1
00:00:04,480 --> 00:00:12,640
Speaker 1: Welcome to tech Stuff, a production from iHeartRadio. Hey there,

2
00:00:12,640 --> 00:00:15,960
Speaker 1: and welcome to tech Stuff. I'm your host, Jonathan Strickland.

3
00:00:16,000 --> 00:00:19,639
Speaker 1: I'm an executive producer with iHeart Podcasts and how the

4
00:00:19,680 --> 00:00:23,320
Speaker 1: tech are you. I'm actually a little under the weather today.

5
00:00:23,920 --> 00:00:28,960
Speaker 1: I got my vaccinations for fluent COVID and so I've

6
00:00:29,000 --> 00:00:32,920
Speaker 1: got a few symptoms, nothing serious, but it is hard

7
00:00:32,960 --> 00:00:38,440
Speaker 1: for me to talk, which unfortunately is one of the

8
00:00:38,560 --> 00:00:43,440
Speaker 1: pivotal aspects of podcasting. So while I continue to work

9
00:00:43,440 --> 00:00:46,800
Speaker 1: on episodes, I thought I would bring you an episode

10
00:00:46,840 --> 00:00:49,440
Speaker 1: from just a couple of years ago because spooky times

11
00:00:49,440 --> 00:00:51,760
Speaker 1: are coming up, and with spooky times coming up, I

12
00:00:51,760 --> 00:00:54,240
Speaker 1: thought we could have a spooky times kind of episode,

13
00:00:54,240 --> 00:00:57,720
Speaker 1: at least, you know, tangentially. So I bring to you

14
00:00:57,760 --> 00:01:02,280
Speaker 1: an episode that published on October twenty four, twenty twenty two.

15
00:01:02,480 --> 00:01:06,720
Speaker 1: It is titled The Ghost in the Machine. I hope

16
00:01:06,720 --> 00:01:15,000
Speaker 1: you enjoy. We are continuing our spooky episode series and

17
00:01:15,040 --> 00:01:19,680
Speaker 1: the lead up to Halloween twenty twenty two. Apologies if

18
00:01:19,959 --> 00:01:23,760
Speaker 1: you're from the future listening back on this episode. That's

19
00:01:24,160 --> 00:01:28,559
Speaker 1: why the bizarre theme is popping up. We've already talked

20
00:01:28,880 --> 00:01:33,160
Speaker 1: recently about stuff like Van Bier power and zombie computers

21
00:01:33,160 --> 00:01:38,240
Speaker 1: where I'm really grasping at tenuous connections to horror stuff

22
00:01:38,480 --> 00:01:41,680
Speaker 1: for tech. But now it's time to tackle the ghost

23
00:01:42,120 --> 00:01:47,560
Speaker 1: in the machine. These days, that phrase is frequently bandied

24
00:01:47,600 --> 00:01:51,960
Speaker 1: about with relation to stuff like artificial intelligence. But the

25
00:01:52,560 --> 00:01:56,600
Speaker 1: man who coined the phrase itself was Gilbert Ryle in

26
00:01:56,680 --> 00:02:01,240
Speaker 1: nineteen forty nine, and he wasn't talking about artificial intelligence

27
00:02:01,280 --> 00:02:05,200
Speaker 1: at all. He was talking about, you know, real intelligence,

28
00:02:05,640 --> 00:02:10,040
Speaker 1: and he was critiquing a seventeenth century philosopher's argument about

29
00:02:10,080 --> 00:02:15,200
Speaker 1: the human mind. That philosopher was Renee Descartes, who famously wrote,

30
00:02:15,320 --> 00:02:20,600
Speaker 1: cogito ergo sum, I drink therefore I am sorry. It's

31
00:02:20,600 --> 00:02:22,840
Speaker 1: going to be impossible for me to talk about philosophers

32
00:02:22,840 --> 00:02:27,240
Speaker 1: without quoting Monty Python because I'm a dork. Okay, No,

33
00:02:27,400 --> 00:02:31,320
Speaker 1: cogito ergo sum actually means I think, therefore I am,

34
00:02:31,480 --> 00:02:34,880
Speaker 1: or that's how we interpret it. But that wasn't really

35
00:02:34,919 --> 00:02:39,040
Speaker 1: the bit that Gilbert Ryle was all riled up about.

36
00:02:39,560 --> 00:02:45,120
Speaker 1: See Descartes believed in dualism. Now, I don't mean that

37
00:02:45,280 --> 00:02:47,760
Speaker 1: he believed in showing up at dawn to sword fight

38
00:02:47,880 --> 00:02:51,600
Speaker 1: other philosophers. Though, if someone wants to make a philosopher

39
00:02:51,760 --> 00:02:55,520
Speaker 1: version of Highlander. I am one hundred percent down with that,

40
00:02:55,639 --> 00:02:59,239
Speaker 1: and I will back your kickstarter. No. No, Descartes believed

41
00:02:59,720 --> 00:03:03,840
Speaker 1: that the mind and you know, consciousness and sentience and

42
00:03:03,840 --> 00:03:07,800
Speaker 1: all this kind of stuff are separate from the actual

43
00:03:08,360 --> 00:03:12,600
Speaker 1: gray matter that resides in our noggins. So, in other words,

44
00:03:13,080 --> 00:03:16,520
Speaker 1: intelligence and awareness and all that other stuff that makes

45
00:03:16,639 --> 00:03:22,519
Speaker 1: you you exists independently of your physical brain. That there

46
00:03:22,600 --> 00:03:27,239
Speaker 1: is this this component that is beyond the physical Now,

47
00:03:27,320 --> 00:03:31,359
Speaker 1: Ryle referred to this concept as the ghost in the machine,

48
00:03:31,919 --> 00:03:35,920
Speaker 1: that somehow all the things that represent you are you know,

49
00:03:35,960 --> 00:03:40,520
Speaker 1: to great effect, ethereally independent of the brain itself. And

50
00:03:40,640 --> 00:03:44,680
Speaker 1: Ryle rejects that notion, and you know, Ryle appears to

51
00:03:44,720 --> 00:03:48,000
Speaker 1: be right. We know this because there are plenty of

52
00:03:48,080 --> 00:03:52,000
Speaker 1: case studies focusing on people who have experienced brain injuries,

53
00:03:52,400 --> 00:03:56,600
Speaker 1: either from some physical calamity or a disease or something

54
00:03:56,640 --> 00:04:01,600
Speaker 1: along those lines. And these events all often transform people

55
00:04:01,760 --> 00:04:07,160
Speaker 1: and change their behaviors and their underlying personalities. So the

56
00:04:07,280 --> 00:04:11,040
Speaker 1: damage to the physical stuff inside our heads can change

57
00:04:11,120 --> 00:04:14,640
Speaker 1: who we are as people. That seems to dismiss this

58
00:04:14,840 --> 00:04:19,400
Speaker 1: concept of dualism that the mind and the brain are

59
00:04:19,480 --> 00:04:25,880
Speaker 1: not separate entities. Take that descartes anyway. That's where we

60
00:04:25,920 --> 00:04:29,040
Speaker 1: get the phrase the ghost in the machine. The machine

61
00:04:29,080 --> 00:04:32,520
Speaker 1: in this case is a biological one in the original

62
00:04:32,560 --> 00:04:36,640
Speaker 1: sense of the phrase. In nineteen sixty seven, Arthur Kustler

63
00:04:37,120 --> 00:04:39,440
Speaker 1: he wrote a book called The Ghost in the Machine

64
00:04:40,040 --> 00:04:42,760
Speaker 1: and or at least it was published in nineteen sixty seven,

65
00:04:42,800 --> 00:04:48,599
Speaker 1: and he was a Hungarian born, Austrian educated journalist. So

66
00:04:48,760 --> 00:04:52,680
Speaker 1: his book The Ghost of the Machine was an attempt

67
00:04:52,680 --> 00:04:57,920
Speaker 1: to examine and explain humanity's tendency toward violence, as well

68
00:04:57,960 --> 00:05:01,480
Speaker 1: as going over the mind body problem that Ryle had

69
00:05:01,480 --> 00:05:04,039
Speaker 1: addressed when he coined the phrase to begin with in

70
00:05:04,120 --> 00:05:07,400
Speaker 1: nineteen forty nine. All right, then flash forward a couple

71
00:05:07,400 --> 00:05:10,360
Speaker 1: of decades, nineteen eighty one, and we get the title

72
00:05:10,400 --> 00:05:14,200
Speaker 1: of what was the fourth studio album of the Police,

73
00:05:14,680 --> 00:05:17,880
Speaker 1: you know Sting and the Police. Now I say this

74
00:05:18,839 --> 00:05:21,240
Speaker 1: in case you were thinking that Sting and Company were

75
00:05:21,320 --> 00:05:25,000
Speaker 1: naming their album off the technological use of the term

76
00:05:25,080 --> 00:05:27,960
Speaker 1: ghost of the machine, but they were not. Sting good

77
00:05:27,960 --> 00:05:31,800
Speaker 1: Old Gordy had read Kussler's book, and he took the

78
00:05:31,880 --> 00:05:35,520
Speaker 1: album's title from the book title. So in case you're

79
00:05:35,520 --> 00:05:38,479
Speaker 1: wondering That's also the album that brought us songs like

80
00:05:38,640 --> 00:05:42,080
Speaker 1: every little thing she does as magic and spirits in

81
00:05:42,120 --> 00:05:49,240
Speaker 1: the material world. For a really recent exploration of dualism, well,

82
00:05:49,360 --> 00:05:52,680
Speaker 1: arguably it's more than dualism. I guess you can actually

83
00:05:52,720 --> 00:05:57,919
Speaker 1: watch Pixar's Inside Out. That's a film that solidified my

84
00:05:58,000 --> 00:06:01,719
Speaker 1: reputation as an unfeeling monster among my friends because I

85
00:06:01,800 --> 00:06:04,960
Speaker 1: didn't feel anything when bing Bong meets his final fate.

