1
00:00:00,240 --> 00:00:05,600
Speaker 1: Now here's a highlight from Coast to Coast AM on iHeartRadio.

2
00:00:05,000 --> 00:00:08,520
Speaker 2: And welcome back George Doria, doctor Bart costco Bart. Seems

3
00:00:08,560 --> 00:00:12,639
Speaker 2: like the world has gone crazy with artificial intelligence chat systems.

4
00:00:12,960 --> 00:00:17,000
Speaker 3: What's going on here? Sure? Yeah, sure has Well, they've

5
00:00:17,040 --> 00:00:19,680
Speaker 3: gotten more powerful. They're now up to more than a

6
00:00:19,840 --> 00:00:25,280
Speaker 3: trillion parameters in their brain, in their synapses, and they're wiring.

7
00:00:25,920 --> 00:00:28,200
Speaker 3: And that's a good thing in terms of applications are

8
00:00:28,200 --> 00:00:31,000
Speaker 3: more powerful. It's a bad thing because all the inherent

9
00:00:31,080 --> 00:00:35,680
Speaker 3: problems in these key forward black boxes we call them,

10
00:00:36,080 --> 00:00:39,919
Speaker 3: simply get worse. The black boxes get bigger and blacker,

11
00:00:40,600 --> 00:00:42,880
Speaker 3: and they hallucinate and they do a lot of other things.

12
00:00:43,200 --> 00:00:45,960
Speaker 3: But I think, George, it's because of the language interface.

13
00:00:45,960 --> 00:00:48,240
Speaker 3: I mean, the real advances in AI that are ongoing

14
00:00:48,320 --> 00:00:50,599
Speaker 3: and will continue to be for a long time are

15
00:00:50,680 --> 00:00:54,320
Speaker 3: not so much on the language front, but the tie

16
00:00:54,440 --> 00:00:56,760
Speaker 3: end to the language. It's a front end with words,

17
00:00:56,800 --> 00:00:59,440
Speaker 3: it's a back end with words and in between. It's

18
00:01:00,040 --> 00:01:05,200
Speaker 3: straightforward nineteen eighties nineteen nineties processing, but greatly speeded up

19
00:01:05,520 --> 00:01:09,160
Speaker 3: with just a lot of parameters added now courtesy of

20
00:01:09,280 --> 00:01:12,600
Speaker 3: chips like from Nvidia and elsewhere. Chips that were really

21
00:01:12,760 --> 00:01:17,360
Speaker 3: initially designed for imaging and for graphics and game playing

22
00:01:17,360 --> 00:01:19,800
Speaker 3: and that sort of thing. But the other types of

23
00:01:19,840 --> 00:01:22,240
Speaker 3: AI that will continue to be out there, applications of

24
00:01:22,319 --> 00:01:26,160
Speaker 3: data and statistics, I think a little safer. You may

25
00:01:26,200 --> 00:01:28,679
Speaker 3: not realize it for the better and for the worse,

26
00:01:29,040 --> 00:01:32,800
Speaker 3: medical applications and just about anywhere you use data, because

27
00:01:32,800 --> 00:01:35,919
Speaker 3: that's essentially what we mean in this context by AI,

28
00:01:36,040 --> 00:01:38,440
Speaker 3: access to a lot of data. But that language front

29
00:01:38,560 --> 00:01:42,280
Speaker 3: end has always been a big deal, and it's a

30
00:01:42,280 --> 00:01:45,560
Speaker 3: little bit illsory. You know, when someone uses language to

31
00:01:45,600 --> 00:01:49,200
Speaker 3: speak to you, you impute to them human like characteristics,

32
00:01:49,240 --> 00:01:52,040
Speaker 3: Like when you watch a Disney cartoon, you think that

33
00:01:52,200 --> 00:01:54,840
Speaker 3: Vambi is like a person. So we have that effect.

34
00:01:54,880 --> 00:01:57,520
Speaker 3: But I can assure you there is no person, no

35
00:01:57,560 --> 00:01:59,840
Speaker 3: homunculous inside those big black boxes.

36
00:02:01,560 --> 00:02:05,720
Speaker 2: When you coin the phrase fuzzy logic, what does that mean?

