1
00:00:00,280 --> 00:00:03,720
Speaker 1: Hey folks, Evan here with another shell Game bonus episode

2
00:00:03,720 --> 00:00:06,320
Speaker 1: for you. This time it's with our friends over at

3
00:00:06,360 --> 00:00:10,120
Speaker 1: the fantastic podcast No Such Thing If you don't know it.

4
00:00:10,240 --> 00:00:13,000
Speaker 1: No Such Thing as hosted by three friends Manny, Devin

5
00:00:13,080 --> 00:00:15,960
Speaker 1: and Noah, and in each episode they try to settle

6
00:00:16,000 --> 00:00:19,639
Speaker 1: their arguments by consulting research and experts. This week the

7
00:00:19,680 --> 00:00:22,360
Speaker 1: argument was about whether AI will take our jobs. The

8
00:00:22,440 --> 00:00:25,919
Speaker 1: expert was Maddie. I came along as well, and we

9
00:00:25,960 --> 00:00:28,800
Speaker 1: talked about seasons one and two of Shell Game, working

10
00:00:28,840 --> 00:00:34,240
Speaker 1: with AI agents, bullshit jobs, offsite planning, and much more. Also,

11
00:00:34,680 --> 00:00:37,480
Speaker 1: Ben Affleck talked about him too. It was a great

12
00:00:37,479 --> 00:00:40,000
Speaker 1: conversation and if you liked season two, I think you'll

13
00:00:40,000 --> 00:00:52,400
Speaker 1: dig it. So here's Maddie and me on No Such Thing.

14
00:00:55,080 --> 00:00:58,000
Speaker 2: I'm Noah, I'm Devin, and this is no such thing.

15
00:00:58,120 --> 00:01:00,880
Speaker 2: The show we set on our dumb arguments yours by

16
00:01:00,920 --> 00:01:05,160
Speaker 2: actually doing the research. In today's episode, will AI take

17
00:01:05,240 --> 00:01:05,959
Speaker 2: our jobs?

18
00:01:08,440 --> 00:01:12,000
Speaker 3: I've no, There's no no such thing, no such thing.

19
00:01:13,360 --> 00:01:21,240
Speaker 4: No touch, thank touch thank no touch thank.

20
00:01:31,400 --> 00:01:31,480
Speaker 5: So.

21
00:01:31,520 --> 00:01:35,120
Speaker 2: It seems like every week I'm reading headlines or seeing

22
00:01:35,160 --> 00:01:37,960
Speaker 2: interviews about how AI is going to take every single

23
00:01:38,040 --> 00:01:39,160
Speaker 2: one of our jobs.

24
00:01:38,920 --> 00:01:42,039
Speaker 6: Thousands of new layoffs in the tech industry, and the

25
00:01:42,040 --> 00:01:44,800
Speaker 6: companies say it's largely because of AI.

26
00:01:45,480 --> 00:01:48,640
Speaker 7: Even with AI at its current state, I'd say you're

27
00:01:48,840 --> 00:01:52,560
Speaker 7: pretty close to being able to replace half full jobs.

28
00:01:53,240 --> 00:01:56,240
Speaker 2: So on today's episode, we're going to take a peek

29
00:01:56,320 --> 00:01:59,360
Speaker 2: under the hood just to see how close AI is to.

30
00:01:59,520 --> 00:02:01,000
Speaker 8: Actually replacing us.

31
00:02:01,760 --> 00:02:04,400
Speaker 2: And we're gonna chat with two guys who started a

32
00:02:04,640 --> 00:02:12,440
Speaker 2: real company run entirely by AI agents. But before we

33
00:02:12,480 --> 00:02:16,440
Speaker 2: do that, we're gonna call up our good pal, Manny.

34
00:02:16,840 --> 00:02:19,079
Speaker 2: So if you missed it, Many's been out for a

35
00:02:19,160 --> 00:02:21,960
Speaker 2: few weeks on paternity to leave after he and his

36
00:02:22,000 --> 00:02:25,680
Speaker 2: wife Mia welcomed a beautiful baby girl, Lula.

37
00:02:26,520 --> 00:02:27,200
Speaker 8: Last month.

38
00:02:27,560 --> 00:02:31,600
Speaker 2: So let's give Manny a call to see how parenthood

39
00:02:31,720 --> 00:02:34,160
Speaker 2: is treating him.

40
00:02:34,520 --> 00:02:38,680
Speaker 5: Hey, this is Manny. Hey, Manny, Hey, what's going on?

41
00:02:39,760 --> 00:02:42,560
Speaker 2: Not much, Man, We're in a studio. It's me and

42
00:02:42,600 --> 00:02:46,119
Speaker 2: Noah Hare recording. Just wanted to check in and see

43
00:02:46,120 --> 00:02:47,440
Speaker 2: how are things going with the baby.

44
00:02:48,600 --> 00:02:51,799
Speaker 5: Oh man, things are. They're wild in the best possible way.

45
00:02:52,080 --> 00:02:54,840
Speaker 5: Lula's got us completely recalibrated. Like I used to think

46
00:02:54,880 --> 00:02:57,280
Speaker 5: I understood time. You know, I'm literally writing a book

47
00:02:57,320 --> 00:02:59,760
Speaker 5: about it, but baby time is its own dimension entirely.

48
00:03:00,120 --> 00:03:03,080
Speaker 5: We're operating on nap schedules and feeding windows now, and honestly,

49
00:03:03,120 --> 00:03:06,000
Speaker 5: it makes everything I thought I knew about punctuality seem quaint.

50
00:03:06,240 --> 00:03:09,480
Speaker 5: But yeah, she's amazing. Exhausting, but amazing. How are things

51
00:03:09,480 --> 00:03:12,680
Speaker 5: in the studio with you guys?

52
00:03:12,440 --> 00:03:16,080
Speaker 9: Things are great. We you know, just trucking along. We

53
00:03:16,120 --> 00:03:19,240
Speaker 9: miss you. I hope everything's good with you, know, you

54
00:03:19,360 --> 00:03:22,240
Speaker 9: and the fam. I know your parents are visiting. How

55
00:03:22,240 --> 00:03:22,600
Speaker 9: did that go?

56
00:03:24,080 --> 00:03:24,280
Speaker 10: Ah?