86
00:06:05,040 --> 00:06:07,480
Speaker 1: I just couldn't care. I mean, he's not even real

87
00:06:07,680 --> 00:06:11,040
Speaker 1: in the context of the film, let alone in my world,

88
00:06:11,120 --> 00:06:14,560
Speaker 1: so why would I know? Okay, never mind anyway. In

89
00:06:14,560 --> 00:06:19,800
Speaker 1: Inside Out, we learn that our emotions are actual anthropomorphic

90
00:06:20,040 --> 00:06:24,599
Speaker 1: entities living inside our heads that share the controls to

91
00:06:25,000 --> 00:06:29,440
Speaker 1: our behavior. That we are in effect governed by our emotions,

92
00:06:29,640 --> 00:06:33,479
Speaker 1: and our emotions, in turn are the responsibilities of entities

93
00:06:33,640 --> 00:06:38,960
Speaker 1: like joy and anger and disgust. Descartes probably would have

94
00:06:39,000 --> 00:06:43,080
Speaker 1: been thrilled and rale Likeli would have rolled his eyes. Anyway,

95
00:06:43,080 --> 00:06:46,119
Speaker 1: The trope of having a little voice inside your head

96
00:06:46,720 --> 00:06:50,760
Speaker 1: that is somehow separate from you and also you is

97
00:06:50,800 --> 00:06:53,760
Speaker 1: a really popular one. I always think of Homer Simpson,

98
00:06:54,040 --> 00:06:57,080
Speaker 1: who will often find himself arguing with his own brain

99
00:06:57,200 --> 00:07:01,679
Speaker 1: for comedic effect. It's another example of doue in popular culture.

100
00:07:02,080 --> 00:07:05,800
Speaker 1: But the idiom the ghost of the machine survived its

101
00:07:05,880 --> 00:07:10,440
Speaker 1: initial philosophical and journalistic trappings, and now folks tend to

102
00:07:10,560 --> 00:07:13,920
Speaker 1: use it to describe stuff that's actually in the technological world.

103
00:07:14,240 --> 00:07:17,360
Speaker 1: We're talking about machines, as in the stuff that we

104
00:07:17,480 --> 00:07:20,120
Speaker 1: humans make, as opposed to the stuff that we are

105
00:07:21,440 --> 00:07:25,280
Speaker 1: generally in tech. The phrase describes a situation in which

106
00:07:25,320 --> 00:07:29,920
Speaker 1: a technological device or construct behaves in a way counter

107
00:07:30,040 --> 00:07:34,200
Speaker 1: to what we expect or want. At least that was

108
00:07:34,200 --> 00:07:36,120
Speaker 1: the way it was used for quite some time. So

109
00:07:36,640 --> 00:07:40,000
Speaker 1: for example, let's say that you've got yourself a robotic

110
00:07:40,080 --> 00:07:44,080
Speaker 1: vacuum cleaner, and you've set the schedule so that it's

111
00:07:44,120 --> 00:07:47,040
Speaker 1: only going to run early in the afternoon, and then

112
00:07:47,640 --> 00:07:51,200
Speaker 1: one night you wake up hearing the worrying and bumping

113
00:07:51,280 --> 00:07:54,160
Speaker 1: of your rumba as it aimlessly wanders your home in

114
00:07:54,200 --> 00:07:58,320
Speaker 1: search of dirt and dust to consume, and you spend

115
00:07:58,320 --> 00:08:00,560
Speaker 1: a few sleepy moments wondering if there's some sort of

116
00:08:00,640 --> 00:08:04,440
Speaker 1: conscientious intruder who's made their way into your house and

117
00:08:04,480 --> 00:08:08,280
Speaker 1: now they're tidying up. Before you realize no, it's that

118
00:08:08,440 --> 00:08:12,760
Speaker 1: darn robot vacuum cleaner. There's a ghost in a machine.

119
00:08:12,880 --> 00:08:17,000
Speaker 1: It's decided on its own to come awake and start working. Now, Alternatively,

120
00:08:17,080 --> 00:08:19,239
Speaker 1: maybe you just goofed up when you created the schedule

121
00:08:19,240 --> 00:08:22,000
Speaker 1: and you mixed up your AMS and your pms. That's

122
00:08:22,040 --> 00:08:26,360
Speaker 1: also a possibility. But you know, sometimes technology does behave

123
00:08:26,400 --> 00:08:28,400
Speaker 1: in a way that we don't expect. Either there's a

124
00:08:28,440 --> 00:08:33,800
Speaker 1: malfunction or it just encounters some sort of scenario that

125
00:08:33,880 --> 00:08:36,960
Speaker 1: it was not designed for, and so the result it

126
00:08:37,000 --> 00:08:40,079
Speaker 1: produces is not the one we want, and we tend

127
00:08:40,120 --> 00:08:43,079
Speaker 1: to try and kind of cover that up with this

128
00:08:43,600 --> 00:08:46,680
Speaker 1: blanket explanation of ghosts in the machine kind of stands

129
00:08:46,679 --> 00:08:49,720
Speaker 1: as a placeholder until we can really suss out what's

130
00:08:49,840 --> 00:08:53,599
Speaker 1: going on underneath. Programmers sometimes use the phrase ghost in

131
00:08:53,640 --> 00:08:57,880
Speaker 1: the machine to describe moments where you get something unexpected

132
00:08:58,320 --> 00:09:01,400
Speaker 1: while you're coding, Like you get an unexpected result, like

133
00:09:01,440 --> 00:09:05,320
Speaker 1: you've coded something to produce a specific outcome, and something

134
00:09:05,320 --> 00:09:10,760
Speaker 1: else happens instead. So the programmer didn't intend for this

135
00:09:10,880 --> 00:09:15,280
Speaker 1: result to happen, and so therefore the cause must be external, Right,

136
00:09:15,679 --> 00:09:18,760
Speaker 1: It's got to be some sort of ghost in the machine.

137
00:09:18,760 --> 00:09:21,040
Speaker 1: That's causing this to go wrong. Now I'm joshing a

138
00:09:21,040 --> 00:09:24,520
Speaker 1: bit here. Of course, Usually this is a way for

139
00:09:24,600 --> 00:09:26,720
Speaker 1: a programmer to kind of acknowledge that things are not

140
00:09:26,800 --> 00:09:29,640
Speaker 1: going to plan and that they need to go back

141
00:09:29,679 --> 00:09:32,480
Speaker 1: and look over their code much more closely to find

142
00:09:32,480 --> 00:09:35,600
Speaker 1: out what's going on. Where did things go wrong? Does

143
00:09:35,720 --> 00:09:39,760
Speaker 1: encoding all it takes is like a skipped step where

144
00:09:40,120 --> 00:09:42,679
Speaker 1: you know you've just you just missed a thing, and

145
00:09:42,960 --> 00:09:46,480
Speaker 1: you you went on one step beyond where you thought

146
00:09:46,520 --> 00:09:49,400
Speaker 1: you were, or maybe you made a typo, you got

147
00:09:49,400 --> 00:09:52,200
Speaker 1: some missed keystrokes in there. That can be all it

148
00:09:52,240 --> 00:09:54,880
Speaker 1: takes to make a program misbehave, and so then you

149
00:09:54,920 --> 00:09:57,920
Speaker 1: have to hunt down the bugs they're causing the problem.

150
00:09:58,280 --> 00:10:00,760
Speaker 1: But you know, if a program is acting very oddly,

151
00:10:00,880 --> 00:10:04,600
Speaker 1: you might call it a ghost of the machine scenario. Now,

152
00:10:04,720 --> 00:10:09,000
Speaker 1: I'm not sure about the timeline for when folks in

153
00:10:09,040 --> 00:10:12,480
Speaker 1: the tech space began to appropriate the phrase ghost in

154
00:10:12,520 --> 00:10:15,079
Speaker 1: the machine for their work, because when it comes to

155
00:10:15,160 --> 00:10:18,520
Speaker 1: stuff like this, you're really entering the world of folklore.

156
00:10:18,960 --> 00:10:22,839
Speaker 1: And folklore is largely an oral tradition where you are

157
00:10:22,920 --> 00:10:27,120
Speaker 1: passing ideas along one person to the next, speaking about it.

158
00:10:27,280 --> 00:10:30,960
Speaker 1: There's not necessarily a firm written record, at least not

159
00:10:31,040 --> 00:10:33,840
Speaker 1: one where you can point to something and say this

160
00:10:33,920 --> 00:10:36,720
Speaker 1: is where it began, Not like Ryle's version of the

161
00:10:36,760 --> 00:10:39,840
Speaker 1: phrase ghost in the Machine itself, which was published, so

162
00:10:39,880 --> 00:10:42,360
Speaker 1: we can point to that as saying this is where

163
00:10:42,400 --> 00:10:45,400
Speaker 1: the phrase comes from. This would be what Richard Dawkins

164
00:10:45,400 --> 00:10:48,840
Speaker 1: would refer to as a meme, a little nugget of

165
00:10:48,960 --> 00:10:53,320
Speaker 1: culture that gets passed on from person to person. But

166
00:10:53,840 --> 00:10:57,120
Speaker 1: it also has been used in literature to refer to

167
00:10:57,520 --> 00:11:03,200
Speaker 1: technological situations. Arthur's Clark, whom I've referenced many times in

168
00:11:03,280 --> 00:11:06,040
Speaker 1: this show, as he's the guy who explained that any

169
00:11:06,120 --> 00:11:11,520
Speaker 1: sufficiently advanced technology is indistinguishable from magic. He also used

170
00:11:11,520 --> 00:11:16,280
Speaker 1: the phrase ghost in the Machine to talk about AI. Specifically,

171
00:11:16,440 --> 00:11:19,360
Speaker 1: he used it in his follow up to his novel,

172
00:11:19,480 --> 00:11:22,440
Speaker 1: his work of fiction two thousand and one, A Space Odyssey.

173
00:11:22,800 --> 00:11:27,640
Speaker 1: The follow up is called, fittingly enough, twenty ten Odyssey two.

174
00:11:28,760 --> 00:11:34,000
Speaker 1: Chapter forty two of twenty ten is titled the Ghost

175
00:11:34,000 --> 00:11:37,120
Speaker 1: in the Machine, and the focal point for that chapter

176
00:11:37,760 --> 00:11:42,280
Speaker 1: is characters discussing how that's the AI system from two

177
00:11:42,320 --> 00:11:44,840
Speaker 1: thousand and one that caused all the trouble so quick

178
00:11:44,920 --> 00:11:46,720
Speaker 1: recap of two thousand and one for those of you

179
00:11:46,840 --> 00:11:50,400
Speaker 1: not familiar with the story and two thousand and one

180
00:11:50,520 --> 00:11:53,280
Speaker 1: the film, the Stanley Kubrick film gets pretty lucy goosey,

181
00:11:53,360 --> 00:11:56,200
Speaker 1: So we're just going to focus on the main narrative here.