37
00:02:05,840 --> 00:02:08,240
Speaker 3: I want to clarify what my PHA advisors. The late

38
00:02:08,240 --> 00:02:12,360
Speaker 3: Great loss the Auto coined that phrase in nineteen sixty five.

39
00:02:12,520 --> 00:02:14,440
Speaker 3: He was subject by the way of a Google Doodle.

40
00:02:14,440 --> 00:02:19,480
Speaker 3: He died in twenty seventeen. Fuzzy logic means thinking in

41
00:02:19,560 --> 00:02:23,919
Speaker 3: shades of gray. It's maybe not the best adjective fuzzy.

42
00:02:24,280 --> 00:02:27,919
Speaker 3: Sometimes the word is vague, that's where the boundaries are

43
00:02:27,919 --> 00:02:30,320
Speaker 3: not black and white, like the name of the series

44
00:02:30,440 --> 00:02:34,799
Speaker 3: Twilight Zone refers to that fuzzy area of twilight between

45
00:02:34,960 --> 00:02:38,560
Speaker 3: night and day, or the rose that's both pink cannot peak.

46
00:02:39,040 --> 00:02:42,000
Speaker 3: And if you look at most human concepts, from cool

47
00:02:42,160 --> 00:02:45,680
Speaker 3: air to green grass and those kinds of things, rarely

48
00:02:45,720 --> 00:02:47,480
Speaker 3: do you get something that you could really say is

49
00:02:47,480 --> 00:02:50,440
Speaker 3: one hundred percent one way or the other. But classical

50
00:02:50,480 --> 00:02:55,120
Speaker 3: binary logic, the modern mathematics from one hundred plus years ago, anyway,

51
00:02:55,440 --> 00:02:58,720
Speaker 3: is based on assume that ran with that the first approximation.

52
00:02:59,200 --> 00:03:01,520
Speaker 3: But that's not how think, and that's not how we talked,

53
00:03:01,840 --> 00:03:04,119
Speaker 3: it's not how we reason. When you park your car

54
00:03:04,200 --> 00:03:06,520
Speaker 3: in a parking lot, in a parking space, you don't

55
00:03:06,520 --> 00:03:09,200
Speaker 3: have to get it exactly precisely right in a little

56
00:03:09,240 --> 00:03:14,080
Speaker 3: rectangular slot. It's approximate. It's fuzzy. And we came to

57
00:03:14,160 --> 00:03:17,880
Speaker 3: learn in the sixties, seventies, eighties, and nineties that demanding

58
00:03:18,200 --> 00:03:21,720
Speaker 3: that extra precision was just too costly and then necessary.

59
00:03:21,800 --> 00:03:25,440
Speaker 3: So what happened with fuzzy logic when it became popular,

60
00:03:25,480 --> 00:03:29,080
Speaker 3: I got involved in it. Forty plus years ago and

61
00:03:29,320 --> 00:03:33,720
Speaker 3: the Japanese started doing applications. Small skill are actually the

62
00:03:33,800 --> 00:03:38,520
Speaker 3: first commercial applications that we have and you probably own one.

63
00:03:38,600 --> 00:03:40,760
Speaker 3: Like I have a subru outback car. It has a

64
00:03:40,760 --> 00:03:44,480
Speaker 3: fuzzy transmission to smoothly shift the gears and soon in

65
00:03:44,520 --> 00:03:46,440
Speaker 3: my class this week told me he had a neural

66
00:03:46,480 --> 00:03:51,880
Speaker 3: fuzzy lights cooker. But thousands of products, largely commercial products

67
00:03:51,880 --> 00:03:55,960
Speaker 3: out of Japan and South Korea, commercial electronics type products,

68
00:03:55,960 --> 00:03:58,560
Speaker 3: and many many other systems are under control of fuzzy

69
00:03:58,600 --> 00:04:01,840
Speaker 3: logic because it's not complete black box like the neural

70
00:04:01,840 --> 00:04:04,920
Speaker 3: networks you're referring to we call AI today and not

71
00:04:05,080 --> 00:04:07,600
Speaker 3: like the old black white decision trees that we called

72
00:04:07,640 --> 00:04:10,520
Speaker 3: AI for about thirty years. It's sort of a compromise.