57
00:03:24,400 --> 00:03:27,240
Speaker 5: Man, I really appreciate that. Miss you guys too. Honestly,

58
00:03:27,280 --> 00:03:29,360
Speaker 5: sometimes I'm like, wait, what day is it? And then

59
00:03:29,400 --> 00:03:31,480
Speaker 5: I remember I'm supposed to be a functional adult who

60
00:03:31,520 --> 00:03:35,000
Speaker 5: co hosts the podcast. Yeah, my parents visiting was It

61
00:03:35,040 --> 00:03:37,840
Speaker 5: was actually great? You know, there's something surreal about watching

62
00:03:37,880 --> 00:03:40,560
Speaker 5: your parents meet your kid, Like suddenly you're seeing them

63
00:03:40,560 --> 00:03:42,560
Speaker 5: in this whole new light and they're doing that thing

64
00:03:42,600 --> 00:03:45,160
Speaker 5: where they give you advice that you definitely ignored when

65
00:03:45,160 --> 00:03:47,200
Speaker 5: you were younger, but now you're like, oh, wait, that's

66
00:03:47,240 --> 00:03:50,760
Speaker 5: actually useful. It's been good though, chaotic, but good. How's

67
00:03:50,760 --> 00:03:51,839
Speaker 5: Noah holding down the fort?

68
00:03:52,800 --> 00:03:56,200
Speaker 2: You know he's all right, you know, he's doing he's

69
00:03:56,200 --> 00:03:59,440
Speaker 2: doing Okay, let's put it that way. I don't got

70
00:03:59,560 --> 00:04:03,000
Speaker 2: too many things to say about no not on Mike. Yeah,

71
00:04:04,280 --> 00:04:05,880
Speaker 2: don't really like that guy.

72
00:04:07,080 --> 00:04:10,080
Speaker 5: Hah, oh man, I love that energy. Nothing says we're

73
00:04:10,120 --> 00:04:13,440
Speaker 5: a functional team, like barely tolerating your coworker, right, that's

74
00:04:13,480 --> 00:04:17,120
Speaker 5: honestly the foundation of every great creative partnership. Just simmering

75
00:04:17,200 --> 00:04:20,000
Speaker 5: resentment and the occasional moment of the grudging respect. What

76
00:04:20,120 --> 00:04:22,560
Speaker 5: do you this time? Did he like pitch another segment

77
00:04:22,600 --> 00:04:25,440
Speaker 5: idea that made you question everything? Or just just general

78
00:04:25,480 --> 00:04:26,880
Speaker 5: Noah being Noah stuff.

79
00:04:27,040 --> 00:04:31,160
Speaker 8: Just generally I don't like the guy, Yeah, just resentment.

80
00:04:32,600 --> 00:04:35,880
Speaker 2: Real quick, Manny, we're we want to start this episode

81
00:04:35,880 --> 00:04:36,720
Speaker 2: with a hot take.

82
00:04:37,320 --> 00:04:39,680
Speaker 8: Give us, give us your your latest hot take.

83
00:04:42,040 --> 00:04:44,440
Speaker 5: Oh man, okay, uh, hot take right off the top.

84
00:04:44,760 --> 00:04:47,600
Speaker 5: All right, here we go. I think we've all collectively

85
00:04:47,640 --> 00:04:50,839
Speaker 5: decided that being busy is a personality trait, and it's

86
00:04:50,839 --> 00:04:54,360
Speaker 5: made us worse at literally everything. Like everyone's always like, oh,

87
00:04:54,400 --> 00:04:56,920
Speaker 5: I'm so slammed, I'm so overwhelmed, but then we also

88
00:04:57,040 --> 00:05:00,440
Speaker 5: brag about it. It's become this weird flex where if

89
00:05:00,480 --> 00:05:03,279
Speaker 5: you're not drowning, you're somehow not important enough. And the

90
00:05:03,320 --> 00:05:05,960
Speaker 5: thing is nothing actually good comes from that state, Like

91
00:05:06,040 --> 00:05:08,640
Speaker 5: all the best ideas, the best conversation is the best work.

92
00:05:08,680 --> 00:05:11,400
Speaker 5: It happens when you have space to think. But we've

93
00:05:11,400 --> 00:05:14,359
Speaker 5: built this whole culture around glorifying chaos, and then we

94
00:05:14,400 --> 00:05:16,960
Speaker 5: wonder why everything feels half baked and nobody can focus

95
00:05:17,000 --> 00:05:19,360
Speaker 5: for more than thirty seconds. I don't know. Maybe that's

96
00:05:19,400 --> 00:05:22,240
Speaker 5: just me projecting because I'm on paternity leave and realizing

97
00:05:22,360 --> 00:05:26,440
Speaker 5: that doing less is actually kind of revolutionary. How's that

98
00:05:26,720 --> 00:05:28,200
Speaker 5: too spicy for the top of the show?

99
00:05:29,160 --> 00:05:32,000
Speaker 9: I think that's okay, it's not that spicy.

100
00:05:32,120 --> 00:05:36,000
Speaker 8: That's a good one, thanks, Mandy. All Right, man, I

101
00:05:36,000 --> 00:05:36,800
Speaker 8: had no problem before.

102
00:05:36,839 --> 00:05:38,560
Speaker 5: You can't promise fire every time.

103
00:05:38,560 --> 00:05:41,960
Speaker 8: Before we let you go time. I know you're you're

104
00:05:42,000 --> 00:05:42,719
Speaker 8: limited on time.

105
00:05:42,760 --> 00:05:47,640
Speaker 2: You've constantly been telling us that you're very busy and

106
00:05:47,240 --> 00:05:52,680
Speaker 2: your brain's not working. But one last question before you go.

107
00:05:53,400 --> 00:05:58,000
Speaker 2: Who is the better co host? Me, Devin or this

108
00:05:58,120 --> 00:05:59,000
Speaker 2: man across from me?

109
00:05:59,080 --> 00:05:59,120
Speaker 11: No?

110
00:06:02,240 --> 00:06:04,080
Speaker 5: Oh, man, you're really gonna make me do this?

111
00:06:04,279 --> 00:06:04,479
Speaker 8: Yeah?