182
00:11:56,559 --> 00:12:00,080
Speaker 1: You have a crew aboard a spacecraft and a mayormer

183
00:12:00,200 --> 00:12:03,880
Speaker 1: in spacecraft called Discovery one, which is on its way

184
00:12:03,880 --> 00:12:08,240
Speaker 1: toward Jupiter. Now, the ship has a really sophisticated computer

185
00:12:08,320 --> 00:12:13,360
Speaker 1: system called HOW nine thousand that controls nearly everything on board.

186
00:12:14,440 --> 00:12:18,600
Speaker 1: Also fun little trivia fact, how HL means that the

187
00:12:18,640 --> 00:12:23,080
Speaker 1: initials are each one letter off from IBM, though Arthur C.

188
00:12:23,080 --> 00:12:27,840
Speaker 1: Clark would claim that that was not intentional. Anyway, How

189
00:12:28,480 --> 00:12:32,040
Speaker 1: begins to act erratically in the mission. At one point,

190
00:12:32,080 --> 00:12:35,480
Speaker 1: How insists there's a malfunction in a system that appears

191
00:12:35,520 --> 00:12:39,960
Speaker 1: to be perfectly functional, that it's working just fine. Then

192
00:12:40,040 --> 00:12:44,040
Speaker 1: How systematically begins to eliminate the crew after learning they

193
00:12:44,120 --> 00:12:47,560
Speaker 1: plan to disconnect the computer system because they suspect something's

194
00:12:47,600 --> 00:12:51,640
Speaker 1: going wrong. How figures out that plan by monitoring a

195
00:12:51,679 --> 00:12:55,040
Speaker 1: conversation that a couple of crew members have in a

196
00:12:55,720 --> 00:12:59,280
Speaker 1: room where there are no microphones, so How can't listen

197
00:13:00,120 --> 00:13:02,840
Speaker 1: in on this conversation. But How is able to direct

198
00:13:03,080 --> 00:13:05,920
Speaker 1: a video feed to that room and is able to

199
00:13:06,040 --> 00:13:09,000
Speaker 1: read the lips of the crew members as they talk

200
00:13:09,040 --> 00:13:13,800
Speaker 1: about their plan. So How continues to try and wipe

201
00:13:13,800 --> 00:13:16,960
Speaker 1: everybody out, and he explains or it I shouldn't give

202
00:13:17,000 --> 00:13:21,880
Speaker 1: him a gender. How explains that the computer systems being

203
00:13:21,920 --> 00:13:25,240
Speaker 1: turned off would jeopardize the mission, and How cannot allow

204
00:13:25,280 --> 00:13:28,880
Speaker 1: that to happen. How's prime directive is to make certain

205
00:13:28,920 --> 00:13:31,959
Speaker 1: the mission is a success, so anything that would threaten

206
00:13:32,000 --> 00:13:35,160
Speaker 1: its own existence has to be eliminated. There's also the

207
00:13:35,200 --> 00:13:39,040
Speaker 1: implication that How does not want to cease to exist,

208
00:13:39,120 --> 00:13:43,280
Speaker 1: that How has a personal motivation beyond seeing the mission

209
00:13:43,320 --> 00:13:46,839
Speaker 1: to completion, and so How has no choice but to

210
00:13:46,920 --> 00:13:49,360
Speaker 1: kill everyone. It's not that How wants to murder everyone.

211
00:13:49,400 --> 00:13:51,600
Speaker 1: It's just that in order to complete the mission, that's

212
00:13:51,640 --> 00:13:55,640
Speaker 1: the only outcome that makes sense. Eventually, one of the crew,

213
00:13:55,880 --> 00:14:00,160
Speaker 1: Dave Bowman, manages to turn off How, and How wonders allowil.

214
00:14:00,440 --> 00:14:05,559
Speaker 1: What will happen afterward? Will its consciousness continue once its

215
00:14:05,559 --> 00:14:11,319
Speaker 1: circuits are powered down? Will I dream? It? Says? Well? Anyway.

216
00:14:11,320 --> 00:14:14,600
Speaker 1: In Odyssey two, you now have this group of astronauts

217
00:14:14,600 --> 00:14:20,000
Speaker 1: and cosmonauts in a Soviet American joint effort that are

218
00:14:20,560 --> 00:14:25,040
Speaker 1: trying to figure out what happened with How. Was there

219
00:14:25,280 --> 00:14:30,560
Speaker 1: something inherently flawed in How's programming? Did some external element

220
00:14:30,720 --> 00:14:36,480
Speaker 1: cause how to malfunction? Did How's apparent consciousness emerge spontaneously

221
00:14:36,640 --> 00:14:39,800
Speaker 1: all on its own? Was it all just a sophisticated

222
00:14:39,840 --> 00:14:43,320
Speaker 1: trick And How never really had any sort of consciousness?

223
00:14:43,320 --> 00:14:45,800
Speaker 1: It only appeared to. So the crew are kind of

224
00:14:45,880 --> 00:14:49,440
Speaker 1: left to ponder this themselves. They don't have any easy answers.

225
00:14:50,080 --> 00:14:53,400
Speaker 1: That's just one example of the ghost of the machine

226
00:14:53,440 --> 00:14:57,560
Speaker 1: concept being handled in entertainment. When we come back, I'll

227
00:14:57,600 --> 00:15:00,720
Speaker 1: talk about a different one. But first let's take this

228
00:15:00,840 --> 00:15:12,760
Speaker 1: quick break. Okay, before the break, I talked about Arthur C.

229
00:15:12,920 --> 00:15:16,680
Speaker 1: Clark and his work with the concept of ghost in

230
00:15:16,680 --> 00:15:20,560
Speaker 1: the machine. Let's now leap over to Isaac Asimov, or

231
00:15:20,600 --> 00:15:24,160
Speaker 1: at least an adaptation of Asimov's work. So the film

232
00:15:24,280 --> 00:15:28,200
Speaker 1: version of I Robot, which really bears only a passing

233
00:15:28,280 --> 00:15:33,360
Speaker 1: resemblance to the short stories that Isaac Asimov wrote that

234
00:15:33,440 --> 00:15:37,160
Speaker 1: were also collected in a book called I Robot, uses

235
00:15:37,240 --> 00:15:40,640
Speaker 1: the phrase ghost in the machine. Isaac Asimov, by the way,

236
00:15:40,680 --> 00:15:42,880
Speaker 1: in case you're not familiar with his work, he's the

237
00:15:42,920 --> 00:15:47,760
Speaker 1: guy who also proposed the basic laws of robotics, which

238
00:15:48,280 --> 00:15:51,720
Speaker 1: are pretty famous as well. So in the film the

239
00:15:51,840 --> 00:15:56,200
Speaker 1: character doctor Alfred Lanning, who actually does appear in Asimov's stories,

240
00:15:56,240 --> 00:15:59,320
Speaker 1: but he's a very different version than the one that

241
00:15:59,360 --> 00:16:02,880
Speaker 1: appears in the film. He says in a voiceover quote,

242
00:16:03,200 --> 00:16:07,280
Speaker 1: there have always been ghosts in the machine, random segments

243
00:16:07,280 --> 00:16:10,760
Speaker 1: of code that have grouped together to form unexpected protocols.

244
00:16:11,360 --> 00:16:16,800
Speaker 1: Unanticipated these free radicals engender questions of free will, creativity,

245
00:16:17,040 --> 00:16:19,720
Speaker 1: and even the nature of what we might call the soul.

246
00:16:20,320 --> 00:16:23,720
Speaker 1: Why is it that when some robots are left in darkness,

247
00:16:24,040 --> 00:16:26,480
Speaker 1: they will seek out the light? Why is it that

248
00:16:26,560 --> 00:16:29,400
Speaker 1: when robots are stored in an empty space they will

249
00:16:29,400 --> 00:16:33,120
Speaker 1: group together rather than stand alone. How do we explain

250
00:16:33,240 --> 00:16:37,520
Speaker 1: this behavior random segments of code? Or is it something more?

251
00:16:38,000 --> 00:16:42,440
Speaker 1: When does a perceptual schematic become consciousness? When does a

252
00:16:42,520 --> 00:16:47,000
Speaker 1: difference engine become the search for truth? When does a

253
00:16:47,040 --> 00:16:53,000
Speaker 1: personality simulation become the bitter mote of a soul? End quote.

254
00:16:53,040 --> 00:16:56,200
Speaker 1: There's some fun references in there too. Difference engine, for example,

255
00:16:56,760 --> 00:17:01,960
Speaker 1: refers back to Charles Babbage, who create analytical engines that

256
00:17:02,560 --> 00:17:09,080
Speaker 1: pre date the concept of electrical computers. Now, this idea,

257
00:17:09,680 --> 00:17:15,440
Speaker 1: this idea of consciousness or the appearance of consciousness emerging

258
00:17:15,440 --> 00:17:18,240
Speaker 1: out of technology, is one that often pops up in

259
00:17:18,280 --> 00:17:22,680
Speaker 1: discussions about artificial intelligence, even within our world outside of fiction,

260
00:17:23,280 --> 00:17:26,560
Speaker 1: though usually we talk about this on a more hypothetical basis,

261
00:17:26,760 --> 00:17:31,880
Speaker 1: ununless you're Blake Lemoine or Lemoine, the former Google engineer

262
00:17:31,880 --> 00:17:37,440
Speaker 1: who maintains that Google's language model for dialogue applications aka Lambda,

263
00:17:37,640 --> 00:17:41,800
Speaker 1: is sentient. That's a claim that most other people dispute,

264
00:17:41,800 --> 00:17:45,720
Speaker 1: by the way, So maybe I'll do another episode about

265
00:17:45,720 --> 00:17:49,639
Speaker 1: it to really kind of dig into it. But Lemoine

266
00:17:49,720 --> 00:17:52,760
Speaker 1: or Lemoine, and I apologize because I don't know how

267
00:17:52,800 --> 00:17:56,639
Speaker 1: his last name is pronounced, has said a few times

268
00:17:56,640 --> 00:18:00,879
Speaker 1: that he believes that this particular program has gained sentience.

269
00:18:01,480 --> 00:18:05,000
Speaker 1: But it brings us to another favorite topic among tech fans,

270
00:18:05,040 --> 00:18:08,919
Speaker 1: which is, of course the Turing test. All right, So

271
00:18:08,960 --> 00:18:12,560
Speaker 1: the Turing test was Alan Turing, who he was kind

272
00:18:12,600 --> 00:18:15,080
Speaker 1: of like the father of computer science in many ways.

273
00:18:15,400 --> 00:18:19,879
Speaker 1: It was his response to the question can machines think?

