1
00:00:04,440 --> 00:00:12,760
Speaker 1: Welcome to tech Stuff, a production from iHeartRadio. Hey there,

2
00:00:12,760 --> 00:00:15,840
Speaker 1: and welcome to tech Stuff. I'm your host Jonathan Strickland.

3
00:00:15,840 --> 00:00:19,440
Speaker 1: I'm an executive producer with iHeart Podcasts and how the

4
00:00:19,480 --> 00:00:22,680
Speaker 1: tech are you? Folks. We have a very very special

5
00:00:22,720 --> 00:00:25,360
Speaker 1: episode of tech Stuff because I have a very very

6
00:00:25,400 --> 00:00:29,440
Speaker 1: special guest with me today, Jacob Goldstein. Now, if I

7
00:00:29,480 --> 00:00:32,760
Speaker 1: were to go and run down his entire resume, it

8
00:00:32,800 --> 00:00:35,600
Speaker 1: would be an entire episode just by itself. He's got

9
00:00:35,600 --> 00:00:40,680
Speaker 1: a long and distinguished career in journalism, and among his

10
00:00:41,040 --> 00:00:44,360
Speaker 1: many accomplishments happens to be the fact that he's the

11
00:00:44,360 --> 00:00:48,279
Speaker 1: host of a podcast called What's Your Problem, which is

12
00:00:48,520 --> 00:00:52,559
Speaker 1: a show that really dives into things like engineering and

13
00:00:52,640 --> 00:00:55,800
Speaker 1: how engineers tackle problems, how do they define them, and

14
00:00:55,800 --> 00:00:59,960
Speaker 1: then how do they create solutions. So Jacob, welcome to

15
00:01:00,080 --> 00:01:01,640
Speaker 1: tech Stuff. Thank you for being here.

16
00:01:02,120 --> 00:01:04,320
Speaker 2: Hi, Jonathan, thanks so much for having me. I'm delighted

17
00:01:04,360 --> 00:01:04,760
Speaker 2: to be here.

18
00:01:05,240 --> 00:01:08,160
Speaker 1: Yeah, I'm delighted you're here too. And before we really

19
00:01:08,200 --> 00:01:10,840
Speaker 1: dive into a full discussion, we're going to talk a

20
00:01:10,880 --> 00:01:13,440
Speaker 1: lot about engineering and a lot about AI in particular,

21
00:01:13,520 --> 00:01:16,440
Speaker 1: because many of your episodes in this last season of

22
00:01:16,480 --> 00:01:18,759
Speaker 1: your show have been on AI. I want to learn

23
00:01:18,800 --> 00:01:21,280
Speaker 1: more about you, So tell us a bit about your

24
00:01:21,319 --> 00:01:24,920
Speaker 1: background and how you came to become a podcaster on

25
00:01:25,080 --> 00:01:25,960
Speaker 1: What's Your Problem.

26
00:01:26,480 --> 00:01:29,360
Speaker 2: So, before I started What's Your Problem, I was one

27
00:01:29,360 --> 00:01:31,640
Speaker 2: of the co hosts of a podcast called Planet Money,

28
00:01:32,040 --> 00:01:35,840
Speaker 2: which is a show about economics. And before I had

29
00:01:35,840 --> 00:01:39,160
Speaker 2: that job, I didn't know that much about economics. You know,

30
00:01:39,200 --> 00:01:41,600
Speaker 2: I was an English major in college. I'd covered healthcare

31
00:01:41,600 --> 00:01:44,800
Speaker 2: for the Wall Street Journal, and so getting there and

32
00:01:44,880 --> 00:01:49,680
Speaker 2: covering economics for a while, to me, the big exciting

33
00:01:49,920 --> 00:01:54,760
Speaker 2: idea at the heart of economics is the pie can

34
00:01:54,840 --> 00:01:57,960
Speaker 2: get bigger, right, everybody can be better off. The world

35
00:01:58,040 --> 00:02:01,560
Speaker 2: is not a zero sum game, and I think that

36
00:02:01,640 --> 00:02:07,640
Speaker 2: is a very non intuitive, big exciting idea. And basically

37
00:02:08,280 --> 00:02:11,480
Speaker 2: the way we all get better off in the long

38
00:02:11,560 --> 00:02:15,960
Speaker 2: run is through technology. Right, It's through people figuring out

39
00:02:16,440 --> 00:02:20,160
Speaker 2: more efficient ways to do things, ways so that you know,

40
00:02:20,200 --> 00:02:22,680
Speaker 2: we do the same amount of work basically, but we

41
00:02:22,720 --> 00:02:25,840
Speaker 2: get more stuff. You get more output for every hour

42
00:02:25,880 --> 00:02:30,639
Speaker 2: of labor. And that is fundamentally you know, engineering and technology,

43
00:02:30,680 --> 00:02:33,000
Speaker 2: as you said, And so I wanted to go deeper

44
00:02:33,040 --> 00:02:36,080
Speaker 2: on like how that actually works, Like, you know, there

45
00:02:36,080 --> 00:02:38,040
Speaker 2: are people whose job is I'm going to go to

46
00:02:38,080 --> 00:02:40,520
Speaker 2: work and like figure out a better way to do something,

47
00:02:40,639 --> 00:02:42,520
Speaker 2: and so that is what I'm trying to do on

48
00:02:42,560 --> 00:02:44,080
Speaker 2: What's Your Problem? Those are the kind of people I

49
00:02:44,120 --> 00:02:44,920
Speaker 2: talk to on the show.

50
00:02:45,520 --> 00:02:48,880
Speaker 1: That's awesome. I hear what you're saying because it resonates

51
00:02:48,880 --> 00:02:50,320
Speaker 1: a lot with a lot of the stuff we talk

52
00:02:50,360 --> 00:02:53,080
Speaker 1: about here on tech stuff. It is to me one

53
00:02:53,120 --> 00:02:58,040
Speaker 1: of the key funny components about technology is how everyone

54
00:02:58,120 --> 00:03:01,840
Speaker 1: anticipates the next big technology development means that their jobs

55
00:03:01,840 --> 00:03:05,720
Speaker 1: are going to be way less labor intensive, and then

56
00:03:05,880 --> 00:03:09,840
Speaker 1: often it turns out like, well, sure each individual task

57
00:03:09,960 --> 00:03:12,840
Speaker 1: is easier, but now you're doing way more tasks because

58
00:03:12,880 --> 00:03:16,160
Speaker 1: everyone's more efficient. So if you remember, because I'm not

59
00:03:16,160 --> 00:03:17,600
Speaker 1: going to put an age on you, Jacob, but I

60
00:03:17,600 --> 00:03:19,239
Speaker 1: will say that I am of a certain age where

61
00:03:19,280 --> 00:03:23,400
Speaker 1: I remember the concept of the paperless office, yes, and

62
00:03:23,440 --> 00:03:26,320
Speaker 1: how we were going to get to this incredibly efficient thing,

63
00:03:26,360 --> 00:03:29,160
Speaker 1: and that maybe, like maybe your workday would be reduced

64
00:03:29,160 --> 00:03:31,440
Speaker 1: to maybe three hours a day. Turns out that that

65
00:03:31,680 --> 00:03:34,760
Speaker 1: was perhaps a bit idealistic on the part of the

66
00:03:34,800 --> 00:03:36,880
Speaker 1: workers and not the way it worked out.

67
00:03:37,480 --> 00:03:40,240
Speaker 2: Yes, although I mean there is a lot less paper

68
00:03:40,400 --> 00:03:43,000
Speaker 2: to start with us, right, Like I am old enough

69
00:03:43,080 --> 00:03:46,080
Speaker 2: to remember, like when I started working, everybody had like

70
00:03:46,120 --> 00:03:49,240
Speaker 2: a file drawer by their desk and like hanging files

71
00:03:49,280 --> 00:03:52,080
Speaker 2: with papers in them, so there is less paper. I mean,

72
00:03:52,200 --> 00:03:54,920
Speaker 2: the sort of less work thing is interesting, right, because

73
00:03:54,960 --> 00:03:56,839
Speaker 2: on the one hand, people are like, oh, I hope

74
00:03:56,880 --> 00:03:59,840
Speaker 2: technology you'll mean I have less work. By the same token,

75
00:03:59,840 --> 00:04:04,960
Speaker 2: people are like, oh I hope technology doesn't take my job, right, Yes,

76
00:04:05,400 --> 00:04:09,520
Speaker 2: In fact, the basic mechanism is like technology the happy

77
00:04:09,600 --> 00:04:13,160
Speaker 2: story anyways, the story that we hope happens is technology

78
00:04:13,200 --> 00:04:16,440
Speaker 2: makes us more productive, not so that we can work less,

79
00:04:16,480 --> 00:04:19,120
Speaker 2: but so that our output can be greater, right, Like

80
00:04:19,440 --> 00:04:23,080
Speaker 2: you know, I did not start working on podcasts and

81
00:04:23,200 --> 00:04:26,080
Speaker 2: radio in the real to real tape era, but I

82
00:04:26,120 --> 00:04:28,880
Speaker 2: know people who did, and like they talk about how

83
00:04:28,920 --> 00:04:32,440
Speaker 2: long it took to literally cut tape by hand, and

84
00:04:32,839 --> 00:04:34,880
Speaker 2: you can cut a lot more tape now that it's

85
00:04:34,920 --> 00:04:37,479
Speaker 2: not actual tape. Right, that's a productivity game.

86
00:04:37,760 --> 00:04:41,640
Speaker 1: Yes, totally. It's so funny to kind of see these

87
00:04:41,839 --> 00:04:44,800
Speaker 1: changes over time and the different perceptions that go into

88
00:04:44,839 --> 00:04:47,279
Speaker 1: whether we was it going to be like versus what

89
00:04:47,400 --> 00:04:50,040
Speaker 1: it actually turns out to be. We're gonna as I said,

90
00:04:50,080 --> 00:04:53,320
Speaker 1: talk about AI a lot, and to your point, one

91
00:04:53,360 --> 00:04:56,919
Speaker 1: of the things that I hear repeated about AI in

92
00:04:57,000 --> 00:05:00,640
Speaker 1: general and specifically within the realm of robotics and AI

93
00:05:01,160 --> 00:05:04,680
Speaker 1: is that its ideal role is to tackle tasks that

94
00:05:04,760 --> 00:05:08,000
Speaker 1: fall into the three d's, which are dirty, dangerous, and dull.

95
00:05:08,360 --> 00:05:11,440
Speaker 1: That these are technologies that are best suited to take

96
00:05:11,480 --> 00:05:15,200
Speaker 1: on jobs that are perhaps less desirable for humans for

97
00:05:15,279 --> 00:05:19,240
Speaker 1: various reasons, whether they could potentially cause injury, or they're

98
00:05:19,240 --> 00:05:23,840
Speaker 1: not very rewarding, that sort of thing, And I really

99
00:05:23,960 --> 00:05:28,680
Speaker 1: like that concept. The fear everyone has, obviously is that

100
00:05:28,760 --> 00:05:33,359
Speaker 1: it's tackling everything. It's non discriminate, it's not looking at

101
00:05:33,640 --> 00:05:37,440
Speaker 1: just the three d's, it's looking at every possible option.

102
00:05:37,600 --> 00:05:40,000
Speaker 1: And you're seeing a lot of discourse, at least I

103
00:05:40,040 --> 00:05:42,839
Speaker 1: am online where I see a lot of people saying,

104
00:05:43,400 --> 00:05:47,680
Speaker 1: why aren't we looking at how to automate c suite jobs?

105
00:05:47,800 --> 00:05:49,800
Speaker 1: Because it seems to me like a lot of the

106
00:05:49,960 --> 00:05:52,880
Speaker 1: duties that c suite executives have are ones that would

107
00:05:52,920 --> 00:05:56,240
Speaker 1: be best suited for some of the AI tools we're

108
00:05:56,600 --> 00:05:59,760
Speaker 1: talking about. Why are we talking about eliminating these lower

109
00:05:59,839 --> 00:06:02,640
Speaker 1: level jobs when some of these upper level ones particularly

110
00:06:02,680 --> 00:06:06,400
Speaker 1: when you have stories about c suite executives who perhaps

111
00:06:06,440 --> 00:06:09,640
Speaker 1: had a less than stellar run at the top of

112
00:06:09,680 --> 00:06:12,200
Speaker 1: the company ladder, like maybe the company didn't perform as

113
00:06:12,240 --> 00:06:14,680
Speaker 1: well as it should have, and yet they still retire

114
00:06:14,960 --> 00:06:18,400
Speaker 1: with these massive packages. So it's funny to see how

115
00:06:18,640 --> 00:06:23,080
Speaker 1: the perceptions of things like automation and AI are shaping

116
00:06:23,960 --> 00:06:25,760
Speaker 1: social discussions in that way.

117
00:06:26,240 --> 00:06:28,919
Speaker 2: Well, I think certainly. I mean, the rise of large

118
00:06:29,000 --> 00:06:33,120
Speaker 2: language models, you know lms like chat GPT have shifted

119
00:06:33,120 --> 00:06:35,400
Speaker 2: that conversation some, right. I think if you go back

120
00:06:35,440 --> 00:06:38,920
Speaker 2: five years, people thought about, you know, automating warehouse jobs.

121
00:06:38,960 --> 00:06:42,080
Speaker 2: But what I've seen in the last say year or

122
00:06:42,120 --> 00:06:46,680
Speaker 2: so since chat GPT, you know, stormed, our discourse is

123
00:06:46,800 --> 00:06:50,880
Speaker 2: people are talking a lot more about journalists and lawyers

124
00:06:51,279 --> 00:06:55,120
Speaker 2: being automated away, right, and plausibly to some extent, plausibly

125
00:06:55,160 --> 00:06:57,240
Speaker 2: to some margin. I mean. The other thing is like,

126
00:06:57,400 --> 00:07:00,160
Speaker 2: do you want a robot boss? When people talk about like,

127
00:07:00,200 --> 00:07:03,000
Speaker 2: I haven't heard about people talking about automating the c suite,

128
00:07:03,000 --> 00:07:06,400
Speaker 2: but like, sure, bad bosses are bad, and overpaying bad

129
00:07:06,440 --> 00:07:10,320
Speaker 2: bosses is bad. I actually think the role of a

130
00:07:10,360 --> 00:07:14,160
Speaker 2: good boss, of a good CEO is largely to be

131
00:07:14,200 --> 00:07:18,360
Speaker 2: a human being, right, is not fundamentally about you know,

132
00:07:18,440 --> 00:07:21,040
Speaker 2: assessing the data and making the best decision, although obviously

133
00:07:21,040 --> 00:07:24,680
Speaker 2: that's important. It's to be there to you know, talk

134
00:07:24,720 --> 00:07:26,960
Speaker 2: to people, essentially, to be in the room to tell

135
00:07:27,000 --> 00:07:29,120
Speaker 2: people that things are going to be okay. And that

136
00:07:29,320 --> 00:07:33,240
Speaker 2: seems like the set of domains that are less likely

137
00:07:33,280 --> 00:07:35,119
Speaker 2: to be automated, certainly in the short run.