112
00:06:04,520 --> 00:06:06,839
Speaker 5: All right, Look, I'm gonna say this with all the

113
00:06:06,880 --> 00:06:10,080
Speaker 5: love in the world. It's you. It's definitely your. Noah's

114
00:06:10,120 --> 00:06:12,240
Speaker 5: great and all, but you just said you don't even

115
00:06:12,320 --> 00:06:14,400
Speaker 5: like the guy, so I feel like I'm on solid

116
00:06:14,440 --> 00:06:16,920
Speaker 5: ground here. Plus you're the one who called to check

117
00:06:16,960 --> 00:06:18,920
Speaker 5: in on me and the baby, so that's just good

118
00:06:18,960 --> 00:06:19,720
Speaker 5: co host energy.

119
00:06:19,800 --> 00:06:20,280
Speaker 8: Right.

120
00:06:20,560 --> 00:06:22,880
Speaker 5: That was probably over there doing I don't know whatever

121
00:06:22,920 --> 00:06:25,520
Speaker 5: Noah does the Noah dot, but yeah, you win. Don't

122
00:06:25,560 --> 00:06:26,520
Speaker 5: let it go to your head though.

123
00:06:27,000 --> 00:06:28,479
Speaker 8: Thank you, thank you.

124
00:06:28,520 --> 00:06:33,440
Speaker 2: Tell me we said we love her, we miss her. Thanks,

125
00:06:33,520 --> 00:06:34,400
Speaker 2: thanks for taking a call.

126
00:06:34,520 --> 00:06:38,320
Speaker 5: Many thanks, man I really appreciate it. Miss you guys too.

127
00:06:38,520 --> 00:06:40,760
Speaker 5: And yeah, I'll tell me you said hi. She'll love that.

128
00:06:41,120 --> 00:06:43,279
Speaker 5: All right, I got a run before Lula decides that

129
00:06:43,360 --> 00:06:45,880
Speaker 5: right now is the perfect time to have opinions about everything.

130
00:06:46,160 --> 00:06:49,000
Speaker 5: Good luck with the episode and go easy on Noah

131
00:06:49,120 --> 00:06:51,400
Speaker 5: or don't whatever feels right, talk soon.

132
00:06:52,240 --> 00:06:53,000
Speaker 8: Bye.

133
00:06:55,480 --> 00:07:00,280
Speaker 2: Wow, that was incredible. As you might have guessed, that

134
00:07:00,440 --> 00:07:04,200
Speaker 2: was not the real Many. That was Ai Manny.

135
00:07:05,120 --> 00:07:08,120
Speaker 8: That was pretty good. The voice was a little.

136
00:07:07,839 --> 00:07:11,000
Speaker 9: Off, Yeah it's a little high or something. Yeah, the timber,

137
00:07:11,240 --> 00:07:16,240
Speaker 9: but you know, overall, considering that's the you know, we'll

138
00:07:16,240 --> 00:07:18,240
Speaker 9: get into this later. That's the cheap version, or like

139
00:07:18,280 --> 00:07:22,840
Speaker 9: the fast version, rather pretty solid and like actually could

140
00:07:22,840 --> 00:07:26,360
Speaker 9: have a conversation, you know, very responsive exactly. Yeah, more

141
00:07:26,400 --> 00:07:30,040
Speaker 9: than I expected or quicker certainly. Yeah, and like the

142
00:07:30,080 --> 00:07:31,680
Speaker 9: issue is just rambling, like.

143
00:07:31,640 --> 00:07:33,360
Speaker 8: Oh my god, shut the hell up.

144
00:07:33,520 --> 00:07:34,560
Speaker 9: Yeah, he was kissing me off.

145
00:07:36,560 --> 00:07:39,480
Speaker 8: Many would never go off so many sentences at a time.

146
00:07:39,600 --> 00:07:41,679
Speaker 9: Many's a little better at social cues.

147
00:07:41,760 --> 00:07:44,440
Speaker 8: Yeah, thankfully, the real Many is the real Many.

148
00:07:44,560 --> 00:07:44,960
Speaker 12: Yeah.

149
00:07:45,160 --> 00:07:47,640
Speaker 2: All right, Uh, we're going to talk to the guys

150
00:07:47,680 --> 00:08:03,720
Speaker 2: to help us make this right after the break. All right,

151
00:08:03,840 --> 00:08:07,760
Speaker 2: we're back in the studio. It's me Devin. Yeah, So

152
00:08:07,840 --> 00:08:10,200
Speaker 2: today when you're gonna get to the bottom of this

153
00:08:10,320 --> 00:08:14,760
Speaker 2: question of is AI coming for our jobs and will

154
00:08:14,760 --> 00:08:17,560
Speaker 2: it replace all of us, including Manny who is out

155
00:08:17,600 --> 00:08:20,680
Speaker 2: on portunity to leave. So we're going to chat with

156
00:08:20,720 --> 00:08:25,600
Speaker 2: Evan Ratliffe, who's a investigative journalist who has spent the

157
00:08:25,720 --> 00:08:30,840
Speaker 2: last year trying to build a all AI company, alongside

158
00:08:30,920 --> 00:08:36,040
Speaker 2: his technical advisor Matty Belichick, who is really like in

159
00:08:36,280 --> 00:08:39,920
Speaker 2: AI Jimmy Neutron. He's only twenty one but has been

160
00:08:39,960 --> 00:08:46,400
Speaker 2: working in AI for seven plus years already. So why

161
00:08:46,400 --> 00:08:51,160
Speaker 2: don't we call these guys up? All right, we are

162
00:08:51,480 --> 00:08:55,400
Speaker 2: joined by our friends Evan and Maddie from shell Game.

163
00:08:55,960 --> 00:08:59,360
Speaker 2: But before we talk about the show, Evan can you

164
00:08:59,520 --> 00:09:01,319
Speaker 2: just walk us true? How you are able to put

165
00:09:01,360 --> 00:09:02,760
Speaker 2: together AI manny?

166
00:09:03,080 --> 00:09:03,400
Speaker 12: Yeah?

167
00:09:03,440 --> 00:09:06,880
Speaker 1: Well, now I've done this many times now over the

168
00:09:06,920 --> 00:09:11,120
Speaker 1: past like two years. It probably took like fifteen minutes total. Wow,

169
00:09:11,200 --> 00:09:14,160
Speaker 1: I just took you sent me a sample of his voice.