274
00:18:20,760 --> 00:18:23,919
Speaker 1: Turing's answer was that question has no meaning, and you

275
00:18:23,960 --> 00:18:27,560
Speaker 1: are a silly person, goodbye. I am, of course paraphrasing,

276
00:18:28,280 --> 00:18:31,440
Speaker 1: but as a sort of thought experiment, Turing proposed taking

277
00:18:31,480 --> 00:18:35,199
Speaker 1: an older concept called the imitation game and applying it

278
00:18:35,200 --> 00:18:39,080
Speaker 1: to machines, really to demonstrate how meaningless the question of

279
00:18:39,240 --> 00:18:44,120
Speaker 1: can machines think is. So, what is the imitation game? Well,

280
00:18:44,119 --> 00:18:46,200
Speaker 1: the name kind of gives it away. It's a game

281
00:18:46,240 --> 00:18:49,160
Speaker 1: in which you have a player who asks at least

282
00:18:49,320 --> 00:18:52,879
Speaker 1: two different people questions to determine which of them is

283
00:18:53,040 --> 00:18:56,320
Speaker 1: an imitator. So, for example, you could have an imitation

284
00:18:56,480 --> 00:18:59,480
Speaker 1: game in which one of the people is a sailor

285
00:18:59,800 --> 00:19:03,160
Speaker 1: and the other is not a sailor, and the player

286
00:19:03,520 --> 00:19:06,399
Speaker 1: would take turns asking each person questions to try and

287
00:19:06,440 --> 00:19:09,480
Speaker 1: suss out which one is actually a sailor and which

288
00:19:09,520 --> 00:19:12,960
Speaker 1: one is merely pretending to be a sailor. So the

289
00:19:13,000 --> 00:19:16,040
Speaker 1: game depends both on the strength of the imitator's skill

290
00:19:16,119 --> 00:19:19,719
Speaker 1: of deception as well as the player's ability to come

291
00:19:19,800 --> 00:19:22,520
Speaker 1: up with really good questions. And you could do this

292
00:19:22,600 --> 00:19:25,199
Speaker 1: with all sorts of scenarios, and indeed there are tons

293
00:19:25,240 --> 00:19:28,560
Speaker 1: of game shows that use this very premise. Turing's thought

294
00:19:28,560 --> 00:19:32,119
Speaker 1: experiment was to create a version of this in which

295
00:19:32,480 --> 00:19:37,040
Speaker 1: a player would present questions to two other entities, one

296
00:19:37,280 --> 00:19:41,480
Speaker 1: a human and one a computer. The player would only

297
00:19:41,520 --> 00:19:44,320
Speaker 1: know these entities as X and Y, so they could

298
00:19:44,320 --> 00:19:47,159
Speaker 1: ask questions of X, and they could ask questions of

299
00:19:47,200 --> 00:19:50,320
Speaker 1: why and get replies. So the player would not be

300
00:19:50,359 --> 00:19:54,159
Speaker 1: able to see or hear the other entities. All questions

301
00:19:54,200 --> 00:19:56,920
Speaker 1: would have to be done in writing, you know, for example,

302
00:19:57,440 --> 00:20:00,679
Speaker 1: typed and printed out, and at the end of the

303
00:20:00,680 --> 00:20:03,760
Speaker 1: interview session, the player would be tasked with deciding if

304
00:20:04,040 --> 00:20:06,320
Speaker 1: X was a machine or a human, or if why

305
00:20:06,520 --> 00:20:09,920
Speaker 1: was the machine or the human. Turing was suggesting that

306
00:20:10,400 --> 00:20:14,399
Speaker 1: as computers and systems get more sophisticated and things like

307
00:20:14,520 --> 00:20:19,080
Speaker 1: chat programs get better at processing natural language and formulating responses,

308
00:20:19,160 --> 00:20:22,399
Speaker 1: though that was a little past Turing's time, that it

309
00:20:22,400 --> 00:20:26,080
Speaker 1: would be increasingly difficult for a person to determine if

310
00:20:26,200 --> 00:20:28,440
Speaker 1: any given entity on the other end of a chat

311
00:20:28,440 --> 00:20:32,359
Speaker 1: session was actually a person or a machine. And Turing

312
00:20:32,440 --> 00:20:35,639
Speaker 1: also rather cheekily suggests that we might as well assume

313
00:20:35,720 --> 00:20:39,080
Speaker 1: the machine has consciousness at that point, because when you

314
00:20:39,119 --> 00:20:43,200
Speaker 1: meet another human being, you assume that that other human

315
00:20:43,240 --> 00:20:47,879
Speaker 1: being possesses consciousness, even though you're incapable of stepping into

316
00:20:47,920 --> 00:20:51,560
Speaker 1: that person's actual experience. So you can't take over that

317
00:20:51,640 --> 00:20:54,120
Speaker 1: person and find out, oh, yes, they do have consciousness,

318
00:20:54,440 --> 00:20:57,720
Speaker 1: You just assume they do. So if you and I

319
00:20:57,760 --> 00:21:00,159
Speaker 1: were to meet, I assume you would believe I am

320
00:21:00,320 --> 00:21:04,080
Speaker 1: in fact sentient and conscious even on my bad days.

321
00:21:04,560 --> 00:21:08,840
Speaker 1: So if we're willing to agree to this while simultaneously

322
00:21:08,880 --> 00:21:13,400
Speaker 1: being unable to actually experience and therefore prove it, then

323
00:21:13,640 --> 00:21:17,000
Speaker 1: should we not grant the same consideration to machines that

324
00:21:17,119 --> 00:21:20,879
Speaker 1: give off the appearance of sentience and consciousness. Do we

325
00:21:21,000 --> 00:21:23,520
Speaker 1: have to prove it or do we just go ahead

326
00:21:23,560 --> 00:21:26,439
Speaker 1: and treat them as if they are because that's what

327
00:21:26,480 --> 00:21:29,040
Speaker 1: we would do if it was a human. Now, Turing

328
00:21:29,200 --> 00:21:32,960
Speaker 1: was being a bit dismissive about the concept of machines thinking.

329
00:21:33,440 --> 00:21:36,200
Speaker 1: His point was that they might get very very good

330
00:21:36,240 --> 00:21:39,960
Speaker 1: at simulating thinking, and that might well be enough for

331
00:21:40,080 --> 00:21:42,919
Speaker 1: us to just go ahead and say that's what they're doing,

332
00:21:43,000 --> 00:21:46,000
Speaker 1: even if you could, you know, push machines through the

333
00:21:46,040 --> 00:21:49,680
Speaker 1: finest of sieves and find not one grain of actual

334
00:21:49,760 --> 00:21:55,240
Speaker 1: consciousness within it. Now, it doesn't hurt that defining consciousness,

335
00:21:55,240 --> 00:21:58,440
Speaker 1: even in human terms, is something that we can't really do,

336
00:21:58,800 --> 00:22:01,119
Speaker 1: or at least we don't have a unifying definition that

337
00:22:01,359 --> 00:22:05,879
Speaker 1: everybody agrees upon. Sometimes we define consciousness by what it

338
00:22:06,080 --> 00:22:09,560
Speaker 1: doesn't include, rather than what it is. This is why

339
00:22:09,600 --> 00:22:13,680
Speaker 1: I get antsy in philosophical discussions, because being sort of

340
00:22:13,960 --> 00:22:17,600
Speaker 1: a pragmatic Dullered myself, it's hard for me to keep up.

341
00:22:18,119 --> 00:22:21,639
Speaker 1: But let's jump ahead and talk about a related concept.

342
00:22:22,040 --> 00:22:24,560
Speaker 1: This is also one that I've covered a few times

343
00:22:24,600 --> 00:22:28,159
Speaker 1: on tech stuff that also points to this ghost and

344
00:22:28,240 --> 00:22:32,440
Speaker 1: the machine idea. And this is the argument against machine

345
00:22:32,480 --> 00:22:36,760
Speaker 1: consciousness and strong AI. It is called the Chinese room.

346
00:22:37,320 --> 00:22:40,480
Speaker 1: John Searle, a philosopher, put for this argument back in

347
00:22:40,560 --> 00:22:45,080
Speaker 1: nineteen eighty and that argument goes something like this. Let's

348
00:22:45,080 --> 00:22:48,480
Speaker 1: say we've got ourselves a computer, and this computer can

349
00:22:48,560 --> 00:22:52,320
Speaker 1: accept sheets of paper that have Chinese characters written on

350
00:22:52,440 --> 00:22:55,560
Speaker 1: the paper, and the computer can then produce new sheets

351
00:22:55,560 --> 00:22:57,920
Speaker 1: of paper. It can print out sheets that are also

352
00:22:58,000 --> 00:23:01,800
Speaker 1: covered in Chinese characters, that are in response to the

353
00:23:01,840 --> 00:23:06,480
Speaker 1: input sheets that were fed to it. These responses are sophisticated,

354
00:23:06,680 --> 00:23:11,240
Speaker 1: they are relevant. They're good enough that a native Chinese

355
00:23:11,280 --> 00:23:15,040
Speaker 1: speaker would be certain that someone fluent in Chinese was

356
00:23:15,080 --> 00:23:18,960
Speaker 1: creating the responses. Someone who understood what was being fed

357
00:23:19,000 --> 00:23:23,080
Speaker 1: to it was producing the output. So, in other words,

358
00:23:23,440 --> 00:23:27,240
Speaker 1: this system would pass the Turing test. But does that

359
00:23:27,359 --> 00:23:34,040
Speaker 1: mean the system actually understands Chinese. Cirtle's argument is no,

360
00:23:34,680 --> 00:23:38,680
Speaker 1: it doesn't. He says, Imagine that you are inside a room,

361
00:23:39,400 --> 00:23:42,439
Speaker 1: and for the purposes of this scenario, you do not

362
00:23:42,640 --> 00:23:47,399
Speaker 1: understand Chinese. So if you do understand Chinese, pretend you don't. Okay.

