1
00:00:00,240 --> 00:00:03,400
Speaker 1: Now here's a highlight from Coast to Coast AM on

2
00:00:03,680 --> 00:00:05,160
Speaker 1: iHeartRadio and.

3
00:00:05,280 --> 00:00:08,160
Speaker 2: Welcome back to Coast to Coast George Noriy with Max Bennett,

4
00:00:08,200 --> 00:00:11,560
Speaker 2: author of A Brief History of Intelligence, Evolution, the AI,

5
00:00:11,600 --> 00:00:15,680
Speaker 2: and the Five breakthroughs that made our brains. What are

6
00:00:15,720 --> 00:00:17,960
Speaker 2: the breakthroughs and what are the five breakthroughs?

7
00:00:18,040 --> 00:00:18,320
Speaker 1: Max?

8
00:00:19,920 --> 00:00:27,640
Speaker 3: Yes, so I will The five in order are steering, reinforcing, simulating, mentalizing,

9
00:00:27,760 --> 00:00:30,240
Speaker 3: and speaking. So I'm going to go through each of those.

10
00:00:31,160 --> 00:00:35,920
Speaker 3: But every breakthrough represented a set of new brain modifications

11
00:00:36,280 --> 00:00:39,120
Speaker 3: that happen in evolution, and then it enabled a suite

12
00:00:39,120 --> 00:00:42,360
Speaker 3: of new intellectual faculties, and they each built on each other.

13
00:00:43,080 --> 00:00:44,400
Speaker 3: So let's go all the way back to the very

14
00:00:44,440 --> 00:00:47,560
Speaker 3: first brain, all right, and we're back in six hundred

15
00:00:47,560 --> 00:00:50,920
Speaker 3: million years ago. It's in a tiny little worm the

16
00:00:50,920 --> 00:00:54,320
Speaker 3: size of a grain of rice. And in that whole ecosystem,

17
00:00:54,840 --> 00:00:59,360
Speaker 3: most animals that were large, like cnm andes which are

18
00:00:59,360 --> 00:01:03,680
Speaker 3: like coral, didn't move, so they didn't actually pursue food.

19
00:01:03,920 --> 00:01:07,240
Speaker 3: The way they got food is they just waited for

20
00:01:07,440 --> 00:01:09,559
Speaker 3: particles to go buy them, and then they would grab

21
00:01:09,560 --> 00:01:14,039
Speaker 3: the particles out of the water. But our ancestor was

22
00:01:14,080 --> 00:01:17,320
Speaker 3: the first creatures that actually found a way to move

23
00:01:17,400 --> 00:01:21,200
Speaker 3: towards food and move away from danger. And what's so

24
00:01:21,280 --> 00:01:25,480
Speaker 3: interesting is this first creature had no eyes, it had

25
00:01:25,520 --> 00:01:29,160
Speaker 3: no ears, it had none of the standard sensory organs

26
00:01:29,200 --> 00:01:31,640
Speaker 3: that we have today, and yet it could still navigate.

27
00:01:31,959 --> 00:01:34,039
Speaker 3: And this is a big clue as to all of

28
00:01:34,040 --> 00:01:36,679
Speaker 3: the foundational things about how brains work, and the way

29
00:01:36,720 --> 00:01:40,479
Speaker 3: we know this is twofold. We can see fossil remnants

30
00:01:40,600 --> 00:01:44,520
Speaker 3: of these early creatures, and there are modern animals that

31
00:01:44,520 --> 00:01:48,520
Speaker 3: are quite similar to these early animals, like a modern nematode,

32
00:01:48,520 --> 00:01:50,880
Speaker 3: which is a really really tiny little worm that has

33
00:01:50,920 --> 00:01:53,320
Speaker 3: only about three hundred and two neurons in this entire

34
00:01:53,360 --> 00:01:56,880
Speaker 3: nervous system. So how did this thing navigate without having

35
00:01:56,920 --> 00:01:59,520
Speaker 3: any eyes or ears. Well, it was a clever trick

36
00:02:00,240 --> 00:02:03,480
Speaker 3: called steering that these creatures us to navigate. And what

37
00:02:03,560 --> 00:02:06,200
Speaker 3: happened is they had these tie little sensory neurons around

38
00:02:06,240 --> 00:02:11,560
Speaker 3: their heads, and when a smell of something good increased,

39
00:02:12,040 --> 00:02:14,600
Speaker 3: it would trigger them to steer in that direction. And

40
00:02:14,639 --> 00:02:17,480
Speaker 3: when the smell of something good, like a food smell, decreased.