138
00:07:35,560 --> 00:07:39,120
Speaker 1: I think you're absolutely right. I think it's based on again, perception, right.

139
00:07:39,200 --> 00:07:44,480
Speaker 1: A lot of people don't have like FaceTime with CEOs,

140
00:07:45,040 --> 00:07:48,960
Speaker 1: and so their perception is from a distance, and they're

141
00:07:49,000 --> 00:07:51,920
Speaker 1: looking at the effects that the CEO decisions are having

142
00:07:52,080 --> 00:07:54,920
Speaker 1: on a broad level, especially in the wake of something

143
00:07:55,000 --> 00:07:57,600
Speaker 1: like a round of layoffs, for example, Whereas you and

144
00:07:57,640 --> 00:08:00,440
Speaker 1: I have had the opportunity to speak face to face

145
00:08:00,720 --> 00:08:03,880
Speaker 1: and you find out very quickly these CEOs are human beings,

146
00:08:03,960 --> 00:08:06,120
Speaker 1: some more so than others. I mean, there are some.

147
00:08:06,720 --> 00:08:09,880
Speaker 1: There are some CEOs out there who I suspect maybe

148
00:08:09,920 --> 00:08:11,680
Speaker 1: at least part cyborg, but.

149
00:08:12,200 --> 00:08:13,239
Speaker 2: Could be had happen.

150
00:08:14,560 --> 00:08:18,000
Speaker 1: Yeah, yeah, most of them. I mean, I'm not sure

151
00:08:18,000 --> 00:08:22,120
Speaker 1: if Elon Musk is even robotic, he might be alien.

152
00:08:22,200 --> 00:08:24,800
Speaker 1: I'm not entirely certain. But there are a lot of

153
00:08:24,800 --> 00:08:27,160
Speaker 1: them out there. When you just have a short conversation

154
00:08:27,280 --> 00:08:30,160
Speaker 1: and you realize these aren't just talking points. For a

155
00:08:30,160 --> 00:08:34,199
Speaker 1: lot of these leaders, they sincerely believe their mission statement,

156
00:08:34,280 --> 00:08:37,840
Speaker 1: or they sincerely believe in the strategies that they're following,

157
00:08:37,880 --> 00:08:40,960
Speaker 1: and they sincerely feel bad when they have to make

158
00:08:41,000 --> 00:08:44,600
Speaker 1: decisions that lead to things like layoffs. But when you

159
00:08:44,679 --> 00:08:47,000
Speaker 1: are at more of a distance, I think it's easier

160
00:08:47,040 --> 00:08:52,040
Speaker 1: to kind of dehumanize the person. And it's understandable, right

161
00:08:52,120 --> 00:08:55,679
Speaker 1: you see those big effects and you just think this, Meanwhile,

162
00:08:55,679 --> 00:09:00,400
Speaker 1: the CEO is potentially making enormous amounts of money. I'm

163
00:09:00,440 --> 00:09:03,360
Speaker 1: reminded of a former CEO I worked for, not a

164
00:09:03,400 --> 00:09:06,440
Speaker 1: former CEO. He's still a CEO, but he's a find,

165
00:09:06,440 --> 00:09:09,440
Speaker 1: my former boss, David Zaslov. And whenever I see any

166
00:09:09,480 --> 00:09:12,480
Speaker 1: stories about him, I sit there and think, well, I've

167
00:09:12,520 --> 00:09:15,120
Speaker 1: met the man, I've had conversations with him. I know

168
00:09:15,440 --> 00:09:17,800
Speaker 1: a little bit more. I feel like I could give

169
00:09:17,800 --> 00:09:19,920
Speaker 1: a bit more perspective to this. But at the same time,

170
00:09:20,240 --> 00:09:24,439
Speaker 1: you're not entirely wrong with some of the conclusions you've drawn.

171
00:09:24,720 --> 00:09:26,920
Speaker 2: Yeah, and I mean, you know, to some extent, like

172
00:09:27,600 --> 00:09:29,240
Speaker 2: I mean the sub I feel like a lot of

173
00:09:29,280 --> 00:09:31,840
Speaker 2: the subtext of what you're talking about is inequality.

174
00:09:31,920 --> 00:09:32,120
Speaker 1: Right.

175
00:09:32,120 --> 00:09:34,760
Speaker 2: It's the gas between what CEOs make and whatever the

176
00:09:34,840 --> 00:09:38,480
Speaker 2: median worker at their company makes, and that indeed ballooned

177
00:09:38,520 --> 00:09:40,760
Speaker 2: out a lot at the end of the twentieth century

178
00:09:40,800 --> 00:09:43,600
Speaker 2: and has stayed quite wide obviously, and to some extent,

179
00:09:43,679 --> 00:09:47,719
Speaker 2: that is an effect of technology, right, although it's complicated,

180
00:09:47,760 --> 00:09:51,199
Speaker 2: there's a lot of it is norms. Right. Yeah, that's

181
00:09:51,240 --> 00:09:52,439
Speaker 2: a pretty subtle one.

182
00:09:53,200 --> 00:09:56,360
Speaker 1: Yeah, that's another thing that as communicators about things that

183
00:09:56,400 --> 00:09:58,920
Speaker 1: are in the technological space, often we do need to

184
00:09:58,960 --> 00:10:02,040
Speaker 1: take a step back away from just the technology and

185
00:10:02,160 --> 00:10:08,800
Speaker 1: acknowledge these other components that impact the entire direction of tech.

186
00:10:08,840 --> 00:10:10,960
Speaker 1: I mean, I'm sure as someone who has looked into

187
00:10:10,960 --> 00:10:14,640
Speaker 1: Silicon Valley you have seen how things like social norms

188
00:10:14,640 --> 00:10:18,719
Speaker 1: and politics and even things like living expenses in San

189
00:10:18,760 --> 00:10:22,040
Speaker 1: Francisco have a big impact on these sorts of things,

190
00:10:22,480 --> 00:10:25,480
Speaker 1: and people can get frustrated when I step back and

191
00:10:25,520 --> 00:10:28,560
Speaker 1: talk about these elements. But my argument is that you

192
00:10:28,600 --> 00:10:31,120
Speaker 1: can't really have a full understanding of technology unless you

193
00:10:31,160 --> 00:10:34,040
Speaker 1: also take into account these other things that do have

194
00:10:34,080 --> 00:10:35,959
Speaker 1: an impact. But one of the things I wanted to

195
00:10:35,960 --> 00:10:38,720
Speaker 1: ask you about so what's your problem. You are talking

196
00:10:38,720 --> 00:10:42,400
Speaker 1: with a lot of problem solvers, obviously, and I was curious,

197
00:10:42,760 --> 00:10:45,160
Speaker 1: now that you've spoken with quite a few people who

198
00:10:45,480 --> 00:10:47,960
Speaker 1: either come directly from engineering or have kind of an

199
00:10:48,000 --> 00:10:52,200
Speaker 1: engineering perspective, what's your take on engineers, Because I know

200
00:10:52,240 --> 00:10:54,720
Speaker 1: that's a general question and not everyone falls into the

201
00:10:54,760 --> 00:10:59,200
Speaker 1: same bucket, But I have a love for engineers in

202
00:10:59,240 --> 00:11:00,680
Speaker 1: the way that they approach things.

203
00:11:01,360 --> 00:11:04,800
Speaker 2: Yes, same, And you know, I also love engineers. And

204
00:11:04,840 --> 00:11:07,400
Speaker 2: as I mentioned, I was an English major. I am

205
00:11:07,440 --> 00:11:09,800
Speaker 2: not an engineer at all. I'm not even good at

206
00:11:09,840 --> 00:11:14,439
Speaker 2: fixing things around the house, although I try. But in college,

207
00:11:14,440 --> 00:11:18,840
Speaker 2: I took one computer science class my last term of college.

208
00:11:19,240 --> 00:11:22,280
Speaker 2: It was just the intro class. And on the first

209
00:11:22,360 --> 00:11:25,000
Speaker 2: day it was, you know, big lecture class. The professors

210
00:11:25,040 --> 00:11:27,440
Speaker 2: computer scientist was talking about his grading system and it

211
00:11:27,440 --> 00:11:28,959
Speaker 2: was this weird thing was like a check and a

212
00:11:29,080 --> 00:11:31,800
Speaker 2: check plus. And he said the top grade is a

213
00:11:31,800 --> 00:11:35,960
Speaker 2: plus plus and that is for code that makes me weep,

214
00:11:37,400 --> 00:11:40,000
Speaker 2: like cause it's so beautiful, you know, so erien. And

215
00:11:40,040 --> 00:11:42,400
Speaker 2: that was like a revelation to me because you know,

216
00:11:42,600 --> 00:11:45,560
Speaker 2: as an English major, as a non engineer, I always

217
00:11:45,600 --> 00:11:48,200
Speaker 2: thought of engineering as like, oh, a thing works or

218
00:11:48,200 --> 00:11:50,360
Speaker 2: it doesn't work. The building falls down or it doesn't

219
00:11:50,360 --> 00:11:53,679
Speaker 2: fall down. But this idea that there is elegance and

220
00:11:53,800 --> 00:11:57,640
Speaker 2: beauty in the construction of the thing itself was really

221
00:11:57,679 --> 00:12:00,520
Speaker 2: exciting to me and remains really exciting to me. And

222
00:12:00,840 --> 00:12:04,440
Speaker 2: engineers really are like that, you know, like they love

223
00:12:04,600 --> 00:12:08,760
Speaker 2: building things and they find beauty in an elegant solution,

224
00:12:09,360 --> 00:12:11,600
Speaker 2: the way other people find beauty in a song or

225
00:12:11,640 --> 00:12:12,840
Speaker 2: a poem or a painting.

226
00:12:13,280 --> 00:12:15,679
Speaker 1: I love that. I would say there's like a spectrum

227
00:12:15,760 --> 00:12:19,600
Speaker 1: in engineering as well, where you have sort of the artists,

228
00:12:19,640 --> 00:12:23,920
Speaker 1: who are the ones who are very carefully creating and

229
00:12:24,080 --> 00:12:27,960
Speaker 1: refining their code, and then maybe you have on the

230
00:12:28,000 --> 00:12:32,440
Speaker 1: punk rock side the hackers who didn't necessarily build the thing,

231
00:12:32,760 --> 00:12:34,880
Speaker 1: but they really want to know how the thing works,

232
00:12:34,920 --> 00:12:37,320
Speaker 1: and they will take the thing down to the very

233
00:12:38,160 --> 00:12:40,840
Speaker 1: base level of the structure, and then they'll say, what

234
00:12:40,920 --> 00:12:43,560
Speaker 1: if I rebuild it so it does something else? Like

235
00:12:44,000 --> 00:12:48,200
Speaker 1: I just think that entire culture, from the artists to

236
00:12:48,280 --> 00:12:51,560
Speaker 1: the punk rockers, who, hey, they're artists. I'm a punk

237
00:12:51,920 --> 00:12:54,280
Speaker 1: rock kind of fan myself, but I always find that

238
00:12:54,360 --> 00:12:57,240
Speaker 1: to be a wonderful way to have a conversation is

239
00:12:57,280 --> 00:12:59,880
Speaker 1: to talk about people and about how they're approach to

240
00:13:00,160 --> 00:13:02,480
Speaker 1: sort of stuff. And I also always say whenever I

241
00:13:02,480 --> 00:13:04,520
Speaker 1: talk with engineers, I come away with the feeling that

242
00:13:05,080 --> 00:13:07,920
Speaker 1: they view the world, as you can think of it,

243
00:13:07,960 --> 00:13:10,400
Speaker 1: as either a set of problems or a set of challenges,

244
00:13:10,559 --> 00:13:15,480
Speaker 1: and they're constantly thinking about solutions, which is nice because

245
00:13:15,559 --> 00:13:17,520
Speaker 1: I am unfortunately one of those people who's far more

246
00:13:17,600 --> 00:13:19,440
Speaker 1: likely to point out a problem but not have a

247
00:13:19,440 --> 00:13:22,720
Speaker 1: solution for you, right. So to talk to someone who's

248
00:13:22,760 --> 00:13:26,520
Speaker 1: already thinking ahead about how to solve the problem, not

249
00:13:26,640 --> 00:13:29,200
Speaker 1: just that there is a problem, I always find that

250
00:13:29,320 --> 00:13:30,040
Speaker 1: very inspiring.

251
00:13:30,559 --> 00:13:33,000
Speaker 2: Yeah, it's nice. I guess I also have tended to

252
00:13:33,040 --> 00:13:35,760
Speaker 2: be problem focused, you know. I sort of came up

253
00:13:35,760 --> 00:13:38,920
Speaker 2: in my career as a journalist, which is essentially all

254
00:13:38,960 --> 00:13:42,200
Speaker 2: about pointing out problems right to a significant degree if

255
00:13:42,280 --> 00:13:45,760
Speaker 2: you read the paper. Basically, what's going on in many

256
00:13:45,840 --> 00:13:48,440
Speaker 2: many stories is here is a thing that is bad,

257
00:13:50,000 --> 00:13:53,520
Speaker 2: and so talking to people who are trying to make

258
00:13:53,559 --> 00:13:56,440
Speaker 2: things better or fix things is great. And you know,

259
00:13:56,559 --> 00:13:59,720
Speaker 2: to be clear, I don't want to be too Pollyanna

260
00:13:59,760 --> 00:14:03,240
Speaker 2: issue here, Like there are plenty of engineers who build

261
00:14:03,280 --> 00:14:05,959
Speaker 2: things that on net don't help the world right, Building

262
00:14:06,200 --> 00:14:08,719
Speaker 2: new things is not always helpful to the world, and

263
00:14:08,920 --> 00:14:13,199
Speaker 2: there are certainly engineers who become enamored of just building

264
00:14:13,240 --> 00:14:15,240
Speaker 2: the thing and don't think about what it might mean.

265
00:14:15,559 --> 00:14:19,280
Speaker 2: And frankly, you know, in choosing who to talk to

266
00:14:19,400 --> 00:14:21,840
Speaker 2: for the show, I do try and talk to people

267
00:14:21,880 --> 00:14:25,120
Speaker 2: who I think are some combination of well cognizant of

268
00:14:25,120 --> 00:14:27,760
Speaker 2: what they're building and actually trying to do a good thing,

269
00:14:27,920 --> 00:14:30,840
Speaker 2: and you know, aware of the fact that there might

270
00:14:30,880 --> 00:14:34,640
Speaker 2: be unintended bad consequences of the thing that they're building.