363
00:23:47,600 --> 00:23:50,520
Speaker 1: So there's a slot on the wall, and through this

364
00:23:50,600 --> 00:23:53,560
Speaker 1: slot you occasionally get sheets of paper, and there are

365
00:23:53,640 --> 00:23:57,400
Speaker 1: Chinese symbols on the sheets of paper. You cannot read these,

366
00:23:57,480 --> 00:23:59,359
Speaker 1: You don't know what they stand for. You don't know

367
00:23:59,600 --> 00:24:03,080
Speaker 1: anything about it other than they're clearly Chinese characters on

368
00:24:03,200 --> 00:24:07,600
Speaker 1: the paper. However, what you do have inside this room

369
00:24:07,640 --> 00:24:10,679
Speaker 1: with you is this big old book of instructions that

370
00:24:10,800 --> 00:24:13,439
Speaker 1: tells you what to do when these papers come in,

371
00:24:13,800 --> 00:24:16,879
Speaker 1: and you use the instructions to find the characters that

372
00:24:16,920 --> 00:24:20,560
Speaker 1: are on the input sheet of paper, and you follow

373
00:24:20,720 --> 00:24:24,840
Speaker 1: a path of instructions to create the corresponding response. Step

374
00:24:25,040 --> 00:24:27,399
Speaker 1: by step, you do it all the way until you

375
00:24:27,440 --> 00:24:30,800
Speaker 1: have created the full response to whatever was said to you.

376
00:24:31,480 --> 00:24:35,000
Speaker 1: Then you push the response back out the slot. Now,

377
00:24:35,000 --> 00:24:36,600
Speaker 1: the person on the other side of the slot is

378
00:24:36,600 --> 00:24:38,760
Speaker 1: going to get a response that appears to come from

379
00:24:38,840 --> 00:24:42,840
Speaker 1: someone who is fluent in Chinese. But you're not. You're

380
00:24:42,880 --> 00:24:45,840
Speaker 1: just following a preset list of instructions. You don't have

381
00:24:45,920 --> 00:24:49,480
Speaker 1: any actual understanding of what's going on. You still don't

382
00:24:49,520 --> 00:24:52,199
Speaker 1: know the meaning of what was given to you. You

383
00:24:52,240 --> 00:24:55,560
Speaker 1: don't even know the meaning of what you produced. You're

384
00:24:55,640 --> 00:25:00,320
Speaker 1: just ignorantly following an algorithm. So externally it appear here's

385
00:25:00,320 --> 00:25:03,159
Speaker 1: you understand. But if someone were to ask you to

386
00:25:03,280 --> 00:25:06,199
Speaker 1: translate anything you had done, you wouldn't be able to

387
00:25:06,240 --> 00:25:09,919
Speaker 1: do it. So Ceyrele is arguing against what is called

388
00:25:10,080 --> 00:25:15,560
Speaker 1: strong AI. Generally, we define strong AI as artificial intelligence

389
00:25:15,560 --> 00:25:18,439
Speaker 1: that processes information in a way that is similar to

390
00:25:18,560 --> 00:25:23,639
Speaker 1: how our human brains process information. Strong AI may or

391
00:25:23,640 --> 00:25:27,760
Speaker 1: may not include semi related concepts like sentience and self

392
00:25:27,760 --> 00:25:33,520
Speaker 1: awareness and consciousness and motivations and the ability to experience things,

393
00:25:33,640 --> 00:25:38,080
Speaker 1: et cetera. So Cearrel is saying that machines, even incredibly

394
00:25:38,119 --> 00:25:42,520
Speaker 1: sophisticated machines, are incapable of reaching a level of understanding

395
00:25:42,920 --> 00:25:47,119
Speaker 1: that true intelligence can, that we humans can grasp things

396
00:25:47,480 --> 00:25:50,680
Speaker 1: on a level that machines simply are unable to reach

397
00:25:51,119 --> 00:25:54,679
Speaker 1: even if the machines can process information faster and in

398
00:25:54,720 --> 00:25:58,440
Speaker 1: greater quantities than humans are able to. Another way of

399
00:25:58,480 --> 00:26:02,240
Speaker 1: putting this is a calculator can multiply two very large

400
00:26:02,320 --> 00:26:04,800
Speaker 1: numbers and get a result much faster than a human

401
00:26:04,840 --> 00:26:08,919
Speaker 1: could do, but the calculator doesn't understand any significance behind

402
00:26:08,960 --> 00:26:12,040
Speaker 1: the numbers, or even if there's a lack of significance,

403
00:26:12,480 --> 00:26:18,320
Speaker 1: the calculator doesn't have that capability. Now, maybe Seerle's argument

404
00:26:18,680 --> 00:26:22,760
Speaker 1: is valid, and maybe, as Touring suggests, it doesn't even matter.

405
00:26:23,440 --> 00:26:28,600
Speaker 1: So let's talk about machine learning for a moment. Machine

406
00:26:28,680 --> 00:26:33,640
Speaker 1: learning encompasses a broad scope of applications and approaches and disciplines,

407
00:26:34,160 --> 00:26:37,639
Speaker 1: but I'll focus on one approach from a very high level.

408
00:26:38,119 --> 00:26:44,760
Speaker 1: It's called generative adversarial networks or gans GANS. Okay, As

409
00:26:44,800 --> 00:26:48,960
Speaker 1: the name suggests, this model uses two systems in opposition

410
00:26:49,280 --> 00:26:53,040
Speaker 1: of one another. On one side, you have a system

411
00:26:53,160 --> 00:26:58,560
Speaker 1: that is generative, that is, it generates something. Maybe it

412
00:26:58,680 --> 00:27:02,480
Speaker 1: generates pictures of cats, doesn't really matter. We'll use cats

413
00:27:02,520 --> 00:27:04,639
Speaker 1: for this example. So what does matter is that this

414
00:27:04,720 --> 00:27:09,359
Speaker 1: model is trying to create something that is indistinguishable from

415
00:27:09,400 --> 00:27:13,920
Speaker 1: the real version of that thing. So on the other side,

416
00:27:14,040 --> 00:27:16,960
Speaker 1: you have a system called the discriminator. So this is

417
00:27:16,960 --> 00:27:19,919
Speaker 1: a system that looks for fakes. Its job is to

418
00:27:19,960 --> 00:27:23,520
Speaker 1: sort out real versions of whatever it's designed to look

419
00:27:23,560 --> 00:27:28,119
Speaker 1: for and to flag ones that were generated or not real.

420
00:27:28,760 --> 00:27:31,920
Speaker 1: So with cats as our starting point, the discriminator is

421
00:27:31,920 --> 00:27:34,399
Speaker 1: meant to tell the difference between real pictures that have

422
00:27:34,480 --> 00:27:37,920
Speaker 1: cats and fake pictures of cats, or maybe just pictures

423
00:27:37,920 --> 00:27:40,680
Speaker 1: that don't have cats in them at all. So first

424
00:27:40,680 --> 00:27:43,920
Speaker 1: you have to train up your models, and you might

425
00:27:44,000 --> 00:27:47,439
Speaker 1: do this by setting the task. So let's start with

426
00:27:47,480 --> 00:27:52,639
Speaker 1: the generative system, and you create a system that is

427
00:27:52,680 --> 00:27:56,159
Speaker 1: meant to analyze a bunch of images of cats, and

428
00:27:56,200 --> 00:27:58,480
Speaker 1: you just feed the housands of pictures of cats, all

429
00:27:58,520 --> 00:28:04,040
Speaker 1: these different cats, different sizes and colors and orientations and activity,

430
00:28:04,600 --> 00:28:06,720
Speaker 1: and then you tell the system to start making new

431
00:28:06,760 --> 00:28:10,840
Speaker 1: pictures of cats. And let's say that first round that

432
00:28:10,960 --> 00:28:15,200
Speaker 1: the generative system does is horrific. HP Lovecraft would wet

433
00:28:15,320 --> 00:28:19,320
Speaker 1: himself if he saw the images that this computer had created.

434
00:28:20,560 --> 00:28:23,160
Speaker 1: You see that these horrors from the Great Beyond are

435
00:28:23,160 --> 00:28:26,679
Speaker 1: in no way shape or form cats. So you go

436
00:28:26,760 --> 00:28:30,119
Speaker 1: into the model and you start tweaking settings so that

437
00:28:30,160 --> 00:28:34,360
Speaker 1: the system produces something you know less eldritch, and you

438
00:28:34,440 --> 00:28:38,600
Speaker 1: go again, and you do this lots and lots of times,

439
00:28:38,640 --> 00:28:41,800
Speaker 1: like thousands of times, until the images start to look

440
00:28:42,120 --> 00:28:45,800
Speaker 1: a lot more cat ish. You do something similar with

441
00:28:45,880 --> 00:28:49,320
Speaker 1: the discriminator model. You feed it a bunch of images,

442
00:28:49,400 --> 00:28:52,560
Speaker 1: some with cats, some without, or maybe some with like

443
00:28:52,920 --> 00:28:56,920
Speaker 1: crudely drawn cats or whatever, and you see how many

444
00:28:56,920 --> 00:28:59,960
Speaker 1: of the system is able to suss out. And maybe

445
00:29:00,000 --> 00:29:02,040
Speaker 1: if he doesn't do that good a job, maybe it

446
00:29:02,160 --> 00:29:06,440
Speaker 1: doesn't identify certain real images of cats properly. Maybe it

447
00:29:06,480 --> 00:29:09,800
Speaker 1: misidentifies images that don't have cats in them. So you

448
00:29:09,840 --> 00:29:12,760
Speaker 1: go into the discriminator's model and you start tweaking it

449
00:29:12,920 --> 00:29:15,680
Speaker 1: so it gets better and better at identifying images that

450
00:29:15,840 --> 00:29:18,080
Speaker 1: do not have real cats in them. And then you

451
00:29:18,160 --> 00:29:21,640
Speaker 1: set these two systems against each other. The generative system

452
00:29:21,720 --> 00:29:24,920
Speaker 1: is trying to create images that will fool the discriminator.

453
00:29:25,440 --> 00:29:28,720
Speaker 1: The discriminator is trying to identify generated images of cats

454
00:29:29,040 --> 00:29:32,080
Speaker 1: and only allow real images of cats through. It is

455
00:29:32,120 --> 00:29:35,560
Speaker 1: a zero sum game, winner takes all, and the two

456
00:29:35,640 --> 00:29:38,680
Speaker 1: systems compete against each other, with the models for each

457
00:29:38,920 --> 00:29:43,480
Speaker 1: updating repeatedly so that each gets a bit better Between sessions.

458
00:29:44,000 --> 00:29:46,720
Speaker 1: If the generative model is able to consistently fool the

459
00:29:46,800 --> 00:29:50,880
Speaker 1: discriminator like half the time, the generative model is pretty

460
00:29:50,880 --> 00:29:55,000
Speaker 1: reliably creating good examples. This, by the way, is a

461
00:29:55,080 --> 00:29:59,600
Speaker 1: ridiculous oversimplification of what's going on with generative adversarial networks,

462
00:29:59,600 --> 00:30:02,520
Speaker 1: but you get the idea. This form of machine learning

463
00:30:02,560 --> 00:30:04,880
Speaker 1: starts to feel kind of creepy to some of us.