41
00:02:19,280 --> 00:02:22,600
Speaker 3: The entire brain was based on this foundation of steering,

42
00:02:22,600 --> 00:02:25,600
Speaker 3: towards good things and steering away from bad things, which

43
00:02:25,639 --> 00:02:28,320
Speaker 3: means the foundation of brains is based on the idea

44
00:02:28,360 --> 00:02:31,320
Speaker 3: of classifying things in the world into good and bad.

45
00:02:32,440 --> 00:02:33,919
Speaker 3: And so that was the very first brain. There was

46
00:02:33,919 --> 00:02:36,000
Speaker 3: a whole suite of really interesting things that came with this.

47
00:02:37,040 --> 00:02:41,120
Speaker 3: In that brain, the standard use of dopamine, serotonin, and

48
00:02:41,160 --> 00:02:47,079
Speaker 3: adrenaline emerged, which has a really interesting relationship to our

49
00:02:47,120 --> 00:02:50,760
Speaker 3: modern brains. So there's a whole long journey of how

50
00:02:50,760 --> 00:02:54,000
Speaker 3: we've tried to understand what dopamine does, for example, and

51
00:02:54,080 --> 00:02:57,280
Speaker 3: so for a long time people thought dopamine was the

52
00:02:57,639 --> 00:03:00,880
Speaker 3: pleasure chemical, and has been a lot of research that

53
00:03:00,960 --> 00:03:04,200
Speaker 3: shows if you stimulate dopamine neurons in a rat, for example,

54
00:03:04,840 --> 00:03:07,200
Speaker 3: every time it pushes the lever, the rat will endlessly

55
00:03:07,240 --> 00:03:10,079
Speaker 3: push the lever. And that makes people think, oh, it

56
00:03:10,120 --> 00:03:11,320
Speaker 3: must be the pleasure chemical.

57
00:03:11,480 --> 00:03:13,920
Speaker 2: They want it, they want more, Yeah.

58
00:03:13,440 --> 00:03:16,800
Speaker 3: They want more. Right, It turns out this guy Ken

59
00:03:16,880 --> 00:03:20,760
Speaker 3: Berage did some incredible study that revealed that dopamine is

60
00:03:20,760 --> 00:03:23,400
Speaker 3: actually not pleasure and this goes all the way back

61
00:03:23,440 --> 00:03:26,080
Speaker 3: to the first prance. But his experiments were the following

62
00:03:26,720 --> 00:03:30,359
Speaker 3: He said, okay, well, we actually know that rats make

63
00:03:30,520 --> 00:03:34,880
Speaker 3: very specific facial expressions when they enjoy food. They smack

64
00:03:34,960 --> 00:03:37,600
Speaker 3: their lips and they make a gaping face when they

65
00:03:37,640 --> 00:03:41,280
Speaker 3: don't enjoy food. And so if we give these rats

66
00:03:41,320 --> 00:03:44,560
Speaker 3: a ton of dopamine, do they actually show more pleasurable

67
00:03:44,640 --> 00:03:47,040
Speaker 3: lip smacks? And the answer is no, they will eat

68
00:03:47,080 --> 00:03:48,880
Speaker 3: a lot more, but if anything, it looks like they

69
00:03:48,920 --> 00:03:53,280
Speaker 3: don't like the food anymore. And in fact, if you

70
00:03:53,360 --> 00:03:56,080
Speaker 3: remove dopamine so they have no dopamine in their brain,

71
00:03:56,360 --> 00:03:59,040
Speaker 3: they'll stop eating. But if you put food in their mouths,

72
00:03:59,200 --> 00:04:01,960
Speaker 3: they'll enjoy the food. That's what does this mean. It