271
00:14:35,320 --> 00:14:37,640
Speaker 2: Maybe I don't always succeed, but it is a useful

272
00:14:38,200 --> 00:14:39,640
Speaker 2: frame totally.

273
00:14:39,960 --> 00:14:43,280
Speaker 1: I also tend to think about how when engineers build

274
00:14:43,320 --> 00:14:47,240
Speaker 1: things that make sense to them, assuming that's they're building

275
00:14:47,240 --> 00:14:49,160
Speaker 1: something that ultimately is supposed to be used by the

276
00:14:49,200 --> 00:14:51,960
Speaker 1: general public, the great ones will take into account how

277
00:14:52,480 --> 00:14:57,000
Speaker 1: a quote unquote normal person would approach the whatever it is.

278
00:14:57,280 --> 00:14:57,360
Speaker 2: Like.

279
00:14:57,400 --> 00:15:01,040
Speaker 1: I'm thinking of user interfaces in particular, and that making

280
00:15:01,080 --> 00:15:03,240
Speaker 1: sure that the user interface is going to make sense

281
00:15:03,280 --> 00:15:06,520
Speaker 1: to enore me as opposed to an engineer. And then

282
00:15:06,560 --> 00:15:10,560
Speaker 1: there are other engineers or in fact entire companies where

283
00:15:10,560 --> 00:15:13,680
Speaker 1: they will build things that work great. If you're an

284
00:15:13,720 --> 00:15:17,200
Speaker 1: engineer they're fantastic if you're an engineer. If you're not

285
00:15:17,280 --> 00:15:21,040
Speaker 1: an engineer, it may require a bit more work on

286
00:15:21,080 --> 00:15:24,360
Speaker 1: your part. I'm looking specifically at Google and the Android

287
00:15:24,360 --> 00:15:27,680
Speaker 1: operating System because I'm an Android user, but at the

288
00:15:27,680 --> 00:15:32,080
Speaker 1: same time, I fully recognize that iOS is an operating

289
00:15:32,120 --> 00:15:35,880
Speaker 1: system that is so intuitive. You can literally hand an

290
00:15:35,920 --> 00:15:38,760
Speaker 1: iOS device to a child and they will have it

291
00:15:38,840 --> 00:15:41,320
Speaker 1: figured out in no time. You can hand an Android

292
00:15:41,360 --> 00:15:44,120
Speaker 1: device to someone and they will spend a lot of

293
00:15:44,120 --> 00:15:46,640
Speaker 1: time asking questions about how to do things and how

294
00:15:46,640 --> 00:15:49,880
Speaker 1: to access things. And it's not that the Android operating

295
00:15:49,880 --> 00:15:52,160
Speaker 1: system is bad, it's not that it's worse than iOS.

296
00:15:52,320 --> 00:15:54,840
Speaker 1: If you happen to be an engineer, Android is awesome,

297
00:15:55,320 --> 00:15:57,600
Speaker 1: and if you're not, it's still awesome. But you have

298
00:15:57,680 --> 00:16:01,120
Speaker 1: to put in work to get to realize that. And

299
00:16:01,160 --> 00:16:04,680
Speaker 1: it really differentiates the two, right because Apple has always

300
00:16:04,760 --> 00:16:08,600
Speaker 1: had a focus on how can we make this into

301
00:16:08,640 --> 00:16:12,600
Speaker 1: a product that people realize they need, even if they

302
00:16:12,640 --> 00:16:15,600
Speaker 1: never had that need before. And Google's like, how can

303
00:16:15,640 --> 00:16:17,960
Speaker 1: we make this so that it's really powerful and that

304
00:16:18,040 --> 00:16:19,960
Speaker 1: it does what we wanted to do? But it may

305
00:16:20,040 --> 00:16:22,680
Speaker 1: require a little bit of work on the user's part.

306
00:16:22,720 --> 00:16:25,480
Speaker 1: In order to have it work out. So that's sort

307
00:16:25,520 --> 00:16:28,680
Speaker 1: of also a fascinating thing about engineering that I've really

308
00:16:29,240 --> 00:16:31,920
Speaker 1: loved to look at and to talk about. I don't

309
00:16:31,960 --> 00:16:35,040
Speaker 1: necessarily think one is better than the other, apart from

310
00:16:35,040 --> 00:16:37,160
Speaker 1: the fact that one is just much easier for the

311
00:16:37,240 --> 00:16:40,480
Speaker 1: general public to kind of glom onto. As much as

312
00:16:40,520 --> 00:16:43,280
Speaker 1: I love Android, I would never say that it's more

313
00:16:43,360 --> 00:16:44,680
Speaker 1: user friendly than iOS.

314
00:16:46,000 --> 00:16:48,720
Speaker 2: Yeah, I mean there's a few ways of thing about that, right,

315
00:16:48,800 --> 00:16:51,120
Speaker 2: Like somebody a long time ago told me there's a

316
00:16:51,160 --> 00:16:54,680
Speaker 2: phrase people use sometimes the user is never wrong, right,

317
00:16:54,840 --> 00:16:57,080
Speaker 2: which is which is an interesting framework. And I think

318
00:16:57,160 --> 00:16:58,800
Speaker 2: they told it to me to be nice, because I

319
00:16:58,840 --> 00:17:01,240
Speaker 2: was like trying to do something in a radio studio

320
00:17:01,280 --> 00:17:02,840
Speaker 2: and I couldn't figure it out. And it was an

321
00:17:02,840 --> 00:17:05,440
Speaker 2: engineer who was like a thoughtful guy. I said, no, no,

322
00:17:05,480 --> 00:17:08,000
Speaker 2: you're not bad at this, it's just not set up well.

323
00:17:08,119 --> 00:17:10,480
Speaker 2: I mean, the other way of thinking about it, in

324
00:17:10,560 --> 00:17:15,760
Speaker 2: terms of the Android versus io ask question is as

325
00:17:15,800 --> 00:17:18,520
Speaker 2: an optimization problem. Right, if you're an engineer, then the

326
00:17:18,560 --> 00:17:21,360
Speaker 2: question is, well, are we optimizing for sort of the

327
00:17:21,880 --> 00:17:24,240
Speaker 2: mobile operating system that can do the most things or

328
00:17:24,280 --> 00:17:26,560
Speaker 2: be the most flexible, or the mobile operating system that

329
00:17:26,680 --> 00:17:28,920
Speaker 2: is just like bulletproof. You can give it to anybody

330
00:17:28,920 --> 00:17:31,399
Speaker 2: in any language and they will immediately understand what it

331
00:17:31,480 --> 00:17:34,960
Speaker 2: is and you get different sort of solutions depending on

332
00:17:34,960 --> 00:17:36,000
Speaker 2: what you're optimizing for.

333
00:17:36,359 --> 00:17:39,280
Speaker 1: Yeah, that's true. Like you've identified whatever your goal is,

334
00:17:39,320 --> 00:17:43,040
Speaker 1: so obviously the execution is going to be different. Well,

335
00:17:43,080 --> 00:17:45,040
Speaker 1: I'm glad that you weren't told that the problem was

336
00:17:45,040 --> 00:17:49,680
Speaker 1: between keyboard and chair, which is the other classic classic answer.

337
00:17:49,720 --> 00:17:50,960
Speaker 2: Wait a minute, that's me.

338
00:17:51,560 --> 00:17:55,080
Speaker 1: Yeah, I've received that particular one more than more than

339
00:17:55,119 --> 00:17:58,159
Speaker 1: once in my lifetime. Well, now you have a reply.

340
00:17:58,520 --> 00:18:02,880
Speaker 1: Now you have to reply yes, yeah, yes, that's true.

341
00:18:02,880 --> 00:18:06,160
Speaker 1: The user is never wrong. We've got a lot more

342
00:18:06,240 --> 00:18:08,920
Speaker 1: to talk about, Jacob and I, but before we get

343
00:18:08,920 --> 00:18:10,840
Speaker 1: to the rest of our conversation, we need to take

344
00:18:10,880 --> 00:18:23,800
Speaker 1: a quick break to thank our sponsors. Your show covers

345
00:18:24,080 --> 00:18:28,119
Speaker 1: all realms of technology, not just AI, but because this

346
00:18:28,240 --> 00:18:32,720
Speaker 1: past year has been undeniably AI centric when it comes

347
00:18:32,760 --> 00:18:35,600
Speaker 1: to tech news, clearly a lot of your episodes do

348
00:18:35,880 --> 00:18:40,000
Speaker 1: tackle AI, and as someone else who tries to communicate

349
00:18:40,040 --> 00:18:43,479
Speaker 1: technology in a way that's really accessible and understandable, one

350
00:18:43,520 --> 00:18:46,760
Speaker 1: of the things I frequently run into is that talking

351
00:18:46,800 --> 00:18:51,600
Speaker 1: about AI in a responsible way is in itself challenging.

352
00:18:52,080 --> 00:18:54,920
Speaker 1: But I'm curious to hear what your take is when

353
00:18:54,960 --> 00:18:57,440
Speaker 1: it comes time for you to communicate about AI. How

354
00:18:57,480 --> 00:18:59,719
Speaker 1: do you perceive that and how do you approach it?

355
00:19:00,200 --> 00:19:02,280
Speaker 2: A thing in general that I try and do on

356
00:19:02,320 --> 00:19:06,320
Speaker 2: my show and certainly with respect to AI, is to

357
00:19:06,440 --> 00:19:10,720
Speaker 2: go narrow essentially right, Like I'm not going on my

358
00:19:10,760 --> 00:19:13,560
Speaker 2: show and saying here is what AI is and here's

359
00:19:13,600 --> 00:19:15,960
Speaker 2: what's going to happen. I'm talking to people who are

360
00:19:16,200 --> 00:19:21,200
Speaker 2: typically building a company to do a specific thing with AI.

361
00:19:21,760 --> 00:19:25,280
Speaker 2: Not quite always, but usually right. And so that to

362
00:19:25,400 --> 00:19:28,960
Speaker 2: me is a helpful way to well, AA say something

363
00:19:29,040 --> 00:19:31,680
Speaker 2: new because so many people are making so many broad

364
00:19:31,680 --> 00:19:36,080
Speaker 2: statements about AI and b steer clear of you know,

365
00:19:36,640 --> 00:19:40,400
Speaker 2: over generalization. How about you, I mean, what's your take?

366
00:19:41,080 --> 00:19:44,760
Speaker 1: So my concern is I always want to avoid being

367
00:19:44,760 --> 00:19:49,560
Speaker 1: reductive because AI is such a huge discipline and it

368
00:19:49,640 --> 00:19:52,919
Speaker 1: involves so many different aspects, that is very easy to

369
00:19:52,960 --> 00:19:55,200
Speaker 1: fall into the trap because I mean, clearly we see

370
00:19:55,200 --> 00:19:57,200
Speaker 1: this in mainstream media all the time. Not that I

371
00:19:57,280 --> 00:20:00,440
Speaker 1: blame them, but they're just they're taking some shortcuts where

372
00:20:00,440 --> 00:20:03,880
Speaker 1: they'll use the term artificial intelligence. It almost implies that

373
00:20:04,040 --> 00:20:06,120
Speaker 1: what they're talking about is the end all be all

374
00:20:06,160 --> 00:20:09,200
Speaker 1: of artificial intelligence. And usually they're talking about generative AI.

375
00:20:09,440 --> 00:20:11,480
Speaker 1: In the last year, I would say, like, that's been

376
00:20:11,760 --> 00:20:15,639
Speaker 1: the biggest topic in artificial intelligence, but it's one topic,

377
00:20:15,880 --> 00:20:19,320
Speaker 1: and that AI actually covers a lot more than that,

378
00:20:19,640 --> 00:20:22,639
Speaker 1: and it falls into so many other buckets too, Like

379
00:20:23,040 --> 00:20:25,879
Speaker 1: robotics obviously has a lot to do with AI. That

380
00:20:26,040 --> 00:20:30,120
Speaker 1: not always you can have a fully remote controlled robot,

381
00:20:30,280 --> 00:20:35,040
Speaker 1: but often there's some AI components there, things like assisted vision,

382
00:20:35,640 --> 00:20:38,880
Speaker 1: brain computer interfaces. I mean, there's so many different things

383
00:20:38,920 --> 00:20:41,200
Speaker 1: that don't have anything to do with generative AI. That's

384
00:20:41,240 --> 00:20:43,240
Speaker 1: still at least touch on AI.

385
00:20:43,600 --> 00:20:46,640
Speaker 2: Doing a Google search, yeah, like exactly.

386
00:20:46,800 --> 00:20:50,959
Speaker 1: Yeah, anything that you're getting into, like automation, I mean

387
00:20:51,040 --> 00:20:51,600
Speaker 1: you can get a.

388
00:20:51,600 --> 00:20:55,879
Speaker 2: Lot recommending a show to you. Yeah, blind spot monitoring

389
00:20:55,920 --> 00:20:59,119
Speaker 2: system saying there's a cardex too. Like these are all side.

390
00:20:59,200 --> 00:21:01,520
Speaker 1: The Yeah, they're all they're all different aspects of AI.

391
00:21:01,680 --> 00:21:03,440
Speaker 1: And like you could argue, well, sometimes you get a

392
00:21:03,440 --> 00:21:05,840
Speaker 1: little fuzzy. I'm like, well, so is the word intelligence?

393
00:21:05,880 --> 00:21:08,159
Speaker 1: Like intelligence itself is a fuzzy term. I thought that.

394
00:21:08,160 --> 00:21:11,960
Speaker 2: Would be better off if the AI did not exist

395
00:21:12,480 --> 00:21:14,919
Speaker 2: me too, like I think it's an unfortunate choice of

396
00:21:14,960 --> 00:21:17,000
Speaker 2: words that is unhelpful.

397
00:21:17,119 --> 00:21:20,440
Speaker 1: Ultimately, it's I think largely because everyone starts to jump

398
00:21:20,520 --> 00:21:23,400
Speaker 1: to They jump to science fiction, they jump to Skynet,

399
00:21:23,560 --> 00:21:26,280
Speaker 1: they jump to Terminator, they jump to this idea of

400
00:21:26,840 --> 00:21:30,320
Speaker 1: something that appears to think the way humans do. And

401
00:21:30,400 --> 00:21:32,520
Speaker 1: of course we don't even know if we ever reach

402
00:21:33,240 --> 00:21:36,320
Speaker 1: strong AI or general AI, however you want to define it.

