1
00:00:00,000 --> 00:00:01,780
Look around you for a second, like literally

2
00:00:01,780 --> 00:00:03,620
right now. Yeah, just take a second and look

3
00:00:03,620 --> 00:00:06,019
around. Whether you are driving your car, sitting

4
00:00:06,019 --> 00:00:09,039
at your desk, going for a walk, or... Just lounging

5
00:00:09,039 --> 00:00:12,019
on the couch. Right. Take a look at the world

6
00:00:12,019 --> 00:00:14,119
you are interacting with. You are absolutely,

7
00:00:14,500 --> 00:00:17,559
completely, and totally saturated with data.

8
00:00:17,780 --> 00:00:20,899
It is everywhere. Everywhere. It is in the device

9
00:00:20,899 --> 00:00:24,199
you're using to stream this deep dive. bouncing

10
00:00:24,199 --> 00:00:26,859
through the invisible cellular networks in the

11
00:00:26,859 --> 00:00:29,460
air right around you. Through the Wi -Fi. Yeah,

12
00:00:29,500 --> 00:00:31,640
and it's in the targeted algorithms analyzing

13
00:00:31,640 --> 00:00:34,479
your behavior online. It's even in the genetic

14
00:00:34,479 --> 00:00:36,600
code of the plants sitting on your windowsill.

15
00:00:36,719 --> 00:00:39,460
We basically swing in an ocean of data. We really

16
00:00:39,460 --> 00:00:41,659
do, to the point where we do not even see the

17
00:00:41,659 --> 00:00:44,219
water anymore. Yeah. But it wasn't always like

18
00:00:44,219 --> 00:00:46,820
this. No, not at all. I want to start off today

19
00:00:46,820 --> 00:00:48,899
by looking at a reflection from the novelist

20
00:00:48,899 --> 00:00:51,140
Zadie Smith. Because she was thinking about the

21
00:00:51,140 --> 00:00:54,859
old days, specifically looking at vintage train

22
00:00:54,859 --> 00:00:57,119
tickets. Oh, that's a great example. Yeah, she

23
00:00:57,119 --> 00:00:59,799
wrote, quote, Anyway, those tickets, the old

24
00:00:59,799 --> 00:01:01,740
ones, they didn't tell you where you were going,

25
00:01:01,880 --> 00:01:04,359
much less where you came from. He couldn't remember

26
00:01:04,359 --> 00:01:06,579
seeing any baits on them either, and there was

27
00:01:06,579 --> 00:01:09,900
certainly no mention of time. Wow. Right. And

28
00:01:09,900 --> 00:01:12,159
she goes on. It was all different now, of course,

29
00:01:12,180 --> 00:01:15,049
all this information. Archie wondered why that

30
00:01:15,049 --> 00:01:17,670
was. Why that was. Exactly. Why is that? How

31
00:01:17,670 --> 00:01:20,109
did we get to this point where absolutely everything

32
00:01:20,109 --> 00:01:23,530
is quantified, date stamped, tracked, and stored?

33
00:01:23,750 --> 00:01:26,849
It is a profound question. And to answer it,

34
00:01:26,890 --> 00:01:29,650
I mean, we really have to recognize that the

35
00:01:29,650 --> 00:01:32,230
very concept of information, as we understand

36
00:01:32,230 --> 00:01:35,129
it today anyway, is a surprisingly recent invention.

37
00:01:35,450 --> 00:01:38,430
Very recent. There is this brilliant quote in

38
00:01:38,430 --> 00:01:40,530
our source material from the theoretical physicist

39
00:01:40,530 --> 00:01:43,439
John Archibald Wheeler. It perfectly frames what

40
00:01:43,439 --> 00:01:44,739
we're going to talk about today. What did he

41
00:01:44,739 --> 00:01:47,780
say? He said, what we call the past is built

42
00:01:47,780 --> 00:01:50,939
on bits. Built on bits. Yeah. History is not

43
00:01:50,939 --> 00:01:53,799
just some fluid sequence of events anymore. It

44
00:01:53,799 --> 00:01:57,900
is a compilation of discrete data. And to understand

45
00:01:57,900 --> 00:02:00,799
how we arrived at this modern, highly quantified

46
00:02:00,799 --> 00:02:03,280
era we live in, we are going to look at the Genesis.

47
00:02:03,930 --> 00:02:06,370
of that exact concept. That brings us to our

48
00:02:06,370 --> 00:02:08,469
mission for this deep dive. Let's hear it. We

49
00:02:08,469 --> 00:02:10,849
are jumping into James Gleick's monumental book,

50
00:02:11,009 --> 00:02:14,789
The Information. A history, a theory, a flood.

51
00:02:14,969 --> 00:02:17,830
Such a massive book. It's huge. And our goal

52
00:02:17,830 --> 00:02:20,909
today is to unpack the fascinating and frankly

53
00:02:20,909 --> 00:02:23,729
deeply counterintuitive history of how information

54
00:02:23,729 --> 00:02:26,949
actually became a quantifiable physical thing.

55
00:02:27,129 --> 00:02:29,120
By taking something totally abstract. Exactly.

56
00:02:29,120 --> 00:02:32,039
Taking a concept as abstract as a thought or

57
00:02:32,039 --> 00:02:34,439
a melody or a conversation and turning it into

58
00:02:34,439 --> 00:02:36,460
something you can measure with mathematical precision,

59
00:02:36,719 --> 00:02:39,439
just like you measure mass or temperature or

60
00:02:39,439 --> 00:02:41,900
energy. That's a wild transition. It really is.

61
00:02:42,020 --> 00:02:44,500
So we are going to explore the rigorous mathematical

62
00:02:44,500 --> 00:02:47,020
underpinnings of this transformation. We will

63
00:02:47,020 --> 00:02:49,340
trace it from the early days of telegraphs, deep

64
00:02:49,340 --> 00:02:52,319
into the biological codes of DNA, outward into

65
00:02:52,319 --> 00:02:54,680
the cultural codes we call memes. Which is a

66
00:02:54,680 --> 00:02:59,129
fascinating shift. Totally. And finally. into

67
00:02:59,129 --> 00:03:01,629
the foundation of algorithmic complexity that

68
00:03:01,629 --> 00:03:03,889
drives today's artificial thinking machines.

69
00:03:04,189 --> 00:03:07,030
We have a really comprehensive roadmap today.

70
00:03:07,189 --> 00:03:09,889
We will start with the origins of modern communication.

71
00:03:10,639 --> 00:03:13,520
specifically looking at the crucial year of 1948.

72
00:03:14,020 --> 00:03:16,960
A huge year. A moment that fundamentally bifurcated

73
00:03:16,960 --> 00:03:19,539
human history. We will examine how engineers

74
00:03:19,539 --> 00:03:21,939
had to completely deconstruct human language

75
00:03:21,939 --> 00:03:24,819
just to send it over wires. They had to strip

76
00:03:24,819 --> 00:03:27,219
away the meaning. Yes, stripping away meaning

77
00:03:27,219 --> 00:03:29,199
to find the mathematical core of communication.

78
00:03:29,680 --> 00:03:32,479
And from there, we will follow that theoretical

79
00:03:32,479 --> 00:03:35,000
thread into biology to see how scientists realized

80
00:03:35,000 --> 00:03:38,340
life itself operates on a digital code. And then

81
00:03:38,340 --> 00:03:40,419
into culture. Right. Then we expand into the

82
00:03:40,419 --> 00:03:42,699
cultural realm to examine how ideas replicate

83
00:03:42,699 --> 00:03:45,520
like viruses. And finally, we will grapple with

84
00:03:45,520 --> 00:03:47,719
the intense mathematical theories of randomness

85
00:03:47,719 --> 00:03:50,539
and complexity that literally define our modern

86
00:03:50,539 --> 00:03:53,419
technological age. It dictates how computers

87
00:03:53,419 --> 00:03:55,900
and AI process the world. Okay, let's unpack

88
00:03:55,900 --> 00:03:57,840
this. Let's start by getting a sense of the grand

89
00:03:57,840 --> 00:03:59,939
scope of Gleick's text. Because the progression

90
00:03:59,939 --> 00:04:03,159
of ideas reads almost like a map of human evolution

91
00:04:03,159 --> 00:04:06,699
itself. It really does. It moves from early physical

92
00:04:06,699 --> 00:04:09,539
communication like drums that talk across African

93
00:04:09,539 --> 00:04:12,419
plains or the persistence of the written word

94
00:04:12,419 --> 00:04:14,740
into the mechanical era. Right, attempting to

95
00:04:14,740 --> 00:04:17,680
throw the powers of thought into wheel work with

96
00:04:17,680 --> 00:04:20,399
those early calculating engines. Exactly. Then

97
00:04:20,399 --> 00:04:23,120
it explodes into the 20th century. New wires,

98
00:04:23,319 --> 00:04:26,459
new logic, the formalization of information theory,

99
00:04:26,680 --> 00:04:29,600
entropy, the discovery of life's own biological

100
00:04:29,600 --> 00:04:33,060
code, the concept of the meme pool, and the mathematical

101
00:04:33,060 --> 00:04:36,060
sense of randomness. It is essentially the story

102
00:04:36,060 --> 00:04:38,939
of how humanity learned to process the universe.

103
00:04:39,120 --> 00:04:41,970
It's a great way to put it. To truly grasp the

104
00:04:41,970 --> 00:04:44,850
climax of that grand story, we really have to

105
00:04:44,850 --> 00:04:47,149
anchor our conversation in the mid -20th century,

106
00:04:47,350 --> 00:04:51,250
specifically the year 1948. 1948. That year is

107
00:04:51,250 --> 00:04:53,610
the fulcrum around which our entire modern digital

108
00:04:53,610 --> 00:04:56,410
world began to turn. The post -World War II era

109
00:04:56,410 --> 00:04:58,649
was a time of massive industrial and scientific

110
00:04:58,649 --> 00:05:01,350
transition, and Bell Telephone Laboratories was

111
00:05:01,350 --> 00:05:03,470
the undisputed epicenter of global technological

112
00:05:03,470 --> 00:05:06,290
research. The environment at Bell Labs during

113
00:05:06,290 --> 00:05:09,089
that period is just fascinating to consider because

114
00:05:09,089 --> 00:05:12,329
in May of 1948, they announced an invention that

115
00:05:12,329 --> 00:05:14,310
would change the physical landscape of technology

116
00:05:14,310 --> 00:05:17,689
forever. They developed a tiny electronic semiconductor.

117
00:05:18,029 --> 00:05:20,189
Right. They described it as a simple device that

118
00:05:20,189 --> 00:05:23,110
could do anything a fragile heat generating vacuum

119
00:05:23,110 --> 00:05:26,370
tube could do, but with vastly more efficiency.

120
00:05:26,769 --> 00:05:29,230
It was just a crystalline sliver. So incredibly

121
00:05:29,230 --> 00:05:31,250
small that you could fit a hundred of them in

122
00:05:31,250 --> 00:05:33,120
the palm of your hand. The corporate process

123
00:05:33,120 --> 00:05:35,660
behind naming this revolutionary device is a

124
00:05:35,660 --> 00:05:37,720
perfect artifact of mid -century engineering

125
00:05:37,720 --> 00:05:40,480
culture, too. Oh, I love this part. Yeah. The

126
00:05:40,480 --> 00:05:42,720
scientists formed a committee. They distributed

127
00:05:42,720 --> 00:05:45,060
paper ballots to the senior engineers at their

128
00:05:45,060 --> 00:05:48,139
facility in Murray Hill, New Jersey. So bureaucratic.

129
00:05:48,379 --> 00:05:51,199
Very. Uh -huh. And the ballot offered a list

130
00:05:51,199 --> 00:05:54,040
of highly technical, somewhat clunky choices.