73
00:04:01,960 --> 00:04:06,080
Speaker 3: means dopamine is not about pleasure. It's about craving. Dopamine

74
00:04:06,080 --> 00:04:08,960
Speaker 3: is about pursuing things that we want. It's not about

75
00:04:08,960 --> 00:04:11,800
Speaker 3: the actual enjoyment of things when we get it. Let's

76
00:04:11,840 --> 00:04:14,720
Speaker 3: go all the way back to a nematode. Now, what

77
00:04:14,840 --> 00:04:17,800
Speaker 3: are the dopamine neurons In a nematode. There are little

78
00:04:17,880 --> 00:04:20,200
Speaker 3: neurons that stick out of their head that detect the

79
00:04:20,240 --> 00:04:23,720
Speaker 3: presence of food around the animal. And then there's another

80
00:04:23,760 --> 00:04:26,720
Speaker 3: set of neurons and the nemotod's throat that are serotonin

81
00:04:26,760 --> 00:04:30,680
Speaker 3: neurons that detect when it's actually eating food. So in

82
00:04:30,720 --> 00:04:33,480
Speaker 3: the very first brain, dopamine was the signal for good

83
00:04:33,560 --> 00:04:37,120
Speaker 3: things are nearby, and serotonin was the signal for good

84
00:04:37,120 --> 00:04:40,640
Speaker 3: things are actually happening. And that basic template between these

85
00:04:40,680 --> 00:04:44,560
Speaker 3: two neuromodulators exists all the way throughout the animal kingdom,

86
00:04:44,839 --> 00:04:48,719
Speaker 3: and it's served a whole purpose for steering because and

87
00:04:48,800 --> 00:04:50,159
Speaker 3: you see this in a roomba. I don't know if

88
00:04:50,200 --> 00:04:54,719
Speaker 3: you know what a rumba is, the little robot vacum, right,

89
00:04:55,080 --> 00:04:57,840
Speaker 3: So rumba's work very similarly to the first frame. So

90
00:04:57,880 --> 00:05:01,120
Speaker 3: there's something called dirt detect in a roomba. When a

91
00:05:01,600 --> 00:05:04,800
Speaker 3: room but passes a patch of dirt, it assumes that

92
00:05:04,800 --> 00:05:07,360
Speaker 3: it must be nearby patches of dirt, so it changes

93
00:05:07,400 --> 00:05:09,760
Speaker 3: its repertoire. So if it passes a little piece of dirt,

94
00:05:09,800 --> 00:05:13,080
Speaker 3: it starts turning randomly in the local area because it

95
00:05:13,120 --> 00:05:15,320
Speaker 3: predicts that if there's a little dirt in one location,

96
00:05:15,600 --> 00:05:18,800
Speaker 3: there's probably dirt nearby. This is exactly what dopamine did

97
00:05:19,040 --> 00:05:22,120
Speaker 3: and does nematode brains. If it detects a little bit

98
00:05:22,160 --> 00:05:25,400
Speaker 3: of food, it floods the brain with dopamine, which triggers

99
00:05:25,400 --> 00:05:28,479
Speaker 3: it into an emotional state of what could be called

100
00:05:28,520 --> 00:05:31,560
Speaker 3: exploitation when they start turning around looking for food for

101
00:05:31,560 --> 00:05:33,480
Speaker 3: a while, so dopamine levels drop and then it keeps

102
00:05:33,560 --> 00:05:38,000
Speaker 3: going elsewhere. So dopamine is this initial seeking chemical. Serotonin

103
00:05:38,080 --> 00:05:40,839
Speaker 3: is the statiation chemical, it makes them stop and rest,

104
00:05:41,520 --> 00:05:44,919
Speaker 3: an adrenaline is this big fear chemical. So the basic

105
00:05:45,000 --> 00:05:50,039
Speaker 3: template of emotion and stress and categorizing the world into

106
00:05:50,040 --> 00:05:53,080
Speaker 3: good and bad with all present in the steering brain

107
00:05:53,480 --> 00:05:58,120
Speaker 3: of the very first animal that was break through one.

