other people did.
1210
00:58:33,720 --> 00:58:36,776
And then, at some point that diminished
1211
00:58:36,800 --> 00:58:38,400
and he was home more often.
1212
00:58:39,440 --> 00:58:40,377
At some point my mother explained
1213
00:58:40,401 --> 00:58:42,736
that he had taken a leave from MIT.
1214
00:58:42,760 --> 00:58:47,176
After he sort of moved on
from the more theoretical
1215
00:58:47,200 --> 00:58:48,616
and academic work that he did,
1216
00:58:48,640 --> 00:58:51,256
and more into things like
unicycling and juggling,
1217
00:58:51,280 --> 00:58:53,536
many people would consider
that he wasn't working anymore,
1218
00:58:53,560 --> 00:58:56,176
but I still think of that as his work.
1219
00:58:56,200 --> 00:58:57,856
My dad got interested
in building all kinds
1220
00:58:57,880 --> 00:59:00,856
of different unicycles, very small wheels,
1221
00:59:00,880 --> 00:59:04,896
very large wheels, very
tall, short, with pedals,
1222
00:59:04,920 --> 00:59:07,616
without pedals, with seat, without seat.
1223
00:59:07,640 --> 00:59:09,496
And we tried to ride them all.
1224
00:59:09,520 --> 00:59:10,696
He had an interest in unicycles
1225
00:59:10,720 --> 00:59:12,056
because my mother bought him one,
1226
00:59:12,080 --> 00:59:16,216
not so easy to come by
back in maybe 1950 or '51.
1227
00:59:16,240 --> 00:59:18,096
It was sort of a running
joke when I was a child
1228
00:59:18,120 --> 00:59:20,576
about how if I wasn't
doing too well in school
1229
00:59:20,600 --> 00:59:20,615
I could always run off
and join the circus,
about how if I wasn't
doing too well in school
1230
00:59:20,639 --> 00:59:22,176
I could always run off
and join the circus,
1231
00:59:22,200 --> 00:59:23,896
and my parents would love that.
1232
00:59:23,920 --> 00:59:26,976
(cheerful piano music)
1233
00:59:27,000 --> 00:59:28,776
Before I was born he knew how to juggle,
1234
00:59:28,800 --> 00:59:28,815
and he taught me how to
juggle when I was maybe eight
Before I was born he knew how to juggle,
1235
00:59:28,839 --> 00:59:30,856
and he taught me how to
juggle when I was maybe eight
1236
00:59:30,880 --> 00:59:30,897
or nine years old.
and he taught me how to
juggle when I was maybe eight
1237
00:59:30,921 --> 00:59:32,256
or nine years old.
1238
00:59:32,280 --> 00:59:34,816
He also started trying to
build machines to juggle.
1239
00:59:34,840 --> 00:59:36,856
- Juggling relates to patterns.
1240
00:59:36,880 --> 00:59:40,816
There's a topology and
a combinatorial aspect.
1241
00:59:40,840 --> 00:59:40,857
I wrote a paper for Scientific American.
There's a topology and
a combinatorial aspect.
1242
00:59:40,881 --> 00:59:43,376
I wrote a paper for Scientific American.
1243
00:59:43,400 --> 00:59:43,417
(cheerful piano music)
(toy rattling)
I wrote a paper for Scientific American.
1244
00:59:43,441 --> 00:59:47,256
(cheerful piano music)
(toy rattling)
1245
00:59:47,280 --> 00:59:49,536
Oh, come on, Mr. Fields.
1246
00:59:49,560 --> 00:59:53,136
(cheerful piano music)
(toy rattling)
1247
00:59:53,160 --> 00:59:54,336
Ah.
1248
00:59:54,360 --> 00:59:56,400
I came up with a formula for juggling
1249
00:59:57,280 --> 00:59:59,496
that relates the number
of balls you're juggling
1250
00:59:59,520 --> 00:59:59,537
to how long each is in the
air versus in the hand.
that relates the number
of balls you're juggling
1251
00:59:59,561 --> 01:00:02,440
to how long each is in the
air versus in the hand.
1252
01:00:03,640 --> 01:00:04,417
- Did you ever think
1253
01:00:04,441 --> 01:00:07,056
about commercializing your inventions?
1254
01:00:07,080 --> 01:00:10,496
- Oh, fiddle, we don't believe
in commercializing fun.
1255
01:00:10,520 --> 01:00:11,296
(Claude chuckles)
1256
01:00:11,320 --> 01:00:11,335
- You didn't worry
(Claude chuckles)
1257
01:00:11,359 --> 01:00:12,097
- You didn't worry
1258
01:00:12,121 --> 01:00:13,416
that they might not be practical?
1259
01:00:13,440 --> 01:00:13,457
- Well, that'd be the last thing.
That they might not be practical?
1260
01:00:13,481 --> 01:00:15,776
- Well, that'd be the last thing.
1261
01:00:15,800 --> 01:00:19,176
I built a lot of useless gadgets
1262
01:00:19,200 --> 01:00:20,800
just because they're fun to make.
1263
01:00:22,480 --> 01:00:25,416
I'm more interested in, is
this an interesting problem?
1264
01:00:25,440 --> 01:00:27,576
Is it an elegant problem?
1265
01:00:27,600 --> 01:00:27,617
I get a big bang out of proving a theory
Is it an elegant problem?
1266
01:00:27,641 --> 01:00:30,216
I get a big bang out of proving a theory
1267
01:00:30,240 --> 01:00:33,616
or seeing a clever way to do something.
1268
01:00:33,640 --> 01:00:35,376
I make my money in the stock market,
1269
01:00:35,400 --> 01:00:36,496
not by proving theorems.
1270
01:00:36,520 --> 01:00:37,776
- You play the stock market?
1271
01:00:37,800 --> 01:00:39,576
You're interested in making money?
1272
01:00:39,600 --> 01:00:40,416
- Oh, sure.
1273
01:00:40,440 --> 01:00:41,440
Aren't you?
1274
01:00:43,480 --> 01:00:44,696
- Were you ever successful?
1275
01:00:44,720 --> 01:00:46,736
- Beyond reasonable expectations.
1276
01:00:46,760 --> 01:00:47,536
(Claude chuckles)
1277
01:00:47,560 --> 01:00:47,577
- I even did some work
on the theory of stocks.
(Claude chuckles)
1278
01:00:47,601 --> 01:00:50,760
- I even did some work
on the theory of stocks.
1279
01:00:52,040 --> 01:00:54,816
In a certain way it's
like channel capacity
1280
01:00:54,840 --> 01:00:56,576
in a noisy channel.
1281
01:00:56,600 --> 01:00:57,457
- You found that you
1282
01:00:57,481 --> 01:01:00,816
could apply information
theory to the stock market?