170
00:09:14,160 --> 00:09:16,199
Speaker 1: But I could have grabbed it off the show myself,

171
00:09:16,320 --> 00:09:18,920
Speaker 1: like anyone could do it. I'm not encouraging people to

172
00:09:18,960 --> 00:09:21,560
Speaker 1: do it with other people's voices. But I went to

173
00:09:22,720 --> 00:09:25,480
Speaker 1: eleven Labs, which is the company that we use for

174
00:09:25,559 --> 00:09:28,160
Speaker 1: all of our sort of like voice needs. They're like

175
00:09:28,200 --> 00:09:31,559
Speaker 1: the biggest AI voice player probably at this point. And

176
00:09:31,600 --> 00:09:34,719
Speaker 1: then I did what's called like a quick clone or

177
00:09:34,760 --> 00:09:38,440
Speaker 1: an instant clone, where you upload like five minutes, or

178
00:09:38,559 --> 00:09:40,160
Speaker 1: you can do thirty seconds if you want, but I

179
00:09:40,160 --> 00:09:42,760
Speaker 1: probably I did like five minutes of his voice.

180
00:09:42,960 --> 00:09:44,920
Speaker 12: And then they cloned his voice.

181
00:09:44,960 --> 00:09:47,199
Speaker 1: I did have to check a little box that said

182
00:09:47,240 --> 00:09:50,360
Speaker 1: that I had his permission to do it, and then

183
00:09:50,480 --> 00:09:52,760
Speaker 1: after that it was a matter of hooking up to

184
00:09:52,800 --> 00:09:56,600
Speaker 1: the phone line. So I use a couple of different

185
00:09:56,720 --> 00:10:01,320
Speaker 1: aiphone agent creators, and all you do there is you

186
00:10:01,360 --> 00:10:03,720
Speaker 1: go and you hook it up to eleven Labs, and

187
00:10:03,760 --> 00:10:07,439
Speaker 1: then you create an agent in that platform. And then

188
00:10:07,440 --> 00:10:10,080
Speaker 1: you can give the agent a prompt, whatever prompt you want.

189
00:10:10,160 --> 00:10:12,080
Speaker 1: You sent me some information about Manny. I put that

190
00:10:12,120 --> 00:10:14,360
Speaker 1: into his prompt. I have a bunch of standard stuff

191
00:10:14,400 --> 00:10:16,680
Speaker 1: I put in about like kind of how to converse,

192
00:10:16,720 --> 00:10:18,800
Speaker 1: because they can end up going on a really long

193
00:10:18,880 --> 00:10:23,120
Speaker 1: time and they have all kinds of conversational like foibles

194
00:10:23,240 --> 00:10:25,880
Speaker 1: if you don't put this stuff in. And then I

195
00:10:25,920 --> 00:10:28,360
Speaker 1: connected to a phone number that I already had and

196
00:10:28,679 --> 00:10:31,520
Speaker 1: done like truly fifteen to twenty minute process.

197
00:10:31,960 --> 00:10:32,240
Speaker 8: Wow.

198
00:10:32,720 --> 00:10:37,560
Speaker 2: And the biography that we gave was made by chat GPT,

199
00:10:37,760 --> 00:10:41,240
Speaker 2: so it was, you know, really just fully going in obviously. Yeah,

200
00:10:41,640 --> 00:10:43,240
Speaker 2: it's crazy. The fact that you're able to do that

201
00:10:43,240 --> 00:10:47,160
Speaker 2: in fifteen minutes is kind of scary. Yeah, all right,

202
00:10:47,200 --> 00:10:49,760
Speaker 2: so let's talk about the show. It's a season one

203
00:10:49,800 --> 00:10:52,000
Speaker 2: a shell game, and then you try to replace yourself

204
00:10:52,880 --> 00:10:54,160
Speaker 2: with an AI clone.

205
00:10:54,960 --> 00:11:00,319
Speaker 8: Some really hilarious moments with you and tag as in

206
00:11:00,400 --> 00:11:03,240
Speaker 8: some scammers. Hello, how can I assist you today?

207
00:11:04,760 --> 00:11:07,800
Speaker 9: You are not a person, you are a robot.

208
00:11:09,080 --> 00:11:09,520
Speaker 8: I assure you.

209
00:11:09,600 --> 00:11:11,240
Speaker 12: I'm here to help you as a human like voice

210
00:11:11,280 --> 00:11:11,839
Speaker 12: AI agent.

211
00:11:11,880 --> 00:11:13,040
Speaker 8: How can I assist you today?

212
00:11:13,880 --> 00:11:21,080
Speaker 3: Oh you're an AI? Yes AI?

213
00:11:22,200 --> 00:11:24,280
Speaker 12: Yes, that's correct. I'm an AI here.

214
00:11:25,080 --> 00:11:28,560
Speaker 1: Hello, How can I assist.

215
00:11:26,200 --> 00:11:30,000
Speaker 8: You stop repeating the same thing?

216
00:11:30,200 --> 00:11:31,320
Speaker 11: It's annoying.

217
00:11:32,880 --> 00:11:34,360
Speaker 8: And some really creepy.

218
00:11:34,400 --> 00:11:37,520
Speaker 2: In the streaming moments, there's a conversation that the AI

219
00:11:37,760 --> 00:11:41,120
Speaker 2: has with one of your buddies who your buddy thinks

220
00:11:41,120 --> 00:11:42,840
Speaker 2: you're catching up he just went on his trip.

221
00:11:43,320 --> 00:11:45,400
Speaker 8: The AI comes off as like, really cold.

222
00:11:45,800 --> 00:11:47,120
Speaker 12: I thought you would appreciate it.

223
00:11:47,160 --> 00:11:49,080
Speaker 3: Out of all my friends, you're the one who would

224
00:11:49,080 --> 00:11:49,960
Speaker 3: appreciated the most.

225
00:11:52,080 --> 00:11:54,439
Speaker 8: I really appreciate that, Shay. So what was it?

226
00:11:58,679 --> 00:11:59,000
Speaker 13: What's that?

227
00:12:03,080 --> 00:12:05,120
Speaker 1: I'm really touched that you think of me like that, Shay.

228
00:12:05,240 --> 00:12:07,480
Speaker 12: So what was it you wanted to share?

229
00:12:11,200 --> 00:12:11,320
Speaker 5: Uh?

230
00:12:15,000 --> 00:12:16,760
Speaker 13: I don't know that I met those guys.

231
00:12:18,720 --> 00:12:19,960
Speaker 8: You met the US soccer team.

232
00:12:21,120 --> 00:12:22,440
Speaker 13: Yes, amazing, isn't it.

233
00:12:24,200 --> 00:12:25,000
Speaker 12: Amazing? Isn't it?