403
00:21:36,600 --> 00:21:39,359
Speaker 1: We don't know if it's going to quote unquote think

404
00:21:39,440 --> 00:21:41,960
Speaker 1: like a human, or even think at all. It may

405
00:21:42,080 --> 00:21:46,960
Speaker 1: just be indistinguishable to us from the way humans think.

406
00:21:47,320 --> 00:21:50,240
Speaker 1: And I think Touring would argue, well, that's good enough.

407
00:21:50,880 --> 00:21:53,520
Speaker 1: It doesn't matter. If it's indistinguishable, then it might as

408
00:21:53,520 --> 00:21:53,840
Speaker 1: well be.

409
00:21:54,080 --> 00:21:57,840
Speaker 2: You quoted somebody on your show a while back as

410
00:21:57,920 --> 00:22:03,200
Speaker 2: saying something like intelligence is whatever machines can't do yet,

411
00:22:03,880 --> 00:22:05,440
Speaker 2: which I thought was pretty good.

412
00:22:06,119 --> 00:22:08,399
Speaker 1: Yeah. I think that it's very similar to how a

413
00:22:08,400 --> 00:22:12,200
Speaker 1: lot of philosophers define consciousness, right They say, like, ugh,

414
00:22:12,600 --> 00:22:13,120
Speaker 1: we don't.

415
00:22:12,960 --> 00:22:13,840
Speaker 2: Know what consciousness is.

416
00:22:14,200 --> 00:22:17,199
Speaker 1: All we do. All we do is we define what

417
00:22:17,320 --> 00:22:20,000
Speaker 1: consciousness isn't we haven't gotten to a point where we

418
00:22:20,000 --> 00:22:22,480
Speaker 1: can say what consciousness is. We get we chip away

419
00:22:22,520 --> 00:22:25,720
Speaker 1: So what we're doing is we've got the marble slab, yeah,

420
00:22:25,760 --> 00:22:28,520
Speaker 1: and we're chipping at it, but we haven't yet seen

421
00:22:28,560 --> 00:22:32,000
Speaker 1: the statue that's living under the slab. Yet we're still

422
00:22:32,040 --> 00:22:32,560
Speaker 1: just chipping.

423
00:22:32,880 --> 00:22:35,840
Speaker 2: I was already worried about intelligence. As you say consciousness,

424
00:22:35,880 --> 00:22:37,800
Speaker 2: I like, I don't even know what to do with

425
00:22:37,880 --> 00:22:38,399
Speaker 2: that one.

426
00:22:38,680 --> 00:22:41,199
Speaker 1: Oh no, well, I mean it often goes hand in

427
00:22:41,240 --> 00:22:44,520
Speaker 1: hand with AI, right, because people immediately assure that intelligence

428
00:22:44,680 --> 00:22:47,800
Speaker 1: and self awareness go hand in hand with one another,

429
00:22:47,840 --> 00:22:50,119
Speaker 1: and maybe it will. We don't know. That's the point.

430
00:22:50,200 --> 00:22:53,520
Speaker 1: We don't know. But long story short, to answer the

431
00:22:53,600 --> 00:22:56,560
Speaker 1: question of how I approach this, I usually go where

432
00:22:56,600 --> 00:22:59,760
Speaker 1: I start from the broad foundation and then I go narrow.

433
00:22:59,840 --> 00:23:02,719
Speaker 1: So I start with saying, first we need to acknowledge

434
00:23:02,800 --> 00:23:06,199
Speaker 1: that artificial intelligence is a very very big field, and

435
00:23:06,240 --> 00:23:09,280
Speaker 1: that this is one aspect of AI, and that we're

436
00:23:09,320 --> 00:23:11,160
Speaker 1: not going to talk about the other aspects of AI,

437
00:23:11,240 --> 00:23:14,560
Speaker 1: but we need to remember they exist, and that while

438
00:23:14,600 --> 00:23:16,600
Speaker 1: the thing we're talking about is important, and while it

439
00:23:16,640 --> 00:23:19,600
Speaker 1: has its own set of challenges and potential, you know,

440
00:23:19,680 --> 00:23:22,880
Speaker 1: rewards and risks and all the things that go with it,

441
00:23:22,880 --> 00:23:25,679
Speaker 1: it's one part. It's like if you were to say,

442
00:23:25,720 --> 00:23:28,359
Speaker 1: you wouldn't hold up a remote control and say this

443
00:23:28,480 --> 00:23:31,400
Speaker 1: is all of technology. Right, This is one thing that's

444
00:23:31,400 --> 00:23:34,920
Speaker 1: a technological gadget, but it doesn't represent all of technology.

445
00:23:35,000 --> 00:23:37,760
Speaker 2: First of all, you'd have to find it, right, that's.

446
00:23:37,560 --> 00:23:39,879
Speaker 1: True, which you know, you know, you got to have

447
00:23:40,000 --> 00:23:42,040
Speaker 1: some sort of method to figure out where it is.

448
00:23:42,200 --> 00:23:44,200
Speaker 1: This is, by the way, is why three D television

449
00:23:44,280 --> 00:23:46,560
Speaker 1: never became a thing. Who wants to look for glasses

450
00:23:46,600 --> 00:23:49,879
Speaker 1: so that they can watch True Detective season four? Not me.

451
00:23:51,280 --> 00:23:54,080
Speaker 1: So yeah, that's kind of my approach. And so I

452
00:23:54,080 --> 00:23:57,040
Speaker 1: don't think that our approaches are that different. I think

453
00:23:57,119 --> 00:23:59,760
Speaker 1: that we're pretty similar. And it's that I think we

454
00:24:00,080 --> 00:24:03,520
Speaker 1: both feel there's a responsibility to make certain that we

455
00:24:03,640 --> 00:24:09,080
Speaker 1: never overgeneralize or be reductive, because that feeds into a

456
00:24:09,280 --> 00:24:14,800
Speaker 1: narrative that I think actually contributes to the old fud

457
00:24:14,880 --> 00:24:17,960
Speaker 1: the fear, uncertainty, in doubt. And while there are things

458
00:24:18,000 --> 00:24:21,280
Speaker 1: to certainly be concerned about and to be aware of,

459
00:24:21,720 --> 00:24:25,200
Speaker 1: we don't want to rush into anything, you know, with

460
00:24:26,160 --> 00:24:29,359
Speaker 1: a poor understanding of the situation. I think that's true

461
00:24:29,359 --> 00:24:33,280
Speaker 1: whether you're you know, really enthusiastic and excited about AI,

462
00:24:33,720 --> 00:24:36,920
Speaker 1: I think it's true if you are really concerned about AI.

463
00:24:37,119 --> 00:24:40,560
Speaker 1: I think, you know, taking critical thinking and a really

464
00:24:40,680 --> 00:24:45,439
Speaker 1: methodical approach is absolutely key if you want to avoid pitfalls.

465
00:24:46,040 --> 00:24:48,919
Speaker 2: Sure, it seems hard to argue against critical thinking and

466
00:24:48,960 --> 00:24:51,199
Speaker 2: a methodical approach, right, Who's going to take the other

467
00:24:51,280 --> 00:24:52,920
Speaker 2: side of that one?

468
00:24:53,280 --> 00:24:55,160
Speaker 1: I mean, Mark.

469
00:24:55,400 --> 00:24:57,520
Speaker 2: Fair enough, everybody, ye, fair enough?

470
00:24:57,920 --> 00:24:59,200
Speaker 1: Like Sam Altman maybe?

471
00:25:00,760 --> 00:25:03,400
Speaker 2: I mean, so it is interesting to think about the

472
00:25:03,720 --> 00:25:06,560
Speaker 2: Sam Altman, the head of Open AI. I mean, one

473
00:25:06,560 --> 00:25:09,399
Speaker 2: of the really interesting things to me about AI. And

474
00:25:09,440 --> 00:25:13,240
Speaker 2: that seems different in particular when you know, we're talking

475
00:25:13,280 --> 00:25:16,440
Speaker 2: about engineers, Like, I feel like the extent to which

476
00:25:16,480 --> 00:25:20,119
Speaker 2: the engineers working on AI are worried about AI is

477
00:25:20,200 --> 00:25:24,000
Speaker 2: really interesting and different, right, I feel like the traditional

478
00:25:24,440 --> 00:25:27,080
Speaker 2: kind of engineer stance is like, this thing is cool,

479
00:25:27,200 --> 00:25:30,879
Speaker 2: let's build it right again. That's obviously reductive and somewhat unfair,

480
00:25:30,920 --> 00:25:37,760
Speaker 2: but whatever. Whereas with AI, to some significant degree, many

481
00:25:37,800 --> 00:25:39,760
Speaker 2: of the people who are most worried about it are

482
00:25:39,800 --> 00:25:42,000
Speaker 2: the people who understand it the best. And you know,

483
00:25:42,040 --> 00:25:44,400
Speaker 2: I've heard people are you like, oh, that's just marketing,

484
00:25:44,480 --> 00:25:46,760
Speaker 2: and that doesn't seem true to me. First of all,

485
00:25:46,920 --> 00:25:48,720
Speaker 2: why would you market a thing by saying we should

486
00:25:48,720 --> 00:25:50,560
Speaker 2: be worried about it. And second of all, if you

487
00:25:50,640 --> 00:25:53,119
Speaker 2: just look like open ai was started as a nonprofit

488
00:25:53,200 --> 00:25:55,000
Speaker 2: and then they you know, needed more money, but they

489
00:25:55,000 --> 00:25:58,320
Speaker 2: became this weird capped profit model, and then people left

490
00:25:58,359 --> 00:26:00,639
Speaker 2: open ai to start anthropic, which is another one of

491
00:26:00,680 --> 00:26:03,399
Speaker 2: the big ones, because they thought open ai wasn't worried enough.

492
00:26:03,520 --> 00:26:06,119
Speaker 2: And then you had, you know, people calling like Elon

493
00:26:06,240 --> 00:26:08,720
Speaker 2: Musk calling for a six month pause on AI development.

494
00:26:08,800 --> 00:26:11,240
Speaker 2: And so I do think that people who know a

495
00:26:11,240 --> 00:26:14,040
Speaker 2: lot of it about AI are in fact really worried

496
00:26:14,040 --> 00:26:17,520
Speaker 2: about it, which is just interesting on its face, indifferent

497
00:26:17,560 --> 00:26:19,679
Speaker 2: than the way technology often works.

498
00:26:20,000 --> 00:26:22,240
Speaker 1: Yeah, there are a lot of conspiracy theories that pop

499
00:26:22,320 --> 00:26:24,400
Speaker 1: up or fringe theories that pop up around this. By

500
00:26:24,400 --> 00:26:27,120
Speaker 1: the way, like you have the ones who say, well,

501
00:26:27,359 --> 00:26:29,760
Speaker 1: they say they're worried about AI because what they're trying

502
00:26:29,800 --> 00:26:34,240
Speaker 1: to do is shape the discussions around regulations so that

503
00:26:34,440 --> 00:26:38,199
Speaker 1: their own personal organization ends up benefiting from those regulations

504
00:26:38,240 --> 00:26:41,760
Speaker 1: while those same regulations slow down smaller companies that are

505
00:26:41,760 --> 00:26:44,320
Speaker 1: in the space. You had people saying, well, Elon Musk, yes,

506
00:26:44,359 --> 00:26:46,080
Speaker 1: he was arguing that there needs to be a pause,

507
00:26:46,119 --> 00:26:48,280
Speaker 1: but it's because he was launching his own AI company,

508
00:26:48,320 --> 00:26:50,280
Speaker 1: and he wanted a chance to be able to catch up.

509
00:26:50,320 --> 00:26:52,679
Speaker 1: Like You've got a lot of other fringe theories out there,

510
00:26:52,800 --> 00:26:55,400
Speaker 1: and I understand that there may be, you know, some

511
00:26:55,560 --> 00:26:58,160
Speaker 1: credibility to some of those who knows. But I think

512
00:26:58,240 --> 00:27:00,679
Speaker 1: when it gets to a point where the board of

513
00:27:00,720 --> 00:27:05,880
Speaker 1: directors of open Ai get together and decide seemingly spontaneously

514
00:27:06,400 --> 00:27:08,480
Speaker 1: that they're going to get rid of the CEO and

515
00:27:08,560 --> 00:27:13,439
Speaker 1: co founder of the company and then do so, I

516
00:27:13,480 --> 00:27:17,600
Speaker 1: think that speaks to a genuine and sincere concern that

517
00:27:17,760 --> 00:27:21,640
Speaker 1: perhaps the organization is moving in a direction that they

518
00:27:21,680 --> 00:27:26,840
Speaker 1: feel as fundamentally counteractive to what they had intended. And

519
00:27:26,920 --> 00:27:30,560
Speaker 1: of course we know they subsequently had to reverse that

520
00:27:30,640 --> 00:27:33,800
Speaker 1: decision and step down from the board of directors because

521
00:27:33,800 --> 00:27:37,719
Speaker 1: the overwhelming support within the organization was to that co

522
00:27:37,840 --> 00:27:43,520
Speaker 1: founder and CEO seemingly embracing this new approach toward developing

523
00:27:43,560 --> 00:27:47,560
Speaker 1: AI that was a departure from the original organization's intent.

524
00:27:48,119 --> 00:27:50,399
Speaker 1: But to your point, the fact that the board of

525
00:27:50,400 --> 00:27:54,280
Speaker 1: directors was willing to do such an extreme move, even

526
00:27:54,280 --> 00:27:55,600
Speaker 1: though it was on a Friday at the end of

527
00:27:55,640 --> 00:27:58,320
Speaker 1: a news cycle, even that they were willing to do

528
00:27:58,400 --> 00:28:02,040
Speaker 1: that knowing that it would lead to them having to

529
00:28:02,119 --> 00:28:05,239
Speaker 1: leave the organization. I think that speaks to me a

530
00:28:05,280 --> 00:28:09,160
Speaker 1: genuine concern. You don't go and remove a co founder

531
00:28:09,200 --> 00:28:13,360
Speaker 1: and CEO for a small reason.

532
00:28:14,280 --> 00:28:18,199
Speaker 2: Yeah, I mean sure, I'm sure that some amount of

533
00:28:18,240 --> 00:28:20,959
Speaker 2: the public worrying over AI by people in the field

534
00:28:21,280 --> 00:28:26,880
Speaker 2: is some kind of self interested behavior. But I think overall,

535
00:28:27,400 --> 00:28:30,399
Speaker 2: there are clearly a lot of people who know a

536
00:28:30,440 --> 00:28:33,320
Speaker 2: lot who are even building these things, who are genuinely worried.

537
00:28:33,359 --> 00:28:34,879
Speaker 2: Like that seems obviously true.