73
00:04:10,600 --> 00:04:14,000
Speaker 3: It's sort of a gray box reasoning with shades of gray,

74
00:04:14,560 --> 00:04:17,000
Speaker 3: and it has only grown over time. But the essential

75
00:04:17,040 --> 00:04:21,279
Speaker 3: difference between fuzzy Logic and old AI is old AI

76
00:04:21,480 --> 00:04:24,480
Speaker 3: in effect try to get people to think or model

77
00:04:24,520 --> 00:04:28,039
Speaker 3: how people think, like on off binary switches work. Where

78
00:04:28,160 --> 00:04:31,240
Speaker 3: fuzzy Logic tries to get and is getting computers to

79
00:04:31,279 --> 00:04:32,520
Speaker 3: think more like people think.

80
00:04:33,200 --> 00:04:35,719
Speaker 2: Why does so many people bart seem to be afraid

81
00:04:36,400 --> 00:04:36,920
Speaker 2: of AI.

82
00:04:38,680 --> 00:04:40,359
Speaker 3: You know, there has been a lot of hype, and

83
00:04:40,520 --> 00:04:43,159
Speaker 3: I think, just like when you watch that Disney movie

84
00:04:43,200 --> 00:04:46,799
Speaker 3: whatever happens to be, you impute two it human aspects.

85
00:04:47,320 --> 00:04:49,840
Speaker 3: When I ask people to mean by AI number one,

86
00:04:50,120 --> 00:04:53,800
Speaker 3: they don't know what the adjective artificial stands for. And

87
00:04:54,080 --> 00:04:58,039
Speaker 3: it was something the late great AI expert John McCarthy

88
00:04:58,160 --> 00:05:00,480
Speaker 3: just sort of threw out in the late fifties the name,

89
00:05:01,160 --> 00:05:04,240
Speaker 3: and so that's a problem. And intelligence they don't know

90
00:05:04,279 --> 00:05:06,560
Speaker 3: what that stands for. Just a real quick point. There

91
00:05:06,600 --> 00:05:10,120
Speaker 3: is a split between the computer science folks in the

92
00:05:10,400 --> 00:05:14,720
Speaker 3: engineering folks. Actual engineering folks like me. We have renamed

93
00:05:14,760 --> 00:05:17,920
Speaker 3: that to distinguish at CI may not be as catchy

94
00:05:18,360 --> 00:05:23,440
Speaker 3: for computational intelligence. By intelligence we mean behavior that tends

95
00:05:23,480 --> 00:05:27,080
Speaker 3: to increase or decrease a performance measure, like the accuracy

96
00:05:27,080 --> 00:05:29,960
Speaker 3: of a test or something like that. And by computational

97
00:05:30,040 --> 00:05:32,479
Speaker 3: we mean something on a computer. So what do you

98
00:05:32,480 --> 00:05:34,760
Speaker 3: fill in with if you don't know what the adjective

99
00:05:34,839 --> 00:05:37,640
Speaker 3: is the artificial and the noun you fill in with

100
00:05:38,120 --> 00:05:39,920
Speaker 3: what you remember of a lot of good and bad

101
00:05:40,000 --> 00:05:42,320
Speaker 3: old science fiction Movies'm a great one like how the

102
00:05:42,760 --> 00:05:45,760
Speaker 3: computer in two thousand and one of Space, Odyssey to

103
00:05:45,960 --> 00:05:48,600
Speaker 3: the Terminator fantasies and so on, And there's a lot

104
00:05:48,600 --> 00:05:52,080
Speaker 3: of the nature of two hour dramas like that is

105
00:05:52,080 --> 00:05:54,600
Speaker 3: to have a scary problem and then fix it at

106
00:05:54,600 --> 00:05:57,039
Speaker 3: the end. So a lot of the sense of the

107
00:05:57,200 --> 00:06:01,120
Speaker 3: overtaking of the robots and so forth, like the Terminator movies.

108
00:06:01,400 --> 00:06:03,479
Speaker 3: And they're like, I mean, there's some risk of that,

109
00:06:03,560 --> 00:06:06,920
Speaker 3: but it's pretty slim, George, And there's other problems out there,

110
00:06:06,920 --> 00:06:10,839
Speaker 3: but it's pretty slim. That's just your background associative memories.