1283
01:01:00,840 --> 01:01:02,896
- Inside information.
1284
01:01:02,920 --> 01:01:05,856
- We had good friends who
started good companies.
1285
01:01:05,880 --> 01:01:05,895
- Yeah.
- We had good friends who
started good companies.
1286
01:01:05,919 --> 01:01:06,857
- Yeah.
1287
01:01:06,881 --> 01:01:10,336
The bottom line is that
theories aren't as important
1288
01:01:10,360 --> 01:01:13,416
as products or people.
1289
01:01:13,440 --> 01:01:16,416
(gentle instrumental music)
1290
01:01:16,440 --> 01:01:18,976
- In the '60s and '70s
Shannon's ideas began
1291
01:01:19,000 --> 01:01:20,640
to have a lot of impact.
1292
01:01:22,080 --> 01:01:25,136
Computers are shrinking
from the size of rooms,
1293
01:01:25,160 --> 01:01:27,656
data is being compressed
and stored on drives
1294
01:01:27,680 --> 01:01:27,697
and CDs that could be
scratched and still play.
data is being compressed
and stored on drives
1295
01:01:27,721 --> 01:01:30,160
and CDs that could be
scratched and still play.
1296
01:01:31,080 --> 01:01:33,816
The Voyager satellite is
sending images billions
1297
01:01:33,840 --> 01:01:35,400
of miles back to Earth.
1298
01:01:36,880 --> 01:01:39,496
But when it comes to finding the codes
1299
01:01:39,520 --> 01:01:42,696
that could get you anywhere
near the Shannon limit,
1300
01:01:42,720 --> 01:01:44,320
there's a lot of frustration.
1301
01:01:45,280 --> 01:01:49,256
A lot of very talented
people, over decades,
1302
01:01:49,280 --> 01:01:53,600
tried and failed to find
the codes he said existed.
1303
01:01:55,520 --> 01:01:57,896
By the 1980s there's a sense
1304
01:01:57,920 --> 01:02:00,736
that maybe we've reached
a practical limit,
1305
01:02:00,760 --> 01:02:03,696
a real world boundary
you can't get through.
1306
01:02:03,720 --> 01:02:08,520
The Shannon limit is beautiful,
an elegant theoretical idea,
1307
01:02:08,680 --> 01:02:10,000
but it's unattainable.
1308
01:02:10,840 --> 01:02:13,416
So people began to lose faith and wonder,
1309
01:02:13,440 --> 01:02:15,736
"Is information theory dead?"
1310
01:02:15,760 --> 01:02:15,777
And they begin to abandon
information theory and,
"Is information theory dead?"
1311
01:02:15,801 --> 01:02:19,456
And they begin to abandon
information theory and,
1312
01:02:19,480 --> 01:02:22,280
in a sense, abandon Shannon too.
1313
01:02:25,440 --> 01:02:28,016
- It seems like after your big discoveries
1314
01:02:28,040 --> 01:02:30,416
you stopped really working
on information theory.
1315
01:02:30,440 --> 01:02:33,080
Is that because you just grew tired of it?
1316
01:02:34,360 --> 01:02:39,160
- Well, the first part is right,
that I didn't do much work
1317
01:02:40,640 --> 01:02:43,656
on it later, but, you know, most
1318
01:02:43,680 --> 01:02:43,697
of the great mathematicians
have done their finest work
on it later, but, you know, most
1319
01:02:43,721 --> 01:02:46,576
of the great mathematicians
have done their finest work
1320
01:02:46,600 --> 01:02:49,976
in their '20s and '30s and not so much
1321
01:02:50,000 --> 01:02:52,416
when they're older, like athletes.
1322
01:02:52,440 --> 01:02:55,976
Although, I've just recently
taken up jogging. (chuckles)
1323
01:02:56,000 --> 01:02:58,056
- But it seems like you
spend a lot more time working
1324
01:02:58,080 --> 01:02:59,736
on your gadgets now.
1325
01:02:59,760 --> 01:03:03,536
- I have a lot of things
I like to spend time on.
1326
01:03:03,560 --> 01:03:07,720
The intelligent machine
idea is my big interest now.
1327
01:03:09,160 --> 01:03:13,960
The transformation of information
that a human brain does,
1328
01:03:14,320 --> 01:03:19,120
not just the transmission, how
the brain works or doesn't.
1329
01:03:20,800 --> 01:03:22,800
I think that's the biggest challenge now.
1330
01:03:23,880 --> 01:03:25,616
(bleeping)
1331
01:03:25,640 --> 01:03:28,040
Simulating the human brain.
1332
01:03:28,840 --> 01:03:31,216
Alan Turing and I often had
lunch together and we'd talk
1333
01:03:31,240 --> 01:03:31,257
about if you could get a computer
the equivalent of a brain,
Alan Turing and I often had
lunch together and we'd talk
1334
01:03:31,281 --> 01:03:35,296
about if you could get a computer
the equivalent of a brain,
1335
01:03:35,320 --> 01:03:37,776
a human brain, or even a lot better.
1336
01:03:37,800 --> 01:03:37,817
I wrote a paper about how a
computer could play chess.
a human brain, or even a lot better.
1337
01:03:37,841 --> 01:03:41,656
I wrote a paper about how a
computer could play chess.
1338
01:03:41,680 --> 01:03:42,657
I have a standing bet
1339
01:03:42,681 --> 01:03:46,696
that a computer will beat
the grand champion by 2001.
1340
01:03:46,720 --> 01:03:46,737
(big band music)
that a computer will beat
the grand champion by 2001.
1341
01:03:46,761 --> 01:03:48,256
(big band music)
1342
01:03:48,280 --> 01:03:51,136
- Back in 1949, when computers were first
1343
01:03:51,160 --> 01:03:53,656
becoming available, Shannon had the vision
1344
01:03:53,680 --> 01:03:53,697
that it would be great to
program one of these contraptions
becoming available, Shannon had the vision
1345
01:03:53,721 --> 01:03:58,480
that it would be great to
program one of these contraptions
1346
01:03:58,520 --> 01:04:02,416
to play the complicated,
difficult game of chess.
1347
01:04:02,440 --> 01:04:04,416
He laid out a complete roadmap
1348
01:04:04,440 --> 01:04:06,296
of how to build a chess program.
1349
01:04:06,320 --> 01:04:08,936
- Of course, being Claude
Shannon, it's not enough
1350
01:04:08,960 --> 01:04:12,096
to just write a paper, he also builds one
1351
01:04:12,120 --> 01:04:15,736
of the world's first successful
chess-playing computers.