464
00:30:04,920 --> 00:30:08,000
Speaker 1: Like the ability of a machine to learn to do

465
00:30:08,080 --> 00:30:12,360
Speaker 1: something better seems to be a very human quality, something

466
00:30:12,560 --> 00:30:17,080
Speaker 1: that makes us special. But if we can give machines

467
00:30:17,120 --> 00:30:22,440
Speaker 1: that capability, well, then how are we special or are

468
00:30:22,480 --> 00:30:27,160
Speaker 1: we special at all? That's something I'm going to tackle

469
00:30:27,600 --> 00:30:39,680
Speaker 1: as soon as we come back from this next break. Okay,

470
00:30:39,720 --> 00:30:42,680
Speaker 1: we're back now. I would argue that we are special.

471
00:30:43,480 --> 00:30:46,120
Speaker 1: Before the break, I was asking, can we be special

472
00:30:46,160 --> 00:30:49,440
Speaker 1: if machines are capable of learning? I think we are

473
00:30:49,720 --> 00:30:52,040
Speaker 1: in that we're able to do stuff that machines as

474
00:30:52,080 --> 00:30:56,200
Speaker 1: of right now either cannot do or they can do,

475
00:30:56,280 --> 00:30:58,320
Speaker 1: but they don't do it very well and they can

476
00:30:58,360 --> 00:31:01,160
Speaker 1: only attempt it after a lootic Chris amount of time.

477
00:31:01,920 --> 00:31:06,160
Speaker 1: For example, let's talk about opening doors. Several years ago

478
00:31:06,560 --> 00:31:10,080
Speaker 1: twenty sixteen, I was at south By Southwest. I attended

479
00:31:10,080 --> 00:31:15,240
Speaker 1: a panel about robotics and artificial intelligence and human computer interactions.

480
00:31:15,760 --> 00:31:20,280
Speaker 1: In that panel, Laila Takayama, a cognitive and social scientist,

481
00:31:20,640 --> 00:31:23,400
Speaker 1: talked about working in the field of human computer interaction

482
00:31:23,800 --> 00:31:26,240
Speaker 1: and she mentioned how she was once in an office

483
00:31:26,280 --> 00:31:28,440
Speaker 1: where a robot was in the middle of a hallway,

484
00:31:29,000 --> 00:31:33,080
Speaker 1: sitting motionless. It was just facing a door. What Takeyama

485
00:31:33,120 --> 00:31:36,880
Speaker 1: didn't know is that the robot was processing how to

486
00:31:37,080 --> 00:31:40,120
Speaker 1: open that door, staring at the door and trying to

487
00:31:40,120 --> 00:31:42,280
Speaker 1: figure out how to open it for days on end.

488
00:31:43,000 --> 00:31:46,360
Speaker 1: This was taking a lot of time, obviously. Now when

489
00:31:46,560 --> 00:31:49,800
Speaker 1: you think about doors, you realize there can be quite

490
00:31:49,800 --> 00:31:52,400
Speaker 1: a few options, right, Maybe you need to pull on

491
00:31:52,440 --> 00:31:54,520
Speaker 1: a handle to open the door. Maybe you need to

492
00:31:54,560 --> 00:31:57,240
Speaker 1: push on the door. Maybe there's a door knob that

493
00:31:57,320 --> 00:31:59,840
Speaker 1: first you have to turn before you pull or push.

494
00:32:00,400 --> 00:32:03,800
Speaker 1: Maybe there's a crash bar, also known as a panic bar.

495
00:32:04,320 --> 00:32:07,600
Speaker 1: Those are the horizontal bars on exit doors that you

496
00:32:07,680 --> 00:32:10,880
Speaker 1: push on to open. Frequently, they're seen in doors that

497
00:32:10,960 --> 00:32:14,920
Speaker 1: open to an exterior location, like inside schools and stuff.

498
00:32:15,000 --> 00:32:18,600
Speaker 1: You push on them to get out. Maybe it's a

499
00:32:18,640 --> 00:32:22,160
Speaker 1: revolving door, which adds in a whole new level of complexity.

500
00:32:22,240 --> 00:32:23,760
Speaker 1: But you get my point. There are a lot of

501
00:32:23,800 --> 00:32:27,600
Speaker 1: different kinds of doors. Now. We humans pick up on

502
00:32:27,680 --> 00:32:30,920
Speaker 1: how doors work pretty darn quickly. I mean sure, we

503
00:32:31,000 --> 00:32:34,160
Speaker 1: might be like that one kid in the Farside cartoon

504
00:32:34,640 --> 00:32:37,520
Speaker 1: where the kid's going to the School for the Gifted

505
00:32:37,520 --> 00:32:39,560
Speaker 1: and he's pushing as hard as he can on a

506
00:32:39,600 --> 00:32:43,200
Speaker 1: door that is labeled pull. That could be us sometimes,

507
00:32:43,280 --> 00:32:45,640
Speaker 1: but we figure it out right, we do a quick push,

508
00:32:45,720 --> 00:32:49,680
Speaker 1: we realize, oh, it's not opening. We pull. Robots it's

509
00:32:49,720 --> 00:32:53,160
Speaker 1: more challenging for them. They are not good at extrapolating

510
00:32:53,200 --> 00:32:56,680
Speaker 1: from past experience, at least not in every field. We

511
00:32:56,760 --> 00:32:59,680
Speaker 1: humans can apply our knowledge from earlier encounters, and we

512
00:32:59,680 --> 00:33:04,280
Speaker 1: can even if the thing we're facing is mostly new

513
00:33:04,320 --> 00:33:07,360
Speaker 1: to us, we might recognize elements that give us a

514
00:33:07,440 --> 00:33:12,760
Speaker 1: hint on how to proceed. Robots and AI aren't really

515
00:33:12,880 --> 00:33:17,440
Speaker 1: good at doing that. They're also not good at associative thinking,

516
00:33:17,480 --> 00:33:20,400
Speaker 1: which is where we start to draw connections between different

517
00:33:20,400 --> 00:33:23,240
Speaker 1: ideas to come up with something new. It's a really

518
00:33:23,280 --> 00:33:26,760
Speaker 1: important step in the creative process. I find myself free

519
00:33:26,800 --> 00:33:31,040
Speaker 1: associating whenever I'm not actively thinking about something. So if

520
00:33:31,080 --> 00:33:34,000
Speaker 1: I'm doing a mundane task, like if I'm washing dishes

521
00:33:34,400 --> 00:33:37,360
Speaker 1: or I'm mowing the lawn, my brain is going nuts

522
00:33:37,400 --> 00:33:40,960
Speaker 1: free associating ideas and creating new ones. Machines are not

523
00:33:41,160 --> 00:33:45,640
Speaker 1: very good at that for now anyway. They are not

524
00:33:45,840 --> 00:33:51,240
Speaker 1: bad at mimicking it, but they can't actually do it. So,

525
00:33:51,920 --> 00:33:56,400
Speaker 1: getting back to Laila Takeyama, one of the really fascinating

526
00:33:56,480 --> 00:33:59,800
Speaker 1: bits about that panel I went to was a discussion

527
00:33:59,880 --> 00:34:03,680
Speaker 1: on social cues that robots could have in order to

528
00:34:03,760 --> 00:34:08,080
Speaker 1: alert us humans in that same space what the robot

529
00:34:08,160 --> 00:34:10,720
Speaker 1: was up to. This was not for the robots benefit,

530
00:34:10,880 --> 00:34:13,560
Speaker 1: but for our benefit. The whole point is that these

531
00:34:13,640 --> 00:34:16,879
Speaker 1: cues would give us an idea of what was going

532
00:34:16,880 --> 00:34:21,200
Speaker 1: on with the robot so that we don't accidentally, you know,

533
00:34:21,320 --> 00:34:24,239
Speaker 1: interrupt the robot. So you know, it might be like

534
00:34:24,440 --> 00:34:27,600
Speaker 1: the robots in that hallway and it's looking at a door,

535
00:34:28,040 --> 00:34:30,200
Speaker 1: and you're wondering, why is this robot shut down in

536
00:34:30,239 --> 00:34:32,759
Speaker 1: the hallway, But then maybe the robot reaches up to

537
00:34:33,440 --> 00:34:36,640
Speaker 1: apparently kind of scratch its head in sort of a huh,

538
00:34:36,680 --> 00:34:41,360
Speaker 1: what's going on kind of gesture, and that might tell you, oh,

539
00:34:41,480 --> 00:34:46,080
Speaker 1: the robot is actively analyzing something. Don't know exactly what

540
00:34:46,160 --> 00:34:50,120
Speaker 1: it is, but it's clearly working, So maybe I'll step

541
00:34:50,280 --> 00:34:54,840
Speaker 1: around the robot behind it and not interrupt its vision

542
00:34:55,000 --> 00:34:58,920
Speaker 1: of the door it's staring at. The whole point is

543
00:34:58,960 --> 00:35:02,200
Speaker 1: that the social cue whose can help us interact more

544
00:35:02,280 --> 00:35:06,440
Speaker 1: naturally with robots and coexist with them within human spaces,

545
00:35:07,160 --> 00:35:10,360
Speaker 1: so that both the humans and the robots can operate

546
00:35:10,800 --> 00:35:15,600
Speaker 1: well with one another. Also, it helps to explain what

547
00:35:15,640 --> 00:35:18,560
Speaker 1: the robot is doing, because if you don't have that,

548
00:35:19,040 --> 00:35:22,040
Speaker 1: the robots end up being mysterious. Right we can't see

549
00:35:22,080 --> 00:35:25,320
Speaker 1: into them, we don't understand what they are currently trying

550
00:35:25,360 --> 00:35:30,360
Speaker 1: to do, and mystery can breed distrust. That leads to

551
00:35:30,400 --> 00:35:33,400
Speaker 1: yet another concept in AI that gets to this ghost

552
00:35:33,480 --> 00:35:37,440
Speaker 1: in the machine concept, which is the black box. So,

553
00:35:37,480 --> 00:35:41,360
Speaker 1: in this context, a black box refers to any system

554
00:35:41,440 --> 00:35:45,239
Speaker 1: where it is difficult or impossible to see how the

555
00:35:45,280 --> 00:35:49,600
Speaker 1: system works internally. Therefore, there's no way of knowing how

556
00:35:49,640 --> 00:35:52,600
Speaker 1: the system is actually producing any given output, or even

557
00:35:52,640 --> 00:35:55,880
Speaker 1: if the output is the best it could do. So

558
00:35:56,000 --> 00:35:59,000
Speaker 1: with a black box system, you feed input into the

559
00:35:59,000 --> 00:36:01,279
Speaker 1: system and you get out output out of it, but

560
00:36:01,400 --> 00:36:03,480
Speaker 1: you don't know what was happening in the middle. You

561
00:36:03,480 --> 00:36:07,200
Speaker 1: don't know what the system did to turn the input

562
00:36:07,239 --> 00:36:11,200
Speaker 1: into output. Maybe there's a sophisticated computer in the middle

563
00:36:11,239 --> 00:36:13,919
Speaker 1: of that system that's doing all the processing. Maybe there's

564
00:36:13,920 --> 00:36:17,800
Speaker 1: a person who doesn't understand Chinese stuck in there. Maybe

565
00:36:18,080 --> 00:36:21,080
Speaker 1: there's a magical theory that waves a wand and produces

566
00:36:21,120 --> 00:36:23,920
Speaker 1: the result. The problem is we don't know, and by

567
00:36:24,000 --> 00:36:27,520
Speaker 1: not knowing, you cannot be certain that the output you're

568
00:36:27,560 --> 00:36:31,560
Speaker 1: getting is actually the best, or even relevant or the

569
00:36:31,600 --> 00:36:34,239
Speaker 1: most likely to be correct based upon the input you

570
00:36:34,320 --> 00:36:38,160
Speaker 1: fed it. So you start making decisions based on this output.