131
00:05:54,300 --> 00:05:57,120
They were voting on names like Semiconductor

132
00:05:57,120 --> 00:06:00,480
Triode. Catchy. Or the Iatatron. Which sounds

133
00:06:00,480 --> 00:06:03,740
distinctly like a villain from a 1950s science

134
00:06:03,740 --> 00:06:07,279
fiction serial. It really does. Beware the Iodotron.

135
00:06:07,480 --> 00:06:10,379
But another option was Transistor, which was

136
00:06:10,379 --> 00:06:13,699
a portmanteau of varistor and transconductance.

137
00:06:13,819 --> 00:06:16,160
And Transistor obviously won the vote. Thankfully.

138
00:06:16,420 --> 00:06:19,060
Bell Labs put out a rather understated press

139
00:06:19,060 --> 00:06:21,660
release about it, declaring that the device,

140
00:06:21,860 --> 00:06:24,750
quote, may have far -reaching significance in

141
00:06:24,750 --> 00:06:27,310
electronics and electrical communication. May

142
00:06:27,310 --> 00:06:29,810
have far -reaching significance. Talk about an

143
00:06:29,810 --> 00:06:32,209
understatement. It's hard to overstate how modest

144
00:06:32,209 --> 00:06:34,910
that sounds. In hindsight, the transistor sparked

145
00:06:34,910 --> 00:06:37,649
the entire revolution in solid -state electronics.

146
00:06:38,129 --> 00:06:40,529
Absolutely. It set technology on the relentless

147
00:06:40,529 --> 00:06:43,509
path of Moore's Law and miniaturization that

148
00:06:43,509 --> 00:06:46,089
led directly to the microscopic processors in

149
00:06:46,089 --> 00:06:48,240
the smart devices you are using right now. It

150
00:06:48,240 --> 00:06:50,459
won the Nobel Prize for its inventors. It was

151
00:06:50,459 --> 00:06:52,439
an absolute triumph of hardware engineering.

152
00:06:52,720 --> 00:06:56,899
But the transistor was arguably only the second

153
00:06:56,899 --> 00:06:59,120
most significant development to emerge from Bell

154
00:06:59,120 --> 00:07:02,160
Labs that exact same year. Exactly. While the

155
00:07:02,160 --> 00:07:04,339
transistor was a revolution in physical hardware,

156
00:07:04,639 --> 00:07:06,839
an intellectual breakthrough that was far more

157
00:07:06,839 --> 00:07:09,500
profound and fundamental was quietly published

158
00:07:09,500 --> 00:07:13,290
in the summer of 1948. And it did not get a flashy

159
00:07:13,290 --> 00:07:15,730
press release. No, there were no committee votes.

160
00:07:15,889 --> 00:07:19,810
It was just a dense, highly mathematical monograph

161
00:07:19,810 --> 00:07:23,870
spread across 79 pages of the Bell System Technical

162
00:07:23,870 --> 00:07:26,470
Journal. The title of the paper was brilliantly

163
00:07:26,470 --> 00:07:29,290
understated, too. It was called A Mathematical

164
00:07:29,290 --> 00:07:32,250
Theory of Communication. Just a theory. Right.

165
00:07:32,329 --> 00:07:34,350
And just like the invention of the transistor,

166
00:07:34,670 --> 00:07:37,790
this paper introduced a brand new word to the

167
00:07:37,790 --> 00:07:40,589
English language, a neologism that would become

168
00:07:40,589 --> 00:07:43,269
the absolute foundation of the modern world.

169
00:07:43,389 --> 00:07:46,910
The word was the bit, D -I -T, short for binary

170
00:07:46,910 --> 00:07:49,740
digit. The bit. But unlike the transistor. This

171
00:07:49,740 --> 00:07:52,199
word wasn't focus grouped by a committee of senior

172
00:07:52,199 --> 00:07:55,439
engineers. It was chosen by the lone author of

173
00:07:55,439 --> 00:07:57,819
that monograph, a 32 -year -old mathematician

174
00:07:57,819 --> 00:08:00,740
named Claude Shannon. And with the publication

175
00:08:00,740 --> 00:08:03,879
of that paper, the bit joined the inch, the pound,

176
00:08:03,959 --> 00:08:06,639
the quart, and the minute as a determinate, universally

177
00:08:06,639 --> 00:08:08,920
applicable quantity. It became a fundamental

178
00:08:08,920 --> 00:08:11,680
unit of measure. A unit of measure for information.

179
00:08:12,139 --> 00:08:14,259
Yes. I really want you to pause and think about

180
00:08:14,259 --> 00:08:17,000
how radical that is. Imagine trying to measure

181
00:08:17,000 --> 00:08:19,170
a conversation you had this morning. How do you

182
00:08:19,170 --> 00:08:21,350
do that? Exactly. How do you weigh a phone call?

183
00:08:21,569 --> 00:08:24,370
Do you measure the decibel level of your voice?

184
00:08:24,529 --> 00:08:27,170
Do you count the syllables? Do you try to quantify

185
00:08:27,170 --> 00:08:29,569
the emotional weight of a secret compared to

186
00:08:29,569 --> 00:08:32,710
a joke? Right. Before Shannon, information was

187
00:08:32,710 --> 00:08:35,649
this squishy philosophical concept. It belonged

188
00:08:35,649 --> 00:08:38,549
to librarians and linguists. Shannon came along

189
00:08:38,549 --> 00:08:41,070
and proved mathematically that information is

190
00:08:41,070 --> 00:08:44,009
a physical, measurable quantity. To understand

191
00:08:44,009 --> 00:08:46,629
how Shannon achieved this massive paradigm shift,

192
00:08:46,769 --> 00:08:48,909
we really have to look closely at his background

193
00:08:48,909 --> 00:08:51,769
and how he operated within Bell Labs. Because

194
00:08:51,769 --> 00:08:54,049
he was an odd guy. He was definitely not your

195
00:08:54,049 --> 00:08:56,929
standard corporate company man. He was ostensibly

196
00:08:56,929 --> 00:08:59,580
part of the mathematical research group. But

197
00:08:59,580 --> 00:09:01,399
when the rest of his department relocated from

198
00:09:01,399 --> 00:09:04,759
New York City to this shiny, sprawling new suburban

199
00:09:04,759 --> 00:09:07,399
campus in Murray Hill, New Jersey. Where they

200
00:09:07,399 --> 00:09:09,980
voted on the transistor. Exactly. When they moved

201
00:09:09,980 --> 00:09:12,120
there, Shannon simply refused to go. He just

202
00:09:12,120 --> 00:09:14,740
said no. He stayed behind in this 12 -story,

203
00:09:14,820 --> 00:09:16,960
sandy brick building on West Street in Manhattan.

204
00:09:17,320 --> 00:09:19,399
The source material paints such an incredible

205
00:09:19,399 --> 00:09:21,980
picture of him in that building. He was a man

206
00:09:21,980 --> 00:09:24,980
who rode a unicycle down the long corridors of

207
00:09:24,980 --> 00:09:27,700
that brick building while juggling. Juggling

208
00:09:27,700 --> 00:09:30,360
on a unicycle in a corporate research lab. Yeah.

209
00:09:30,700 --> 00:09:33,259
He stayed in Manhattan because he hated commuting.

210
00:09:33,500 --> 00:09:35,980
And he wanted to be able to walk out of his laboratory

211
00:09:35,980 --> 00:09:39,360
and hear late -night jazz clarinetists playing

212
00:09:39,360 --> 00:09:41,440
in Greenwich Village clubs. And he was also spending

213
00:09:41,440 --> 00:09:43,480
time flirting with a young woman who worked in

214
00:09:43,480 --> 00:09:45,740
the microwave research group, housed in an old

215
00:09:45,740 --> 00:09:48,419
Nabisco factory across the street. He was deeply

216
00:09:48,419 --> 00:09:50,820
eccentric, and Bell Labs management essentially

217
00:09:50,820 --> 00:09:53,259
just left him alone in his cubbyhole to stare

218
00:09:53,259 --> 00:09:54,789
at his blackboard. Well, they left him alone

219
00:09:54,789 --> 00:09:56,889
because his intellectual track record was already

220
00:09:56,889 --> 00:09:59,850
unassailable. I mean, fresh out of MIT, Shannon

221
00:09:59,850 --> 00:10:02,450
had plunged into the laboratory's vital, highly

222
00:10:02,450 --> 00:10:04,850
classified work during World War II. Right, the

223
00:10:04,850 --> 00:10:07,009
war effort. He initially worked on developing

224
00:10:07,009 --> 00:10:10,110
automatic fire control directors for anti -aircraft

225
00:10:10,110 --> 00:10:12,590
guns, which required these incredibly complex

226
00:10:12,590 --> 00:10:15,570
mathematical predictions of flight paths. And

227
00:10:15,570 --> 00:10:18,250
from there, he moved deeply into cryptography.

228
00:10:18,450 --> 00:10:20,860
Making things secure. Working on the theoretical

229
00:10:20,860 --> 00:10:23,980
underpinnings of secure communication. And this

230
00:10:23,980 --> 00:10:27,059
is where he made his first monumental contribution.

231
00:10:27,379 --> 00:10:29,440
He worked out the mathematical proof of security

232
00:10:29,440 --> 00:10:32,000
for the X system. The X system. For context,

233
00:10:32,080 --> 00:10:34,139
the X system, it was also known as Six Ailey,

234
00:10:34,240 --> 00:10:37,240
was the highly classified telephone hotline used

235
00:10:37,240 --> 00:10:39,580
by Winston Churchill and President Franklin Roosevelt

236
00:10:39,580 --> 00:10:42,820
to coordinate the Allied war effort. And the

237
00:10:42,820 --> 00:10:46,000
X system was an absolute marvel of early digital

238
00:10:46,000 --> 00:10:48,659
engineering. It didn't just transmit an analog

239
00:10:48,659 --> 00:10:51,899
voice signal. No. It took the continuous wave

240
00:10:51,899 --> 00:10:55,539
of human speech, sampled it, and quantized it

241
00:10:55,539 --> 00:10:58,980
into discrete steps. Then it encrypted that digitized

242
00:10:58,980 --> 00:11:01,679
voice using a one -time pad. A one -time pad,

243
00:11:01,799 --> 00:11:03,559
which is fascinating. Yeah, it was a massive

244
00:11:03,559 --> 00:11:06,779
set of randomly generated keys stored on actual

245
00:11:06,779 --> 00:11:08,539
phonograph records that had to be physically

246
00:11:08,539 --> 00:11:10,840
transported and synchronized in Washington and

247
00:11:10,840 --> 00:11:13,299
London. Think about the logistics of that, shipping

248
00:11:13,299 --> 00:11:15,740
records across the Atlantic just to have a phone

249
00:11:15,740 --> 00:11:18,340
call. And Shana's job wasn't just to build it,

250
00:11:18,399 --> 00:11:20,399
but to prove that it could not be broken. To

251
00:11:20,399 --> 00:11:23,559
prove it mathematically. Yes. He demonstrated

252
00:11:23,559 --> 00:11:27,159
that if the encryption key is truly random, is

253
00:11:27,159 --> 00:11:29,379
at least as long as the message, and is never

254
00:11:29,379 --> 00:11:32,940
reused, the encrypted audio contains absolutely

255
00:11:32,940 --> 00:11:35,399
zero information about the original message.

256
00:11:35,659 --> 00:11:39,159
Zero. The ciphertext is mathematically indistinguishable

257
00:11:39,159 --> 00:11:42,039
from pure noise. The Nazis could intercept the

258
00:11:42,039 --> 00:11:44,279
signal all day long, but without the specific

259
00:11:44,279 --> 00:11:46,779
phonograph record acting as the key, they were

260
00:11:46,779 --> 00:11:48,820
just listening to static. That is brilliant.