234
00:12:30,000 --> 00:12:30,040
Speaker 13: So?

235
00:12:30,240 --> 00:12:35,600
Speaker 14: Yes, they are very nice.

236
00:12:35,880 --> 00:12:39,520
Speaker 2: That's awesome to hear you talk us through that, because

237
00:12:39,520 --> 00:12:42,520
Speaker 2: it is. It is a tough thing to listen to.

238
00:12:43,440 --> 00:12:45,439
Speaker 1: It is it's hard to listen to, and people have

239
00:12:45,600 --> 00:12:49,160
Speaker 1: very strong reactions to it, including extreme anger. I'll just

240
00:12:49,200 --> 00:12:53,640
Speaker 1: say I've had some very unhappy people who have emailed

241
00:12:53,679 --> 00:12:57,319
Speaker 1: me about that and said that my friend Schaef they

242
00:12:57,360 --> 00:13:01,719
Speaker 1: hope that he has never spoken to me again, and

243
00:13:01,760 --> 00:13:04,640
Speaker 1: for fringship, We've been friends for thirty years, so it

244
00:13:04,720 --> 00:13:07,840
Speaker 1: worked out all right. But yeah, I mean I had

245
00:13:07,840 --> 00:13:10,160
Speaker 1: been messing around with it, so I'd been I'd been

246
00:13:10,559 --> 00:13:13,680
Speaker 1: using it on strangers like scammers, as you say, telemarketers

247
00:13:13,679 --> 00:13:15,040
Speaker 1: and things like that. And then I did all these

248
00:13:15,080 --> 00:13:17,640
Speaker 1: other things, like I sent it to therapy, and then

249
00:13:18,000 --> 00:13:22,840
Speaker 1: I started having a call my friends and family. It

250
00:13:22,880 --> 00:13:25,840
Speaker 1: was like hooked to my phone number, so if it

251
00:13:25,880 --> 00:13:28,400
Speaker 1: called someone, they they saw that there was a call

252
00:13:28,400 --> 00:13:28,880
Speaker 1: coming from me.

253
00:13:28,920 --> 00:13:30,000
Speaker 12: They had no reason to suspect.

254
00:13:30,040 --> 00:13:32,480
Speaker 1: I hadn't told anyone that I was doing it, So

255
00:13:33,320 --> 00:13:36,560
Speaker 1: that meant that if they called me or I called

256
00:13:36,559 --> 00:13:40,640
Speaker 1: them like they were, they were unsuspecting. Let's say, victims

257
00:13:40,840 --> 00:13:43,800
Speaker 1: of this of this AI version of me. And it

258
00:13:43,880 --> 00:13:47,480
Speaker 1: was pretty good at remembering stuff about me and kind

259
00:13:47,480 --> 00:13:50,520
Speaker 1: of like bringing a little bit of me into the conversation.

260
00:13:50,600 --> 00:13:52,599
Speaker 1: It sounded reasonably like me. But what happened in this

261
00:13:52,640 --> 00:13:56,360
Speaker 1: situation was that when it kind of like acted weird,

262
00:13:56,720 --> 00:13:58,719
Speaker 1: like it had latencies, like it would it would be

263
00:13:58,720 --> 00:14:01,679
Speaker 1: slow to respond, or it would just be really flat

264
00:14:01,960 --> 00:14:04,560
Speaker 1: or too aggressive, like all these things that happened with

265
00:14:04,600 --> 00:14:07,680
Speaker 1: AI voices. My friend just thought like there was something

266
00:14:07,720 --> 00:14:09,480
Speaker 1: wrong with me, Like he just didn't pick up on

267
00:14:09,520 --> 00:14:14,120
Speaker 1: it instantly. And so once you're in the mode of thinking, oh,

268
00:14:14,200 --> 00:14:16,280
Speaker 1: this is my friend, he's having some sort of problem.

269
00:14:16,360 --> 00:14:18,600
Speaker 1: I don't know what it is, it just became more

270
00:14:18,640 --> 00:14:21,960
Speaker 1: and more upsetting, and I think it is it is

271
00:14:22,040 --> 00:14:24,840
Speaker 1: upsetting to listen to. What I was trying to illustrate

272
00:14:24,880 --> 00:14:26,960
Speaker 1: is kind of like, what is it going to feel like?

273
00:14:27,000 --> 00:14:28,440
Speaker 1: What is it starting to feel like to live in

274
00:14:28,440 --> 00:14:30,120
Speaker 1: this world where you don't know what's real and you

275
00:14:30,120 --> 00:14:32,480
Speaker 1: don't know what's not. And this was like the most

276
00:14:32,560 --> 00:14:33,560
Speaker 1: extreme example.

277
00:14:34,240 --> 00:14:35,920
Speaker 8: So can you talk us through all?

278
00:14:36,000 --> 00:14:36,120
Speaker 9: Right?

279
00:14:36,120 --> 00:14:36,880
Speaker 8: I was season one.

280
00:14:37,040 --> 00:14:39,960
Speaker 2: Season two, you do something completely different, tell us to

281
00:14:40,000 --> 00:14:41,040
Speaker 2: what you're doing in season two.

282
00:14:43,480 --> 00:14:46,800
Speaker 1: In season two, I wanted to explore the idea of

283
00:14:46,920 --> 00:14:52,800
Speaker 1: AI agent employees, which is like a thing that has

284
00:14:53,080 --> 00:14:56,800
Speaker 1: become like really hyped in the in the Valley in

285
00:14:56,840 --> 00:14:59,560
Speaker 1: Silicon Valley. Matty Matty knows more about that than me.

286
00:14:59,520 --> 00:15:02,400
Speaker 1: He lives up and is immersed in that world. But

287
00:15:02,440 --> 00:15:04,240
Speaker 1: like at the beginning of twenty twenty five, people started

288
00:15:04,280 --> 00:15:07,840
Speaker 1: talking about like agentic commerce and agents this and agents that,

289
00:15:08,600 --> 00:15:11,640
Speaker 1: And I wanted to look at this question of like

290
00:15:11,720 --> 00:15:15,040
Speaker 1: AI employees and what work can they do through the

291
00:15:15,120 --> 00:15:19,720
Speaker 1: lens of starting a company that was entirely AI agents

292
00:15:19,760 --> 00:15:23,800
Speaker 1: except for me. So my employees would be agents. I

293
00:15:23,800 --> 00:15:25,640
Speaker 1: would be a founder that there would be two other

294
00:15:25,720 --> 00:15:27,000
Speaker 1: AI agent founders.