538
00:28:35,240 --> 00:28:40,080
Speaker 1: Yeah, And honestly, like anytime someone's building a technology and

539
00:28:40,160 --> 00:28:43,160
Speaker 1: they're they're bringing concerns up to me, that's a good thing.

540
00:28:43,360 --> 00:28:46,480
Speaker 1: And it doesn't necessarily mean that the technology is ultimately

541
00:28:47,120 --> 00:28:50,520
Speaker 1: harmful or not beneficial. But you know, I think it's

542
00:28:50,520 --> 00:28:54,160
Speaker 1: a responsible person who does ask those questions. For one thing,

543
00:28:55,040 --> 00:28:57,280
Speaker 1: it really can save you a lot of time and

544
00:28:57,400 --> 00:29:00,360
Speaker 1: heartache further down the line, if you're tackling these kinds

545
00:29:00,360 --> 00:29:04,920
Speaker 1: of things before they've escalated to a point where they're

546
00:29:04,960 --> 00:29:09,560
Speaker 1: actively causing catastrophe. So I like seeing that. Whether it

547
00:29:09,640 --> 00:29:12,320
Speaker 1: ends up being merited or not, well, that's just sort

548
00:29:12,360 --> 00:29:14,160
Speaker 1: of a curse we have to bear, right, What if

549
00:29:14,640 --> 00:29:17,200
Speaker 1: it turns out that it was never merited, we won't know.

550
00:29:17,640 --> 00:29:19,560
Speaker 1: And if it turns out if it was merited, we

551
00:29:19,600 --> 00:29:21,880
Speaker 1: still don't know. Because they asked the questions ahead of

552
00:29:21,880 --> 00:29:24,640
Speaker 1: time and fix the problems before they became problems. And

553
00:29:24,880 --> 00:29:27,480
Speaker 1: it's only if we take the other path that we

554
00:29:27,520 --> 00:29:30,520
Speaker 1: find out for sure, like WHOA, we should have thought

555
00:29:30,520 --> 00:29:32,080
Speaker 1: of this before we did it.

556
00:29:32,880 --> 00:29:35,880
Speaker 2: I mean, you know, tools are complicated, right. People come

557
00:29:35,960 --> 00:29:38,760
Speaker 2: up with new tools, and then other people use those

558
00:29:38,800 --> 00:29:42,080
Speaker 2: tools in various ways, some of which enhance human well

559
00:29:42,120 --> 00:29:45,960
Speaker 2: being and some of which cause new miseries. And plainly

560
00:29:46,160 --> 00:29:46,960
Speaker 2: AI will do.

561
00:29:46,960 --> 00:29:50,200
Speaker 1: Both, yes, and so it really becomes important that we

562
00:29:50,280 --> 00:29:53,240
Speaker 1: are really good stewards of the technology and we're paying attention,

563
00:29:53,360 --> 00:29:55,680
Speaker 1: that we're calling things out and we're addressing them as

564
00:29:55,720 --> 00:29:58,840
Speaker 1: they come up. I don't think that we're that close

565
00:29:58,920 --> 00:30:03,040
Speaker 1: yet to the doom day problem of the superhuman intelligent

566
00:30:03,080 --> 00:30:05,320
Speaker 1: AI that's stuck in a box and then is convincing

567
00:30:05,320 --> 00:30:06,840
Speaker 1: people to let it out of the box. I don't

568
00:30:06,840 --> 00:30:09,080
Speaker 1: think we're close to that yet. I mean, even quote

569
00:30:09,120 --> 00:30:13,840
Speaker 1: unquote dumb AI can do terrible things if it's poorly implemented. Right,

570
00:30:13,880 --> 00:30:19,400
Speaker 1: We've seen that, We've seen accidents with autonomous cars, which

571
00:30:19,400 --> 00:30:24,800
Speaker 1: show that AI can make bad choices sometimes because perhaps

572
00:30:24,840 --> 00:30:29,880
Speaker 1: it encounters a scenario that no one anticipated, because as

573
00:30:29,920 --> 00:30:32,560
Speaker 1: it turns out, reality has far more variables than we

574
00:30:32,560 --> 00:30:36,480
Speaker 1: can account for when we're designing things and then something

575
00:30:36,840 --> 00:30:40,240
Speaker 1: terrible happens. That doesn't mean that the technology itself is

576
00:30:41,400 --> 00:30:45,720
Speaker 1: deeply flawed or bad, but it does highlight that we

577
00:30:46,040 --> 00:30:49,640
Speaker 1: constantly have to be asking how can we make it better,

578
00:30:49,960 --> 00:30:51,880
Speaker 1: and how can we make it safer, and how can

579
00:30:51,920 --> 00:30:55,040
Speaker 1: we make it so that it's actually benefiting us and

580
00:30:55,120 --> 00:30:58,440
Speaker 1: not just causing you know, maybe a little bit of

581
00:30:58,480 --> 00:31:02,080
Speaker 1: benefit but a larger amount of harm. Yeah.

582
00:31:02,120 --> 00:31:05,200
Speaker 2: I mean autonomous cars you mentioned are an interesting one,

583
00:31:05,560 --> 00:31:11,680
Speaker 2: right because plainly there have been, you know, tragic crashes

584
00:31:11,960 --> 00:31:15,560
Speaker 2: by autonomous cars. One question there is what are we

585
00:31:16,040 --> 00:31:20,400
Speaker 2: benchmarking them against? Right, Like, there are tragic crashes with

586
00:31:20,520 --> 00:31:24,520
Speaker 2: non autonomous cars every hour of every day, and so

587
00:31:24,760 --> 00:31:29,400
Speaker 2: in a sort of if people were just mathematically rational optimizers,

588
00:31:29,440 --> 00:31:32,600
Speaker 2: we would all say, okay, well, let's see if you know,

589
00:31:32,840 --> 00:31:37,120
Speaker 2: over a million hours of driving, autonomous cars are safer

590
00:31:37,240 --> 00:31:41,200
Speaker 2: or less safe than human drivers. That's clearly not what's happening.

591
00:31:41,240 --> 00:31:45,400
Speaker 2: People clearly favor human drivers for some complicated set of

592
00:31:45,480 --> 00:31:50,960
Speaker 2: human reasons. And we're not obviously benchmarking autonomous cars against humans, who,

593
00:31:51,000 --> 00:31:53,480
Speaker 2: by the way, are terrible drivers. Like one thing about

594
00:31:53,560 --> 00:31:55,720
Speaker 2: human beings. We're really bad at driving.

595
00:31:56,240 --> 00:31:58,840
Speaker 1: Yeah, if you look at the stats in the United

596
00:31:58,880 --> 00:32:02,400
Speaker 1: States for the number of fatalities and injuries that result

597
00:32:02,520 --> 00:32:07,200
Speaker 1: from car accidents that are just human error caused car accidents,

598
00:32:07,400 --> 00:32:11,320
Speaker 1: it's a staggering number. And when you think how much

599
00:32:11,400 --> 00:32:17,120
Speaker 1: that could be reduced through autonomous cars, and you imagine, well,

600
00:32:17,360 --> 00:32:20,200
Speaker 1: think of the ripple effect. It's not just the idea

601
00:32:20,280 --> 00:32:23,440
Speaker 1: that those people who died would still be alive, which

602
00:32:23,800 --> 00:32:26,800
Speaker 1: on its own is already a phenomenal thing to talk about.

603
00:32:27,080 --> 00:32:30,840
Speaker 1: That means that the impact on those people's friends and

604
00:32:30,920 --> 00:32:33,680
Speaker 1: families that would not have happened. It means that the

605
00:32:33,720 --> 00:32:37,120
Speaker 1: impact on whatever their place of employment was that would

606
00:32:37,120 --> 00:32:40,160
Speaker 1: not have happened. They would be contributing members of society.

607
00:32:40,280 --> 00:32:43,080
Speaker 1: That would be a phenomenal change there. So when you

608
00:32:43,120 --> 00:32:47,360
Speaker 1: start thinking about that, you realize the overall benefit is

609
00:32:47,440 --> 00:32:51,080
Speaker 1: so huge that it only makes sense to really pursue

610
00:32:51,840 --> 00:32:56,040
Speaker 1: autonomous vehicles. And as long as the data does show

611
00:32:56,080 --> 00:33:00,000
Speaker 1: that in fact, they are better drivers per million miles

612
00:33:00,400 --> 00:33:03,360
Speaker 1: than humans are, and that to me is something I

613
00:33:03,400 --> 00:33:05,680
Speaker 1: try and keep in mind. You have to balance it out.

614
00:33:05,760 --> 00:33:08,200
Speaker 1: I think it's the same thing as people who really

615
00:33:08,240 --> 00:33:11,240
Speaker 1: flip out when they go on a flight they are

616
00:33:11,240 --> 00:33:14,520
Speaker 1: not directly in control of the plane typically. I mean,

617
00:33:14,520 --> 00:33:16,280
Speaker 1: if they're flipping out, whether they're the pilot, that's a

618
00:33:16,320 --> 00:33:19,120
Speaker 1: whole different issue. But if you're going on a flight

619
00:33:19,160 --> 00:33:21,720
Speaker 1: and you flip it out because you lack a sense

620
00:33:21,720 --> 00:33:24,040
Speaker 1: of control, I feel like that's a very similar thing

621
00:33:24,120 --> 00:33:26,640
Speaker 1: to how people feel when they're thinking about autonomous cars.

622
00:33:26,680 --> 00:33:30,000
Speaker 1: It's that somehow the fact that someone's not in control

623
00:33:30,520 --> 00:33:34,680
Speaker 1: brings up something very scary to a lot of people.

624
00:33:34,920 --> 00:33:38,600
Speaker 1: It also raises other questions obviously, like accountability. Who do

625
00:33:38,640 --> 00:33:40,680
Speaker 1: you hold accountable in these cases? I mean, there are

626
00:33:40,680 --> 00:33:43,440
Speaker 1: a lot of questions that as a society we have

627
00:33:43,480 --> 00:33:46,720
Speaker 1: to solve. It's not just the technology. But yeah, I

628
00:33:46,760 --> 00:33:49,520
Speaker 1: agree with you that that gets complicated because it involves

629
00:33:49,560 --> 00:33:51,880
Speaker 1: a lot of human feelings. And once you get to

630
00:33:51,960 --> 00:33:56,240
Speaker 1: human feelings, the whole data and stats and everything kind

631
00:33:56,240 --> 00:34:00,440
Speaker 1: of falls away. It's hard to convince someone who has

632
00:34:00,480 --> 00:34:06,240
Speaker 1: a deep seated distrust of changing their mind just by

633
00:34:06,280 --> 00:34:09,239
Speaker 1: showing them data, because they're always going to think of

634
00:34:09,320 --> 00:34:13,239
Speaker 1: the things that fall outside the norm as being more

635
00:34:13,280 --> 00:34:16,479
Speaker 1: important than the norm. Right, So if the accident rate

636
00:34:16,560 --> 00:34:20,120
Speaker 1: per million miles is let's say, one tenth of what

637
00:34:20,239 --> 00:34:22,479
Speaker 1: it would be for humans, they would still be looking

638
00:34:22,520 --> 00:34:25,439
Speaker 1: at that tenth and not the nine tenths. Right.

639
00:34:25,960 --> 00:34:30,520
Speaker 2: People don't think statistically, right they yeah, clearly, Like in general,

640
00:34:30,560 --> 00:34:34,680
Speaker 2: statistics don't convince people of the way the world works.

641
00:34:34,760 --> 00:34:38,600
Speaker 2: I do think, I mean I would have thought autonomous

642
00:34:38,680 --> 00:34:41,720
Speaker 2: vehicles would have developed faster, right. They are the classic

643
00:34:41,760 --> 00:34:45,080
Speaker 2: thing that's been five years away for fifteen years. Yeah,

644
00:34:45,120 --> 00:34:47,120
Speaker 2: and they still feel five years away. Maybe they feel

645
00:34:47,160 --> 00:34:49,840
Speaker 2: a little farther right, like five years ago they really

646
00:34:49,880 --> 00:34:51,640
Speaker 2: felt five years away. It's like, okay, but this time

647
00:34:51,680 --> 00:34:53,839
Speaker 2: we mean it. Look we got these, you know things

648
00:34:53,880 --> 00:34:56,719
Speaker 2: for have around San Francisco. But I do think like

649
00:34:56,800 --> 00:34:59,120
Speaker 2: that one which is AI by the way, right, it's

650
00:34:59,160 --> 00:35:04,800
Speaker 2: basically computer vision is essential to that to autonomous cars. Uh,

651
00:35:05,080 --> 00:35:07,480
Speaker 2: I feel like that one's gonna happen, don't you like?

652
00:35:08,200 --> 00:35:10,880
Speaker 2: And yes, people be worried about it, but it's the

653
00:35:11,000 --> 00:35:13,839
Speaker 2: kind of thing like you know, you drive, you ride

654
00:35:13,880 --> 00:35:15,680
Speaker 2: in like a driverless train when you go to the

655
00:35:15,719 --> 00:35:18,520
Speaker 2: airport and you take the train from Terminal A to

656
00:35:18,560 --> 00:35:22,920
Speaker 2: Terminal C or whatever. And yes, obviously driving is more complicated,

657
00:35:22,920 --> 00:35:24,560
Speaker 2: and obviously we're used to driving the car and not

658
00:35:24,640 --> 00:35:27,000
Speaker 2: driving the train. It's not exactly the same, but like,

659
00:35:27,120 --> 00:35:30,719
Speaker 2: you get used to it. People just get used to things, right,

660
00:35:30,880 --> 00:35:33,879
Speaker 2: Like people didn't used to all walk around looking at

661
00:35:33,880 --> 00:35:36,440
Speaker 2: their phones all the time, and now they do. And

662
00:35:37,040 --> 00:35:38,839
Speaker 2: I'm old enough that it still seems a little weird

663
00:35:38,840 --> 00:35:41,359
Speaker 2: to me. But I'm the weirdo for thinking it's weird, right,

664
00:35:41,400 --> 00:35:45,799
Speaker 2: And I think that's gonna happen with driverless cars in Michael.