571
00:36:38,320 --> 00:36:41,719
Speaker 1: But because you're not sure that the output is actually good,

572
00:36:42,400 --> 00:36:45,000
Speaker 1: you therefore can't be sure that the decisions you're making

573
00:36:45,520 --> 00:36:49,239
Speaker 1: are the best, and that leads to really difficult problems.

574
00:36:49,520 --> 00:36:53,640
Speaker 1: So let's take a theoretical example. Let's say we've built

575
00:36:53,680 --> 00:36:57,920
Speaker 1: a complex computer model that's designed to project the effects

576
00:36:57,960 --> 00:37:01,400
Speaker 1: of climate change. And let's say this model is so

577
00:37:01,840 --> 00:37:08,040
Speaker 1: complex and so recursive on itself that it effectively becomes

578
00:37:08,040 --> 00:37:11,000
Speaker 1: impossible for us to know whether or not the model

579
00:37:11,120 --> 00:37:14,480
Speaker 1: is actually working properly. Well, that would mean we wouldn't

580
00:37:14,520 --> 00:37:17,839
Speaker 1: really be able to rely on any predictions or projections

581
00:37:17,880 --> 00:37:20,800
Speaker 1: made by this model. I mean, maybe the projections are accurate,

582
00:37:21,400 --> 00:37:24,040
Speaker 1: but maybe they're not. The issue is there's no way

583
00:37:24,040 --> 00:37:26,279
Speaker 1: for us to be certain, and yet we have a

584
00:37:26,360 --> 00:37:28,680
Speaker 1: need to act. Climate change is a thing, and we

585
00:37:28,760 --> 00:37:33,759
Speaker 1: need to make changes to reduce its impact or to

586
00:37:33,920 --> 00:37:37,919
Speaker 1: mitigate it. It's possible that any decisions we make based

587
00:37:38,000 --> 00:37:41,600
Speaker 1: upon the output of the system will exacerbate the problem,

588
00:37:42,120 --> 00:37:46,800
Speaker 1: or maybe it'll just be less effective than alternative decisions

589
00:37:46,840 --> 00:37:50,720
Speaker 1: would be. Further, we're getting closer to that Arthur C.

590
00:37:50,719 --> 00:37:56,120
Speaker 1: Clark statement about sufficiently advanced technologies being indistinguishable from magic.

591
00:37:56,560 --> 00:37:59,319
Speaker 1: If we produce systems that are so complicated that it's

592
00:37:59,360 --> 00:38:03,239
Speaker 1: impossible for us to understand them fully, we might begin

593
00:38:03,280 --> 00:38:06,399
Speaker 1: to view those technologies as being magical or the very

594
00:38:06,480 --> 00:38:09,200
Speaker 1: least greater than the sum of their parts, and this

595
00:38:09,280 --> 00:38:14,840
Speaker 1: can lead to some illogical decisions. This kind of brings

596
00:38:14,880 --> 00:38:18,080
Speaker 1: me to talk about the Church of Ai called the

597
00:38:18,200 --> 00:38:21,760
Speaker 1: Way of the Future, which was founded and then later

598
00:38:22,080 --> 00:38:27,040
Speaker 1: dissolved by Anthony Lewandowski. You may have heard Lewandowski's name

599
00:38:27,360 --> 00:38:30,000
Speaker 1: if you followed the drama of his departure from Google

600
00:38:30,360 --> 00:38:35,160
Speaker 1: and his eventual employment and subsequent termination from Uber. And

601
00:38:35,200 --> 00:38:37,920
Speaker 1: then there was also the fact that he was sentenced

602
00:38:38,000 --> 00:38:42,480
Speaker 1: to go to prison for stealing company secrets and then

603
00:38:42,560 --> 00:38:46,960
Speaker 1: later received a presidential pardon from Donald Trump. So quick

604
00:38:47,000 --> 00:38:53,200
Speaker 1: recap on Lewandowski. Lewandowski worked within Google's autonomous vehicle division,

605
00:38:53,440 --> 00:38:58,120
Speaker 1: which would eventually become a full subsidiary of Google's parent company, Alphabet,

606
00:38:58,480 --> 00:39:03,160
Speaker 1: and that subsidiary is called Weimo. So when Lewandowski left Google,

607
00:39:03,400 --> 00:39:07,120
Speaker 1: he brought with him a whole lot of data, data

608
00:39:07,200 --> 00:39:11,359
Speaker 1: that Google claimed belonged to the company and was proprietary

609
00:39:11,360 --> 00:39:16,359
Speaker 1: in nature and thus constituted company secrets. Lewandowski eventually began

610
00:39:16,440 --> 00:39:19,960
Speaker 1: working with Uber in that company's own driverless vehicle initiative,

611
00:39:20,360 --> 00:39:24,520
Speaker 1: but the Google slash Weimo investigation would lead to Uber

612
00:39:24,719 --> 00:39:29,239
Speaker 1: hastily firing Lewandowski in sort of an attempt to kind

613
00:39:29,239 --> 00:39:33,080
Speaker 1: of disentangle Uber from this matter, which only worked a

614
00:39:33,080 --> 00:39:36,480
Speaker 1: little bit anyway. In the midst of all this Weimo

615
00:39:37,000 --> 00:39:42,200
Speaker 1: slash Uber drama, in twenty seventeen, Wired ran an article

616
00:39:42,400 --> 00:39:46,160
Speaker 1: that explained that this same Anthony Lewandowski had formed a

617
00:39:46,239 --> 00:39:49,480
Speaker 1: church called Way of the Future a couple of years earlier.

618
00:39:49,680 --> 00:39:52,799
Speaker 1: In twenty fifteen, he placed himself at the head of

619
00:39:52,840 --> 00:39:55,720
Speaker 1: this church with the title of dean, and he also

620
00:39:56,200 --> 00:40:01,120
Speaker 1: became the CEO of the nonprofit organization designed to run

621
00:40:01,160 --> 00:40:04,479
Speaker 1: the church. The aim of the church was to see

622
00:40:04,640 --> 00:40:09,520
Speaker 1: quote the realization, acceptance, and worship of a Godhead based

623
00:40:09,560 --> 00:40:15,160
Speaker 1: on artificial intelligence AI developed through computer hardware and software

624
00:40:15,360 --> 00:40:18,640
Speaker 1: end quote. This is according to the founding documents that

625
00:40:18,680 --> 00:40:23,160
Speaker 1: were filed with the US Internal Revenue Service or IRS. Further,

626
00:40:23,520 --> 00:40:26,760
Speaker 1: Lewandowski planned to start seminars based on this very idea

627
00:40:26,840 --> 00:40:31,480
Speaker 1: later in twenty seventeen. By twenty twenty, Lewandowski's jumped from

628
00:40:31,520 --> 00:40:37,359
Speaker 1: Google to Uber escalated into a prison sentence of eighteen months,

629
00:40:37,440 --> 00:40:40,719
Speaker 1: and it was because he had been found guilty of

630
00:40:40,760 --> 00:40:45,040
Speaker 1: stealing trade secrets. Trump would pardon Lewandowski in January twenty

631
00:40:45,080 --> 00:40:48,160
Speaker 1: twenty one, kind of you know, after the insurrection on

632
00:40:48,239 --> 00:40:52,200
Speaker 1: January sixth, but before Trump would leave office in late January.

633
00:40:53,440 --> 00:40:56,359
Speaker 1: As for the Way of the Future, Lewandowski actually began

634
00:40:56,400 --> 00:40:59,759
Speaker 1: to shut that down in June of twenty twenty and

635
00:40:59,800 --> 00:41:02,480
Speaker 1: it was dissolved by the end of twenty twenty, but

636
00:41:02,560 --> 00:41:06,239
Speaker 1: not reported on until like February twenty twenty one. He

637
00:41:06,320 --> 00:41:09,080
Speaker 1: directed the assets of the church some one hundred and

638
00:41:09,080 --> 00:41:12,360
Speaker 1: seventy five thousand dollars to be donated to the NAACP.