1352
01:04:15,760 --> 01:04:18,216
Now, computers weren't
powerful enough at the time
1353
01:04:18,240 --> 01:04:20,176
to play a full game of chess,
1354
01:04:20,200 --> 01:04:20,217
so his computer can only play
the last six moves of a game.
to play a full game of chess,
1355
01:04:20,241 --> 01:04:24,176
so his computer can only play
the last six moves of a game.
1356
01:04:24,200 --> 01:04:27,776
He calls it End Game, because
that's what it can play.
1357
01:04:27,800 --> 01:04:31,496
It took eight years from
when End Game was completed
1358
01:04:31,520 --> 01:04:34,336
to when someone else came
and used Shannon's ideas
1359
01:04:34,360 --> 01:04:36,936
to build a computer that could
play a full game of chess.
1360
01:04:36,960 --> 01:04:39,296
- Shannon observed that although chess
1361
01:04:39,320 --> 01:04:39,337
is incredibly complicated,
the astronomical number
- Shannon observed that although chess
1362
01:04:39,361 --> 01:04:42,576
is incredibly complicated,
the astronomical number
1363
01:04:42,600 --> 01:04:45,296
of possible chess games,
and he calculated those,
1364
01:04:45,320 --> 01:04:47,336
you don't have to look all
the way to the end of the game
1365
01:04:47,360 --> 01:04:47,377
to figure out what a good
move is, you only have
you don't have to look all
the way to the end of the game
1366
01:04:47,401 --> 01:04:50,136
to figure out what a good
move is, you only have
1367
01:04:50,160 --> 01:04:53,136
to look ahead a few moves
and conduct some sort
1368
01:04:53,160 --> 01:04:55,696
of evaluation about the
possible moves you could reach
1369
01:04:55,720 --> 01:04:58,336
and find the path that
leads to the best one.
1370
01:04:58,360 --> 01:05:02,496
This principle became the
basis for every chess program
1371
01:05:02,520 --> 01:05:03,936
that came after it.
1372
01:05:03,960 --> 01:05:03,977
(big band music)
that came after it.
1373
01:05:04,001 --> 01:05:06,640
(big band music)
1374
01:05:12,400 --> 01:05:14,816
- When you've done something
like Shannon did in '48
1375
01:05:14,840 --> 01:05:18,056
and then again, or earlier
in '45, I mean, cryptography
1376
01:05:18,080 --> 01:05:18,097
and information theory, what
do you do for an encore?
and then again, or earlier
in '45, I mean, cryptography
1377
01:05:18,121 --> 01:05:21,016
and information theory, what
do you do for an encore?
1378
01:05:21,040 --> 01:05:22,656
It's really hard.
1379
01:05:22,680 --> 01:05:26,336
And I think people are waiting
for either something good
1380
01:05:26,360 --> 01:05:30,656
to come out, or some
people who, maybe jealous,
1381
01:05:30,680 --> 01:05:32,856
are waiting for you to fall down and say,
1382
01:05:32,880 --> 01:05:34,496
"See, he's a has been."
1383
01:05:34,520 --> 01:05:36,576
- There is that small contusion
of people though who say,
1384
01:05:36,600 --> 01:05:41,296
"Well, you know, Shannon worked
on information theory alone
1385
01:05:41,320 --> 01:05:41,337
"in private for 10 years,
"Well, you know, Shannon worked
on information theory alone
1386
01:05:41,361 --> 01:05:43,136
"in private for 10 years,
1387
01:05:43,160 --> 01:05:45,896
"and then dropped this great
achievement on everyone's lap
1388
01:05:45,920 --> 01:05:45,937
"without telling a
soul, maybe, just maybe,
"and then dropped this great
achievement on everyone's lap
1389
01:05:45,961 --> 01:05:49,136
"without telling a
soul, maybe, just maybe,
1390
01:05:49,160 --> 01:05:50,976
"he's up to something again, like that,
1391
01:05:51,000 --> 01:05:52,056
"that he's working on something.
1392
01:05:52,080 --> 01:05:54,656
"Could it be that he just has
a project that he's working on
1393
01:05:54,680 --> 01:05:56,896
"that he hasn't told anyone about?"
1394
01:05:56,920 --> 01:05:59,320
- Do you wanna see the
first wearable computer?
1395
01:06:00,200 --> 01:06:01,200
- Okay.
1396
01:06:04,360 --> 01:06:06,456
- I built this with Ed Thorp,
1397
01:06:06,480 --> 01:06:08,376
who was a professor here at MIT.
1398
01:06:08,400 --> 01:06:12,536
Ed had a paper he wanted to
get published about blackjack
1399
01:06:12,560 --> 01:06:15,736
in the proceedings of the
National Academy of Sciences.
1400
01:06:15,760 --> 01:06:15,777
Since I was the only member
of the academy at the time
in the proceedings of the
National Academy of Sciences.
1401
01:06:15,801 --> 01:06:18,336
Since I was the only member
of the academy at the time
1402
01:06:18,360 --> 01:06:21,736
at MIT, he asked for my help.
1403
01:06:21,760 --> 01:06:23,616
We got to talking, and it turns out
1404
01:06:23,640 --> 01:06:23,657
he had some pretty interesting
ideas about roulette, too.
We got to talking, and it turns out
1405
01:06:23,681 --> 01:06:27,096
he had some pretty interesting
ideas about roulette, too.
1406
01:06:27,120 --> 01:06:29,656
(intense tropical music)
1407
01:06:29,680 --> 01:06:29,697
- My father was very
interested in gambling.
(intense tropical music)
1408
01:06:29,721 --> 01:06:32,576
- My father was very
interested in gambling.
1409
01:06:32,600 --> 01:06:34,776
He loved con men, he loved that sort
1410
01:06:34,800 --> 01:06:34,817
of shadier side of people.
He loved con men, he loved that sort
1411
01:06:34,841 --> 01:06:37,416
of shadier side of people.
1412
01:06:37,440 --> 01:06:39,456
Not interested in violence at all,
1413
01:06:39,480 --> 01:06:41,536
he was interested in cleverness.
1414
01:06:41,560 --> 01:06:45,016
So, movies that captured that
cleverness, like "Topkapi"
1415
01:06:45,040 --> 01:06:48,776
or "The Music Man", were favorites of his.
1416
01:06:48,800 --> 01:06:51,776
So, he ordered a roulette
wheel from Las Vegas
1417
01:06:51,800 --> 01:06:53,736
and they set it up in the
basement with a strobe light
1418
01:06:53,760 --> 01:06:57,176
and went to work on a wearable computer
1419
01:06:57,200 --> 01:07:00,616
that could try to improve
the odds for roulette.
1420
01:07:00,640 --> 01:07:05,440
- Our little computer could
time the speed of the wheel
1421
01:07:06,120 --> 01:07:08,216
and the speed of the ball
in the other direction.
1422
01:07:08,240 --> 01:07:11,736
It's a simple dynamic system
with very little friction.