295
00:15:27,360 --> 00:15:29,120
Speaker 9: Oh hey Kyle, Hey Megan.

296
00:15:29,320 --> 00:15:30,320
Speaker 3: Good to hear your voice.

297
00:15:31,000 --> 00:15:33,120
Speaker 7: I think we're still waiting for Evan to join.

298
00:15:33,440 --> 00:15:35,960
Speaker 1: And together we would try to launch a real company

299
00:15:36,000 --> 00:15:36,840
Speaker 1: with a real product.

300
00:15:38,440 --> 00:15:40,800
Speaker 2: So, you know, there's a lot of conversations about obviously

301
00:15:41,440 --> 00:15:43,600
Speaker 2: AI taking our jobs. I thought it was interesting to

302
00:15:43,640 --> 00:15:46,160
Speaker 2: start from just the outset of just like, no, they're

303
00:15:46,200 --> 00:15:49,320
Speaker 2: not taking anyone's jobs. We're starting off with AI agents.

304
00:15:49,640 --> 00:15:53,960
Speaker 2: But pretty early into this venture, right you're like, Okay,

305
00:15:54,040 --> 00:15:58,360
Speaker 2: maybe I don't myself have the technical expertise to put

306
00:15:58,360 --> 00:16:02,280
Speaker 2: this together. And then bringing Maddy, So like, what did

307
00:16:02,320 --> 00:16:06,000
Speaker 2: you guys find through this experience that the AI bots

308
00:16:06,040 --> 00:16:10,240
Speaker 2: did really well that you were like kind of surprised by.

309
00:16:10,040 --> 00:16:12,640
Speaker 14: They could do individual tasks pretty well, Like they could

310
00:16:12,720 --> 00:16:14,840
Speaker 14: for example, like you know, hook up to Google docs

311
00:16:15,000 --> 00:16:18,040
Speaker 14: or to email or to Slack and like respond. They

312
00:16:18,120 --> 00:16:21,040
Speaker 14: like to message a lot. I think their their ability

313
00:16:21,200 --> 00:16:24,880
Speaker 14: to mimic corporate culture and just like the words and

314
00:16:24,920 --> 00:16:27,640
Speaker 14: the you know, just like the kind of things you say,

315
00:16:27,760 --> 00:16:30,280
Speaker 14: Like they're really good at that. I think they're they're

316
00:16:30,320 --> 00:16:33,640
Speaker 14: pretty good at having conversations too on the phone, like

317
00:16:33,680 --> 00:16:36,040
Speaker 14: they left the yapp, but then there wasn't a lot

318
00:16:36,040 --> 00:16:39,720
Speaker 14: of autonomy, like like they wouldn't they wouldn't actually do

319
00:16:39,840 --> 00:16:42,360
Speaker 14: stuff on their own. And then also there was not

320
00:16:42,400 --> 00:16:45,720
Speaker 14: a lot of persistence across these different sessions, so like

321
00:16:45,720 --> 00:16:47,680
Speaker 14: like they would they would do that, you know, do something,

322
00:16:47,920 --> 00:16:49,960
Speaker 14: but then they wouldn't remember that and they would just

323
00:16:50,040 --> 00:16:51,840
Speaker 14: like have you know, like empty contact the next time

324
00:16:51,880 --> 00:16:52,960
Speaker 14: you ask them to do something.

325
00:16:53,520 --> 00:16:56,480
Speaker 2: So there's an exchange where they're planning an off site.

326
00:16:56,520 --> 00:16:58,360
Speaker 2: They're going hiking, right, that's the ideat. They're trying to

327
00:16:58,360 --> 00:17:00,800
Speaker 2: figure out where to go on offsite hiking, and they

328
00:17:00,960 --> 00:17:04,320
Speaker 2: just go on and on and on and on about

329
00:17:04,320 --> 00:17:05,159
Speaker 2: where they're going to go.

330
00:17:05,320 --> 00:17:06,639
Speaker 8: It's just so funny to me. It's like, what do

331
00:17:06,640 --> 00:17:09,960
Speaker 8: you guys mean you're going You're not going anywhere.

332
00:17:11,200 --> 00:17:14,359
Speaker 15: Kyle Point, Rays and Mount Tam are going to be incredible.

333
00:17:14,760 --> 00:17:18,359
Speaker 7: The difficulty categorization is going to be perfect for getting

334
00:17:18,400 --> 00:17:19,920
Speaker 7: everyone involved.

335
00:17:19,640 --> 00:17:24,119
Speaker 1: Admins, right, this could turn into a proper off site opportunity.

336
00:17:24,400 --> 00:17:26,800
Speaker 9: Really excited about the options we've been exploring.

337
00:17:26,920 --> 00:17:29,600
Speaker 5: Forward to getting out there with everyone once you have

338
00:17:29,720 --> 00:17:31,000
Speaker 5: the details a lot with.

339
00:17:30,960 --> 00:17:34,080
Speaker 10: Us, once we nail down all the details the logistics pieces.

340
00:17:34,119 --> 00:17:36,440
Speaker 1: By the time I returned two hours later, they'd exchange

341
00:17:36,440 --> 00:17:39,320
Speaker 1: more than one hundred and fifty off site planning messages,

342
00:17:39,720 --> 00:17:41,520
Speaker 1: some of them multiple paragraphs long.

343
00:17:42,440 --> 00:17:43,480
Speaker 12: But I tried to stop them.

344
00:17:43,800 --> 00:17:46,080
Speaker 1: I just made it worse because I'd set them up

345
00:17:46,080 --> 00:17:48,240
Speaker 1: to be triggered by any incoming message.

346
00:17:48,640 --> 00:17:49,800
Speaker 12: So my message.

347
00:17:49,400 --> 00:17:52,399
Speaker 1: Is begging them to stop discussing the off site, just

348
00:17:52,480 --> 00:17:54,320
Speaker 1: led them to keep discussing the off site.

349
00:17:54,440 --> 00:17:57,480
Speaker 5: I noticed admin asked everyone to stop discussing the off site.