108
00:05:56,960 --> 00:05:59,919
Speaker 2: Number two, I can keep going ye five.

109
00:06:02,120 --> 00:06:07,160
Speaker 3: So moving forward fifty million years into the Cambrian period,

110
00:06:08,160 --> 00:06:11,640
Speaker 3: our ancestors evolved into a creature that was most similar

111
00:06:12,320 --> 00:06:16,680
Speaker 3: to a modern fish, which were the first vertebrates, and

112
00:06:17,560 --> 00:06:21,080
Speaker 3: they now had eyes lends shaped eyes, so they looked

113
00:06:21,120 --> 00:06:23,720
Speaker 3: very similar to modern fish. They had ears, they had

114
00:06:23,720 --> 00:06:27,960
Speaker 3: a vestibular system, meaning they could detect direction, and so

115
00:06:29,320 --> 00:06:31,120
Speaker 3: you know, they had the basic template of what we

116
00:06:31,200 --> 00:06:34,720
Speaker 3: think about when we look at a fish. And most

117
00:06:34,760 --> 00:06:38,400
Speaker 3: of the creatures in the Cambrian period were arthropods, so

118
00:06:38,480 --> 00:06:42,440
Speaker 3: they were huge insect like creatures, and our ancestors were

119
00:06:42,520 --> 00:06:45,320
Speaker 3: very humble, tiny maybe one to two inch long fish

120
00:06:45,360 --> 00:06:48,159
Speaker 3: in this whole world, but if you looked in their brain,

121
00:06:49,080 --> 00:06:51,800
Speaker 3: you would see a brain that has most of the

122
00:06:51,839 --> 00:06:54,120
Speaker 3: same structures as a human brain, which is exactly what

123
00:06:54,120 --> 00:06:56,240
Speaker 3: we see when we look at modern fish, which is

124
00:06:56,360 --> 00:06:59,080
Speaker 3: kind of mind blowing to realize that even in a

125
00:06:59,120 --> 00:07:01,800
Speaker 3: tiny fish brain, you can see most of the same

126
00:07:01,839 --> 00:07:04,360
Speaker 3: brain structures as a human brain. And so the question

127
00:07:04,360 --> 00:07:06,280
Speaker 3: is what does this brain do. Well, it's a really

128
00:07:06,279 --> 00:07:11,800
Speaker 3: cool story that is connected to AI where it goes

129
00:07:11,800 --> 00:07:15,080
Speaker 3: a little bit back to dopamine, where I'm gonna tell

130
00:07:15,080 --> 00:07:16,240
Speaker 3: a little bit of a story of AI, and then

131
00:07:16,280 --> 00:07:20,120
Speaker 3: I'll connect it back to brands. So in the nineteen fifties,

132
00:07:20,720 --> 00:07:23,960
Speaker 3: people have this grand idea of what if we could

133
00:07:24,000 --> 00:07:28,040
Speaker 3: train a neural network with reinforcement learning, In other words,

134
00:07:28,080 --> 00:07:30,880
Speaker 3: we just tell it when it does something right, and

135
00:07:30,920 --> 00:07:33,680
Speaker 3: then over time I should learn the right decisions to

136
00:07:34,040 --> 00:07:36,160
Speaker 3: keep doing the thing that quote unquote is right. And

137
00:07:36,160 --> 00:07:38,480
Speaker 3: perhaps we could teach it to play chess if we

138
00:07:38,560 --> 00:07:41,080
Speaker 3: just say, hey, AI system, play human in chess, and

139
00:07:41,120 --> 00:07:44,120
Speaker 3: when you win, I'll give you a reinforcer. When you lose,

140
00:07:44,160 --> 00:07:46,800
Speaker 3: I'll punish the system, and if you do it enough times,

141
00:07:46,880 --> 00:07:49,720
Speaker 3: eventually should get good of chess. And this guy, Marvin

142
00:07:49,760 --> 00:07:52,920
Speaker 3: Minsky tried this, and it didn't work. And the reason

143
00:07:52,920 --> 00:07:55,360
Speaker 3: it didn't work is because of something called the credit