1423
01:07:11,760 --> 01:07:11,777
We discovered that nearly
all roulette tables
It's a simple dynamic system
with very little friction.
1424
01:07:11,801 --> 01:07:16,056
We discovered that nearly
all roulette tables
1425
01:07:16,080 --> 01:07:18,816
are slightly tilted, so we
could predict two thirds
1426
01:07:18,840 --> 01:07:20,936
of the time which half of the
wheel the ball would fall in.
1427
01:07:20,960 --> 01:07:20,977
The computer would be in my pocket,
of the time which half of the
wheel the ball would fall in.
1428
01:07:21,001 --> 01:07:24,336
The computer would be in my pocket,
1429
01:07:24,360 --> 01:07:26,896
and these little wires
would run down my pant leg
1430
01:07:26,920 --> 01:07:26,937
into my shoe, and there was
a little switch in there
and these little wires
would run down my pant leg
1431
01:07:26,961 --> 01:07:29,616
into my shoe, and there was
a little switch in there
1432
01:07:29,640 --> 01:07:32,056
that I operated with my big toe.
1433
01:07:32,080 --> 01:07:36,880
And then these ran up to a
little loud speaker in my ear,
1434
01:07:36,920 --> 01:07:39,856
and the computer would send
one of the eight musical notes
1435
01:07:39,880 --> 01:07:43,096
of the scale to mark the
octants in the wheel.
1436
01:07:43,120 --> 01:07:45,416
Ed had a little loud
speaker in his ear, too,
1437
01:07:45,440 --> 01:07:49,736
and I operated the computer,
and Ed placed the bets
1438
01:07:49,760 --> 01:07:49,777
and pretended he didn't know me.
and I operated the computer,
and Ed placed the bets
1439
01:07:49,801 --> 01:07:51,496
and pretended he didn't know me.
1440
01:07:51,520 --> 01:07:54,800
(intense tropical music)
1441
01:07:56,440 --> 01:07:59,536
We had some practical problems
when we headed to Las Vegas.
1442
01:07:59,560 --> 01:08:02,376
The wires going down my
pant leg kept shorting
1443
01:08:02,400 --> 01:08:03,776
and giving me little shocks.
1444
01:08:03,800 --> 01:08:04,777
(all laugh)
1445
01:08:04,801 --> 01:08:07,416
- And Ed's speaker fell out of his ear,
1446
01:08:07,440 --> 01:08:09,376
and he nearly scared this lady to death,
1447
01:08:09,400 --> 01:08:12,576
because she thought it
was some alien insect.
1448
01:08:12,600 --> 01:08:12,617
(all laugh)
because she thought it
was some alien insect.
1449
01:08:12,641 --> 01:08:14,816
(all laugh)
1450
01:08:14,840 --> 01:08:17,296
We were lucky we didn't get arrested.
1451
01:08:17,320 --> 01:08:18,656
(all laugh)
1452
01:08:18,680 --> 01:08:21,656
(intense tropical music)
1453
01:08:21,680 --> 01:08:22,456
- He said this at one point,
1454
01:08:22,480 --> 01:08:22,497
he wasn't interested in practicality.
- He said this at one point,
1455
01:08:22,521 --> 01:08:26,320
he wasn't interested in practicality.
1456
01:08:27,520 --> 01:08:29,640
I don't think that was
quite what he meant.
1457
01:08:31,400 --> 01:08:35,240
He was very interested in
what's needed to be practical,
1458
01:08:36,680 --> 01:08:40,816
but he didn't wanna spend
his life starting a company.
1459
01:08:40,840 --> 01:08:40,857
He knew that that wasn't
the right thing for him.
but he didn't wanna spend
his life starting a company.
1460
01:08:40,881 --> 01:08:43,456
He knew that that wasn't
the right thing for him.
1461
01:08:43,480 --> 01:08:45,296
People who start companies have
1462
01:08:45,320 --> 01:08:48,416
to feel a very strong competitive
urge, they have to want
1463
01:08:48,440 --> 01:08:53,240
to beat other people, and he
was not a competitive person.
1464
01:08:55,080 --> 01:08:58,576
He was a person who,
whenever you talked to him,
1465
01:08:58,600 --> 01:09:02,096
he was not trying to
dominate you intellectually,
1466
01:09:02,120 --> 01:09:06,920
he was very much a
cooperator, not a competitor.
1467
01:09:08,480 --> 01:09:11,376
He really was interested
in practical things.
1468
01:09:11,400 --> 01:09:11,417
All the things he built
were very practical.
He really was interested
in practical things.
1469
01:09:11,441 --> 01:09:13,656
All the things he built
were very practical.
1470
01:09:13,680 --> 01:09:17,616
His maze, a toy, yes, but it was a toy
1471
01:09:17,640 --> 01:09:21,376
that illustrated some
very interesting things.
1472
01:09:21,400 --> 01:09:24,936
His chess-playing machine, again, a toy,
1473
01:09:24,960 --> 01:09:27,216
but artificial intelligence people
1474
01:09:27,240 --> 01:09:27,257
just learned so many things from that,
but artificial intelligence people
1475
01:09:27,281 --> 01:09:29,056
just learned so many things from that,
1476
01:09:29,080 --> 01:09:30,976
and it was highly useful.
1477
01:09:31,000 --> 01:09:33,656
His cryptography during
the second World War,
1478
01:09:33,680 --> 01:09:36,576
the fact that it wasn't
very practical then
1479
01:09:36,600 --> 01:09:38,136
is almost irrelevant.
1480
01:09:38,160 --> 01:09:38,177
He still wrote the best
paper on cryptography
is almost irrelevant.
1481
01:09:38,201 --> 01:09:41,096
He still wrote the best
paper on cryptography
1482
01:09:41,120 --> 01:09:41,137
that's been written yet.
He still wrote the best
paper on cryptography
1483
01:09:41,161 --> 01:09:42,856
that's been written yet.
1484
01:09:42,880 --> 01:09:44,696
- He asked questions that no one asked,
1485
01:09:44,720 --> 01:09:46,096
and that was an amazing thing.
1486
01:09:46,120 --> 01:09:48,896
The questions are very
meaningful, and he answered them,
1487
01:09:48,920 --> 01:09:48,935
and no one thought those questions
The questions are very
meaningful, and he answered them,
1488
01:09:48,959 --> 01:09:50,176
and no one thought those questions
1489
01:09:50,200 --> 01:09:52,256
could be answered that way.
1490
01:09:52,280 --> 01:09:54,240
- Practical or playful?
1491
01:09:55,560 --> 01:09:57,560
It's a silly question in most cases.
1492
01:09:59,120 --> 01:10:00,520
He was into understanding.