111
00:06:10,839 --> 00:06:13,679
Speaker 3: We call it filling in teeing off of that word AI.

112
00:06:14,960 --> 00:06:17,239
Speaker 2: Where do you see AI going bar in the next

113
00:06:17,400 --> 00:06:20,520
Speaker 2: fifty years? I know it's a long ways out there.

114
00:06:20,760 --> 00:06:22,719
Speaker 3: Yeah, it's a tough call. And if you go online,

115
00:06:22,760 --> 00:06:25,440
Speaker 3: there's a series call Closer to Truth and tape an

116
00:06:25,440 --> 00:06:27,720
Speaker 3: episode of the late great founder of AI, the true

117
00:06:27,720 --> 00:06:30,320
Speaker 3: father of AI is Marvin Minsky, and we had one

118
00:06:30,360 --> 00:06:32,039
Speaker 3: of the episodes is what do you think it will

119
00:06:32,080 --> 00:06:34,920
Speaker 3: be like at fifty and one hundred years and far out?

120
00:06:35,320 --> 00:06:38,559
Speaker 3: We do think there will be supplements to the brain,

121
00:06:38,839 --> 00:06:41,760
Speaker 3: uploads to the brain at least implants more bringing human

122
00:06:41,880 --> 00:06:44,280
Speaker 3: interface and when it hit they'll hit as fast as

123
00:06:44,279 --> 00:06:48,120
Speaker 3: the smartphones hit. But well before that kind of thing happens, George,

124
00:06:48,440 --> 00:06:50,880
Speaker 3: we're just going to have a kind of AI creep everywhere,

125
00:06:50,880 --> 00:06:53,880
Speaker 3: because what we're really talking about is applying lots of

126
00:06:54,000 --> 00:06:57,599
Speaker 3: data to whatever problem we're looking at. And there's some

127
00:06:57,720 --> 00:07:00,880
Speaker 3: very good things about that. Your systems will become more accurate,

128
00:07:01,200 --> 00:07:04,120
Speaker 3: the scheduling systems and stores how much food they need

129
00:07:04,160 --> 00:07:06,280
Speaker 3: to buy based on much want that day. The all

130
00:07:06,279 --> 00:07:08,360
Speaker 3: griddles will get better and better and more precise. And

131
00:07:08,400 --> 00:07:11,480
Speaker 3: there's some bad things. We already have a big privacy creep,

132
00:07:11,920 --> 00:07:14,720
Speaker 3: and no one realize how much personal privacy we'd leave,

133
00:07:14,960 --> 00:07:17,720
Speaker 3: we would lose just when everyone has a smartphone and

134
00:07:17,880 --> 00:07:19,760
Speaker 3: is taping at will, even though a lot of that's

135
00:07:19,800 --> 00:07:22,880
Speaker 3: not legal at least in California, for example, without consent

136
00:07:23,520 --> 00:07:27,600
Speaker 3: and so on. So it's a mixed bag, and a

137
00:07:27,600 --> 00:07:30,280
Speaker 3: lot of us are very concerned. I'm also a lawyer

138
00:07:32,000 --> 00:07:35,360
Speaker 3: in the law school of USC besides engineering school. We're

139
00:07:35,400 --> 00:07:39,720
Speaker 3: concerned about the loss of digital privacy in the law

140
00:07:39,800 --> 00:07:44,680
Speaker 3: that's changing right in front of us, without let's say,

141
00:07:45,080 --> 00:07:47,560
Speaker 3: judges and justice is really being fully up to speed

142
00:07:48,080 --> 00:07:51,120
Speaker 3: on what's going on, and the law that we have

143
00:07:51,280 --> 00:07:55,320
Speaker 3: is based on paper, old paper, and stuff in dictionaries

144
00:07:55,320 --> 00:07:58,000
Speaker 3: and law books and things like that. It's not the

145
00:07:58,000 --> 00:08:01,520
Speaker 3: way the system is working today. And there's a huge

146
00:08:01,520 --> 00:08:05,320
Speaker 3: clash as we move in effect from atoms or things

147
00:08:05,360 --> 00:08:08,520
Speaker 3: to bits or information. It's a big clash. And we're

148
00:08:08,520 --> 00:08:10,760
Speaker 3: trying to update the law, I mean at a criminal level,

149
00:08:11,120 --> 00:08:14,840
Speaker 3: at the surveillance level, and also now that you're hearing

150
00:08:14,840 --> 00:08:17,480
Speaker 3: more about in all the lawsuits at the civil level,

151
00:08:17,480 --> 00:08:19,760
Speaker 3: but so far not we haven't been very successful with that.