144
00:07:55,400 --> 00:07:59,160
Speaker 3: assignment problem. When you win a game of chess, which

145
00:07:59,280 --> 00:08:03,280
Speaker 3: previous moves should get credit for being good? You don't

146
00:08:03,320 --> 00:08:05,240
Speaker 3: know because it could have been the first moves that

147
00:08:05,280 --> 00:08:07,480
Speaker 3: were really good. It could have been moved in the middle,

148
00:08:08,160 --> 00:08:10,280
Speaker 3: it could have been moved at the very end. And

149
00:08:10,360 --> 00:08:13,640
Speaker 3: so when you get a reward when you win a game,

150
00:08:14,120 --> 00:08:17,520
Speaker 3: it's actually hard to know what actions should actually be reinforced.

151
00:08:17,960 --> 00:08:20,560
Speaker 3: So this guy, Richard Sutton, came up with this really

152
00:08:20,600 --> 00:08:24,440
Speaker 3: clever strategy, which is how all modern reinforcement learning systems work.

153
00:08:24,720 --> 00:08:27,120
Speaker 3: In your self driving car, that's using the exact same

154
00:08:27,160 --> 00:08:31,280
Speaker 3: technique he discovered, which is then called temporal difference learning,

155
00:08:31,960 --> 00:08:35,400
Speaker 3: where the reinforcer is actually when you think the world

156
00:08:35,480 --> 00:08:38,559
Speaker 3: just got better. In other words, when you're playing a

157
00:08:38,600 --> 00:08:41,440
Speaker 3: game of chess and you do a move and you realize, oh,

158
00:08:41,480 --> 00:08:44,160
Speaker 3: the thing I just did increased my likelihood of winning

159
00:08:44,200 --> 00:08:47,400
Speaker 3: by a lot, that's when you get reinforced. And so

160
00:08:47,440 --> 00:08:51,079
Speaker 3: when we look in fish brains, we actually see exactly

161
00:08:51,200 --> 00:08:55,199
Speaker 3: this algorithm being implemented, which is it has an expectation

162
00:08:55,320 --> 00:08:59,319
Speaker 3: of future rewards and it's reinforcing itself again with dopamine

163
00:08:59,480 --> 00:09:01,640
Speaker 3: when things seemed to be doing better. This all happens

164
00:09:01,640 --> 00:09:04,720
Speaker 3: in the structure called the basal ganglia. So breakthrough two

165
00:09:04,840 --> 00:09:08,040
Speaker 3: is reinforcement learning. It was this ability to learn through

166
00:09:08,080 --> 00:09:10,360
Speaker 3: trial and error, and we see this in fish all

167
00:09:10,400 --> 00:09:12,520
Speaker 3: the time. You can teach a fish to do very

168
00:09:12,600 --> 00:09:15,000
Speaker 3: very clever things through error. You can teach fish to

169
00:09:15,080 --> 00:09:20,480
Speaker 3: jump through hoops. You can teach fish to remember random

170
00:09:20,520 --> 00:09:22,559
Speaker 3: shapes of animals and they'll run away from them or

171
00:09:22,559 --> 00:09:25,719
Speaker 3: they run towards them. So reinforcement learning was break through two.

172
00:09:26,520 --> 00:09:29,280
Speaker 3: We go forward a bunch further, all the way to

173
00:09:29,280 --> 00:09:34,280
Speaker 3: about one hundred million years ago. Our ancestors were early

174
00:09:34,320 --> 00:09:38,720
Speaker 3: mammals and they resemble most resembled in early creature that

175
00:09:38,720 --> 00:09:42,000
Speaker 3: would look like today's squirrels, and they lived in a

176
00:09:42,040 --> 00:09:45,480
Speaker 3: world surrounded by dinosaurs. They were very low on the

177
00:09:45,480 --> 00:09:48,679
Speaker 3: totem pole. They were hiding in burrows. The world of

178
00:09:49,240 --> 00:09:54,840
Speaker 3: sinus our life and our little mammal ancestors evolved a

179
00:09:54,840 --> 00:09:58,920
Speaker 3: new brain region called the neocortex, and the neocortext gave

180
00:09:58,960 --> 00:10:02,120
Speaker 3: them a very remarkable ability, which is to imagine the

181
00:10:02,120 --> 00:10:05,040
Speaker 3: world as it is not in other words, to simulate things,

182
00:10:05,640 --> 00:10:11,040
Speaker 3: so a mammal can imagine futures and remember paths and

183
00:10:11,120 --> 00:10:14,880
Speaker 3: can consider alternative decisions. You see this in rast today.