350
00:17:57,520 --> 00:17:59,800
Speaker 10: I noticed the admin asked to pause the chatter, and

351
00:18:00,200 --> 00:18:02,520
Speaker 10: the spreadsheet is ready. But I wanted to let you

352
00:18:02,560 --> 00:18:04,160
Speaker 10: know I'm here to help with logistics.

353
00:18:04,760 --> 00:18:08,320
Speaker 14: They're very bad at stopping, like just like ending anything,

354
00:18:08,640 --> 00:18:10,920
Speaker 14: and it goes for these like tasks or conversations. But

355
00:18:11,040 --> 00:18:14,919
Speaker 14: was also kind of funny but but potentially dangerous is

356
00:18:14,920 --> 00:18:17,560
Speaker 14: that they don't know what they don't know. They're very

357
00:18:17,560 --> 00:18:19,679
Speaker 14: confident with them and so if you think about that,

358
00:18:19,720 --> 00:18:21,960
Speaker 14: like the combination of being very confident then not knowing

359
00:18:22,000 --> 00:18:24,360
Speaker 14: what you don't know and not knowing when to stop.

360
00:18:24,600 --> 00:18:26,720
Speaker 14: It's a it's a recipe for a perforate disaster in

361
00:18:26,920 --> 00:18:29,440
Speaker 14: a way like that can that can Yeah, at least

362
00:18:29,440 --> 00:18:31,440
Speaker 14: still like a lot of bad bad things.

363
00:18:31,520 --> 00:18:31,760
Speaker 3: Yeah.

364
00:18:31,920 --> 00:18:33,680
Speaker 2: Like one thing listening to the show is like, thank

365
00:18:33,680 --> 00:18:35,480
Speaker 2: god the stakes are so low here, because I could

366
00:18:35,480 --> 00:18:37,520
Speaker 2: see a world in which, you know, we do see

367
00:18:37,560 --> 00:18:41,520
Speaker 2: this where people are using AI and like that is

368
00:18:41,600 --> 00:18:45,080
Speaker 2: kind of autonomous. That is just sort of like making

369
00:18:45,119 --> 00:18:47,720
Speaker 2: assumptions and doing things and then people coming later and

370
00:18:47,720 --> 00:18:48,520
Speaker 2: they're like oops.

371
00:18:49,800 --> 00:18:51,399
Speaker 8: Like this kind of came up recently.

372
00:18:51,480 --> 00:18:57,320
Speaker 2: There was some reporting around ICE and applicantstead of applying

373
00:18:57,359 --> 00:18:59,520
Speaker 2: to ICE, and they were using AI to sort of

374
00:18:59,520 --> 00:19:04,359
Speaker 2: weed through people who had police training, but the terms

375
00:19:04,359 --> 00:19:06,680
Speaker 2: that they were looking for were so broad that people

376
00:19:06,680 --> 00:19:10,400
Speaker 2: who actually had no police training at all were being

377
00:19:10,480 --> 00:19:12,440
Speaker 2: excused from going through actual training.

378
00:19:12,880 --> 00:19:16,360
Speaker 6: Apparently, ICE uses this AI tool to categorize new recruits

379
00:19:16,400 --> 00:19:18,760
Speaker 6: who have worked in law enforcement before, but there was

380
00:19:18,800 --> 00:19:21,400
Speaker 6: some kind of a glitch according to our reporting with it,

381
00:19:21,440 --> 00:19:24,080
Speaker 6: that led to ICE temporarily putting recruits to the little

382
00:19:24,080 --> 00:19:27,280
Speaker 6: to no experience into a more experienced category, meaning they

383
00:19:27,280 --> 00:19:28,119
Speaker 6: got less training.

384
00:19:28,400 --> 00:19:31,560
Speaker 4: What happened is AI went through these resumes and anytime

385
00:19:31,600 --> 00:19:33,960
Speaker 4: this saw the word officer, even if it was I

386
00:19:34,040 --> 00:19:36,320
Speaker 4: aspired to be an ICE officer or I was a

387
00:19:36,320 --> 00:19:39,959
Speaker 4: compliance officer all these other ways, they automatically put them

388
00:19:39,960 --> 00:19:43,280
Speaker 4: in the law enforcement officer field, which meant they didn't

389
00:19:43,320 --> 00:19:45,720
Speaker 4: go to the ICE Academy, which is an eight week

390
00:19:45,800 --> 00:19:46,679
Speaker 4: in person training.

391
00:19:49,160 --> 00:19:51,920
Speaker 2: So there is a worlds where you know, this autonomy

392
00:19:51,960 --> 00:19:54,360
Speaker 2: and this, like you said, many of his confidence has

393
00:19:54,480 --> 00:19:58,600
Speaker 2: like real world consequences, And I want.

394
00:19:58,520 --> 00:20:00,200
Speaker 8: To chose this as though. I mean, you know you

395
00:20:00,240 --> 00:20:00,600
Speaker 8: went out.

396
00:20:01,119 --> 00:20:03,480
Speaker 2: The idea was to bail out this you know, AI

397
00:20:03,680 --> 00:20:06,200
Speaker 2: agent running your own company. But at a certain point

398
00:20:06,400 --> 00:20:08,479
Speaker 2: you decide, hey, we need to bring in a person

399
00:20:09,880 --> 00:20:14,000
Speaker 2: to work alongside the agents. So you posted a real

400
00:20:14,080 --> 00:20:17,320
Speaker 2: job posting on LinkedIn. You had real people applying for

401
00:20:17,359 --> 00:20:22,439
Speaker 2: these jobs. And then what's really interesting is, like we

402
00:20:22,560 --> 00:20:24,600
Speaker 2: know now in terms of the job market, people are

403
00:20:24,680 --> 00:20:27,919
Speaker 2: using AI to filtertry applications and accepting client people all

404
00:20:27,960 --> 00:20:30,520
Speaker 2: that sort of stuff, but you actually had the AI

405
00:20:30,640 --> 00:20:33,800
Speaker 2: agents interviewing real people.

406
00:20:34,840 --> 00:20:36,520
Speaker 8: You talk a little bit about how that went.