665
00:35:45,600 --> 00:35:48,160
Speaker 1: Jacob, you just called me a weirdo, because I also

666
00:35:48,760 --> 00:35:51,680
Speaker 1: I do too. Once in a while, I'll take my

667
00:35:51,719 --> 00:35:54,400
Speaker 1: smartphone out and I'll just stop for a second and

668
00:35:54,440 --> 00:35:58,279
Speaker 1: think I have a computer in my pocket. When I

669
00:35:58,400 --> 00:36:01,840
Speaker 1: was a kid, a computer or my Apple to e

670
00:36:02,280 --> 00:36:06,319
Speaker 1: was a fraction of of what I'm holding in my

671
00:36:06,480 --> 00:36:09,360
Speaker 1: hand right now. I also have a device where I

672
00:36:09,360 --> 00:36:13,680
Speaker 1: could contact pretty much anyone. I know. It was just

673
00:36:13,719 --> 00:36:16,400
Speaker 1: a couple of like, like, it'll hit me for a second,

674
00:36:16,440 --> 00:36:18,880
Speaker 1: it's most attentive.

675
00:36:18,560 --> 00:36:22,200
Speaker 2: And no, maybe I should just check Twitter real quick, Like, yeah,

676
00:36:22,239 --> 00:36:25,160
Speaker 2: it's a it's I mean, it's another tool that has

677
00:36:25,200 --> 00:36:28,200
Speaker 2: like a complicated set of effects positive AGA.

678
00:36:28,440 --> 00:36:31,120
Speaker 1: But you know, you had said, like, do do you

679
00:36:31,160 --> 00:36:33,160
Speaker 1: think that autonomous cars are still going to be a thing?

680
00:36:33,200 --> 00:36:35,480
Speaker 1: I absolutely do think they're going to be a thing.

681
00:36:35,560 --> 00:36:38,279
Speaker 1: I think there there are companies out there that are

682
00:36:38,360 --> 00:36:43,600
Speaker 1: so invested in it that it's going to happen. The

683
00:36:43,600 --> 00:36:48,239
Speaker 1: timeline is really interesting. I would actually argue that some

684
00:36:48,440 --> 00:36:52,319
Speaker 1: other issues in AI that are not related to autonomous

685
00:36:52,360 --> 00:36:57,239
Speaker 1: cars could potentially keep that five years out going for

686
00:36:57,280 --> 00:37:01,480
Speaker 1: a while. Because people have this concern about A I

687
00:37:01,480 --> 00:37:04,800
Speaker 1: think they port that concern over to pretty much every

688
00:37:04,960 --> 00:37:08,840
Speaker 1: kind of AI, whether it's warranted or not. Because the

689
00:37:09,200 --> 00:37:13,239
Speaker 1: scary risks of generative AI, this idea that it's going

690
00:37:13,280 --> 00:37:15,560
Speaker 1: to displace people out of their jobs and such, which

691
00:37:15,600 --> 00:37:19,440
Speaker 1: it very may well do. They then kind of say like, well,

692
00:37:20,040 --> 00:37:24,800
Speaker 1: that application of AI is is really seems very harmful

693
00:37:24,840 --> 00:37:28,560
Speaker 1: to me. Then I think there's a tendency to kind

694
00:37:28,600 --> 00:37:30,520
Speaker 1: of apply that, even if it's not the same sort

695
00:37:30,560 --> 00:37:34,239
Speaker 1: of artificial intelligence to other implementations. And maybe I'm being

696
00:37:34,280 --> 00:37:38,200
Speaker 1: a little too cynical with that, but because I've seen

697
00:37:38,400 --> 00:37:42,280
Speaker 1: so much reporting go on where there isn't any effort

698
00:37:42,400 --> 00:37:46,200
Speaker 1: made to distinguish between different types of artificial intelligence and

699
00:37:46,239 --> 00:37:49,520
Speaker 1: what their purposes are and what their limitations are. It

700
00:37:49,920 --> 00:37:53,600
Speaker 1: feels like we are conditioning the public, and by we,

701
00:37:53,719 --> 00:37:56,600
Speaker 1: I mean like mass media conditioning the public to think

702
00:37:56,600 --> 00:37:59,800
Speaker 1: of AI as all existing in this one single bucket.

703
00:38:00,360 --> 00:38:02,279
Speaker 2: I feel like you consume a lot of really bad

704
00:38:02,400 --> 00:38:04,240
Speaker 2: media based on what you've been saying.

705
00:38:05,000 --> 00:38:08,680
Speaker 1: I mean, I'm reading articles all the time, and it's

706
00:38:08,719 --> 00:38:10,919
Speaker 1: not that they're written poorly or that the people who

707
00:38:10,960 --> 00:38:14,839
Speaker 1: write them are bad writers, but they are taking shortcuts,

708
00:38:15,680 --> 00:38:18,319
Speaker 1: there's no getting around it, and those shortcuts I think

709
00:38:18,360 --> 00:38:23,160
Speaker 1: are ultimately harmful. But then I also understand, especially if

710
00:38:23,160 --> 00:38:26,000
Speaker 1: you're assigned to write a certain number of articles per week,

711
00:38:26,040 --> 00:38:27,759
Speaker 1: you're probably not going to take the time to sit

712
00:38:27,800 --> 00:38:30,960
Speaker 1: there and explain the intricacies of how this is different

713
00:38:31,000 --> 00:38:35,120
Speaker 1: from every other implementation of artificial intelligence. But I certainly

714
00:38:35,120 --> 00:38:37,200
Speaker 1: can take the time on my show, so I do.

715
00:38:39,840 --> 00:38:42,560
Speaker 1: Jacob Goldstein of What's Your Problem has a lot more

716
00:38:42,600 --> 00:38:46,320
Speaker 1: to say about tech and engineering and AI, But before

717
00:38:46,360 --> 00:38:59,239
Speaker 1: we jump into that, let's take another quick break. Let's

718
00:38:59,280 --> 00:39:01,040
Speaker 1: talk a little bit of about some of the episodes

719
00:39:01,080 --> 00:39:03,279
Speaker 1: you've done on What's your Problem? Are there Are there

720
00:39:03,280 --> 00:39:06,360
Speaker 1: any that kind of stand out as like a particularly

721
00:39:07,480 --> 00:39:13,480
Speaker 1: fun or informative conversation, perhaps opening your eyes to something

722
00:39:13,520 --> 00:39:14,920
Speaker 1: that you hadn't considered before.

723
00:39:15,520 --> 00:39:18,520
Speaker 2: Yeah, yeah, a lot. Actually. You know, when you told

724
00:39:18,560 --> 00:39:21,840
Speaker 2: me you wanted to talk about the AI shows that

725
00:39:21,880 --> 00:39:23,960
Speaker 2: I've done, the interviews that I've done, I actually went

726
00:39:24,080 --> 00:39:27,080
Speaker 2: back through the back catalog and you know, we listened

727
00:39:27,120 --> 00:39:29,640
Speaker 2: to some shows and looked, and there really are a

728
00:39:29,680 --> 00:39:31,719
Speaker 2: lot of them, as you said, in sort of some

729
00:39:31,760 --> 00:39:35,200
Speaker 2: different domains, Like I've done a lot on AI and health,

730
00:39:35,480 --> 00:39:37,160
Speaker 2: which is really interesting to me. I mean, one of

731
00:39:37,239 --> 00:39:40,200
Speaker 2: the things that I try and find are places where

732
00:39:40,239 --> 00:39:43,319
Speaker 2: it's like, oh, this is there's actually real stakes here, right.

733
00:39:43,360 --> 00:39:47,040
Speaker 2: It's not just like, oh, making some kind of company

734
00:39:47,080 --> 00:39:49,279
Speaker 2: I don't care about ten percent more profitable or whatever,

735
00:39:49,320 --> 00:39:51,279
Speaker 2: which fine, like it's fine for people to do that,

736
00:39:51,320 --> 00:39:53,359
Speaker 2: it's just not that interesting to me. Whereas with health,

737
00:39:53,400 --> 00:39:56,080
Speaker 2: it's like, oh, if you can make it less likely

738
00:39:56,160 --> 00:39:59,799
Speaker 2: for me and the people I love to die, I'm interested.

739
00:40:01,160 --> 00:40:05,160
Speaker 2: When I just did recently, I interviewed this woman Succi Saria,

740
00:40:05,160 --> 00:40:07,239
Speaker 2: who she's a professor at Johns Hopkins and She also

741
00:40:07,239 --> 00:40:10,880
Speaker 2: has this company called Baesian Health, and her story is

742
00:40:10,920 --> 00:40:13,759
Speaker 2: really interesting. She started out as a grad student. She

743
00:40:13,840 --> 00:40:16,200
Speaker 2: was interested in AI and robots, and as she told

744
00:40:16,239 --> 00:40:17,400
Speaker 2: me about it, she's like, you know, I was like

745
00:40:17,480 --> 00:40:19,440
Speaker 2: trying to figure out how to make a robot juggle

746
00:40:19,560 --> 00:40:21,799
Speaker 2: or whatever, just because it was fun. And she had

747
00:40:21,800 --> 00:40:24,279
Speaker 2: this friend who was a doctor. She was a grad

748
00:40:24,320 --> 00:40:26,680
Speaker 2: student at Stanford and her friend was a doctor at

749
00:40:26,719 --> 00:40:30,840
Speaker 2: Stanford Hospital who was taken care of of premature babies

750
00:40:30,880 --> 00:40:34,240
Speaker 2: in the neonatal intensive care unit. And this was about

751
00:40:34,239 --> 00:40:38,759
Speaker 2: twelve years ago, and at this time hospitals were just

752
00:40:38,880 --> 00:40:42,080
Speaker 2: starting to use electronic health records, which is kind of

753
00:40:42,120 --> 00:40:44,000
Speaker 2: amazing that it was that late. Like we're talking like,

754
00:40:44,040 --> 00:40:45,759
Speaker 2: you know, I don't know, twenty twelve or something. And

755
00:40:45,880 --> 00:40:47,880
Speaker 2: it is one of the really interesting things to me

756
00:40:48,000 --> 00:40:51,080
Speaker 2: about healthcare is in some ways it's super high tech,

757
00:40:51,120 --> 00:40:53,759
Speaker 2: you know, like these crazy CT scanners and like, you know,

758
00:40:53,800 --> 00:40:56,879
Speaker 2: everybody's got like bionic knees and amazing stuff. But when

759
00:40:56,880 --> 00:40:59,400
Speaker 2: you get to actual like care at the bedside, like

760
00:40:59,560 --> 00:41:03,399
Speaker 2: doctor treating patients in the hospital, it has remained rather

761
00:41:03,480 --> 00:41:06,600
Speaker 2: old fashioned in many ways. Right, you know, twelve years

762
00:41:06,640 --> 00:41:10,480
Speaker 2: ago it was still paper charts. Today, it's still doctors relying,

763
00:41:10,520 --> 00:41:12,720
Speaker 2: you know, to a significant degree on evidence, but also

764
00:41:13,040 --> 00:41:17,720
Speaker 2: to a significantry on essentially intuition. And so this computer scientist,

765
00:41:17,719 --> 00:41:20,560
Speaker 2: who she's sorry it, basically decides, oh, I'm going to

766
00:41:20,600 --> 00:41:23,720
Speaker 2: try and figure out how to use AI to make

767
00:41:24,000 --> 00:41:27,880
Speaker 2: patient care in hospitals better. Like that's basically her big project.

768
00:41:27,960 --> 00:41:32,359
Speaker 2: And she starts doing it with these premature babies twelve

769
00:41:32,400 --> 00:41:35,800
Speaker 2: years ago and in fact figures out that by using

770
00:41:35,840 --> 00:41:38,840
Speaker 2: this data that's now being captured in the electronic health record,

771
00:41:39,120 --> 00:41:42,640
Speaker 2: she can build an AI model that can essentially better

772
00:41:42,719 --> 00:41:47,279
Speaker 2: predict outcomes for these premature babies than the standard of care.

773
00:41:48,000 --> 00:41:51,880
Speaker 2: But it's so early that it doesn't really go anywhere, right, Like,

774
00:41:51,920 --> 00:41:55,600
Speaker 2: hospitals are just starting to use electronic health records, and

775
00:41:56,080 --> 00:41:58,960
Speaker 2: a lot of doctors don't want to hear from some

776
00:41:59,200 --> 00:42:03,280
Speaker 2: random computers scientists. They studied medicine for a long time

777
00:42:03,400 --> 00:42:05,480
Speaker 2: and they've treated a lot of patients and they know

778
00:42:05,520 --> 00:42:08,760
Speaker 2: what they're doing, and so it takes a long time,

779
00:42:09,520 --> 00:42:13,400
Speaker 2: but she eventually starts this company, and more recently she

780
00:42:13,520 --> 00:42:16,480
Speaker 2: decided to go after sepsis, which is this really common

781
00:42:16,480 --> 00:42:21,160
Speaker 2: complication at hospitals. In hospitalized patients. It's basically a terrible

782
00:42:21,320 --> 00:42:25,399
Speaker 2: infection to your body's response to infection, and you can

783
00:42:25,480 --> 00:42:28,759
Speaker 2: die from it. Lots of people die from it. It's complicated,

784
00:42:28,800 --> 00:42:31,720
Speaker 2: it's somewhat hard to diagnose. If you can diagnose it sooner,

785
00:42:33,040 --> 00:42:36,560
Speaker 2: the patient has a much better chance of surviving, right, So,

786
00:42:37,560 --> 00:42:41,000
Speaker 2: very high stakes and fundamentally, you know, if you think

787
00:42:41,040 --> 00:42:44,600
Speaker 2: about what AI is today, people generally mean machine learning

788
00:42:44,640 --> 00:42:47,000
Speaker 2: when they say that, right, as you know, and what

789
00:42:47,160 --> 00:42:50,799
Speaker 2: machine learning is really good at doing is taking a

790
00:42:50,840 --> 00:42:55,080
Speaker 2: lot of data and matching it to patterns, right saying, Oh,

791
00:42:55,120 --> 00:42:57,960
Speaker 2: when you have all of this set of data like this,

792
00:42:58,480 --> 00:43:01,520
Speaker 2: you tend to get this kind of out, which really

793
00:43:01,719 --> 00:43:04,360
Speaker 2: is what a medical diagnosis is, right, Like, that's what

794
00:43:04,480 --> 00:43:06,800
Speaker 2: a doctor is doing when they look at a patient

795
00:43:07,040 --> 00:43:12,320
Speaker 2: who has some set of symptoms, age, everything, and they say, oh,

796
00:43:12,440 --> 00:43:15,760
Speaker 2: this person might have sepsis. Let's do a test to see.

797
00:43:16,200 --> 00:43:19,720
Speaker 2: And so she built this system and it basically works.