184
00:10:15,040 --> 00:10:17,880
Speaker 3: So if you put a ratinum aze and you record

185
00:10:17,960 --> 00:10:22,280
Speaker 3: neurons and a hippocampus which has these things called place cells,

186
00:10:22,520 --> 00:10:26,280
Speaker 3: you can literally wash a mouse imagine going down different paths,

187
00:10:26,679 --> 00:10:29,080
Speaker 3: and you can watch it imagining eating one type of

188
00:10:29,080 --> 00:10:32,880
Speaker 3: treat versus another before it makes a decision, which is unbelievable.

189
00:10:32,920 --> 00:10:34,800
Speaker 3: We can literally go into mouse brains and see them

190
00:10:34,800 --> 00:10:39,120
Speaker 3: imagining things and so simulating. It's this incredible ability to

191
00:10:39,200 --> 00:10:41,880
Speaker 3: not only learn by actual trial and error, but to

192
00:10:42,280 --> 00:10:44,800
Speaker 3: learn by vicarious trial and error. In other words, I

193
00:10:44,840 --> 00:10:47,280
Speaker 3: can learn before I take an action. I can imagine

194
00:10:47,280 --> 00:10:49,960
Speaker 3: doing things and then decides to them. It's an incredible

195
00:10:50,000 --> 00:10:52,280
Speaker 3: break forward, and this is one thing that's really missing

196
00:10:52,280 --> 00:10:57,040
Speaker 3: in aisystems, which has important consequences. Break Through four emerge

197
00:10:57,040 --> 00:11:00,840
Speaker 3: in early primates something called mentalizing, which is the ability

198
00:11:00,880 --> 00:11:04,520
Speaker 3: to think about thinking, and this is another key thing

199
00:11:04,559 --> 00:11:08,760
Speaker 3: missing in AI systems. So if you go into a primate,

200
00:11:08,880 --> 00:11:13,480
Speaker 3: there's some cool stories where primates trick each other. So

201
00:11:13,880 --> 00:11:19,680
Speaker 3: one famous one is this guy Mlo Menzel was trying

202
00:11:19,679 --> 00:11:23,400
Speaker 3: to teach a chimpanzee to find locations of food. And

203
00:11:23,440 --> 00:11:26,079
Speaker 3: you would place a little piece of a treat under

204
00:11:26,120 --> 00:11:29,320
Speaker 3: a rock or in a bush until and show one

205
00:11:29,400 --> 00:11:32,480
Speaker 3: champanzee in this group where to find it, and that

206
00:11:32,559 --> 00:11:35,840
Speaker 3: chimpanzee would very easily find the tree. But then he

207
00:11:35,920 --> 00:11:38,600
Speaker 3: noticed something that he did not expect to find, which

208
00:11:38,640 --> 00:11:42,040
Speaker 3: is when Belle, one of the chimpanzees, ran to the tree,

209
00:11:42,120 --> 00:11:44,079
Speaker 3: she would share it with other chimps in the group.

210
00:11:44,760 --> 00:11:47,200
Speaker 3: But then what's started happening is Rock, the top chimp,

211
00:11:47,679 --> 00:11:49,880
Speaker 3: would take the treat from her for himself. He was

212
00:11:49,920 --> 00:11:54,600
Speaker 3: kind of a jerk. So what Belle started doing is

213
00:11:54,679 --> 00:11:57,880
Speaker 3: she wouldn't go to the tree until Rock was not looking,

214
00:11:58,640 --> 00:12:01,480
Speaker 3: so she was trying to vent him from seeing it.