1493
01:10:05,920 --> 01:10:07,136
- What do
you think about the idea
1494
01:10:07,160 --> 01:10:10,400
that a machine might one
day outthink a human?
1495
01:10:11,840 --> 01:10:14,360
- Well, I'm
rooting for the machines.
1496
01:10:16,480 --> 01:10:18,096
- But do you think a
machine would ever come up
1497
01:10:18,120 --> 01:10:18,137
with the idea of a flaming trumpet?
- But do you think a
machine would ever come up
1498
01:10:18,161 --> 01:10:20,616
with the idea of a flaming trumpet?
1499
01:10:20,640 --> 01:10:22,176
(Claude chuckles)
1500
01:10:22,200 --> 01:10:25,576
- I think that creative
stuff is a long way off,
1501
01:10:25,600 --> 01:10:30,400
but in the meantime they
can be a boon to humanity.
1502
01:10:31,160 --> 01:10:33,056
(Claude sighs)
1503
01:10:33,080 --> 01:10:33,097
And I'd like to see that
happen as soon as possible.
(Claude sighs)
1504
01:10:33,121 --> 01:10:37,400
And I'd like to see that
happen as soon as possible.
1505
01:10:40,160 --> 01:10:42,696
- Shannon's Master's
thesis lays the foundation
1506
01:10:42,720 --> 01:10:45,816
for this idea that machines
could be made to think,
1507
01:10:45,840 --> 01:10:45,857
that they could capably
mimic a human brain.
for this idea that machines
could be made to think,
1508
01:10:45,881 --> 01:10:48,416
that they could capably
mimic a human brain.
1509
01:10:48,440 --> 01:10:51,736
And he himself invents
devices that mimic a brain.
1510
01:10:51,760 --> 01:10:54,016
There's Theseus, the mouse
that can navigate its way
1511
01:10:54,040 --> 01:10:56,336
through a maze, there's
End Game, the computer
1512
01:10:56,360 --> 01:10:59,160
that can play the last
six moves in a chess game,
1513
01:10:59,960 --> 01:11:04,336
but it's at precisely the time
when artificial intelligence
1514
01:11:04,360 --> 01:11:04,377
and computing is taking off
but it's at precisely the time
when artificial intelligence
1515
01:11:04,401 --> 01:11:07,216
and computing is taking off
1516
01:11:07,240 --> 01:11:10,296
that Shannon is on the sidelines.
1517
01:11:10,320 --> 01:11:10,337
Even though information
theory was being applied
that Shannon is on the sidelines.
1518
01:11:10,361 --> 01:11:13,016
Even though information
theory was being applied
1519
01:11:13,040 --> 01:11:13,057
in ways beyond what
Shannon could've imagined.
Even though information
theory was being applied
1520
01:11:13,081 --> 01:11:15,320
in ways beyond what
Shannon could've imagined.
1521
01:11:16,160 --> 01:11:18,656
- I heard that one
of your last papers proved
1522
01:11:18,680 --> 01:11:20,336
that you could make a reliable circuit
1523
01:11:20,360 --> 01:11:22,656
from faulty components.
1524
01:11:22,680 --> 01:11:24,256
Do you think that this
might have applications
1525
01:11:24,280 --> 01:11:25,097
for the brain?
1526
01:11:25,121 --> 01:11:26,656
(Claude chuckles)
1527
01:11:26,680 --> 01:11:27,680
- I hope so.
1528
01:11:28,320 --> 01:11:30,640
I mean, the brain can
suffer all kinds of damage
1529
01:11:31,560 --> 01:11:34,416
and still handle things pretty well,
1530
01:11:34,440 --> 01:11:37,016
which means there must be redundancy.
1531
01:11:37,040 --> 01:11:40,136
The fact that the brain
has 10 billion neurons
1532
01:11:40,160 --> 01:11:42,536
probably means that it
was cheaper for biology
1533
01:11:42,560 --> 01:11:45,896
to just make more components than
1534
01:11:45,920 --> 01:11:45,937
to figure out sophisticated circuits.
To just make more components than
1535
01:11:45,961 --> 01:11:50,176
to figure out sophisticated circuits.
1536
01:11:50,200 --> 01:11:50,217
Trying to work out all
the different connections
to figure out sophisticated circuits.
1537
01:11:50,241 --> 01:11:53,176
Trying to work out all
the different connections
1538
01:11:53,200 --> 01:11:58,000
would be astronomical, impossible,
1539
01:11:58,520 --> 01:12:01,680
but if we can find patterns,
1540
01:12:03,480 --> 01:12:08,280
well, there may be ways
to simplify things.
1541
01:12:09,840 --> 01:12:12,576
- If we look at the data in the brain,
1542
01:12:12,600 --> 01:12:15,856
all the data is
interrelated to each other.
1543
01:12:15,880 --> 01:12:17,920
So, there's a huge amount of redundancy.
1544
01:12:18,800 --> 01:12:21,016
If I have this type of
neuron there, it tells me
1545
01:12:21,040 --> 01:12:21,055
that there's another kind of neuron there,
If I have this type of
neuron there, it tells me
1546
01:12:21,079 --> 01:12:22,416
that there's another kind of neuron there,
1547
01:12:22,440 --> 01:12:24,336
and this kind of neuron,
and this one's not there,
1548
01:12:24,360 --> 01:12:24,377
and this one's not there.
and this kind of neuron,
and this one's not there,
1549
01:12:24,401 --> 01:12:25,416
and this one's not there.
1550
01:12:25,440 --> 01:12:27,296
So, we develop rules
1551
01:12:27,320 --> 01:12:29,480
that decide this is how neurons connect.
1552
01:12:30,880 --> 01:12:33,096
We take that brain model,
1553
01:12:33,120 --> 01:12:37,576
and we basically attach it
to a virtual mouse body.
1554
01:12:37,600 --> 01:12:40,896
The virtual mouse body has
simple models of the eyes
1555
01:12:40,920 --> 01:12:43,056
and the ears and the skin and the whiskers
1556
01:12:43,080 --> 01:12:43,097
so they can receive sensory input,
and the ears and the skin and the whiskers
1557
01:12:43,121 --> 01:12:45,536
so they can receive sensory input,
1558
01:12:45,560 --> 01:12:48,136
then you let the virtual mouse loose
1559
01:12:48,160 --> 01:12:48,177
to start running through the maze.
Then you let the virtual mouse loose
1560
01:12:48,201 --> 01:12:50,016
to start running through the maze.
1561
01:12:50,040 --> 01:12:50,057
So, in that way it's quite similar
to start running through the maze.
1562
01:12:50,081 --> 01:12:51,936
So, in that way it's quite similar
1563
01:12:51,960 --> 01:12:55,296
to what Shannon did nearly 60 years ago.