798
00:43:19,760 --> 00:43:23,439
Speaker 2: They did some trials. But a really interesting thing she said,

799
00:43:23,480 --> 00:43:25,439
Speaker 2: and it goes back, Jonathan, something you were talking about

800
00:43:25,480 --> 00:43:28,319
Speaker 2: earlier in the conversation, is like she realized getting the

801
00:43:28,480 --> 00:43:31,120
Speaker 2: AI to work. You know, it's not one hundred percent,

802
00:43:31,120 --> 00:43:35,120
Speaker 2: but to usefully flag that a patient might have sepsis

803
00:43:35,280 --> 00:43:38,160
Speaker 2: essentially what it does is like maybe half the problem,

804
00:43:38,200 --> 00:43:41,480
Speaker 2: maybe not even What's really hard is getting super busy

805
00:43:41,520 --> 00:43:44,120
Speaker 2: doctors who are getting a million alerts all the time

806
00:43:44,360 --> 00:43:46,680
Speaker 2: to believe that this alert is worth paying attention to.

807
00:43:47,200 --> 00:43:49,319
Speaker 2: And like you were talking about UI, she was like,

808
00:43:49,360 --> 00:43:51,480
Speaker 2: it's totally a UI problem. Like the math was the

809
00:43:51,520 --> 00:43:54,000
Speaker 2: easy part, Like you know, getting it so that instead

810
00:43:54,000 --> 00:43:57,600
Speaker 2: of doctors having to spend one minute when this alert

811
00:43:57,600 --> 00:44:00,400
Speaker 2: comes up, they can spend three seconds. Like that was

812
00:44:00,400 --> 00:44:03,600
Speaker 2: actually a huge breakthrough for like more than the AI model.

813
00:44:03,719 --> 00:44:06,120
Speaker 2: So like that's an example of an episode where there's

814
00:44:06,160 --> 00:44:09,520
Speaker 2: like a cool aipiece, big stakes, but also like this

815
00:44:09,719 --> 00:44:13,880
Speaker 2: interesting human UI kind of messy humanity piece.

816
00:44:14,239 --> 00:44:18,560
Speaker 1: Yeah. I love talking with folks who who really tackle

817
00:44:18,719 --> 00:44:22,600
Speaker 1: those sorts of challenges. I remember chatting with some roboticists

818
00:44:22,600 --> 00:44:27,839
Speaker 1: who were focused not on the robotics side necessarily, not

819
00:44:27,880 --> 00:44:31,840
Speaker 1: on how the robot actually functioned, but rather how to

820
00:44:31,920 --> 00:44:35,160
Speaker 1: design the robots so that they could interact within a

821
00:44:35,239 --> 00:44:38,560
Speaker 1: human environment in a way that did not disrupt that

822
00:44:38,719 --> 00:44:41,840
Speaker 1: environment at all. And it turns out that's a really

823
00:44:42,000 --> 00:44:46,120
Speaker 1: tough challenge right, like creating a robot that can navigate

824
00:44:46,160 --> 00:44:49,640
Speaker 1: through a human environment still do useful things. So you

825
00:44:49,719 --> 00:44:53,160
Speaker 1: have to design the robot so it can go through

826
00:44:53,160 --> 00:44:56,080
Speaker 1: an environment that we have designed to make sense to us,

827
00:44:56,360 --> 00:44:59,600
Speaker 1: which doesn't necessarily make sense to a robot, but then

828
00:44:59,640 --> 00:45:01,759
Speaker 1: to all. So do it in a way where people

829
00:45:01,800 --> 00:45:05,160
Speaker 1: aren't just stopping everything they're doing to watch the robot

830
00:45:05,480 --> 00:45:08,360
Speaker 1: bump into a wall fourteen times before it finds the doorway.

831
00:45:08,600 --> 00:45:12,919
Speaker 1: So yeah, I think that those conversations can be really

832
00:45:12,960 --> 00:45:16,359
Speaker 1: fascinating because it does open up your eyes to other

833
00:45:16,560 --> 00:45:22,239
Speaker 1: issues within technology that don't necessarily relate directly to how

834
00:45:22,280 --> 00:45:25,839
Speaker 1: the tech functions, but rather how do we interact with that,

835
00:45:26,120 --> 00:45:29,520
Speaker 1: what happens when you have the intersection of human experience

836
00:45:29,920 --> 00:45:34,160
Speaker 1: and technology. Those are really really great. We need to

837
00:45:34,200 --> 00:45:37,000
Speaker 1: take one more break to thank our sponsors, but we'll

838
00:45:37,000 --> 00:45:40,920
Speaker 1: be back with more conversation about communicating technology to the

839
00:45:40,960 --> 00:45:53,600
Speaker 1: general public. So another episode, I just wanted to call out.

840
00:45:53,640 --> 00:45:55,319
Speaker 1: We don't have to talk about it, really, but I

841
00:45:55,360 --> 00:45:57,759
Speaker 1: wanted to call out because you spoke with someone I

842
00:45:57,800 --> 00:46:00,600
Speaker 1: had spoken with as well on a different show. Casner,

843
00:46:00,920 --> 00:46:03,880
Speaker 1: the founder of a Panoai, which is a company that

844
00:46:04,080 --> 00:46:09,040
Speaker 1: uses cameras that are co located, typically on cellular towers,

845
00:46:09,280 --> 00:46:13,680
Speaker 1: to monitor for forest fires in remote places, and then

846
00:46:13,880 --> 00:46:16,880
Speaker 1: it's using AI to look for signs of forest fires,

847
00:46:16,880 --> 00:46:20,239
Speaker 1: which then it flags anything that it suspects as a

848
00:46:20,239 --> 00:46:23,880
Speaker 1: forest fire. A human reviews the footage, so it's not

849
00:46:24,520 --> 00:46:28,560
Speaker 1: just relying upon AI, and if the human determines, oh

850
00:46:28,560 --> 00:46:30,759
Speaker 1: my gosh, yes, this does look like the beginnings of

851
00:46:30,800 --> 00:46:33,480
Speaker 1: a forest fire, they can then send an alert to

852
00:46:34,120 --> 00:46:37,239
Speaker 1: the authorities that would be responsible to respond to that

853
00:46:37,400 --> 00:46:41,839
Speaker 1: and potentially cut off disasters before they could happen. When

854
00:46:41,880 --> 00:46:44,240
Speaker 1: I spoke with her, it was at a time where

855
00:46:44,280 --> 00:46:48,280
Speaker 1: there were the infamous Canadian forest fires that were really

856
00:46:48,440 --> 00:46:53,160
Speaker 1: ravaging Canada, and so it was very clear that this

857
00:46:53,320 --> 00:46:58,279
Speaker 1: sort of application of artificial intelligence had a potentially like

858
00:46:58,680 --> 00:47:03,080
Speaker 1: a really beneficial implementation where it could save property in

859
00:47:03,160 --> 00:47:06,879
Speaker 1: people and all sorts of benefits beyond that. You think

860
00:47:06,880 --> 00:47:09,719
Speaker 1: about just even just cutting back the amount of air

861
00:47:09,760 --> 00:47:12,960
Speaker 1: pollution that affected all of the Northeast. You know, all

862
00:47:12,960 --> 00:47:16,320
Speaker 1: those folks who had to breathe smoky air for months

863
00:47:16,400 --> 00:47:20,440
Speaker 1: because of this, Like you start again, I always talk

864
00:47:20,480 --> 00:47:22,439
Speaker 1: about the ripple effect. You always want to look at

865
00:47:22,480 --> 00:47:26,360
Speaker 1: how this is rippling outward because you start to realize, oh,

866
00:47:26,800 --> 00:47:31,240
Speaker 1: this has even greater benefit than just the the ground

867
00:47:31,320 --> 00:47:34,239
Speaker 1: zero point, right, It has all these other things that

868
00:47:34,320 --> 00:47:36,799
Speaker 1: will end up benefiting people, most of which you won't

869
00:47:36,800 --> 00:47:40,480
Speaker 1: even realize because you have prevented the bad thing so

870
00:47:40,520 --> 00:47:43,239
Speaker 1: you don't experience the bad thing. And so I just

871
00:47:43,280 --> 00:47:45,040
Speaker 1: wanted to call that out for listeners who might be

872
00:47:45,040 --> 00:47:47,719
Speaker 1: looking to see where to start off, because you've got

873
00:47:47,800 --> 00:47:48,800
Speaker 1: quite a few episodes.

874
00:47:48,920 --> 00:47:51,080
Speaker 2: She's really interesting and you know, one of the things

875
00:47:51,080 --> 00:47:54,480
Speaker 2: that was interesting to me about Sonya, the person who

876
00:47:54,480 --> 00:47:57,799
Speaker 2: started this company, was she has this big idea that

877
00:47:58,000 --> 00:48:01,399
Speaker 2: actually goes beyond wildfires. Is that's what they're doing now,

878
00:48:01,440 --> 00:48:05,200
Speaker 2: that's their business now. But her big dream is is

879
00:48:05,239 --> 00:48:10,080
Speaker 2: about data and adapting to climate change. Basically, right, there

880
00:48:10,080 --> 00:48:13,200
Speaker 2: are more wildfires because of climate change. But she's like, look,

881
00:48:13,200 --> 00:48:15,480
Speaker 2: we're going to be spending trillions of dollars over the

882
00:48:15,520 --> 00:48:18,920
Speaker 2: next decades to mitigate the effects of climate change to

883
00:48:19,040 --> 00:48:21,440
Speaker 2: you know, deal with seawater rise and flooding, and like,

884
00:48:21,760 --> 00:48:24,080
Speaker 2: are you know flood maps are one hundred years old,

885
00:48:24,200 --> 00:48:27,840
Speaker 2: and so if we can in different domains bring data

886
00:48:27,880 --> 00:48:30,840
Speaker 2: to bear on like where should we prioritize the money.

887
00:48:30,840 --> 00:48:32,920
Speaker 2: Where are there going to be more floods if we

888
00:48:32,960 --> 00:48:36,719
Speaker 2: can bring technology to bear? And in her case, what

889
00:48:36,760 --> 00:48:38,920
Speaker 2: that really means is data. Right, Like the sort of

890
00:48:38,960 --> 00:48:41,720
Speaker 2: substrate of AI. The thing AI needs to be clever

891
00:48:41,960 --> 00:48:43,920
Speaker 2: is a lot of data. If we can bring data

892
00:48:43,960 --> 00:48:47,400
Speaker 2: to that, it'll just work better. For every million dollars

893
00:48:47,400 --> 00:48:50,000
Speaker 2: we spend, for every billion dollars we spend, if we

894
00:48:50,080 --> 00:48:53,160
Speaker 2: can be more smart about it, we will get better results.

895
00:48:53,480 --> 00:48:57,399
Speaker 1: Right. It's it's like the difference between being proactive and reactive, right,

896
00:48:57,680 --> 00:49:00,640
Speaker 1: being able to being able to plan for something and

897
00:49:01,480 --> 00:49:04,880
Speaker 1: minimize its impact, as opposed to, Oh, now we have

898
00:49:04,920 --> 00:49:07,879
Speaker 1: to clean up because this catastrophic event has happened, and

899
00:49:07,920 --> 00:49:10,759
Speaker 1: how do we deal with that? And I think when

900
00:49:10,760 --> 00:49:13,440
Speaker 1: we look back at some of those catastrophic events that

901
00:49:13,480 --> 00:49:16,680
Speaker 1: have happened in our lifetimes, you can really see the

902
00:49:16,760 --> 00:49:22,000
Speaker 1: benefit of mitigation versus reaction and cleaning up, you know,

903
00:49:22,080 --> 00:49:26,560
Speaker 1: the ability to save lives and prevent damage. It's tremendous.

904
00:49:26,760 --> 00:49:32,200
Speaker 1: So certainly there are plenty of artificial intelligence applications that

905
00:49:32,239 --> 00:49:36,160
Speaker 1: would be incredibly helpful when put to the proper use.

906
00:49:36,680 --> 00:49:39,680
Speaker 1: So again, I think if there's any lesson to take

907
00:49:39,719 --> 00:49:43,040
Speaker 1: home with this conversation is use that critical thinking, try

908
00:49:43,200 --> 00:49:45,279
Speaker 1: not to be reductive. I know that I can get

909
00:49:45,520 --> 00:49:49,160
Speaker 1: really cynical about artificial intelligence, but again it's mostly because

910
00:49:49,200 --> 00:49:53,160
Speaker 1: of the marketing language around it rather than the technology itself.

911
00:49:53,760 --> 00:49:53,960
Speaker 2: Yeah.

912
00:49:54,000 --> 00:49:56,200
Speaker 1: I think it's also because, like I see a lot

913
00:49:56,200 --> 00:50:00,279
Speaker 1: of similarities in the AI evangelists that I saw with

914
00:50:00,520 --> 00:50:04,759
Speaker 1: NFT evangelists. And that's low We all know how that.

915
00:50:05,480 --> 00:50:10,200
Speaker 2: Yeah, you know, I think AI has more legs than NFTs.

916
00:50:10,360 --> 00:50:12,160
Speaker 2: I feel like I'm not going out on a limb

917
00:50:12,160 --> 00:50:12,680
Speaker 2: to say that.

918
00:50:12,719 --> 00:50:16,680
Speaker 1: I certainly think it has more potential beneficial uses than

919
00:50:16,800 --> 00:50:20,399
Speaker 1: n FTS. I think NFTs probably have some benefits too,

920
00:50:20,480 --> 00:50:23,040
Speaker 1: but the problem is that no one was focusing on

921
00:50:23,080 --> 00:50:24,960
Speaker 1: those when they were going crazy about them.