261
00:11:48,879 --> 00:11:51,539
And working on the X system forced Shannon to

262
00:11:51,539 --> 00:11:53,899
think about communication in a completely novel

263
00:11:53,899 --> 00:11:56,159
way. Completely different. He was looking at

264
00:11:56,159 --> 00:11:59,169
human speech. Not as sound, but as a sequence

265
00:11:59,169 --> 00:12:01,970
of discrete, quantifiable states that could be

266
00:12:01,970 --> 00:12:04,850
randomized, transmitted, and perfectly reconstructed.

267
00:12:04,909 --> 00:12:07,690
And after the war, AT &T, which was operating

268
00:12:07,690 --> 00:12:10,730
as a massive mid -century monopoly, allowed brilliant

269
00:12:10,730 --> 00:12:13,370
minds like Shannon to pursue these theoretical

270
00:12:13,370 --> 00:12:15,649
tangents without demanding immediate commercial

271
00:12:15,649 --> 00:12:17,610
products. Which makes sense when you consider

272
00:12:17,610 --> 00:12:20,169
the scale of what AT &T was dealing with. By

273
00:12:20,169 --> 00:12:23,009
1948, the Bell system was routing more than 125

274
00:12:23,009 --> 00:12:26,110
million conversations every single day through

275
00:12:26,110 --> 00:12:29,330
138 million miles of cable. The scale is almost

276
00:12:29,330 --> 00:12:31,710
hard to comprehend. And the U .S. Census Bureau

277
00:12:31,710 --> 00:12:34,330
tracked these metrics under the broad category

278
00:12:34,330 --> 00:12:37,190
of communications in the United States. But their

279
00:12:37,190 --> 00:12:39,450
metrics were incredibly primitive. What were

280
00:12:39,450 --> 00:12:41,529
they tracking? They counted the number of physical

281
00:12:41,529 --> 00:12:44,490
radio stations. They counted newspapers printed

282
00:12:44,490 --> 00:12:47,350
and parcels delivered by the post office. Very

283
00:12:47,350 --> 00:12:50,190
physical things. Exactly. But the engineers at

284
00:12:50,190 --> 00:12:53,110
Bell Labs faced a deeper, almost existential

285
00:12:53,110 --> 00:12:56,909
crisis. What exactly was the fundamental substance

286
00:12:56,909 --> 00:12:59,710
flowing through those copper wires? Right, because

287
00:12:59,710 --> 00:13:01,669
they knew they weren't actually transmitting

288
00:13:01,669 --> 00:13:03,710
conversations. They were electrical engineers

289
00:13:03,710 --> 00:13:06,669
dealing with voltage and current. Not words.

290
00:13:07,009 --> 00:13:08,909
Exactly. Yeah. They were transmitting electrical

291
00:13:08,909 --> 00:13:11,870
waveforms that acted as an analog surrogate for

292
00:13:11,870 --> 00:13:15,309
sound. The human vocal cords create variations

293
00:13:15,309 --> 00:13:18,610
in air pressure. Those waves vibrate a diaphragm

294
00:13:18,610 --> 00:13:20,889
in the telephone mouthpiece. And that mechanical

295
00:13:20,889 --> 00:13:23,169
movement is converted into an electrical current

296
00:13:23,169 --> 00:13:25,750
that mimics the shape of the sound wave. That

297
00:13:25,750 --> 00:13:27,889
analog conversion is what made the telephone

298
00:13:27,889 --> 00:13:30,549
such a massive leap over its predecessor, the

299
00:13:30,549 --> 00:13:32,990
telegraph. But the telegraph actually provides

300
00:13:32,990 --> 00:13:35,490
a much clearer window into the chain of abstraction

301
00:13:35,490 --> 00:13:38,230
that Shannon was exploring. Because the telegraph

302
00:13:38,230 --> 00:13:40,669
didn't use the physics of sound. No, it used

303
00:13:40,669 --> 00:13:44,059
a discrete code. Dots and dashes. The source

304
00:13:44,059 --> 00:13:47,399
text meticulously breaks down this human chain

305
00:13:47,399 --> 00:13:50,059
of abstraction, and it's so important to grasp.

306
00:13:50,440 --> 00:13:53,919
Consider the layers involved in a telegram. The

307
00:13:53,919 --> 00:13:56,639
physical electrical pulses, the long and short

308
00:13:56,639 --> 00:13:58,919
bursts of current represent dots and dashes.

309
00:13:58,919 --> 00:14:00,899
Okay, layer one. The dots and dashes represent

310
00:14:00,899 --> 00:14:04,230
letters of the alphabet, layer two. Those letters

311
00:14:04,230 --> 00:14:06,830
represent phonetic sounds. Layer three. Those

312
00:14:06,830 --> 00:14:09,730
sounds combine to form words, and those words

313
00:14:09,730 --> 00:14:12,590
are constructed to represent some ultimate substrate

314
00:14:12,590 --> 00:14:15,789
of human meaning. Okay, I want to push back on

315
00:14:15,789 --> 00:14:17,529
this a bit, or at least play devil's advocate

316
00:14:17,529 --> 00:14:20,909
here, because this is where Shannon's leap feels

317
00:14:20,909 --> 00:14:23,570
deeply counterintuitive to most people. Sure,

318
00:14:23,610 --> 00:14:25,889
go ahead. If I send a telegram that says, the

319
00:14:25,889 --> 00:14:28,350
house is on fire. The meaning is literally the

320
00:14:28,350 --> 00:14:30,409
only thing that matters. Obviously. If the person

321
00:14:30,409 --> 00:14:32,269
on the other end doesn't understand the meaning,

322
00:14:32,429 --> 00:14:34,710
the communication is a failure. My house burns

323
00:14:34,710 --> 00:14:37,429
down. So how could Shannon build a universal

324
00:14:37,429 --> 00:14:39,750
theory of communication by ignoring the very

325
00:14:39,750 --> 00:14:42,190
reason we communicate in the first place? That

326
00:14:42,190 --> 00:14:45,250
is the exact conceptual hurdle that most people,

327
00:14:45,350 --> 00:14:47,629
even the senior engineers at the time, struggled

328
00:14:47,629 --> 00:14:51,269
to clear. What's fascinating here is that Shannon

329
00:14:51,269 --> 00:14:53,309
recognized that meaning is a function of human

330
00:14:53,309 --> 00:14:57,309
psychology and shared context. It's not a function

331
00:14:57,309 --> 00:14:59,389
of the transmission system itself. It's separate.

332
00:14:59,750 --> 00:15:02,009
Entirely separate. To an engineer designing a

333
00:15:02,009 --> 00:15:04,389
telephone network, it does not matter if the

334
00:15:04,389 --> 00:15:06,230
electrical signals are carrying a teenager's

335
00:15:06,230 --> 00:15:09,549
gossip, a declaration of war, a recipe for pie,

336
00:15:09,870 --> 00:15:13,470
or pure gibberish. The network's job is agnostic

337
00:15:13,470 --> 00:15:16,049
to the semantics. Let me bring in Shanna's direct

338
00:15:16,049 --> 00:15:18,730
words from the 1948 paper because he states this

339
00:15:18,730 --> 00:15:21,500
so cleanly. Yeah, read it. He wrote, quote, The

340
00:15:21,500 --> 00:15:23,659
fundamental problem of communication is that

341
00:15:23,659 --> 00:15:26,039
of reproducing at one point either exactly or

342
00:15:26,039 --> 00:15:28,799
approximately a message selected at another point.

343
00:15:29,100 --> 00:15:31,120
Selecting a message. Right. And he continues.

344
00:15:31,559 --> 00:15:34,139
Frequently, the messages have meaning. That is,

345
00:15:34,139 --> 00:15:36,320
they refer to or are correlated according to

346
00:15:36,320 --> 00:15:38,519
some system with certain physical or conceptual

347
00:15:38,519 --> 00:15:41,299
entities. These semantic aspects of communication

348
00:15:41,299 --> 00:15:43,840
are irrelevant to the engineering problem. Boom.

349
00:15:44,399 --> 00:15:46,659
Irrelevant. By severing the meaning from the

350
00:15:46,659 --> 00:15:49,789
message. Shannon was able to isolate the purely

351
00:15:49,789 --> 00:15:52,809
mathematical core of communication. He defined

352
00:15:52,809 --> 00:15:56,330
information not as knowledge, but as the resolution

353
00:15:56,330 --> 00:15:59,450
of uncertainty. That's the key phrase, resolution

354
00:15:59,450 --> 00:16:01,990
of uncertainty. He stated that the amount of

355
00:16:01,990 --> 00:16:04,190
information in a message is directly tied to

356
00:16:04,190 --> 00:16:06,570
the number of possible alternative messages that

357
00:16:06,570 --> 00:16:08,970
could have been sent. If we break down this concept

358
00:16:08,970 --> 00:16:11,730
of resolving uncertainty, the brilliance of the

359
00:16:11,730 --> 00:16:14,990
bit becomes so clear. Imagine I am holding a

360
00:16:14,990 --> 00:16:17,029
coin behind my back. Okay, I'm imagining it.

361
00:16:17,330 --> 00:16:19,789
You know, it will either be heads or tails. You

362
00:16:19,789 --> 00:16:22,330
have a specific amount of uncertainty. When I

363
00:16:22,330 --> 00:16:24,690
reveal the coin and you see it as heads, your

364
00:16:24,690 --> 00:16:27,149
uncertainty is cut in half. It's resolved. That

365
00:16:27,149 --> 00:16:29,789
single binary choice, the resolution of a two

366
00:16:29,789 --> 00:16:32,450
-state system, is the definition of one bit of

367
00:16:32,450 --> 00:16:35,309
information. Mathematically, Shannon used a base

368
00:16:35,309 --> 00:16:38,230
-two logarithmic function to quantify this. So

369
00:16:38,230 --> 00:16:41,049
if you have four possible choices, say, picking

370
00:16:41,049 --> 00:16:43,809
a suit from a deck of cards, hearts, diamonds,

371
00:16:44,090 --> 00:16:47,889
clubs, or spades. Four choices. Right. You need

372
00:16:47,889 --> 00:16:50,269
two bits of information to resolve the uncertainty.

373
00:16:50,929 --> 00:16:53,750
The first bit narrows it down from four to two.

374
00:16:53,909 --> 00:16:56,090
Like, is it a red suit or a black suit? Okay,

375
00:16:56,149 --> 00:16:57,909
that's one bit. And the second bit narrows it

376
00:16:57,909 --> 00:17:00,549
down from two to one. Is it a heart or a diamond?

377
00:17:00,850 --> 00:17:02,970
That's the second bit. Every time you add a bit,

378
00:17:03,070 --> 00:17:05,450
you double the number of possible states you

379
00:17:05,450 --> 00:17:08,170
can distinguish. This means that absolutely any

380
00:17:08,170 --> 00:17:11,130
message, no matter how complex, a photograph,

381
00:17:11,470 --> 00:17:15,069
a symphony, a moving video, a novel, can be systematically

382
00:17:15,069 --> 00:17:18,630
reduced. to a sequence of binary choices yes

383
00:17:18,630 --> 00:17:21,569
or no true or false one or zero shannon proved

384
00:17:21,569 --> 00:17:23,930
that all communication could be translated into

385
00:17:23,930 --> 00:17:26,589
a universal currency and this realization didn't

386
00:17:26,589 --> 00:17:29,170
happen in a total vacuum either while shannon

387
00:17:29,170 --> 00:17:31,410
worked largely alone in his manhattan office

388
00:17:31,410 --> 00:17:34,089
he had access to some of the greatest minds of

389
00:17:34,089 --> 00:17:37,059
his generation oh absolutely Most notably, his

390
00:17:37,059 --> 00:17:39,299
interactions with Alan Turing during the war

391
00:17:39,299 --> 00:17:41,700
were incredibly catalytic. Turing was amazing.