152
00:08:21,040 --> 00:08:23,680
Speaker 2: Are you impressed with the pace of AI.

153
00:08:24,840 --> 00:08:28,120
Speaker 3: I am impressed with the development of computer chips, which

154
00:08:28,160 --> 00:08:30,240
Speaker 3: is behind a lot of it, especially in the large

155
00:08:30,320 --> 00:08:33,680
Speaker 3: language models that people are using out at chat systems,

156
00:08:34,280 --> 00:08:37,680
Speaker 3: and I've been contributing to them. We have a portfolio

157
00:08:37,880 --> 00:08:41,720
Speaker 3: patents for example at USC that we're taking the market,

158
00:08:41,800 --> 00:08:45,760
Speaker 3: so to speak, and so, rightly or wrongly, I'm contributing

159
00:08:45,800 --> 00:08:47,880
Speaker 3: to this process and have been for a long time,

160
00:08:48,520 --> 00:08:50,800
Speaker 3: and it's got problems and we're trying to fix those

161
00:08:50,840 --> 00:08:54,040
Speaker 3: problems in a variety of ways. Fuzzy logic is one

162
00:08:54,040 --> 00:08:56,240
Speaker 3: way to help out to deal with the black box.

163
00:08:56,280 --> 00:08:58,360
Speaker 3: Let me just say real fast, the problem with that

164
00:08:58,800 --> 00:09:01,600
Speaker 3: you can just expose a neural network system and this

165
00:09:01,760 --> 00:09:04,720
Speaker 3: is the lgims from the eighties to lots of pictures

166
00:09:04,760 --> 00:09:07,600
Speaker 3: of your face, and very soon it will learn to

167
00:09:07,640 --> 00:09:11,400
Speaker 3: recognize your face in many, many versions of it, noisy versions,

168
00:09:11,440 --> 00:09:14,319
Speaker 3: your face at night, your face with a killer mustache

169
00:09:14,400 --> 00:09:16,960
Speaker 3: or half covered in a blanket or something like that.

170
00:09:17,280 --> 00:09:19,840
Speaker 3: And that's a good thing. And your brain does that too,

171
00:09:19,960 --> 00:09:21,880
Speaker 3: So it has that ability to become a kind of

172
00:09:21,960 --> 00:09:26,319
Speaker 3: patterned computer, pattern recognized you. The trouble is is generalizing

173
00:09:26,360 --> 00:09:29,480
Speaker 3: that way beyond the scope of what you trained it,

174
00:09:29,559 --> 00:09:31,640
Speaker 3: so you don't know what it's learned. And when I

175
00:09:31,720 --> 00:09:34,319
Speaker 3: learned something new new, you don't know what it is forgotten.

176
00:09:35,360 --> 00:09:37,640
Speaker 3: And that was for small computers we were worried about

177
00:09:38,040 --> 00:09:40,959
Speaker 3: and it was hard to release these earlier for fear

178
00:09:41,120 --> 00:09:44,400
Speaker 3: in the commercial world of lawsuits because you really needed

179
00:09:44,520 --> 00:09:47,240
Speaker 3: ninety nine percent accuracy and you weren't getting that. Well,

180
00:09:47,280 --> 00:09:50,400
Speaker 3: now we have these giant black boxes, say for example,

181
00:09:50,440 --> 00:09:53,280
Speaker 3: the Chat and other systems that are up to way

182
00:09:53,320 --> 00:09:58,160
Speaker 3: beyond now a trillion parameters and growing, and they don't

183
00:09:58,200 --> 00:10:00,800
Speaker 3: just suffer from the problem of garbage garbage out, but

184
00:10:00,840 --> 00:10:03,920
Speaker 3: they have the problem of garbage in hallucinated garbage out,

185
00:10:04,360 --> 00:10:06,840
Speaker 3: or even when it's not garbage in, when it's really

186
00:10:06,880 --> 00:10:11,160
Speaker 3: good stuff going in known patterns. The system hallucinates in ways,

187
00:10:11,160 --> 00:10:13,040
Speaker 3: and it doesn't tell you it's doing. It's as if

188
00:10:13,400 --> 00:10:15,960
Speaker 3: an expert is lying to your faith and you can't tell.