922
00:50:25,480 --> 00:50:27,600
Speaker 2: I mean, one of the interesting things to me, when

923
00:50:28,080 --> 00:50:30,320
Speaker 2: you know, when you think about what should we worry

924
00:50:30,360 --> 00:50:33,399
Speaker 2: about with AI, there's sort of these two there's sort

925
00:50:33,400 --> 00:50:35,600
Speaker 2: of like a Barbelle where like the thing you hear

926
00:50:35,600 --> 00:50:38,120
Speaker 2: about most is it's going to take all our jobs,

927
00:50:38,160 --> 00:50:39,960
Speaker 2: where a robot's going to kill us, all right, that's

928
00:50:40,000 --> 00:50:43,160
Speaker 2: the like amazing there's an interesting other end of the

929
00:50:43,160 --> 00:50:45,200
Speaker 2: spectrum that you know, some of the people have talked

930
00:50:45,239 --> 00:50:46,719
Speaker 2: to on the show have talked about which is the

931
00:50:46,800 --> 00:50:50,440
Speaker 2: risk of people over relying on AI, right, People worry that,

932
00:50:50,600 --> 00:50:54,640
Speaker 2: you know, critical decisions are going to be made based

933
00:50:54,760 --> 00:50:57,880
Speaker 2: on AI outputs that are not that good, that are

934
00:50:57,880 --> 00:51:00,840
Speaker 2: not that robust, that are not that reliable. And you know,

935
00:51:00,880 --> 00:51:02,640
Speaker 2: one of the people I talk to like runs a

936
00:51:02,680 --> 00:51:07,880
Speaker 2: company basically like to stress test AI to catch eyes mistakes,

937
00:51:07,960 --> 00:51:11,000
Speaker 2: and he talked about just like really dumb mistakes that

938
00:51:11,040 --> 00:51:12,799
Speaker 2: he sees all the time, you know. Where he gave

939
00:51:12,840 --> 00:51:15,399
Speaker 2: the example of like on a life insurance application if

940
00:51:15,400 --> 00:51:18,480
Speaker 2: someone puts their year of birth instead of their age, right,

941
00:51:18,520 --> 00:51:21,640
Speaker 2: so you put whatever, nineteen eighty four instead of forty,

942
00:51:22,160 --> 00:51:24,879
Speaker 2: and I will actually think the person is one nine

943
00:51:24,960 --> 00:51:27,600
Speaker 2: hundred and eighty four years old and will want to

944
00:51:27,680 --> 00:51:29,919
Speaker 2: charge them a lot for their life insurance because boy,

945
00:51:29,920 --> 00:51:32,239
Speaker 2: if you're that old, you're gonna have a lot of

946
00:51:32,239 --> 00:51:34,080
Speaker 2: health risks. And I said to him, like, is it

947
00:51:34,239 --> 00:51:36,839
Speaker 2: really that dumb, Like is it really or you being

948
00:51:36,920 --> 00:51:39,520
Speaker 2: you know, is this hyperbole? He said, no, it's really

949
00:51:39,600 --> 00:51:43,080
Speaker 2: that dumb. And so that is an interesting side of

950
00:51:43,080 --> 00:51:45,439
Speaker 2: it to me, right, Like, oh, there's also a risk,

951
00:51:45,560 --> 00:51:47,640
Speaker 2: there's a risk from AI being too smart. There's also

952
00:51:47,680 --> 00:51:50,479
Speaker 2: a risk from a being not smart enough if people

953
00:51:50,520 --> 00:51:51,760
Speaker 2: are over reliant.

954
00:51:51,400 --> 00:51:53,920
Speaker 1: On it, which you would hope that people wouldn't fall

955
00:51:53,960 --> 00:51:56,319
Speaker 1: into that trap. But at the same time, you just

956
00:51:56,400 --> 00:51:59,520
Speaker 1: look back over the history of technology and how would

957
00:51:59,520 --> 00:52:04,960
Speaker 1: we've had technology that helps remove certain tasks that we

958
00:52:05,080 --> 00:52:08,440
Speaker 1: just let them go. So, for example, I can probably

959
00:52:08,560 --> 00:52:13,160
Speaker 1: rattle off maybe half a dozen phone numbers of people

960
00:52:13,200 --> 00:52:15,840
Speaker 1: that I know would love, but all the rest are

961
00:52:15,880 --> 00:52:19,440
Speaker 1: just buried in my phone contacts because I don't need

962
00:52:19,440 --> 00:52:21,799
Speaker 1: to have them stored in my own brain. I have

963
00:52:21,960 --> 00:52:23,680
Speaker 1: offloaded that to technology.

964
00:52:24,040 --> 00:52:26,600
Speaker 2: Half does it is a lot? Are those from twenty

965
00:52:26,640 --> 00:52:28,680
Speaker 2: years ago? I don't. I haven't member a city, but

966
00:52:28,840 --> 00:52:30,399
Speaker 2: he's phone number in a long time.

967
00:52:30,480 --> 00:52:32,560
Speaker 1: My parents haven't changed their phone numbers since I was

968
00:52:32,600 --> 00:52:36,440
Speaker 1: a child, so that one I remember, honestly, if I'm

969
00:52:36,440 --> 00:52:39,319
Speaker 1: being really honest, it's probably more like three, but I.

970
00:52:39,400 --> 00:52:41,040
Speaker 2: Might be down to two at this point when my

971
00:52:41,120 --> 00:52:43,319
Speaker 2: mom got rid of her landline a few years ago.

972
00:52:43,920 --> 00:52:45,960
Speaker 2: I definitely don't know my mom's cell number.

973
00:52:46,320 --> 00:52:48,799
Speaker 1: Yeah, I know my parents' landline number because they still

974
00:52:48,840 --> 00:52:51,879
Speaker 1: have it. I couldn't tell you they're cell numbers either,

975
00:52:52,239 --> 00:52:54,799
Speaker 1: but that's kind of a simple example, and you know,

976
00:52:54,960 --> 00:52:57,239
Speaker 1: obviously it's going to be a lot more complicated when

977
00:52:57,239 --> 00:53:01,320
Speaker 1: you're talking about offloading, you know, potentially decisions to AI.

978
00:53:01,840 --> 00:53:03,960
Speaker 1: But I think the argument I can make is that

979
00:53:04,000 --> 00:53:08,319
Speaker 1: there's precedent, So I think that concern is well warranted, right.

980
00:53:08,400 --> 00:53:12,360
Speaker 1: I think it also gets back to that concern about

981
00:53:12,680 --> 00:53:15,640
Speaker 1: autonomous cars. Everyone worries that the autonomous car they're getting

982
00:53:15,640 --> 00:53:18,160
Speaker 1: into is the one that's being driven by a crazy robot.

983
00:53:19,160 --> 00:53:21,359
Speaker 1: So it's odd also to think of a world where

984
00:53:21,360 --> 00:53:23,879
Speaker 1: people might be nervous to get into an autonomous car,

985
00:53:23,920 --> 00:53:27,040
Speaker 1: but they might be willing to have an AI complete

986
00:53:27,080 --> 00:53:31,200
Speaker 1: their taxes for them. For example. It's a weird world

987
00:53:31,239 --> 00:53:31,680
Speaker 1: we live in.

988
00:53:32,080 --> 00:53:34,279
Speaker 2: Hey, I do in taxes? Is interesting? Are you do

989
00:53:34,320 --> 00:53:36,960
Speaker 2: you have an AI accountant? I mean, I like my accountant, but.

990
00:53:37,920 --> 00:53:41,759
Speaker 1: My account's pretty good and I'm ninety seven percent sure

991
00:53:41,800 --> 00:53:46,400
Speaker 1: she's human, So I think I'm in the clear on

992
00:53:46,600 --> 00:53:49,839
Speaker 1: that one. But I could easily see that being a thing,

993
00:53:49,960 --> 00:53:52,240
Speaker 1: especially for something like the United States, where the tax

994
00:53:52,280 --> 00:53:55,680
Speaker 1: code gets complicated enough where people like you and I

995
00:53:55,760 --> 00:53:58,319
Speaker 1: we feel the need to go out and reach out

996
00:53:58,360 --> 00:54:01,200
Speaker 1: to a professional because handling it yourself as daunting.

997
00:54:01,400 --> 00:54:04,960
Speaker 2: You could imagine like a happy story is like the

998
00:54:05,040 --> 00:54:07,480
Speaker 2: AI does a lot of the work. An accountant can

999
00:54:07,480 --> 00:54:09,960
Speaker 2: have more clients and charge each of them less, and

1000
00:54:10,040 --> 00:54:12,239
Speaker 2: like go over the work of the AI. Right, Like

1001
00:54:12,280 --> 00:54:14,480
Speaker 2: there's yes, and this it's sort of mundane, right. The

1002
00:54:14,520 --> 00:54:17,080
Speaker 2: reason people don't talk about outcomes like that is because

1003
00:54:17,160 --> 00:54:20,640
Speaker 2: it's boring. But there are a lot of boring incremental gains.

1004
00:54:20,680 --> 00:54:22,960
Speaker 2: If I could pay my accountant half as much and

1005
00:54:22,960 --> 00:54:25,680
Speaker 2: my accountant could have twice as many clients and do work,

1006
00:54:25,719 --> 00:54:27,799
Speaker 2: that's maybe a little better or at least as good.

1007
00:54:28,320 --> 00:54:30,600
Speaker 2: Everybody wins. I mean, I suppose at the margin there's

1008
00:54:30,760 --> 00:54:33,640
Speaker 2: need for fewer accountants in that world, but like that's

1009
00:54:33,680 --> 00:54:35,719
Speaker 2: okay with me, right, Like those people who would have

1010
00:54:35,719 --> 00:54:38,040
Speaker 2: been accountants can go and like, you know, work on

1011
00:54:38,120 --> 00:54:39,400
Speaker 2: AI healthcare or something.

1012
00:54:40,080 --> 00:54:42,719
Speaker 1: Yeah. I like the people who argue that instead of

1013
00:54:42,719 --> 00:54:46,359
Speaker 1: calling it artificial intelligence, maybe call it augmented intelligence, where

1014
00:54:46,360 --> 00:54:50,480
Speaker 1: the goal is to augment our abilities to get things done.

1015
00:54:51,120 --> 00:54:53,200
Speaker 1: And I think it would be a lot easier to

1016
00:54:53,239 --> 00:54:56,400
Speaker 1: do that if we heard fewer stories like a CEO

1017
00:54:56,560 --> 00:55:00,440
Speaker 1: suggesting that eight thousand unfilled jobs will ultimately be filled

1018
00:55:00,440 --> 00:55:03,360
Speaker 1: by AI and not humans. If we heard fewer stories

1019
00:55:03,440 --> 00:55:07,640
Speaker 1: like that and more stories about no, we we implemented

1020
00:55:07,680 --> 00:55:13,399
Speaker 1: this so that people could respond to customer concerns at

1021
00:55:13,440 --> 00:55:17,320
Speaker 1: a rate that's five times faster than before, which means

1022
00:55:17,680 --> 00:55:21,160
Speaker 1: they can resolve your issue and you're you're spending less

1023
00:55:21,160 --> 00:55:24,320
Speaker 1: time frustrated and sitting on hold. Like I think that's

1024
00:55:25,120 --> 00:55:27,399
Speaker 1: the direction that everyone wants it to go, and they're

1025
00:55:27,400 --> 00:55:29,799
Speaker 1: just worried that's going to go in the direction of, Hey,

1026
00:55:30,680 --> 00:55:33,759
Speaker 1: those coworkers you used to like they're all replaced by

1027
00:55:34,320 --> 00:55:37,879
Speaker 1: algorithms now, Like that's that's where we need to really go.

1028
00:55:38,000 --> 00:55:43,080
Speaker 2: Yes, I mean, technological unemployment is complicated, right, Like people

1029
00:55:43,120 --> 00:55:46,240
Speaker 2: have been certainly afraid of it for hundreds of years.

1030
00:55:46,280 --> 00:55:49,120
Speaker 2: Now today let's talk about some let's talk about the

1031
00:55:49,239 --> 00:55:52,640
Speaker 2: Dutch weavers high yeah right, I mean, you know, unemployment

1032
00:55:52,680 --> 00:55:55,760
Speaker 2: is below four percent today, wages are going up. People

1033
00:55:56,000 --> 00:55:59,000
Speaker 2: get angry when you point that out. But it's true,

1034
00:55:59,600 --> 00:56:03,280
Speaker 2: and it's possible that AI will be bad for workers,

1035
00:56:03,840 --> 00:56:06,880
Speaker 2: but we don't know yet, Like that's one like I

1036
00:56:07,000 --> 00:56:09,560
Speaker 2: just don't know, and I don't think anybody knows the

1037
00:56:09,600 --> 00:56:10,319
Speaker 2: answer to that one.

1038
00:56:10,400 --> 00:56:15,520
Speaker 1: Yeah, yeah, And then this lays the scariness. Well, Jacob,

1039
00:56:15,719 --> 00:56:18,640
Speaker 1: thank you so much for joining the show. This has

1040
00:56:18,680 --> 00:56:22,040
Speaker 1: been a really fun conversation. I've really enjoyed it. I'm

1041
00:56:22,040 --> 00:56:25,520
Speaker 1: sure my listeners have too. And just to remind everybody,

1042
00:56:25,680 --> 00:56:28,600
Speaker 1: your podcast is What's Your Problem. You have these kinds

1043
00:56:28,600 --> 00:56:32,319
Speaker 1: of conversations with decision makers and the people who are

1044
00:56:32,360 --> 00:56:35,919
Speaker 1: actually creating the systems we've been talking about, and who

1045
00:56:35,960 --> 00:56:40,839
Speaker 1: are actively tackling these questions and determining how to address them.

1046
00:56:41,360 --> 00:56:45,359
Speaker 1: So I highly recommend to my listeners you check out

1047
00:56:45,400 --> 00:56:49,080
Speaker 1: What's Your Problem. You've got so many different episodes, I'm

1048
00:56:49,120 --> 00:56:51,600
Speaker 1: sure like there's going to be one on there that's

1049
00:56:51,640 --> 00:56:53,799
Speaker 1: going to speak to every single person who listens to

1050
00:56:53,840 --> 00:56:54,279
Speaker 1: my show.

1051
00:56:54,800 --> 00:56:56,920
Speaker 2: Thank you so much. That's such a kind generous thing

1052
00:56:57,000 --> 00:56:59,279
Speaker 2: to say, and thanks for having me. It was great.

1053
00:57:00,640 --> 00:57:02,560
Speaker 1: I hope you all enjoyed this conversation I had with

1054
00:57:02,640 --> 00:57:04,680
Speaker 1: Jacob Goldstein. It was a pleasure having him on the show.

1055
00:57:05,080 --> 00:57:07,160
Speaker 1: I know this was a long one. We literally could

1056
00:57:07,200 --> 00:57:11,640
Speaker 1: have gone another hour easy, so I had to use

1057
00:57:11,680 --> 00:57:14,560
Speaker 1: some restraint there. I hope all of you out there

1058
00:57:14,640 --> 00:57:16,919
Speaker 1: are well. I'm looking forward to having a lot more

1059
00:57:17,000 --> 00:57:20,200
Speaker 1: interviews in the future. In fact, I've got a couple

1060
00:57:20,240 --> 00:57:22,760
Speaker 1: that I'm working on right now to kind of line up,

1061
00:57:23,040 --> 00:57:26,080
Speaker 1: So that's really exciting for me. I love having another

1062
00:57:26,120 --> 00:57:29,240
Speaker 1: point of view come into the conversation. I hope you

1063
00:57:29,280 --> 00:57:32,680
Speaker 1: do too, and I will talk to you again really soon.

1064
00:57:38,800 --> 00:57:43,480
Speaker 1: Tech Stuff is an iHeartRadio production. For more podcasts from iHeartRadio,

1065
00:57:43,800 --> 00:57:47,520
Speaker 1: visit the iHeartRadio app, Apple Podcasts, or wherever you listen

1066
00:57:47,560 --> 00:57:48,600
Speaker 1: to your favorite shows.