392
00:17:41,980 --> 00:17:44,500
Turing, the British mathematical genius who laid

393
00:17:44,500 --> 00:17:46,660
the theoretical groundwork for modern computer

394
00:17:46,660 --> 00:17:50,039
science. He was visiting Bell Labs on a cryptographic

395
00:17:50,039 --> 00:17:52,920
mission. The text describes these lunch meetings

396
00:17:52,920 --> 00:17:55,799
in the Bell Labs cafeteria in the early 1940s

397
00:17:55,799 --> 00:17:58,990
between Turing and Shannon. It is staggering

398
00:17:58,990 --> 00:18:00,869
to think about. He's sitting there eating sandwiches.

399
00:18:01,230 --> 00:18:03,630
Right. You have the man who conceptualized the

400
00:18:03,630 --> 00:18:06,470
universal Turing machine. a theoretical device

401
00:18:06,470 --> 00:18:09,690
that could compute any computable sequence, sitting

402
00:18:09,690 --> 00:18:12,670
across from the man who was formalizing the universal

403
00:18:12,670 --> 00:18:15,230
currency of information. They were trading wild

404
00:18:15,230 --> 00:18:17,430
speculations about artificial thinking machines.

405
00:18:17,869 --> 00:18:20,670
And Turing was so animated by what Shannon was

406
00:18:20,670 --> 00:18:23,609
proposing, because Turing's machines were originally

407
00:18:23,609 --> 00:18:26,549
conceived to crunch numbers and solve mathematical

408
00:18:26,549 --> 00:18:29,970
logic. Giant calculators. Exactly. But if Shannon

409
00:18:29,970 --> 00:18:32,799
was right, and information was a universal binary

410
00:18:32,799 --> 00:18:35,619
currency, then the machine's capabilities were

411
00:18:35,619 --> 00:18:38,220
boundless. Turing excitedly noted that Shannon

412
00:18:38,220 --> 00:18:41,140
wanted to feed not just raw data to a mechanical

413
00:18:41,140 --> 00:18:43,680
brain, but cultural things. Cultural things.

414
00:18:43,759 --> 00:18:45,900
He realized Shannon wanted to play music to it.

415
00:18:46,000 --> 00:18:48,240
In the 1940s, I mean, in the 1940s, computers

416
00:18:48,240 --> 00:18:51,440
were massive room -sized artillery calculators

417
00:18:51,440 --> 00:18:53,720
built from clicking relays and fragile vacuum

418
00:18:53,720 --> 00:18:56,609
tubes. They took up whole floors. The idea that

419
00:18:56,609 --> 00:18:59,329
you could take the continuous analog wave of

420
00:18:59,329 --> 00:19:02,670
a Beethoven symphony, map it into discrete binary

421
00:19:02,670 --> 00:19:06,069
choices, feed it into a machine, and have the

422
00:19:06,069 --> 00:19:09,509
machine process it as pure mathematics was decades

423
00:19:09,509 --> 00:19:12,170
ahead of its time. So far ahead. Shannon realized

424
00:19:12,170 --> 00:19:15,069
that once information is digitized, the hardware

425
00:19:15,069 --> 00:19:17,069
manipulating it doesn't need to know what the

426
00:19:17,069 --> 00:19:20,349
data represents. A logic gate processing an audio

427
00:19:20,349 --> 00:19:22,829
file operates exactly the same way as a logic

428
00:19:22,829 --> 00:19:25,859
gate processing a text document. This incredibly

429
00:19:25,859 --> 00:19:29,539
rigorous, mathematically pure approach put Shannon

430
00:19:29,539 --> 00:19:31,740
somewhat at odds with other major intellectual

431
00:19:31,740 --> 00:19:34,000
movements of the time. Oh, the academic drama.

432
00:19:34,200 --> 00:19:36,740
Oh, yeah. This brings us to his former professor

433
00:19:36,740 --> 00:19:39,859
at MIT, Norbert Wiener. Norbert Wiener. Wiener

434
00:19:39,859 --> 00:19:41,920
was a titan of mathematics who had spent the

435
00:19:41,920 --> 00:19:44,059
war working on the same anti -aircraft prediction

436
00:19:44,059 --> 00:19:46,500
problems as Shannon. But he approached them from

437
00:19:46,500 --> 00:19:48,319
a completely different angle. Wiener was the

438
00:19:48,319 --> 00:19:50,660
champion of cybernetics, a term he coined from

439
00:19:50,660 --> 00:19:52,880
the Greek word for steersman. Right. He defined

440
00:19:52,880 --> 00:19:55,220
it broadly as the study of communication and

441
00:19:55,220 --> 00:19:58,160
control in the animal and the machine. And cybernetics

442
00:19:58,160 --> 00:20:00,900
was largely focused on continuous analog systems

443
00:20:00,900 --> 00:20:04,619
and feedback loops. Think of a thermostat regulating

444
00:20:04,619 --> 00:20:07,839
a room's temperature or a ship's captain continuously

445
00:20:07,839 --> 00:20:10,779
adjusting the rudder based on the visual feedback

446
00:20:10,779 --> 00:20:13,180
of the waves. Right. It's a continuous flowing

447
00:20:13,180 --> 00:20:16,009
adjustment. Exactly. Wiener saw these continuous

448
00:20:16,009 --> 00:20:18,190
feedback loops as the key to understanding both

449
00:20:18,190 --> 00:20:21,990
biological systems and automated machinery. Shannon's

450
00:20:21,990 --> 00:20:26,369
focus, conversely, was on discrete digital distinct

451
00:20:26,369 --> 00:20:29,410
states, the ones and zeros. And the academic

452
00:20:29,410 --> 00:20:31,690
friction between the two is a classic example

453
00:20:31,690 --> 00:20:34,640
of scientific ego. Oh, it really is. In 1950,

454
00:20:34,799 --> 00:20:36,539
when Shannon's information theory was beginning

455
00:20:36,539 --> 00:20:39,400
to gain massive traction, Norbert Wiener penned

456
00:20:39,400 --> 00:20:41,339
a review of Shannon's book in Physics Today.

457
00:20:41,640 --> 00:20:44,500
And it is a master class in academic condescension.

458
00:20:44,740 --> 00:20:47,039
It really is. Wiener's review begins with a rhetorical

459
00:20:47,039 --> 00:20:49,359
maneuver designed entirely to establish a hierarchy.

460
00:20:49,619 --> 00:20:52,299
To put Shannon in his place. Yes. He writes,

461
00:20:52,440 --> 00:20:55,240
some 15 years ago, a very bright young student

462
00:20:55,240 --> 00:20:57,500
came to the authorities at MIT with an idea for

463
00:20:57,500 --> 00:20:59,539
a theory of electric switching dependent on the

464
00:20:59,539 --> 00:21:02,960
algebra of logic. The student was Claude E. Shannon.

465
00:21:03,319 --> 00:21:05,400
By referring to a globally recognized genius

466
00:21:05,400 --> 00:21:08,940
as a very bright young student, Wiener is loudly

467
00:21:08,940 --> 00:21:11,180
reminding the scientific community who the professor

468
00:21:11,180 --> 00:21:13,240
was. Exactly. Look at me. I was the teacher.

469
00:21:13,420 --> 00:21:16,769
But he doesn't stop there. Weiner summarizes

470
00:21:16,769 --> 00:21:19,109
Shannon's work, specifically focusing on the

471
00:21:19,109 --> 00:21:22,150
concept of information as negative entropy, and

472
00:21:22,150 --> 00:21:25,769
then explicitly claims that he, Weiner, had independently

473
00:21:25,769 --> 00:21:29,009
developed the exact same mathematical ideas at

474
00:21:29,009 --> 00:21:31,529
the exact same time. He heavily implies that

475
00:21:31,529 --> 00:21:34,569
Shannon's theory was just a subset of the grander,

476
00:21:34,569 --> 00:21:37,650
overarching discipline of cybernetics. We should

477
00:21:37,650 --> 00:21:40,130
unpack that specific term, entropy, because it

478
00:21:40,130 --> 00:21:42,650
is vital to the text. Definitely. In classical

479
00:21:42,650 --> 00:21:45,170
thermodynamics, entropy is a measure of disorder

480
00:21:45,170 --> 00:21:48,589
or randomness in a physical system. Shannon borrowed

481
00:21:48,589 --> 00:21:51,029
the term, supposedly at the suggestion of mathematician

482
00:21:51,029 --> 00:21:53,690
John von Neumann, to describe the measure of

483
00:21:53,690 --> 00:21:56,309
uncertainty in a message. Right. High entropy

484
00:21:56,309 --> 00:21:59,109
means a message is highly unpredictable, which

485
00:21:59,140 --> 00:22:01,319
means it carries a lot of information. Lots is

486
00:22:01,319 --> 00:22:03,880
for price. Yes. Low entropy means the message

487
00:22:03,880 --> 00:22:06,220
is highly predictable, carrying very little actual

488
00:22:06,220 --> 00:22:09,079
information. Now, Wiener's continuous feedback

489
00:22:09,079 --> 00:22:11,599
loops of cybernetics became a popular cultural

490
00:22:11,599 --> 00:22:15,759
buzzword in the 1950s and 60s. But as a rigid

491
00:22:15,759 --> 00:22:18,619
scientific discipline, it eventually fractured.

492
00:22:18,619 --> 00:22:21,240
It didn't hold up in the same way. No. Shannon's

493
00:22:21,240 --> 00:22:23,859
discrete information theory, however, provided

494
00:22:23,859 --> 00:22:27,140
the unbreakable mathematical bedrock upon which

495
00:22:27,140 --> 00:22:29,519
all digital computing and telecommunications

496
00:22:29,519 --> 00:22:32,460
were subsequently built. And Shannon's math was

497
00:22:32,460 --> 00:22:35,319
so universally applicable. that it inevitably

498
00:22:35,319 --> 00:22:37,539
spilled over the boundaries of electrical engineering.

499
00:22:37,799 --> 00:22:40,180
It went everywhere. This initiates a massive

500
00:22:40,180 --> 00:22:43,059
conceptual transition in the text. For centuries,

501
00:22:43,240 --> 00:22:46,119
the hard sciences like physics, chemistry, were

502
00:22:46,119 --> 00:22:47,900
primarily concerned with the study of energy

503
00:22:47,900 --> 00:22:51,079
and matter, the forces of nature, thermodynamics,

504
00:22:51,200 --> 00:22:53,359
atomic structures. Stuff you could touch and

505
00:22:53,359 --> 00:22:56,319
measure physically. Exactly. But information

506
00:22:56,319 --> 00:22:59,539
theory triggered a profound shift. Scientists

507
00:22:59,539 --> 00:23:01,700
across various disciplines began looking at the

508
00:23:01,700 --> 00:23:05,000
natural world through a new lens. The lens of

509
00:23:05,000 --> 00:23:07,200
information processing. This shift hit the biological

510
00:23:07,200 --> 00:23:09,680
sciences precisely when biologists were grappling

511
00:23:09,680 --> 00:23:11,640
with the most profound mystery of their field.

512
00:23:11,880 --> 00:23:14,460
Heredity. How do we pass things on? Exactly.

513
00:23:14,680 --> 00:23:18,140
How exactly do complex traits pass from a parent

514
00:23:18,140 --> 00:23:21,359
to an offspring? Biologists knew there was a

515
00:23:21,359 --> 00:23:23,859
hypothetical set of instructions located somewhere

516
00:23:23,859 --> 00:23:27,480
within the chromosomes. The physicist Erwin Schrödinger

517
00:23:27,480 --> 00:23:29,900
had published a famous book called What is Life?