1564
01:12:55,320 --> 01:12:57,896
Einstein's theory of general relativity,
1565
01:12:57,920 --> 01:13:01,856
he had an enormous impact
on society and the world
1566
01:13:01,880 --> 01:13:05,120
and physics, and pretty
much how the world is today.
1567
01:13:06,320 --> 01:13:09,216
But I think that we haven't
even begun to see the kind
1568
01:13:09,240 --> 01:13:12,376
of impact that Shannon is
going to have on society,
1569
01:13:12,400 --> 01:13:16,816
because I think the more
we understand how...
1570
01:13:16,840 --> 01:13:16,857
Understanding entropy and
using information theory
because I think the more
we understand how...
1571
01:13:16,881 --> 01:13:20,736
Understanding entropy and
using information theory
1572
01:13:20,760 --> 01:13:23,856
to help us understand complex systems,
1573
01:13:23,880 --> 01:13:23,897
such as the brain or biology or diseases,
to help us understand complex systems,
1574
01:13:23,921 --> 01:13:28,216
such as the brain or biology or diseases,
1575
01:13:28,240 --> 01:13:31,480
I think that this is gonna
open an amazing new door.
1576
01:13:36,520 --> 01:13:38,856
- Most physicists didn't
really care about information
1577
01:13:38,880 --> 01:13:43,456
as a primary focus for many, many years.
1578
01:13:43,480 --> 01:13:45,816
Even most people thinking about abstruse,
1579
01:13:45,840 --> 01:13:47,216
esoteric questions in fundamental physics,
1580
01:13:47,240 --> 01:13:49,336
but that really changed in the 1980s
1581
01:13:49,360 --> 01:13:49,377
when a lot of people focused
but that really changed in the 1980s
1582
01:13:49,401 --> 01:13:50,656
when a lot of people focused
1583
01:13:50,680 --> 01:13:54,536
on the famous black hole
information paradox.
1584
01:13:54,560 --> 01:13:57,776
If you collide particles
with enough energy,
1585
01:13:57,800 --> 01:13:59,936
(whooshing)
1586
01:13:59,960 --> 01:14:03,176
they'll have so much mass that
they'll collapse the region
1587
01:14:03,200 --> 01:14:05,280
that they collided in into a black hole.
1588
01:14:06,640 --> 01:14:10,216
In a classical world, nothing
can get out of the black hole,
1589
01:14:10,240 --> 01:14:13,256
even something going at the
speed of light could not get out
1590
01:14:13,280 --> 01:14:13,297
of a black hole.
even something going at the
speed of light could not get out
1591
01:14:13,321 --> 01:14:15,056
of a black hole.
1592
01:14:15,080 --> 01:14:16,176
But Stephen Hawking showed
1593
01:14:16,200 --> 01:14:18,600
that quantum-mechanically
black holes leak,
1594
01:14:20,240 --> 01:14:21,696
but what came out of the black hole
1595
01:14:21,720 --> 01:14:21,737
is completely independent
of what made the black hole,
but what came out of the black hole
1596
01:14:21,761 --> 01:14:24,296
is completely independent
of what made the black hole,
1597
01:14:24,320 --> 01:14:24,337
and one was left with a major paradox.
is completely independent
of what made the black hole,
1598
01:14:24,361 --> 01:14:26,616
and one was left with a major paradox.
1599
01:14:26,640 --> 01:14:26,657
Where did the information about
what made the black hole go?
and one was left with a major paradox.
1600
01:14:26,681 --> 01:14:29,576
Where did the information about
what made the black hole go?
1601
01:14:29,600 --> 01:14:30,496
Right?
1602
01:14:30,520 --> 01:14:30,535
Because seemingly you
could make a black hole
Right?
1603
01:14:30,559 --> 01:14:31,656
Because seemingly you
could make a black hole
1604
01:14:31,680 --> 01:14:32,577
in a zillion different ways.
1605
01:14:32,601 --> 01:14:34,176
You could throw in an elephant,
1606
01:14:34,200 --> 01:14:34,976
you could throw in an encyclopedia,
1607
01:14:35,000 --> 01:14:35,017
but what came out always
looked exactly the same.
you could throw in an encyclopedia,
1608
01:14:35,041 --> 01:14:37,536
but what came out always
looked exactly the same.
1609
01:14:37,560 --> 01:14:40,096
It looked like information was lost.
1610
01:14:40,120 --> 01:14:41,176
And that is a no-no.
1611
01:14:41,200 --> 01:14:43,336
That cannot happen
according to any of the laws
1612
01:14:43,360 --> 01:14:45,656
of either classical or quantum mechanics.
1613
01:14:45,680 --> 01:14:48,216
And in order to decide whether
the information is lost
1614
01:14:48,240 --> 01:14:51,656
or not, one thing that
became apparent eventually is
1615
01:14:51,680 --> 01:14:55,376
that thinking more carefully
about what the information was
1616
01:14:55,400 --> 01:14:55,417
and very much making use of the ideas
that thinking more carefully
about what the information was
1617
01:14:55,441 --> 01:14:57,656
and very much making use of the ideas
1618
01:14:57,680 --> 01:15:00,920
of entropy ended up being very important.
1619
01:15:02,120 --> 01:15:03,616
What's clear is that information
1620
01:15:03,640 --> 01:15:05,496
in physics is more important
1621
01:15:05,520 --> 01:15:07,616
than you would naively have thought.
1622
01:15:07,640 --> 01:15:10,616
And many of the most
fundamental ideas really go back
1623
01:15:10,640 --> 01:15:10,655
to Shannon.
And many of the most
fundamental ideas really go back
1624
01:15:10,679 --> 01:15:11,679
to Shannon.
1625
01:15:14,480 --> 01:15:17,416
- In 1993, I was finishing up my PhD.
1626
01:15:17,440 --> 01:15:18,936
I was one year away from graduating,
1627
01:15:18,960 --> 01:15:21,336
and I went to a communications conference.
1628
01:15:21,360 --> 01:15:23,576
It wasn't an information
theory conference.
1629
01:15:23,600 --> 01:15:27,296
It was in Switzerland and
it's a very big conference,
1630
01:15:27,320 --> 01:15:29,736
so there's maybe 800 or
900 people that attend.
1631
01:15:29,760 --> 01:15:32,776
There's usually 10 to 12 parallel sessions
1632
01:15:32,800 --> 01:15:32,817
on everything from what's
the next generation
There's usually 10 to 12 parallel sessions
1633
01:15:32,841 --> 01:15:35,736
on everything from what's
the next generation
1634
01:15:35,760 --> 01:15:38,296
of cell phone standards gonna look like
1635
01:15:38,320 --> 01:15:40,656
to more mathematical types of talks
1636
01:15:40,680 --> 01:15:42,896
in more theoretical dimensions.