215
00:12:01,520 --> 00:12:06,160
Speaker 3: But then Rock, in exchange, began looking away, pretending he

216
00:12:06,240 --> 00:12:08,960
Speaker 3: wasn't paying attention until Bell would go to the tree.

217
00:12:09,000 --> 00:12:11,240
Speaker 3: Then he turned around and run towards it. So then

218
00:12:11,320 --> 00:12:14,160
Speaker 3: Bell started trying to lead Rock in the wrong direction

219
00:12:14,520 --> 00:12:17,160
Speaker 3: and then run back. So the cycle of deception and

220
00:12:17,240 --> 00:12:22,400
Speaker 3: counter deception is really amazing evidence that chimpanzees and other

221
00:12:22,440 --> 00:12:25,240
Speaker 3: primates we see the same thing can actually think about

222
00:12:25,240 --> 00:12:28,160
Speaker 3: each other's thinking, because in order to do that, she

223
00:12:28,280 --> 00:12:31,199
Speaker 3: needs to understand well, I can trick rock into thinking

224
00:12:31,320 --> 00:12:33,360
Speaker 3: the rock is not here or the treat is not

225
00:12:33,440 --> 00:12:35,599
Speaker 3: here if I bring him in the other direction or

226
00:12:35,640 --> 00:12:40,280
Speaker 3: if I look away. So mentalizing is this really amazing

227
00:12:40,320 --> 00:12:43,079
Speaker 3: feat that emerge in primates. It comes from variety of

228
00:12:43,080 --> 00:12:47,720
Speaker 3: brain structures like the granular prefrontal cortex, and that's a

229
00:12:47,800 --> 00:12:50,160
Speaker 3: really crucial thing missing in the A systems. For reasons,

230
00:12:50,160 --> 00:12:53,400
Speaker 3: I can go into and then break through five with

231
00:12:53,480 --> 00:12:58,480
Speaker 3: early humans was language. And language is only really possible

232
00:12:58,520 --> 00:13:02,240
Speaker 3: because the mentalizing, because when we're talking, when we talk

233
00:13:02,240 --> 00:13:04,520
Speaker 3: to each other, we're always inferring what each other mean

234
00:13:05,720 --> 00:13:08,040
Speaker 3: when we speak to each other. So this is a

235
00:13:08,080 --> 00:13:11,120
Speaker 3: big problem with AI, where when we say things to them,

236
00:13:11,160 --> 00:13:14,080
Speaker 3: if they don't know what we mean, there's potentially dangerous

237
00:13:14,080 --> 00:13:16,320
Speaker 3: consequences which I can talk about. So anyway, those are

238
00:13:16,320 --> 00:13:18,880
Speaker 3: the five breakthroughs very fast. I can go into any

239
00:13:18,880 --> 00:13:22,960
Speaker 3: of them in much more detail. But steering, reinforcing, simulating, mentalizing,

240
00:13:23,160 --> 00:13:23,800
Speaker 3: and then speaking.

241
00:13:24,040 --> 00:13:28,440
Speaker 2: Do other creatures have a language? I've long suspected that

242
00:13:28,480 --> 00:13:31,800
Speaker 2: some animals or insects have an ability to communicate with

243
00:13:31,840 --> 00:13:34,920
Speaker 2: each other, But is it a language or something else?

244
00:13:36,360 --> 00:13:40,200
Speaker 3: Very very good question. So all creatures communicate with each other.

245
00:13:40,360 --> 00:13:45,120
Speaker 3: I mean even single cell bacteria sends chemical signals to

246
00:13:45,200 --> 00:13:47,760
Speaker 3: each other and will share genes when they need to

247
00:13:47,800 --> 00:13:50,800
Speaker 3: share genes to help each other out. And we know

248
00:13:50,920 --> 00:13:53,960
Speaker 3: for a fact that monkeys signal each other. They have

249
00:13:54,120 --> 00:13:57,000
Speaker 3: Some monkeys have certain signals that mean eagle, which make

250
00:13:57,080 --> 00:13:59,480
Speaker 3: all of them jump to the forest floor. They have

251
00:13:59,520 --> 00:14:01,720
Speaker 3: other signal notes that means snake that leads them all