518
00:23:30,349 --> 00:23:32,869
where he speculated that genetic material had

519
00:23:32,869 --> 00:23:34,910
to be some kind of a periodic crystal. And a

520
00:23:34,910 --> 00:23:36,849
periodic crystal. Yeah, a physical structure

521
00:23:36,849 --> 00:23:39,349
that contained a non -repeating coded script.

522
00:23:39,670 --> 00:23:42,130
But biologists did not know what the physical

523
00:23:42,130 --> 00:23:45,190
tape of this code looked like. Some scientists

524
00:23:45,190 --> 00:23:48,089
were attempting to apply information theory directly

525
00:23:48,089 --> 00:23:51,029
to whole organisms, trying to calculate the total

526
00:23:51,029 --> 00:23:54,069
bit complexity of a human being by estimating

527
00:23:54,069 --> 00:23:56,680
the information content of... every single cell.

528
00:23:56,799 --> 00:23:58,980
Which is just impossible. It resulted in massive,

529
00:23:59,039 --> 00:24:02,500
clunky, completely unhelpful numbers. They couldn't

530
00:24:02,500 --> 00:24:04,539
make use of Shannon's discrete mathematics because

531
00:24:04,539 --> 00:24:06,799
they hadn't found the actual wires carrying the

532
00:24:06,799 --> 00:24:10,180
biological data. That brings us to 1953, across

533
00:24:10,180 --> 00:24:12,599
the Atlantic at the Cavendish Laboratory in Cambridge.

534
00:24:13,140 --> 00:24:15,660
James Watson and Francis Crick were racing to

535
00:24:15,660 --> 00:24:17,519
decipher the physical structure of a substance

536
00:24:17,519 --> 00:24:20,819
found in the nucleus of cells. Deoxyribonucleic

537
00:24:20,819 --> 00:24:23,819
acid. DNA. They were utilizing the X -ray diffraction

538
00:24:23,819 --> 00:24:26,740
images captured by Rosalind Franklin to understand

539
00:24:26,740 --> 00:24:30,279
the physical geometry of the molecule. But Gleick's

540
00:24:30,279 --> 00:24:33,160
text focuses intensely on the realization of

541
00:24:33,160 --> 00:24:35,500
what that geometry represented conceptually.

542
00:24:35,680 --> 00:24:38,079
Watson and Crick deduced the double helix structure.

543
00:24:38,359 --> 00:24:40,640
But the true breakthrough was... Understanding

544
00:24:40,640 --> 00:24:43,339
the rungs of that helical ladder. The parts connecting

545
00:24:43,339 --> 00:24:46,339
the sides. Exactly. They realized that DNA is

546
00:24:46,339 --> 00:24:48,839
constructed from smaller nucleotide units, and

547
00:24:48,839 --> 00:24:51,440
each unit contains one of four specific chemical

548
00:24:51,440 --> 00:24:56,680
bases. Adenine, cytosine, guanine, and thymine.

549
00:24:56,940 --> 00:25:01,119
A, C, G, and T. If we map this directly back

550
00:25:01,119 --> 00:25:03,420
to Shannon, the implications are just staggering.

551
00:25:03,700 --> 00:25:05,839
These four chemical bases were not just structural

552
00:25:05,839 --> 00:25:08,319
support. They were discrete informational states.

553
00:25:08,809 --> 00:25:11,390
They were the letters of a biological code. Yes.

554
00:25:11,609 --> 00:25:14,190
It is the exact mathematical equivalent of Shannon's

555
00:25:14,190 --> 00:25:16,789
binary bit. But instead of a base -2 system of

556
00:25:16,789 --> 00:25:19,750
ones and zeros, biology utilizes a base -4 digital

557
00:25:19,750 --> 00:25:21,990
system. Nature's own hard drive. The sequence

558
00:25:21,990 --> 00:25:24,609
of A, C, G, and T carries the exact instructions

559
00:25:24,609 --> 00:25:27,170
required to build a human, a cuckoo bird, or

560
00:25:27,170 --> 00:25:29,609
a bacterium. It is an algorithmic construction

561
00:25:29,609 --> 00:25:32,250
set written in physical chemistry. The way the

562
00:25:32,250 --> 00:25:34,430
math works out is beautiful, really. Because

563
00:25:34,430 --> 00:25:37,289
it's a base -4 system, the biological machinery

564
00:25:37,289 --> 00:25:40,289
reads the letters in sequences of three, known

565
00:25:40,289 --> 00:25:44,069
as codons. Three -letter words. Right. Four letters

566
00:25:44,069 --> 00:25:46,569
arranged in three -letter words gives you 64

567
00:25:46,569 --> 00:25:50,170
possible combinations. Those 64 combinations

568
00:25:50,170 --> 00:25:53,349
map directly to the 20 amino acids that form

569
00:25:53,349 --> 00:25:55,910
all the proteins in the human body. With a few

570
00:25:55,910 --> 00:25:58,069
combinations acting as start and stop commands.

571
00:25:58,390 --> 00:26:00,329
Just like a computer program. Exactly like a

572
00:26:00,329 --> 00:26:03,180
computer program. But in the early 1950s, the

573
00:26:03,180 --> 00:26:05,720
established language of biology was completely

574
00:26:05,720 --> 00:26:08,880
unequipped to describe this reality. The scientists

575
00:26:08,880 --> 00:26:11,519
were observing digital data processing, but they

576
00:26:11,519 --> 00:26:13,799
were still using the terminology of fluid dynamics

577
00:26:13,799 --> 00:26:16,930
and chemical mixing. There is a fantastic anecdote

578
00:26:16,930 --> 00:26:19,130
in the source material that illustrates this

579
00:26:19,130 --> 00:26:22,250
collision of vocabularies. In the spring of 1953,

580
00:26:22,650 --> 00:26:25,089
just before Watson and Crick officially published

581
00:26:25,089 --> 00:26:28,069
their double helix model, an unusual letter arrived

582
00:26:28,069 --> 00:26:30,390
at the offices of the journal Nature in London.

583
00:26:30,670 --> 00:26:32,769
Oh, the prank letter. The prank letter. It was

584
00:26:32,769 --> 00:26:35,029
signed by several prominent scientists, including

585
00:26:35,029 --> 00:26:38,230
the French geneticist Boris Afrousi and a 25

586
00:26:38,230 --> 00:26:40,730
-year -old James Watson. The letter was essentially

587
00:26:40,730 --> 00:26:44,019
a high -level academic prank. written after a

588
00:26:44,019 --> 00:26:46,279
wine -soaked lunch near a lake in Switzerland.

589
00:26:46,619 --> 00:26:48,680
Sounds like a good lunch. They were complaining

590
00:26:48,680 --> 00:26:51,380
about the chaotic and confusing terminology used

591
00:26:51,380 --> 00:26:53,900
in literature regarding genetic recombination

592
00:26:53,900 --> 00:26:57,019
in bacteria. Biologists were using all these

593
00:26:57,019 --> 00:27:00,319
muddy terms like transformation, induction, transduction,

594
00:27:00,420 --> 00:27:02,819
and infection to describe what was happening

595
00:27:02,819 --> 00:27:06,539
when bacteria exchanged traits. So as a joke,

596
00:27:06,539 --> 00:27:08,599
they proposed discarding all those confusing

597
00:27:08,599 --> 00:27:11,420
terms and replacing them with one clean, unified

598
00:27:11,420 --> 00:27:15,710
phrase. Interbacterial information. Interbacterial

599
00:27:15,710 --> 00:27:18,210
information. You were being facetious, but conceptually

600
00:27:18,210 --> 00:27:20,390
they were hitting the absolute center of the

601
00:27:20,390 --> 00:27:22,829
target. They explicitly noted in their letter

602
00:27:22,829 --> 00:27:25,049
that the term does not imply necessarily the

603
00:27:25,049 --> 00:27:28,049
transfer of material substances and recognizes

604
00:27:28,049 --> 00:27:30,230
the possible future importance of cybernetics

605
00:27:30,230 --> 00:27:32,950
at the bacterial level. Even as a joke, they

606
00:27:32,950 --> 00:27:35,029
were acknowledging that life at the microscopic

607
00:27:35,029 --> 00:27:38,230
level is not merely a chemical soup. It is an

608
00:27:38,230 --> 00:27:40,960
active data processing network. And the very

609
00:27:40,960 --> 00:27:43,599
next issue of Nature carried the famous paper

610
00:27:43,599 --> 00:27:47,240
detailing the structure of DNA. The lunchtime

611
00:27:47,240 --> 00:27:49,680
joke became the fundamental reality of biology

612
00:27:49,680 --> 00:27:52,240
practically overnight. Just like that. The physicist

613
00:27:52,240 --> 00:27:54,960
George Gamow immediately recognized the implications,

614
00:27:55,380 --> 00:28:01,940
writing, If we connect this to the broader arc

615
00:28:01,940 --> 00:28:04,279
of our discussion, it means that biological reproduction

616
00:28:04,279 --> 00:28:07,700
is, at its core, a problem of communication engineering.

617
00:28:08,119 --> 00:28:10,869
Exactly what Shannon was studying. Yes. How does

618
00:28:10,869 --> 00:28:13,910
nature reproduce a highly complex message, the

619
00:28:13,910 --> 00:28:16,950
blueprint of a living organism, at another point

620
00:28:16,950 --> 00:28:19,450
in time and space, maintaining accuracy in the

621
00:28:19,450 --> 00:28:21,369
presence of environmental noise and mutation?

622
00:28:21,769 --> 00:28:24,170
That is Shannon's fundamental problem of communication,

623
00:28:24,609 --> 00:28:27,710
playing out not in Bell Labs' copper wires, but

624
00:28:27,710 --> 00:28:30,250
in the organic machinery of the cell. And it

625
00:28:30,250 --> 00:28:32,710
proves Shannon's point about meaning being irrelevant

626
00:28:32,710 --> 00:28:34,950
to the mechanics of reproduction. Oh, this is

627
00:28:34,950 --> 00:28:37,579
a great point. the dna polymerase enzymes that

628
00:28:37,579 --> 00:28:40,000
unzip and replicate the double helix do not know

629
00:28:40,000 --> 00:28:41,799
what they are building they do not understand

630
00:28:41,799 --> 00:28:44,259
the concept of an eye or a leaf or a heartbeat

631
00:28:44,259 --> 00:28:47,380
they have no idea they are just blind mechanical

632
00:28:47,380 --> 00:28:50,220
logic gates matching an a to a t and a c to a

633
00:28:50,220 --> 00:28:54,579
g they are processing pure information this realization

634
00:28:54,579 --> 00:28:57,099
prompts an inevitable expanding question though

635
00:28:57,099 --> 00:29:00,839
if the biological hardware of our bodies is driven

636
00:29:00,839 --> 00:29:03,440
entirely by the blind replication of encoded

637
00:29:03,440 --> 00:29:06,609
information What about the software of our minds?

638
00:29:06,710 --> 00:29:09,230
Right. What about human culture, language, and

639
00:29:09,230 --> 00:29:12,230
thought? This shifts our focus from the biological

640
00:29:12,230 --> 00:29:14,930
realm into the cultural realm, introducing the

641
00:29:14,930 --> 00:29:18,150
concept of the meme pool. In 1976, the evolutionary

642
00:29:18,150 --> 00:29:21,210
biologist Richard Dawkins published The Selfish

643
00:29:21,210 --> 00:29:24,170
Gene. It was a book that violently reoriented

644
00:29:24,170 --> 00:29:26,609
evolutionary theory. It really shook things up.