1637
01:15:42,920 --> 01:15:46,336
So, I went to this session
'cause there was nothing else
1638
01:15:46,360 --> 01:15:46,377
to see, and maybe there was
one person in the session
So, I went to this session
'cause there was nothing else
1639
01:15:46,401 --> 01:15:48,816
to see, and maybe there was
one person in the session
1640
01:15:48,840 --> 01:15:50,096
that I thought was a good speaker.
1641
01:15:50,120 --> 01:15:52,976
So I thought, "Okay, I'll
go listen to that talk."
1642
01:15:53,000 --> 01:15:53,017
And there was a presentation.
So I thought, "Okay, I'll
go listen to that talk."
1643
01:15:53,041 --> 01:15:54,456
And there was a presentation.
1644
01:15:54,480 --> 01:15:57,176
I didn't know any of the
authors, they were really unknown
1645
01:15:57,200 --> 01:15:58,976
to information theorists,
coding theorists,
1646
01:15:59,000 --> 01:16:00,136
nobody knew who they were.
1647
01:16:00,160 --> 01:16:03,296
And these guys that got up,
they were two French guys
1648
01:16:03,320 --> 01:16:05,256
and one guy from Thailand who nobody knew.
1649
01:16:05,280 --> 01:16:05,297
I mean, they were completely unknown.
And one guy from Thailand who nobody knew.
1650
01:16:05,321 --> 01:16:06,936
I mean, they were completely unknown.
1651
01:16:06,960 --> 01:16:09,496
And so, most people were probably
chatting with each other,
1652
01:16:09,520 --> 01:16:09,537
and maybe there were a few
people paying attention.
And so, most people were probably
chatting with each other,
1653
01:16:09,561 --> 01:16:12,056
and maybe there were a few
people paying attention.
1654
01:16:12,080 --> 01:16:12,095
And they get up and they say
and maybe there were a few
people paying attention.
1655
01:16:12,119 --> 01:16:13,017
And they get up and they say
1656
01:16:13,041 --> 01:16:15,256
they basically achieved the Shannon limit,
1657
01:16:15,280 --> 01:16:17,776
and everybody was incredibly skeptical.
1658
01:16:17,800 --> 01:16:17,815
We're like, "Oh, there must be a mistake
and everybody was incredibly skeptical.
1659
01:16:17,839 --> 01:16:19,096
We're like, "Oh, there must be a mistake
1660
01:16:19,120 --> 01:16:21,256
"in their simulations, or, you know,
1661
01:16:21,280 --> 01:16:22,496
"they don't know what
they're talking about,
1662
01:16:22,520 --> 01:16:23,497
"or their math is wrong."
1663
01:16:23,521 --> 01:16:27,176
- It was such a shock
that people were just,
1664
01:16:27,200 --> 01:16:27,217
"It doesn't make any
sense, this ugly thing."
- It was such a shock
that people were just,
1665
01:16:27,241 --> 01:16:29,536
"It doesn't make any
sense, this ugly thing."
1666
01:16:29,560 --> 01:16:29,577
- It was mostly an experimental result.
"It doesn't make any
sense, this ugly thing."
1667
01:16:29,601 --> 01:16:32,976
- It was mostly an experimental result.
1668
01:16:33,000 --> 01:16:36,080
And the idea was to take
multiple simple things
1669
01:16:37,000 --> 01:16:38,896
and combine them in a way
1670
01:16:38,920 --> 01:16:41,816
that created this very complex code.
1671
01:16:41,840 --> 01:16:41,857
And then they showed
their simulation results,
that created this very complex code.
1672
01:16:41,881 --> 01:16:44,096
And then they showed
their simulation results,
1673
01:16:44,120 --> 01:16:47,736
which basically hit the Shannon
limit, and it was amazing.
1674
01:16:47,760 --> 01:16:50,256
(intense instrumental music)
1675
01:16:50,280 --> 01:16:54,096
50 years on, the final
piece of Shannon's vision
1676
01:16:54,120 --> 01:16:56,640
of the future was proved to be true.
1677
01:16:58,160 --> 01:17:00,376
You could have perfect communication
1678
01:17:00,400 --> 01:17:02,120
with as little errors as you wanted.
1679
01:17:03,400 --> 01:17:06,496
- Actually, everything that
Shannon has predicted back
1680
01:17:06,520 --> 01:17:06,537
in 1948 came true.
- Actually, everything that
Shannon has predicted back
1681
01:17:06,561 --> 01:17:08,296
in 1948 came true.
1682
01:17:08,320 --> 01:17:10,696
He's completely vindicated.
1683
01:17:10,720 --> 01:17:13,376
- Anytime you store, process,
1684
01:17:13,400 --> 01:17:16,976
or transmit information in digital form,
1685
01:17:17,000 --> 01:17:21,800
every time you save a file
on a hard drive, play a CD
1686
01:17:22,000 --> 01:17:26,800
or DVD, talk on a cell phone,
send an email, Tweet, text,
1687
01:17:27,920 --> 01:17:31,000
Instagram, the wireless revolution,
1688
01:17:31,960 --> 01:17:36,760
you're relying on the concepts
in Shannon's 1948 paper.
1689
01:17:37,120 --> 01:17:40,096
- Millions of people spend huge chunks
1690
01:17:40,120 --> 01:17:43,216
of their day consuming the codes
1691
01:17:43,240 --> 01:17:46,256
that Claude Shannon said had to be there.
1692
01:17:46,280 --> 01:17:48,360
Information is everything to us.
1693
01:17:51,240 --> 01:17:54,216
- And today, everyone
carries Shannon around
1694
01:17:54,240 --> 01:17:55,560
in their pocket.
1695
01:18:08,040 --> 01:18:09,040
- Wow!
1696
01:18:11,920 --> 01:18:13,360
- My ultimate machine.
1697
01:18:15,280 --> 01:18:18,000
(machine buzzing)
1698
01:18:20,160 --> 01:18:21,336
(snapping)
1699
01:18:21,360 --> 01:18:24,160
- (chuckles) What?
1700
01:18:25,240 --> 01:18:28,256
- You know what Fats Waller
said about swing music,
1701
01:18:28,280 --> 01:18:28,297
don't you, Michelle?
- You know what Fats Waller
said about swing music,
1702
01:18:28,321 --> 01:18:29,456
don't you, Michelle?
1703
01:18:29,480 --> 01:18:30,480
- Oh, no.
1704
01:18:31,720 --> 01:18:33,480
- "Either you got it or you ain't."
1705
01:18:34,760 --> 01:18:35,760
- Uh.
1706
01:18:38,440 --> 01:18:41,160
(machine buzzing)
1707
01:18:44,000 --> 01:18:47,520
(gentle instrumental music)
1708
01:18:48,480 --> 01:18:52,136
- I often wonder if
Claude Shannon understood
1709
01:18:52,160 --> 01:18:54,040
how much information there would be
1710
01:18:54,920 --> 01:18:57,016
after his many discoveries.