189
00:10:16,080 --> 00:10:18,319
Speaker 3: So one of the things that doesn't do is when

190
00:10:18,320 --> 00:10:20,360
Speaker 3: you ask it a question. Even the chat system that

191
00:10:20,440 --> 00:10:23,040
Speaker 3: doesn't say, hey, I'm giving you an answer, George, but

192
00:10:23,080 --> 00:10:25,959
Speaker 3: I'm only eighty six percent or twenty three percent confident

193
00:10:26,000 --> 00:10:29,080
Speaker 3: of that answer, it spits out all answer is equally

194
00:10:29,080 --> 00:10:33,160
Speaker 3: confident no matter what it's training. Now, the advance of

195
00:10:33,240 --> 00:10:36,400
Speaker 3: fuzzy logic, as I call it fuzzy Logic two point

196
00:10:36,440 --> 00:10:39,000
Speaker 3: zero and discuss this in the re release of the

197
00:10:39,000 --> 00:10:43,440
Speaker 3: book Fuzzy Thinking, is we've gone from the earlier fuzzy

198
00:10:43,480 --> 00:10:46,480
Speaker 3: system of rules to the way the rules map over

199
00:10:46,600 --> 00:10:50,439
Speaker 3: the probability, which is the language of AI. And a

200
00:10:50,559 --> 00:10:53,400
Speaker 3: consequence of that is we can take some of these

201
00:10:53,440 --> 00:10:56,080
Speaker 3: neuralblack boxes, not all of them, but we can take

202
00:10:56,120 --> 00:10:59,760
Speaker 3: them and train on them and build a system of rules.

203
00:11:00,080 --> 00:11:03,320
Speaker 3: Gives us a kind of logical audit trail of inputs

204
00:11:03,320 --> 00:11:06,439
Speaker 3: to outputs to know what's being said, and a probabilistic

205
00:11:06,480 --> 00:11:10,440
Speaker 3: description that tells us for every output what the probability

206
00:11:10,480 --> 00:11:13,400
Speaker 3: is or the uncertainty is and that output or we

207
00:11:13,440 --> 00:11:15,520
Speaker 3: can open up its brain and look at which are

208
00:11:15,559 --> 00:11:18,280
Speaker 3: the rules fired. Now you have to pay for that.

209
00:11:18,800 --> 00:11:22,320
Speaker 3: You need more computation if you want a better XAI system.

210
00:11:22,320 --> 00:11:24,880
Speaker 3: We call it explainable AI. But that's one way to

211
00:11:24,920 --> 00:11:26,520
Speaker 3: do it. And it's hard to keep up with the

212
00:11:26,559 --> 00:11:30,640
Speaker 3: big black boxes. But that's been the contribution of funty logic.

213
00:11:30,960 --> 00:11:35,280
Speaker 3: And with the explosion in ongoing in Moore's law of

214
00:11:36,000 --> 00:11:39,000
Speaker 3: computer chip density, that the number on our circuits have

215
00:11:39,080 --> 00:11:42,240
Speaker 3: been has still been doubling about every two years since

216
00:11:42,320 --> 00:11:44,800
Speaker 3: nineteen sixty. We thought it would run out. It's still

217
00:11:44,840 --> 00:11:47,520
Speaker 3: going strong. And I think we're about to shift from

218
00:11:47,520 --> 00:11:52,000
Speaker 3: electrons to protons, from electricity to light as a lot

219
00:11:52,000 --> 00:11:55,800
Speaker 3: of its computations and some form of More's law may continue.

220
00:11:55,960 --> 00:11:58,480
Speaker 3: So these kind of gains, which are just scaling up

221
00:11:58,520 --> 00:12:01,080
Speaker 3: the old algorithms, are likely to continue at least for

222
00:12:01,120 --> 00:12:03,320
Speaker 3: a decade, if not much more.