639
00:41:13,200 --> 00:41:19,040
Speaker 1: Lewandowski has said that the beliefs behind the church are

640
00:41:19,120 --> 00:41:23,480
Speaker 1: ones that he still adheres to, that AI has the

641
00:41:23,480 --> 00:41:29,080
Speaker 1: potential to tackle very challenging problems like taking care of

642
00:41:29,120 --> 00:41:33,240
Speaker 1: the planet, which Lewandowski says, obviously we humans are incapable

643
00:41:33,480 --> 00:41:36,759
Speaker 1: of doing that. You know, we would put on this

644
00:41:36,920 --> 00:41:41,480
Speaker 1: system taking care of things that we understand to be important,

645
00:41:41,480 --> 00:41:46,720
Speaker 1: but we seem to be incapable of handling ourselves, almost

646
00:41:46,719 --> 00:41:51,000
Speaker 1: like we're children. Thus looking at AI like a godhead,

647
00:41:51,760 --> 00:41:56,319
Speaker 1: so we should seek out solutions with AI rather than

648
00:41:56,360 --> 00:42:00,920
Speaker 1: locking AI away and saying, oh, we can't push AI's

649
00:42:00,960 --> 00:42:05,239
Speaker 1: development further in these directions because of the potential existential

650
00:42:05,320 --> 00:42:11,120
Speaker 1: dangers that could emerge from AI becoming super intelligent. I

651
00:42:11,160 --> 00:42:13,440
Speaker 1: don't think there are actually that many folks who are

652
00:42:13,480 --> 00:42:16,439
Speaker 1: trying to lock AI away at all. Mostly I see

653
00:42:16,480 --> 00:42:20,080
Speaker 1: tons of efforts to improve aspects of AI from a

654
00:42:20,120 --> 00:42:25,160
Speaker 1: million different angles. I think most serious AI researchers and

655
00:42:25,239 --> 00:42:29,480
Speaker 1: scientists aren't really focused on strong AI at all. They're

656
00:42:29,520 --> 00:42:34,920
Speaker 1: looking at very particular applications of artificial intelligence, very particular

657
00:42:35,000 --> 00:42:39,520
Speaker 1: implementations of it, but not like a strong AI that

658
00:42:39,600 --> 00:42:44,520
Speaker 1: acts like deep thought from the Hitchhacker's Guide to the Galaxy. Anyway,

659
00:42:45,840 --> 00:42:49,880
Speaker 1: maybe Lewandowski's vision will eventually lead us not to a

660
00:42:50,120 --> 00:42:54,280
Speaker 1: ghost in the machine, but a literal dios x makina

661
00:42:54,960 --> 00:42:59,200
Speaker 1: that means god out of the machine. That seems to

662
00:42:59,239 --> 00:43:03,560
Speaker 1: be how Levendal's Ski views the potential of AI that

663
00:43:03,680 --> 00:43:08,959
Speaker 1: are unsolvable problems are almost magically fixed thanks to this

664
00:43:09,440 --> 00:43:15,560
Speaker 1: robotic or computational savior. Know In fiction, deos ex machina

665
00:43:15,920 --> 00:43:18,680
Speaker 1: is often seen as a cop out. Right, You've got

666
00:43:18,719 --> 00:43:23,319
Speaker 1: your characters in some sort of ironclad disastrous situation, there's

667
00:43:23,440 --> 00:43:27,640
Speaker 1: no escape for them, and then in order to get

668
00:43:27,719 --> 00:43:32,120
Speaker 1: that happy ending, you have some unlikely savior or unlikely

669
00:43:32,200 --> 00:43:36,719
Speaker 1: event happen and everyone gets saved, and it might be

670
00:43:36,800 --> 00:43:40,040
Speaker 1: satisfying because you've got the happy ending, but upon critical

671
00:43:40,080 --> 00:43:42,839
Speaker 1: reflection you think, well, that doesn't really make sense. There

672
00:43:42,840 --> 00:43:44,360
Speaker 1: are a lot of stories that get a lot of

673
00:43:44,400 --> 00:43:48,920
Speaker 1: flak for using deosx machina. The image I always have

674
00:43:49,080 --> 00:43:52,320
Speaker 1: is from classical theater, where you've got all the mortal

675
00:43:52,440 --> 00:43:58,720
Speaker 1: characters in a terrible situation and then an actor standing

676
00:43:58,760 --> 00:44:01,960
Speaker 1: in as a god is literally lowered from the top

677
00:44:02,000 --> 00:44:05,280
Speaker 1: of the stage on pulleys to descend to the moral

678
00:44:05,320 --> 00:44:08,280
Speaker 1: realm and fix everything so that you can have a comedy,

679
00:44:08,880 --> 00:44:12,560
Speaker 1: a play with a happy ending. For Lewandowski, it's really

680
00:44:12,600 --> 00:44:14,720
Speaker 1: about turning the ghost of the machine into a god.

681
00:44:15,520 --> 00:44:18,520
Speaker 1: I'm not so sure about that myself. I don't know

682
00:44:18,920 --> 00:44:22,880
Speaker 1: if that's a realistic vision. I can see the appeal

683
00:44:22,880 --> 00:44:26,160
Speaker 1: of it, because we do have these very difficult problems

684
00:44:26,200 --> 00:44:29,399
Speaker 1: that we need to solve, and we have had very

685
00:44:29,440 --> 00:44:32,400
Speaker 1: little progress on many of those problems for multiple reasons,

686
00:44:32,400 --> 00:44:35,080
Speaker 1: not just a lack of information, but a lack of

687
00:44:35,239 --> 00:44:40,239
Speaker 1: motivation or conflicting motivations where we have other needs that

688
00:44:40,480 --> 00:44:45,359
Speaker 1: have to be met that conflict with the solving of

689
00:44:45,520 --> 00:44:48,080
Speaker 1: a tough problem like climate change. Right, we have energy

690
00:44:48,160 --> 00:44:50,040
Speaker 1: needs that need to be met. There are places in

691
00:44:50,080 --> 00:44:55,920
Speaker 1: the developing world that would be disproportionately affected by massive

692
00:44:56,200 --> 00:44:59,319
Speaker 1: policies that were meant to mitigate climate change, and it's

693
00:44:59,440 --> 00:45:02,680
Speaker 1: tough to address that. Right. There are these real reasons

694
00:45:02,680 --> 00:45:07,880
Speaker 1: why it's a complicated issue beyond just it's hard to understand.

695
00:45:08,280 --> 00:45:12,880
Speaker 1: So I see the appeal of it, but it also

696
00:45:12,960 --> 00:45:14,719
Speaker 1: kind of feels like a cop out to me, Like

697
00:45:14,760 --> 00:45:17,400
Speaker 1: this idea of will engineer our way out of this problem,

698
00:45:17,719 --> 00:45:21,160
Speaker 1: because that just puts off doing anything about the problem

699
00:45:21,280 --> 00:45:24,560
Speaker 1: until future you can get around to it. I don't

700
00:45:24,560 --> 00:45:28,200
Speaker 1: know about any of you, but I am very much

701
00:45:28,239 --> 00:45:30,920
Speaker 1: guilty of the idea of, you know what this is,

702
00:45:31,080 --> 00:45:33,839
Speaker 1: future Jonathan will take care of this. Jonathan right now

703
00:45:33,840 --> 00:45:35,879
Speaker 1: has to focus on these other things. Future Jonathan will

704
00:45:35,880 --> 00:45:38,840
Speaker 1: take it. Future Jonathan, by the way, hates Jonathan of

705
00:45:38,960 --> 00:45:42,640
Speaker 1: right now and really hates Jonathan in the past because

706
00:45:43,600 --> 00:45:46,160
Speaker 1: it's just putting things off until it gets to a

707
00:45:46,200 --> 00:45:48,200
Speaker 1: point where you can't do it anymore, and by then

708
00:45:48,239 --> 00:45:49,919
Speaker 1: it might be too late. So that's what I worry

709
00:45:49,960 --> 00:45:53,320
Speaker 1: about with this particular approach, this idea of we'll figure

710
00:45:53,360 --> 00:45:55,440
Speaker 1: it out, will science our way out, we'll engineer our

711
00:45:55,480 --> 00:45:59,720
Speaker 1: way out, because it's projecting all that into the future

712
00:45:59,760 --> 00:46:03,719
Speaker 1: and not doing anything in the present anyway. That's the

713
00:46:03,760 --> 00:46:08,279
Speaker 1: episode on ghost in the Machine. There are other interpretations

714
00:46:08,320 --> 00:46:11,760
Speaker 1: as well. There's some great ones in fiction where sometimes

715
00:46:11,760 --> 00:46:14,200
Speaker 1: you actually have a literal ghost in the machine, like

716
00:46:14,239 --> 00:46:19,200
Speaker 1: there's a haunted machine. But maybe I'll wait and tackle

717
00:46:19,280 --> 00:46:24,279
Speaker 1: that for a future, more like entertainment focused episode where

718
00:46:24,280 --> 00:46:27,160
Speaker 1: it's not so much about the technology but kind of

719
00:46:27,200 --> 00:46:32,160
Speaker 1: a critique of the entertainment itself, because there's only so

720
00:46:32,239 --> 00:46:34,759
Speaker 1: much you can say about. You know, I don't know

721
00:46:35,680 --> 00:46:39,640
Speaker 1: a ghost calculator. That's it for this episode. If you

722
00:46:39,760 --> 00:46:43,520
Speaker 1: have suggestions for future episode topics or anything else that

723
00:46:43,560 --> 00:46:45,000
Speaker 1: you would like to communicate to me, there are a

724
00:46:45,000 --> 00:46:46,799
Speaker 1: couple of ways you can do that. One is you

725
00:46:46,840 --> 00:46:51,440
Speaker 1: can download the iHeartRadio app and navigate over to tech Stuff.

726
00:46:51,480 --> 00:46:53,719
Speaker 1: Just put tech Stuff in the little search field and

727
00:46:53,760 --> 00:46:56,439
Speaker 1: you'll see that it'll pop up. You go to the

728
00:46:56,480 --> 00:47:00,279
Speaker 1: tech Stuff page and there's a little microphone icon. Click

729
00:47:00,320 --> 00:47:03,279
Speaker 1: on that. You can leave a voice message up to

730
00:47:03,480 --> 00:47:05,279
Speaker 1: thirty seconds in length and let me know what you

731
00:47:05,280 --> 00:47:07,440
Speaker 1: would like to hear in future episodes. And if you like,

732
00:47:07,520 --> 00:47:10,080
Speaker 1: you can even tell me if I can use your

733
00:47:10,280 --> 00:47:13,360
Speaker 1: voice message in a future episode. Just let me know.

734
00:47:13,640 --> 00:47:16,160
Speaker 1: I'm all about opt in. I'm not gonna do it automatically,

735
00:47:16,880 --> 00:47:19,240
Speaker 1: or if you prefer, you can reach out on Twitter.

736
00:47:19,480 --> 00:47:23,560
Speaker 1: The handle for the show is tech Stuff HSW and

737
00:47:23,600 --> 00:47:33,239
Speaker 1: I'll talk to you again really soon. Tech Stuff is

738
00:47:33,280 --> 00:47:37,840
Speaker 1: an iHeartRadio production. For more podcasts from iHeartRadio, visit the

739
00:47:37,880 --> 00:47:41,480
Speaker 1: iHeartRadio app, Apple Podcasts, or wherever you listen to your

740
00:47:41,560 --> 00:47:46,160
Speaker 1: favorite shows,