645
00:29:26,849 --> 00:29:28,869
Dawkins argued that we had been looking at evolution

646
00:29:28,869 --> 00:29:30,930
from the wrong perspective. We thought natural

647
00:29:30,930 --> 00:29:33,089
selection was about the survival of the fittest

648
00:29:33,089 --> 00:29:36,049
individual organism or the survival of the species.

649
00:29:36,369 --> 00:29:38,650
The whole animal. Right. But Dawkins argued that

650
00:29:38,650 --> 00:29:40,369
the physical organism is largely irrelevant.

651
00:29:40,789 --> 00:29:43,710
The actual unit of selection is the gene itself.

652
00:29:44,089 --> 00:29:46,869
We are just the vehicles. We are simply the temporary

653
00:29:46,869 --> 00:29:50,049
fleshy vehicles that genes construct to ensure

654
00:29:50,049 --> 00:29:51,970
their own replication into the next generation.

655
00:29:52,670 --> 00:29:55,329
It was a stark, uncompromising application of

656
00:29:55,329 --> 00:29:58,230
information theory to biology. But at the very

657
00:29:58,230 --> 00:30:01,450
end of the book, Dawkins took a massive speculative

658
00:30:01,450 --> 00:30:03,950
leap. He asked if there might be another kind

659
00:30:03,950 --> 00:30:06,630
of replicator, a unit of transmission that operates

660
00:30:06,630 --> 00:30:09,230
outside of DNA. He proposed a unit of cultural

661
00:30:09,230 --> 00:30:12,349
inheritance. Yes, he called it the meme. Now,

662
00:30:12,349 --> 00:30:14,990
when you hear the word meme today, you probably

663
00:30:14,990 --> 00:30:17,890
think of a viral image with bold text on the

664
00:30:17,890 --> 00:30:20,250
Internet, a cat picture. Right, the Internet

665
00:30:20,250 --> 00:30:22,970
meme. But Dawkins was proposing something far

666
00:30:22,970 --> 00:30:26,509
more fundamental and pervasive. A meme, in Dawkins'

667
00:30:26,630 --> 00:30:29,369
original formulation, is any unit of information

668
00:30:29,369 --> 00:30:31,829
that can propagate from one human brain to another.

669
00:30:32,049 --> 00:30:34,789
From brain to brain. It could be a specific idea,

670
00:30:35,029 --> 00:30:38,230
a fashion trend, a method of building an arch,

671
00:30:38,490 --> 00:30:41,990
a catchy tune, a religious dogma, or a conspiracy

672
00:30:41,990 --> 00:30:45,390
theory. Just like genes, memes are subject to

673
00:30:45,390 --> 00:30:48,420
the brutal laws of Darwinian evolution. Variation,

674
00:30:48,440 --> 00:30:51,180
mutation, competition for resources, which in

675
00:30:51,180 --> 00:30:53,019
this case is human attention and inheritance.

676
00:30:53,279 --> 00:30:55,039
I think we need to draw a sharp distinction here

677
00:30:55,039 --> 00:30:57,180
because treating a catchy tune exactly like a

678
00:30:57,180 --> 00:30:59,500
strand of DNA seems a bit problematic. How so?

679
00:30:59,849 --> 00:31:02,589
Well, genes have a grounding in physical substance.

680
00:31:02,950 --> 00:31:06,009
I can put a cell under an electron microscope

681
00:31:06,009 --> 00:31:09,269
and literally read the physical sequence of As,

682
00:31:09,430 --> 00:31:12,849
Cs, Gs, and Ts. Furthermore, genes replicate

683
00:31:12,849 --> 00:31:16,109
with incredibly high fidelity. Biological mutations

684
00:31:16,109 --> 00:31:18,589
are relatively rare, which allows species to

685
00:31:18,589 --> 00:31:21,109
remain physically stable over thousands of years.

686
00:31:21,349 --> 00:31:23,490
That is the crucial difference highlighted in

687
00:31:23,490 --> 00:31:26,789
the text. Memes are entirely abstract. They're

688
00:31:26,789 --> 00:31:29,269
intangible patterns of neurological activity.

689
00:31:29,549 --> 00:31:31,670
Right. They aren't physical objects. And unlike

690
00:31:31,670 --> 00:31:34,250
genes, memes are almost never copied with exact

691
00:31:34,250 --> 00:31:36,670
fidelity. Their boundaries are inherently fuzzy

692
00:31:36,670 --> 00:31:39,329
and they mutate with a wild, chaotic flexibility

693
00:31:39,329 --> 00:31:42,349
that would result in immediate biological death

694
00:31:42,349 --> 00:31:44,710
if applied to genetic replication. The examples

695
00:31:44,710 --> 00:31:46,829
the book uses to illustrate this informational

696
00:31:46,829 --> 00:31:49,950
ecology are fantastic. Dawkins himself pointed

697
00:31:49,950 --> 00:31:51,829
out that a meme could be the first four notes

698
00:31:51,829 --> 00:31:55,490
of Beethoven's Fifth Symphony. The moment you

699
00:31:55,490 --> 00:31:58,309
hear it, it lodges in your auditory cortex. You

700
00:31:58,309 --> 00:32:00,549
find yourself humming it later that day. Someone

701
00:32:00,549 --> 00:32:02,789
else hears you hum it, and the pattern replicates

702
00:32:02,789 --> 00:32:05,269
into their brain. The meme has successfully used

703
00:32:05,269 --> 00:32:08,609
you as a host to reproduce. The book also examines

704
00:32:08,609 --> 00:32:11,190
dietary fads through the lens of memetics. Oh,

705
00:32:11,190 --> 00:32:13,609
this is fascinating. Think about the South Beach

706
00:32:13,609 --> 00:32:16,210
diet, the Atkins diet, the Scarsdale diet, the

707
00:32:16,210 --> 00:32:18,470
cabbage soup diet. They come and go in waves.

708
00:32:19,009 --> 00:32:21,390
These protocols propagate through populations

709
00:32:21,390 --> 00:32:24,009
according to dynamics that have virtually nothing

710
00:32:24,009 --> 00:32:26,829
to do with the rigorous science of human metabolism.

711
00:32:27,690 --> 00:32:30,569
They are highly contagious informational viruses

712
00:32:30,569 --> 00:32:34,529
that infect populations, adapt to cultural anxieties

713
00:32:34,529 --> 00:32:37,029
about health, and spread rapidly through social

714
00:32:37,029 --> 00:32:39,890
networks. Another fascinating and somewhat darker

715
00:32:39,890 --> 00:32:42,650
example mentioned in the text is the phenomenon

716
00:32:42,650 --> 00:32:46,210
of iatroepidemics, which are epidemics caused

717
00:32:46,210 --> 00:32:49,450
not by biological pathogens, but by fashions

718
00:32:49,450 --> 00:32:52,009
in medical treatment. Medical memes. Exactly.

719
00:32:52,009 --> 00:32:54,130
In the mid -20th century, there was a massive

720
00:32:54,130 --> 00:32:56,690
sweeping craze in the United States and parts

721
00:32:56,690 --> 00:32:59,210
of Europe for performing tonsillectomies on children.

722
00:32:59,410 --> 00:33:01,750
Just taking them out routinely. Millions of kids

723
00:33:01,750 --> 00:33:03,890
had their tonsils surgically removed despite

724
00:33:03,890 --> 00:33:05,970
the fact that the actual medical benefit was

725
00:33:05,970 --> 00:33:08,190
negligible or non -existent for the vast majority

726
00:33:08,190 --> 00:33:11,779
of them. It was herd behavior. It was a medical

727
00:33:11,779 --> 00:33:14,339
meme that infected the establishment, sounded

728
00:33:14,339 --> 00:33:16,980
plausible, and spread rapidly before eventually

729
00:33:16,980 --> 00:33:19,819
dying out as a new generation of doctors rejected

730
00:33:19,819 --> 00:33:22,119
the pattern. The speed and vector of transmission

731
00:33:22,119 --> 00:33:25,599
are what make memes so powerful. Genes are restricted

732
00:33:25,599 --> 00:33:27,720
to vertical transmission. They can only pass

733
00:33:27,720 --> 00:33:30,319
from parent to offspring, which means a genetic

734
00:33:30,319 --> 00:33:32,680
mutation takes an entire generation to propagate.

735
00:33:32,900 --> 00:33:35,710
Decades. Right. Memes, however... operate via

736
00:33:35,710 --> 00:33:38,490
horizontal transmission. They can pass from person

737
00:33:38,490 --> 00:33:41,569
to person across continents almost instantaneously.

738
00:33:41,690 --> 00:33:43,789
To see horizontal transmission in action, the

739
00:33:43,789 --> 00:33:45,809
text points to the sudden explosion of those

740
00:33:45,809 --> 00:33:48,609
yellow diamond -shaped baby -on -board signs.

741
00:33:48,970 --> 00:33:51,750
None of those signs. In 1984, they appeared almost

742
00:33:51,750 --> 00:33:54,369
out of nowhere in the United States. It was a

743
00:33:54,369 --> 00:33:57,130
meme capitalizing on a mass panic of protective

744
00:33:57,130 --> 00:34:00,170
parenting. It replicated furiously, spreading

745
00:34:00,170 --> 00:34:02,430
to Europe and Japan. It was everywhere on cars.

746
00:34:02,960 --> 00:34:05,640
But then, almost as rapidly as the original meme

747
00:34:05,640 --> 00:34:08,340
saturated the environment, it spawned ironic

748
00:34:08,340 --> 00:34:11,320
mutations to survive in a changing cultural climate.

749
00:34:11,420 --> 00:34:14,500
It had to adapt. Right. Suddenly you saw signs

750
00:34:14,500 --> 00:34:18,480
that said, baby, I'm bored, or X in trunk. The

751
00:34:18,480 --> 00:34:21,280
core idea mutated, capitalizing on the established

752
00:34:21,280 --> 00:34:24,019
visual pattern, but altering the semantic payload

753
00:34:24,019 --> 00:34:26,380
to capture the attention of a more cynical demographic.

754
00:34:26,900 --> 00:34:29,280
If we connect this to the bigger picture. We

755
00:34:29,280 --> 00:34:31,880
can see that human technological progress has

756
00:34:31,880 --> 00:34:34,059
basically been an exercise in creating a world

757
00:34:34,059 --> 00:34:36,719
perfectly optimized for the rapid replication

758
00:34:36,719 --> 00:34:39,320
of memes. The internet is just a meme machine.

759
00:34:39,599 --> 00:34:41,960
The philosopher Daniel Dennett noted that we

760
00:34:41,960 --> 00:34:44,320
live within a physical biosphere, but we have

761
00:34:44,320 --> 00:34:47,469
simultaneously constructed an infosphere. an

762
00:34:47,469 --> 00:34:50,250
invisible parallel environment composed entirely

763
00:34:50,250 --> 00:34:53,190
of propagating information. The infosphere? In

764
00:34:53,190 --> 00:34:55,110
the competition for the limited bandwidth of

765
00:34:55,110 --> 00:34:57,710
human attention, it is the most robust, highly

766
00:34:57,710 --> 00:35:00,449
adapted memes that survive and shape our reality.

767
00:35:00,650 --> 00:35:02,530
And this is regardless of whether those memes

768
00:35:02,530 --> 00:35:04,630
are factually true or beneficial to the host.