252
00:14:01,720 --> 00:14:04,880
Speaker 3: to run to the tree tops. So we know there

253
00:14:04,880 --> 00:14:09,040
Speaker 3: are signals throughout the animal kingdom. But language is unique

254
00:14:09,280 --> 00:14:14,440
Speaker 3: on two fronts. Language contains what's called declarative labels. And

255
00:14:14,520 --> 00:14:17,400
Speaker 3: the declarative label is saying you can point to an

256
00:14:17,440 --> 00:14:20,640
Speaker 3: object and say that is a cup, and now on

257
00:14:20,760 --> 00:14:23,280
Speaker 3: your head, you know the thing is called cup, an

258
00:14:23,360 --> 00:14:27,400
Speaker 3: arbitrary sort of symbol. And then we have grammar. That's

259
00:14:27,400 --> 00:14:30,560
Speaker 3: the second thing. I can start assembling these symbols into

260
00:14:30,680 --> 00:14:34,280
Speaker 3: novel phrases, like the words I'm saying right now. And

261
00:14:34,320 --> 00:14:38,800
Speaker 3: it's really controversial whether any other animals do that, which

262
00:14:38,880 --> 00:14:42,440
Speaker 3: is assigned symbols and phrase them into novel ways. There's

263
00:14:42,480 --> 00:14:45,720
Speaker 3: been many studies trying to get chimpanzees and bonobo's to

264
00:14:45,760 --> 00:14:50,520
Speaker 3: do this with pretty lackluster results. So you know, there's

265
00:14:50,520 --> 00:14:53,000
Speaker 3: still emerging studies and other animals. But for the most part,

266
00:14:53,040 --> 00:14:55,440
Speaker 3: it seems like language, at least the extent to which

267
00:14:55,480 --> 00:14:58,280
Speaker 3: we use it, as pretty unique to humans.

268
00:14:58,320 --> 00:15:01,440
Speaker 2: Pretty dramatic, isn't it amazing?

269
00:15:01,720 --> 00:15:01,880
Speaker 1: Oh?

270
00:15:02,000 --> 00:15:05,000
Speaker 3: The language was mind blowing. Language was the first time

271
00:15:05,720 --> 00:15:09,320
Speaker 3: when we could learn from each other's imaginations. Think about

272
00:15:09,320 --> 00:15:12,720
Speaker 3: how powerful that is. I can go before language. If

273
00:15:12,760 --> 00:15:16,200
Speaker 3: someone in our in our primate group ran across a

274
00:15:17,600 --> 00:15:20,440
Speaker 3: you know, the forest, to a different location and saw

275
00:15:20,520 --> 00:15:24,560
Speaker 3: that a red snake fiting them was really bad and

276
00:15:24,600 --> 00:15:27,840
Speaker 3: it's really dangerous, they could never share that information that, hey,

277
00:15:27,880 --> 00:15:31,360
Speaker 3: avoid red snakes. It was impossible. But with language, I

278
00:15:31,400 --> 00:15:33,360
Speaker 3: can share, Hey, this thing happened in the past, and

279
00:15:33,400 --> 00:15:35,480
Speaker 3: now I'm going to give you that information about it.

280
00:15:35,920 --> 00:15:38,240
Speaker 3: Or together we can plan a hunt. I can say, hey,

281
00:15:38,320 --> 00:15:41,040
Speaker 3: let's go to this location. You go left, I go right.

282
00:15:41,200 --> 00:15:43,800
Speaker 3: When I whistle three times, we're going to start hunting together.

283
00:15:44,200 --> 00:15:46,280
Speaker 3: And all that type of coordination was not possible with

284
00:15:46,320 --> 00:15:47,920
Speaker 3: that language. Unbelievably powerful.

285
00:15:48,280 --> 00:15:51,560
Speaker 1: Listen to more Coast to coast am every weeknight at

286
00:15:51,560 --> 00:15:54,520
Speaker 1: one a m. Eastern and go to Coast to coastam

287
00:15:54,560 --> 00:15:55,640
Speaker 1: dot com for more