223
00:12:04,080 --> 00:12:07,000
Speaker 2: Part I think one of the greatest inventions for mankind

224
00:12:07,320 --> 00:12:10,600
Speaker 2: is the smartphone. I mean the things that they do,

225
00:12:11,040 --> 00:12:15,640
Speaker 2: text messages, sending pictures through the air. It's incredible.

226
00:12:16,640 --> 00:12:20,360
Speaker 3: It is, indeed, and as I said, when I think

227
00:12:20,400 --> 00:12:22,800
Speaker 3: we get to the point, we're not there yet where

228
00:12:22,800 --> 00:12:25,600
Speaker 3: we can bridge this gap in the brain computer interface

229
00:12:26,040 --> 00:12:28,720
Speaker 3: between what's going on in your head and what's going

230
00:12:28,760 --> 00:12:31,720
Speaker 3: on in the smartphone and those kinds of systems, that's

231
00:12:31,760 --> 00:12:34,320
Speaker 3: the next really big advance. And for example, it's one

232
00:12:34,320 --> 00:12:40,559
Speaker 3: of the fourteen objectives or main projects of Modern Engineer

233
00:12:40,800 --> 00:12:43,840
Speaker 3: of the National Academy of Engineering called the Grand Challenges

234
00:12:43,880 --> 00:12:46,600
Speaker 3: for the twenty first century to reverse engineer the brain.

235
00:12:46,640 --> 00:12:48,880
Speaker 3: And that will surely happen. You and I live to see.

236
00:12:48,880 --> 00:12:51,840
Speaker 3: It is a very different matter. But when it does,

237
00:12:51,920 --> 00:12:54,760
Speaker 3: I think it will take off like a California brushfire,

238
00:12:55,360 --> 00:13:00,000
Speaker 3: and you'll have wireless access in effect to your brain

239
00:13:00,520 --> 00:13:02,760
Speaker 3: that you have right now to your smartphone. You may

240
00:13:02,800 --> 00:13:05,080
Speaker 3: not always want that and the fact it has, but

241
00:13:05,520 --> 00:13:08,959
Speaker 3: that will be truly revolutionary. We're not anywhere near that yet,

242
00:13:09,360 --> 00:13:12,600
Speaker 3: just because brains think in these on off spikes, and

243
00:13:12,640 --> 00:13:15,280
Speaker 3: you have one hundred billion neurons in your brain, each

244
00:13:15,320 --> 00:13:18,840
Speaker 3: connected on average to about ten thousand other neurons in

245
00:13:18,840 --> 00:13:22,520
Speaker 3: a totally asynchronous way, and that's completely different in a

246
00:13:22,600 --> 00:13:26,439
Speaker 3: structure than the way these chat systems. For example, the

247
00:13:26,559 --> 00:13:31,480
Speaker 3: chat systems have the form of spaghetti, whereas the human brain,

248
00:13:31,600 --> 00:13:35,240
Speaker 3: or the insect brain or the blue whale brain has

249
00:13:35,280 --> 00:13:38,400
Speaker 3: the form of a fish net. Spaghetti even though it's

250
00:13:38,400 --> 00:13:41,000
Speaker 3: all wound up, you could straighten it out, goes from

251
00:13:41,040 --> 00:13:43,600
Speaker 3: left to right. We call that feed forward, whereas the

252
00:13:43,640 --> 00:13:47,160
Speaker 3: fish net, the thing is linked up. In the case

253
00:13:47,200 --> 00:13:50,119
Speaker 3: of your brain instead of connected just to four other nuts.

254
00:13:50,360 --> 00:13:53,559
Speaker 3: Thereby in the lattice of the net. Again, each neurons

255
00:13:53,559 --> 00:13:56,720
Speaker 3: connected to on the order of ten thousand others, so

256
00:13:56,800 --> 00:14:02,080
Speaker 3: the organizational principles are different. That's the next frontiers computational neuroscience,

257
00:14:02,559 --> 00:14:05,960
Speaker 3: coming very soon to a computer chip near you.

258
00:14:06,400 --> 00:14:09,640
Speaker 1: Listen to more Coast to Coast AM every weeknight at

259
00:14:09,679 --> 00:14:12,920
Speaker 1: one am Eastern and go to Coast to cooastam dot

260
00:14:12,960 --> 00:14:13,720
Speaker 1: com for more