769
00:35:04,869 --> 00:35:08,670
Wow. So, we have traced information from copper

770
00:35:08,670 --> 00:35:12,719
wires, to the biological hardware of DNA, to

771
00:35:12,719 --> 00:35:15,840
the cultural software of memes propagating through

772
00:35:15,840 --> 00:35:18,840
our minds. Quite a journey. But this raises a

773
00:35:18,840 --> 00:35:21,480
fundamental mathematical issue. Shannon's original

774
00:35:21,480 --> 00:35:24,000
theory was brilliant for calculating the bandwidth

775
00:35:24,000 --> 00:35:26,960
needed for telephone wires, but it relied on

776
00:35:26,960 --> 00:35:29,670
probability. Right, ensembles of messages. How

777
00:35:29,670 --> 00:35:31,769
do we mathematically measure the information

778
00:35:31,769 --> 00:35:35,469
content of a single, complex idea? How do we

779
00:35:35,469 --> 00:35:37,590
quantify the difference between a random string

780
00:35:37,590 --> 00:35:39,909
of nonsense syllables and a masterpiece like

781
00:35:39,909 --> 00:35:42,309
Tolstoy's Warren piece? You can't use probability

782
00:35:42,309 --> 00:35:45,250
for that. That brings us to the deepest mathematical

783
00:35:45,250 --> 00:35:47,869
concepts in the text, exploring the sense of

784
00:35:47,869 --> 00:35:50,269
randomness. We're moving past the engineering

785
00:35:50,269 --> 00:35:53,030
problem of transmitting data and looking at the

786
00:35:53,030 --> 00:35:55,610
intrinsic mathematical nature of complexity itself.

787
00:35:56,159 --> 00:35:59,500
To do this, the text introduces a brilliant Soviet

788
00:35:59,500 --> 00:36:02,500
mathematician named Andrei Kolmogorov and an

789
00:36:02,500 --> 00:36:04,500
American mathematician named Gregory Chaitin.

790
00:36:04,599 --> 00:36:06,960
They both arrived at the same conclusions. Independently,

791
00:36:06,980 --> 00:36:09,579
yes. The environment Kolmogorov was working in

792
00:36:09,579 --> 00:36:11,780
provides a stark contrast to Shannon's world

793
00:36:11,780 --> 00:36:14,760
at Bell Labs, too. While America was rapidly

794
00:36:14,760 --> 00:36:17,219
laying the groundwork for the modern telecommunications

795
00:36:17,219 --> 00:36:21,099
grid, the Soviet infrastructure was pragmatically

796
00:36:21,099 --> 00:36:24,280
dismal. Very different worlds. The book notes

797
00:36:24,280 --> 00:36:27,079
that even well into the 1960s, there was no such

798
00:36:27,079 --> 00:36:29,559
thing as direct long -distance dialing in the

799
00:36:29,559 --> 00:36:32,380
Soviet Union. The volume of toll calls had not

800
00:36:32,380 --> 00:36:34,860
even surpassed the volume of physical telegrams.

801
00:36:34,880 --> 00:36:37,460
Yet theoretically, Kornigorov and his mathematics

802
00:36:37,460 --> 00:36:40,320
students in Moscow were operating at the absolute

803
00:36:40,320 --> 00:36:43,300
cutting edge. They even launched a highly influential

804
00:36:43,300 --> 00:36:46,119
journal called Problems of Information Transmission.

805
00:36:46,360 --> 00:36:48,860
They were focused on the math, not the hardware.

806
00:36:49,329 --> 00:36:51,349
Kolmogorov was attempting to solve the exact

807
00:36:51,349 --> 00:36:53,610
conceptual gap in Shannon's theory that you just

808
00:36:53,610 --> 00:36:56,210
mentioned. Right. Because Shannon's entire mathematical

809
00:36:56,210 --> 00:36:58,730
framework was built on the concept of probability

810
00:36:58,730 --> 00:37:01,929
ensembles. The information contained in a specific

811
00:37:01,929 --> 00:37:04,349
message is calculated based on the probability

812
00:37:04,349 --> 00:37:07,250
of that message being selected from a known set

813
00:37:07,250 --> 00:37:09,969
of possible messages. If you draw one card from

814
00:37:09,969 --> 00:37:11,929
a standard check, the information is calculated

815
00:37:11,929 --> 00:37:15,420
based on the 1 in 52 probability. But Kolmogorov

816
00:37:15,420 --> 00:37:18,559
recognized that this probabilistic approach completely

817
00:37:18,559 --> 00:37:21,539
fails when you're dealing with a singular, highly

818
00:37:21,539 --> 00:37:25,420
complex, finite object. Like a book. Right. He

819
00:37:25,420 --> 00:37:27,949
asked in his notes, How do you mathematically

820
00:37:27,949 --> 00:37:30,610
define the information content of a specific

821
00:37:30,610 --> 00:37:33,849
genetic sequence or a singular work of literature?

822
00:37:34,230 --> 00:37:36,710
You can't. You cannot calculate the probability

823
00:37:36,710 --> 00:37:39,469
of war and peace being drawn from a hypothetical

824
00:37:39,469 --> 00:37:42,190
ensemble of all possible novels ever written.

825
00:37:42,389 --> 00:37:45,230
The probability is essentially zero, which makes

826
00:37:45,230 --> 00:37:47,690
Shannon's equations useless for analyzing the

827
00:37:47,690 --> 00:37:50,329
object itself. Here's where it gets really interesting.

828
00:37:50,769 --> 00:37:53,190
Kolmogorov, simultaneously with Chaitin in the

829
00:37:53,190 --> 00:37:55,730
U .S., invented an entirely new mathematical

830
00:37:55,730 --> 00:37:58,050
framework. They called it algorithmic information

831
00:37:58,050 --> 00:38:00,510
theory. Algorithmic information theory. They

832
00:38:00,510 --> 00:38:03,570
decided to abandon probability entirely and redefine

833
00:38:03,570 --> 00:38:05,969
the concept of complexity. Kolmogorov stated

834
00:38:05,969 --> 00:38:08,550
that the complexity of any specific object, whether

835
00:38:08,550 --> 00:38:10,650
it is a string of numbers, a text file, or an

836
00:38:10,650 --> 00:38:13,550
image, is exactly equal to the length of the

837
00:38:13,550 --> 00:38:15,949
shortest possible computer program required to

838
00:38:15,949 --> 00:38:17,829
generate that object. The shortest algorithm.

839
00:38:18,110 --> 00:38:20,210
The shortest algorithm is the true measure of

840
00:38:20,210 --> 00:38:22,650
its information content. Let me make sure I'm

841
00:38:22,650 --> 00:38:24,690
visualizing this correctly because it completely

842
00:38:24,690 --> 00:38:28,090
inverts our standard intuition about data. Okay,

843
00:38:28,110 --> 00:38:30,150
let's walk through it. Imagine I want to send

844
00:38:30,150 --> 00:38:32,409
you a text file and the file consists of the

845
00:38:32,409 --> 00:38:35,510
letter A repeated exactly one million times,

846
00:38:35,750 --> 00:38:39,070
triple A, and so on. A million times. Intuitively,

847
00:38:39,070 --> 00:38:41,889
that looks like a large amount of data. It is

848
00:38:41,889 --> 00:38:44,030
a million characters long. It takes up space

849
00:38:44,030 --> 00:38:46,750
on a hard drive. Under Shannon's framework, treating

850
00:38:46,750 --> 00:38:49,369
each character independently, you might calculate

851
00:38:49,369 --> 00:38:52,469
a certain amount of entropy. But Kolmogorov would

852
00:38:52,469 --> 00:38:55,329
look at that string and declare that it possesses

853
00:38:55,329 --> 00:38:58,469
almost zero algorithmic complexity. Because it's

854
00:38:58,469 --> 00:39:00,869
a simple pattern. Exactly. Because the computer

855
00:39:00,869 --> 00:39:03,489
program, the algorithm required to output that

856
00:39:03,489 --> 00:39:06,530
string, is incredibly short. You only need a

857
00:39:06,530 --> 00:39:09,650
few lines of code. For n equals 1 to 1 million,

858
00:39:09,849 --> 00:39:12,909
print a. Print a 1 million times. Because the

859
00:39:12,909 --> 00:39:15,150
algorithm is tiny, the object is mathematically

860
00:39:15,150 --> 00:39:17,670
simple. It contains almost no actual information.

861
00:39:18,239 --> 00:39:21,099
It is entirely predictable. It is totally ordered.

862
00:39:21,219 --> 00:39:24,440
Now, consider the alternative. Imagine a different

863
00:39:24,440 --> 00:39:26,980
string of a million characters, but this time

864
00:39:26,980 --> 00:39:29,420
the characters were generated by rolling dice

865
00:39:29,420 --> 00:39:32,659
or measuring radioactive decay. Pure randomness.

866
00:39:32,760 --> 00:39:35,400
It is a million characters of pure, unstructured,

867
00:39:35,599 --> 00:39:38,219
unpredictable noise. If you attempt to write

868
00:39:38,219 --> 00:39:40,920
a computer program to generate that exact specific

869
00:39:40,920 --> 00:39:44,000
random string, you will immediately encounter

870
00:39:44,000 --> 00:39:46,940
a mathematical wall because there is no pattern.

871
00:39:47,420 --> 00:39:49,360
There is no shortcut. You can't write a loop.

872
00:39:49,440 --> 00:39:51,320
There is no loop you can write to save space.

873
00:39:51,719 --> 00:39:54,320
The only possible way to write an algorithm to

874
00:39:54,320 --> 00:39:57,280
produce that specific random output is to write

875
00:39:57,280 --> 00:40:00,300
a program that says print, followed by the entire

876
00:40:00,300 --> 00:40:03,059
literal million character string. The algorithm

877
00:40:03,059 --> 00:40:05,400
cannot be compressed. It must be exactly as long

878
00:40:05,400 --> 00:40:07,659
as the object it is trying to produce. Exactly.

879
00:40:07,920 --> 00:40:10,179
And because the algorithm is maximally long,

880
00:40:10,639 --> 00:40:12,719
Kolmogorov's theory dictates that the random

881
00:40:12,719 --> 00:40:14,940
string possesses maximal algorithmic complexity.

882
00:40:15,559 --> 00:40:18,239
Therefore, mathematically speaking, the random

883
00:40:18,239 --> 00:40:20,679
string contains the maximum possible amount of

884
00:40:20,679 --> 00:40:24,099
information. This concept is staggering. Maximum

885
00:40:24,099 --> 00:40:26,719
randomness equals maximum information. It flips

886
00:40:26,719 --> 00:40:28,739
everything on its head. We are culturally conditioned

887
00:40:28,739 --> 00:40:32,420
to think of information as order. as meaning,

888
00:40:32,639 --> 00:40:37,099
as a clear and structured signal. But algorithmic

889
00:40:37,099 --> 00:40:39,239
information theory proves that if a signal is

890
00:40:39,239 --> 00:40:41,119
perfectly ordered and perfectly predictable,

891
00:40:41,400 --> 00:40:44,320
it carries no new information because the receiver

892
00:40:44,320 --> 00:40:46,519
can easily guess what is coming next. Right.

893
00:40:46,539 --> 00:40:49,780
It's boring data. It is the unexpected, the unpredictable,

894
00:40:50,139 --> 00:40:53,619
the highly complex anomalies in a message that

895
00:40:53,619 --> 00:40:56,289
actually carry the informational payload. And

896
00:40:56,289 --> 00:40:58,289
this is not just a philosophical parlor trick.

897
00:40:58,449 --> 00:41:01,309
This rigorous mathematical definition of complexity

898
00:41:01,309 --> 00:41:04,429
and randomness is the exact operational foundation

899
00:41:04,429 --> 00:41:08,230
of how modern computing works, specifically regarding

900
00:41:08,230 --> 00:41:11,449
data compression. Oh, like ZIP files. Exactly.