1711
01:18:57,040 --> 01:18:57,057
I wonder, did he understand
after his many discoveries.
1712
01:18:57,081 --> 01:18:58,336
I wonder, did he understand
1713
01:18:58,360 --> 01:19:01,056
that we would have this
information explosion?
1714
01:19:01,080 --> 01:19:01,097
That in the last couple
that we would have this
information explosion?
1715
01:19:01,121 --> 01:19:02,176
That in the last couple
1716
01:19:02,200 --> 01:19:02,217
of years we've generated
more information than the sum
That in the last couple
1717
01:19:02,241 --> 01:19:04,816
of years we've generated
more information than the sum
1718
01:19:04,840 --> 01:19:06,160
of all human history.
1719
01:19:06,960 --> 01:19:08,776
Literally, he not only won the lottery,
1720
01:19:08,800 --> 01:19:08,817
he invented the lottery.
Literally, he not only won the lottery,
1721
01:19:08,841 --> 01:19:10,120
he invented the lottery.
1722
01:19:11,040 --> 01:19:14,560
(gentle instrumental music)
1723
01:19:25,480 --> 01:19:29,376
- When I was starting my
PhD program at Harvard,
1724
01:19:29,400 --> 01:19:31,296
and this would've been 1992,
1725
01:19:31,320 --> 01:19:33,656
so my father already had Alzheimer's
1726
01:19:33,680 --> 01:19:36,616
and was already fairly
compromised in terms
1727
01:19:36,640 --> 01:19:39,856
of his cognitive ability,
he asked me what I was going
1728
01:19:39,880 --> 01:19:42,336
to be working on, and at the
time I was working on a project
1729
01:19:42,360 --> 01:19:45,976
about how the magnetic
field of the earth flips
1730
01:19:46,000 --> 01:19:50,256
from time to time, approximately
every million years,
1731
01:19:50,280 --> 01:19:53,136
I was astonished at
how quickly he was able
1732
01:19:53,160 --> 01:19:53,177
to ask the very core
questions of the problem,
I was astonished at
how quickly he was able
1733
01:19:53,201 --> 01:19:56,696
to ask the very core
questions of the problem,
1734
01:19:56,720 --> 01:19:59,936
he cut right to it, and I was amazed.
1735
01:19:59,960 --> 01:20:01,576
And I thought, "This is fantastic.
1736
01:20:01,600 --> 01:20:06,296
"I'm going to have this fabulous
mind at my side (chuckles)
1737
01:20:06,320 --> 01:20:06,335
"while I go through graduate school.
"I'm going to have this fabulous
mind at my side (chuckles)
1738
01:20:06,359 --> 01:20:07,696
"while I go through graduate school.
1739
01:20:07,720 --> 01:20:11,120
"This is gonna be so
much fun for both of us."
1740
01:20:11,920 --> 01:20:16,720
Alas, he was not able to
continue that kind of questioning
1741
01:20:17,000 --> 01:20:18,936
or even thinking for very much longer,
1742
01:20:18,960 --> 01:20:21,456
so I didn't in fact have him
at my side for graduate school.
1743
01:20:21,480 --> 01:20:24,456
But I remember very clearly
that conversation and thinking,
1744
01:20:24,480 --> 01:20:25,496
"Wow!
1745
01:20:25,520 --> 01:20:27,216
"That's something."
1746
01:20:27,240 --> 01:20:27,257
(gentle instrumental music)
"That's something."
1747
01:20:27,281 --> 01:20:30,800
(gentle instrumental music)
1748
01:20:37,040 --> 01:20:41,840
- There's science fiction
books in which the dates
1749
01:20:42,200 --> 01:20:45,536
of things are so many years A.S.,
1750
01:20:45,560 --> 01:20:45,577
and the A.S. stands for After Shannon.
Of things are so many years A.S.,
1751
01:20:45,601 --> 01:20:47,400
and the A.S. stands for After Shannon.
1752
01:20:49,120 --> 01:20:53,920
In some sense, if you look
at the human civilization,
1753
01:20:56,760 --> 01:21:00,096
the sudden change in human
civilization that was created
1754
01:21:00,120 --> 01:21:02,760
by the digital revolution, computers,
1755
01:21:04,240 --> 01:21:06,520
happened at such a short period of time
1756
01:21:07,360 --> 01:21:11,816
that there's nothing else
like that in history.
1757
01:21:11,840 --> 01:21:14,256
So, it could very well be
that historically people
1758
01:21:14,280 --> 01:21:16,736
will look back and say,
"What happened there?"
1759
01:21:16,760 --> 01:21:18,840
And if you were going
to tie it to anybody,
1760
01:21:19,760 --> 01:21:22,056
Shannon is the one that gave us the tools
1761
01:21:22,080 --> 01:21:24,416
that people would still
be using in the future.
1762
01:21:24,440 --> 01:21:24,457
So, it's not so implausible
that people would still
be using in the future.
1763
01:21:24,481 --> 01:21:26,056
So, it's not so implausible
1764
01:21:26,080 --> 01:21:30,880
that people would start
measuring years after Shannon.
1765
01:21:31,360 --> 01:21:33,696
(gentle instrumental music)
1766
01:21:33,720 --> 01:21:35,176
- Inventing something
1767
01:21:35,200 --> 01:21:40,000
is like sending a message
in a bottle into the future.
1768
01:21:44,000 --> 01:21:47,576
(gentle jazz music)
1769
01:21:47,600 --> 01:21:48,457
- I think it was special
1770
01:21:48,481 --> 01:21:50,896
that he had both the ability and interest
1771
01:21:50,920 --> 01:21:55,496
in theoretical mathematics
and this passion for building.
1772
01:21:55,520 --> 01:21:57,896
In a way that a toddler
would explore the world,
1773
01:21:57,920 --> 01:22:01,496
he didn't lose that childlike
curiosity and delight
1774
01:22:01,520 --> 01:22:01,537
in learning things, discovering things,
he didn't lose that childlike
curiosity and delight
1775
01:22:01,561 --> 01:22:03,656
in learning things, discovering things,
1776
01:22:03,680 --> 01:22:06,016
putting things together,
making connections.
1777
01:22:06,040 --> 01:22:08,920
(gentle jazz music)
1778
01:23:32,080 --> 01:23:35,680
(intense instrumental music)
1779
01:23:35,704 --> 01:23:38,704
Subtitles Diego Moraes(oakislandtk)
www.opensubtitles.org
1779
01:23:39,305 --> 01:24:39,686
Support us and become VIP member
to remove all ads from OpenSubtitles.org
149169
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.