901
00:41:11,650 --> 00:41:14,409
When you download a ZIP file or stream a highly

902
00:41:14,409 --> 00:41:16,750
compressed 4K video over your cellular network,

903
00:41:16,969 --> 00:41:19,849
the software routing that data is actively utilizing

904
00:41:19,849 --> 00:41:22,389
Kolmogorov's concepts. It's shrinking the algorithm.

905
00:41:23,110 --> 00:41:25,710
The compression algorithm scans the raw data

906
00:41:25,710 --> 00:41:28,909
looking for ordered patterns, the digital equivalent

907
00:41:28,909 --> 00:41:32,570
of print A a million times. It removes the redundancy,

908
00:41:32,670 --> 00:41:35,210
collapsing the predictable data into short algorithms.

909
00:41:35,710 --> 00:41:38,349
What is left over, the incompressible, highly

910
00:41:38,349 --> 00:41:42,889
complex, unpredictable data, is the true, irreducible

911
00:41:42,889 --> 00:41:44,869
information that must be transmitted. We can

912
00:41:44,869 --> 00:41:46,769
connect this directly to the architecture of

913
00:41:46,769 --> 00:41:49,150
modern artificial intelligence, too. Oh, absolutely.

914
00:41:49,449 --> 00:41:52,929
LLMs. When engineers train massive large language

915
00:41:52,929 --> 00:41:55,670
models, they are not teaching the machine meaning

916
00:41:55,670 --> 00:41:58,250
in the human sense. They are feeding the network

917
00:41:58,250 --> 00:42:01,170
petabytes of human text and utilizing mathematical

918
00:42:01,170 --> 00:42:03,710
algorithms to find the statistical redundancies

919
00:42:03,710 --> 00:42:05,969
in our language. They are compressing human thought.

920
00:42:06,150 --> 00:42:07,869
They are compressing the algorithmic complexity

921
00:42:07,869 --> 00:42:10,909
of human thought into a massive web of predictable

922
00:42:10,909 --> 00:42:13,920
probabilities. But as Kolmogorov showed, the

923
00:42:13,920 --> 00:42:16,159
most informative, insightful human thoughts are

924
00:42:16,159 --> 00:42:18,460
mathematically the most unpredictable ones. The

925
00:42:18,460 --> 00:42:20,360
ones you can't guess. The AI is essentially a

926
00:42:20,360 --> 00:42:22,619
massive engine designed to minimize surprise,

927
00:42:22,739 --> 00:42:25,420
while true information is the surprise itself.

928
00:42:26,059 --> 00:42:28,079
It is incredible to trace this narrative line

929
00:42:28,079 --> 00:42:30,900
from beginning to end. It really is. We started

930
00:42:30,900 --> 00:42:33,440
in a sandy brick building on West Street in Manhattan

931
00:42:33,440 --> 00:42:36,119
with Claude Shannon looking at the encryption

932
00:42:36,119 --> 00:42:38,820
of the Churchill Roosevelt hotline and deciding

933
00:42:38,820 --> 00:42:42,179
that a simple coin flip represents one quantifiable

934
00:42:42,179 --> 00:42:44,639
bit of information. We followed that mathematical

935
00:42:44,639 --> 00:42:47,440
unit across the Atlantic to the Cavendish Laboratory.

936
00:42:48,309 --> 00:42:50,570
where Watson and Crick realized that the physical

937
00:42:50,570 --> 00:42:54,070
structure of DNA is actually a base for digital

938
00:42:54,070 --> 00:42:56,969
storage medium, processing information through

939
00:42:56,969 --> 00:43:00,070
triplet codons. We then watched as Richard Dawkins

940
00:43:00,070 --> 00:43:02,650
took that concept of the replicating code and

941
00:43:02,650 --> 00:43:05,030
applied it to human culture, demonstrating how

942
00:43:05,030 --> 00:43:08,369
memes, ideas, tunes, and trends strip away physical

943
00:43:08,369 --> 00:43:11,489
form and utilize our neurological networks to

944
00:43:11,489 --> 00:43:13,449
propagate themselves horizontally through the

945
00:43:13,449 --> 00:43:15,590
infosphere. Finally, we dove into the rigorous

946
00:43:15,590 --> 00:43:17,949
Soviet mathematics of Kolmogorov to discover

947
00:43:17,949 --> 00:43:20,409
that complexity and randomness are the ultimate

948
00:43:20,409 --> 00:43:22,730
measures of information, providing the theoretical

949
00:43:22,730 --> 00:43:25,130
blueprints for data compression and the statistical

950
00:43:25,130 --> 00:43:27,809
mechanics of artificial intelligence. It is exactly

951
00:43:27,809 --> 00:43:29,989
as Gleick titled his book, A Flood of History

952
00:43:29,989 --> 00:43:32,269
and Theory. The impact of this informational

953
00:43:32,269 --> 00:43:36,269
turn on human society has been absolute. We no

954
00:43:36,269 --> 00:43:39,030
longer just live in the physical world. We live

955
00:43:39,030 --> 00:43:41,630
in a quantified abstraction of it. A world made

956
00:43:41,630 --> 00:43:44,789
of bits. The book quotes the novelist David Mitchell,

957
00:43:44,929 --> 00:43:48,289
who captures the slight dizzying sense of alienation

958
00:43:48,289 --> 00:43:51,030
this paradigm shift brings. What's the quote?

959
00:43:51,210 --> 00:43:54,050
He writes, the human world is made of stories,

960
00:43:54,170 --> 00:43:57,110
not people. The people the stories use to tell

961
00:43:57,110 --> 00:44:00,150
themselves are not to be blamed. It frames humans

962
00:44:00,150 --> 00:44:03,269
as the biological hardware. While the propagating

963
00:44:03,269 --> 00:44:05,829
information acts as the software running on us.

964
00:44:05,929 --> 00:44:08,510
We are just the processors. Margaret Atwood captures

965
00:44:08,510 --> 00:44:11,250
a similar sentiment regarding how quickly we

966
00:44:11,250 --> 00:44:14,889
adapted to this new reality. She wrote, as with

967
00:44:14,889 --> 00:44:17,190
all knowledge. Once you knew it, you couldn't

968
00:44:17,190 --> 00:44:19,190
imagine how it was that you hadn't known it before.

969
00:44:19,469 --> 00:44:21,869
Like stage magic, knowledge before you knew it

970
00:44:21,869 --> 00:44:24,230
took place, before your very eyes, but you were

971
00:44:24,230 --> 00:44:26,110
looking elsewhere. We just didn't see the bit.

972
00:44:26,170 --> 00:44:28,730
The bits, the codes, the memes, the algorithms,

973
00:44:28,829 --> 00:44:31,210
they were always there. But until Shannon gave

974
00:44:31,210 --> 00:44:33,230
us the mathematics, we didn't know how to look

975
00:44:33,230 --> 00:44:36,710
at them. So as you navigate this completely data

976
00:44:36,710 --> 00:44:39,710
saturated world, what does this ultimately mean

977
00:44:39,710 --> 00:44:42,900
for you? I want to leave you with a final, highly

978
00:44:42,900 --> 00:44:45,539
provocative thought drawn directly from a philosopher

979
00:44:45,539 --> 00:44:48,840
of mine named Fred Dretzky. His ideas kind of

980
00:44:48,840 --> 00:44:51,199
echo near the end of the text. Dretzky wrote

981
00:44:51,199 --> 00:44:54,480
a profound sentence in 1981. He said, In the

982
00:44:54,480 --> 00:44:56,960
beginning, there was information. The word came

983
00:44:56,960 --> 00:44:59,070
later. In the beginning, there was information.

984
00:44:59,150 --> 00:45:01,550
He argued that the evolutionary transition to

985
00:45:01,550 --> 00:45:04,230
human consciousness was achieved solely by the

986
00:45:04,230 --> 00:45:06,909
development of biological organisms that had

987
00:45:06,909 --> 00:45:09,829
the capacity to selectively extract and exploit

988
00:45:09,829 --> 00:45:11,929
information from their environment in order to

989
00:45:11,929 --> 00:45:15,000
survive. We evolved to use data as a survival

990
00:45:15,000 --> 00:45:17,960
tool. Yes. We inherently assume an anthropocentric

991
00:45:17,960 --> 00:45:20,019
view of history. We assume that human beings

992
00:45:20,019 --> 00:45:22,500
are the main characters of the universe. We evolved,

993
00:45:22,739 --> 00:45:24,679
we became highly intelligent, and then we invented

994
00:45:24,679 --> 00:45:26,860
language, mathematics, telegraphs, and fiber

995
00:45:26,860 --> 00:45:28,900
optic networks to serve our needs. We think we

996
00:45:28,900 --> 00:45:31,719
are in charge? We treat information as our subordinate

997
00:45:31,719 --> 00:45:34,789
invention. But if we take Dawkins' theory of

998
00:45:34,789 --> 00:45:37,090
the replicating meme and combine it with the

999
00:45:37,090 --> 00:45:39,989
relentless, mathematically driven expansion of

1000
00:45:39,989 --> 00:45:42,630
the digital infosphere, we have to ask a terrifying

1001
00:45:42,630 --> 00:45:46,619
question. What if we flip that hierarchy entirely?

1002
00:45:47,260 --> 00:45:49,239
What if we aren't the main characters? What if

1003
00:45:49,239 --> 00:45:51,860
the information itself is the primary evolutionary

1004
00:45:51,860 --> 00:45:55,130
driver? It is a chilling perspective. What if

1005
00:45:55,130 --> 00:45:57,389
information achieved its own evolutionary leap

1006
00:45:57,389 --> 00:46:00,349
by selectively exploiting us? Exploiting us.

1007
00:46:00,469 --> 00:46:03,170
What if data is simply using biological organisms

1008
00:46:03,170 --> 00:46:06,849
to invent better, faster, more complex networks

1009
00:46:06,849 --> 00:46:09,789
from written language to the printing press to

1010
00:46:09,789 --> 00:46:12,150
the telegraph and finally to silicon processors

1011
00:46:12,150 --> 00:46:14,769
just to facilitate its own massive replication?

1012
00:46:15,340 --> 00:46:17,159
Are you consuming this deep dive right now because

1013
00:46:17,159 --> 00:46:19,280
you have an independent desire to learn, or is

1014
00:46:19,280 --> 00:46:21,239
the information utilizing your attention as a

1015
00:46:21,239 --> 00:46:23,760
vehicle, infecting your auditory cortex with

1016
00:46:23,760 --> 00:46:26,179
the concepts of Claude Shannon, Alan Turing,

1017
00:46:26,320 --> 00:46:28,800
and Andrei Kolmogorov, specifically so that you

1018
00:46:28,800 --> 00:46:30,659
will repeat these ideas to someone else tomorrow?

1019
00:46:30,920 --> 00:46:33,599
Spreading the virus. Thereby successfully replicating

1020
00:46:33,599 --> 00:46:36,280
the data further into the infosphere. Are you

1021
00:46:36,280 --> 00:46:38,400
processing the information, or is the information

1022
00:46:38,400 --> 00:46:41,659
processing you? It forces a profound reevaluation

1023
00:46:41,659 --> 00:46:44,380
of human autonomy in the algorithmic age. It

1024
00:46:44,380 --> 00:46:46,780
absolutely does. Something for you to mull over

1025
00:46:46,780 --> 00:46:49,139
the next time you blindly hit share on a viral

1026
00:46:49,139 --> 00:46:52,260
video or ask an AI to summarize a complex topic.

1027
00:46:52,639 --> 00:46:54,980
Thank you for streaming this deep dive into the

1028
00:46:54,980 --> 00:46:57,199
physical, mathematical, and cultural history

1029
00:46:57,199 --> 00:46:59,760
of information. Keep questioning the data around

1030
00:46:59,760 --> 00:47:00,099
you.