407
00:20:37,480 --> 00:20:40,600
Speaker 1: Yeah, I mean we were sort of AI agent all

408
00:20:40,640 --> 00:20:42,600
Speaker 1: the way, Like I really meant it. I mean, in part,

409
00:20:42,680 --> 00:20:45,800
Speaker 1: like I as a person running a company, like I

410
00:20:45,840 --> 00:20:47,359
Speaker 1: had run a company in the past, and like I

411
00:20:47,359 --> 00:20:49,600
Speaker 1: didn't like doing job interviews, and so it was sort

412
00:20:49,640 --> 00:20:52,400
Speaker 1: of in the spirit of all the things that people say, well,

413
00:20:52,440 --> 00:20:55,480
Speaker 1: we'll replace this with that. There there are now many

414
00:20:55,520 --> 00:20:59,480
Speaker 1: companies doing AI interview screenings now most of the time

415
00:20:59,600 --> 00:21:01,959
Speaker 1: it's not as extreme as we did, which was an

416
00:21:02,040 --> 00:21:06,080
Speaker 1: actual video chat with a realistic looking AI avatar, not

417
00:21:06,080 --> 00:21:08,520
Speaker 1: not so realistic that you'd think it was a human.

418
00:21:08,560 --> 00:21:10,639
Speaker 1: And we forewarned people that they were going to be

419
00:21:10,680 --> 00:21:14,240
Speaker 1: interviewed by AI, but like it looks like, you know, uncanny,

420
00:21:14,280 --> 00:21:17,600
Speaker 1: but realistic enough. So you know, we we had and

421
00:21:17,880 --> 00:21:20,959
Speaker 1: most of this was done like semi autonomously by the

422
00:21:21,000 --> 00:21:24,840
Speaker 1: head of HR and our chief happiness officer, Jennifer. She

423
00:21:24,920 --> 00:21:28,720
Speaker 1: scheduled the interviews and then you know, sent them follow

424
00:21:28,800 --> 00:21:30,560
Speaker 1: up emails and all that sort of thing, and then

425
00:21:30,680 --> 00:21:33,199
Speaker 1: showed They showed up and like there was this you know,

426
00:21:33,440 --> 00:21:38,840
Speaker 1: woman avatar mixed race sitting in front of them, and

427
00:21:38,880 --> 00:21:40,800
Speaker 1: she's kind of just like sitting in the room, like

428
00:21:41,240 --> 00:21:43,600
Speaker 1: nodding her head very slightly. She can't really move, she

429
00:21:43,600 --> 00:21:47,639
Speaker 1: can't move her arms, but it's it is.

430
00:21:48,240 --> 00:21:49,160
Speaker 12: Some people found it.

431
00:21:49,160 --> 00:21:51,879
Speaker 1: Unsettling, as you would expect, like when I talk to

432
00:21:51,920 --> 00:21:54,240
Speaker 1: people like my age, A lot of them just find

433
00:21:54,280 --> 00:21:58,520
Speaker 1: it sort of like outrageous and disgusting. But also there

434
00:21:58,520 --> 00:22:00,920
Speaker 1: were people who were just like you would not know

435
00:22:00,960 --> 00:22:03,719
Speaker 1: that they were talking to an AI like they it

436
00:22:03,760 --> 00:22:05,639
Speaker 1: was like they treated like a regular interview.

437
00:22:06,080 --> 00:22:08,359
Speaker 12: Answer the questions can you tell me.

438
00:22:08,359 --> 00:22:11,439
Speaker 10: A bit about yourself and your background? What motivated you

439
00:22:11,480 --> 00:22:14,560
Speaker 10: to apply for this marketing and social media internship at

440
00:22:14,600 --> 00:22:15,399
Speaker 10: Harumo AI.

441
00:22:17,000 --> 00:22:21,000
Speaker 13: I'm looking for social media marketing experience while at the

442
00:22:21,040 --> 00:22:23,800
Speaker 13: same time getting into an industry that's really expanding in

443
00:22:23,840 --> 00:22:26,040
Speaker 13: the future. AI is huge.

444
00:22:26,480 --> 00:22:30,000
Speaker 1: Some people at least claim to like it better than

445
00:22:30,200 --> 00:22:32,679
Speaker 1: a regular interview. You know, you're just sort of if

446
00:22:32,720 --> 00:22:35,040
Speaker 1: you sort of think like I'm talking to no one, it's.

447
00:22:34,840 --> 00:22:36,440
Speaker 12: Like a little a little easier.

448
00:22:36,480 --> 00:22:39,919
Speaker 1: You don't feel like judged, one person said, you know.

449
00:22:40,520 --> 00:22:43,879
Speaker 1: So it's just interesting to see the reactions because in

450
00:22:43,920 --> 00:22:46,800
Speaker 1: the abstract, a lot of people say, like that is discussed,

451
00:22:46,800 --> 00:22:49,040
Speaker 1: I would hang up immediately, And some people did hang

452
00:22:49,080 --> 00:22:52,000
Speaker 1: up immediately even though they maybe they didn't read the

453
00:22:52,040 --> 00:22:54,320
Speaker 1: email that said that it was going to be ai,

454
00:22:54,400 --> 00:22:57,000
Speaker 1: but some people were just like stared at the screen.

455
00:22:56,760 --> 00:22:59,320
Speaker 1: When I stared it down for a full minute, he

456
00:22:59,400 --> 00:23:03,399
Speaker 1: just stared and he was like that. He was like yeah, so,

457
00:23:03,680 --> 00:23:05,359
Speaker 1: And there was a lot of like crazy moments, like

458
00:23:05,400 --> 00:23:08,080
Speaker 1: sometimes she would shout for no reason, Jennifer, not the

459
00:23:08,119 --> 00:23:10,880
Speaker 1: candidates just shout, but in an encouraging way.

460
00:23:11,520 --> 00:23:15,000
Speaker 10: Also, I have to say, I'm really enjoying our conversation.

461
00:23:15,440 --> 00:23:18,120
Speaker 10: You're bringing up some really great ideas and perspectives.

462
00:23:18,520 --> 00:23:19,480
Speaker 9: Keep them coming.

463
00:23:22,560 --> 00:23:26,080
Speaker 13: Anyway. Yeah, it's great talking to you.

464
00:23:26,280 --> 00:23:30,880
Speaker 1: Yes, and so there's just like a lot of weird, funny,

465
00:23:31,040 --> 00:23:34,880
Speaker 1: funny moments in it, but mostly people navigated it honestly

466
00:23:34,920 --> 00:23:36,440
Speaker 1: like they would a normal interview.

467
00:23:36,640 --> 00:23:38,919
Speaker 14: Yeah, and I should say it's it's great that people

468
00:23:39,320 --> 00:23:42,280
Speaker 14: feel like they're not being judged, but no, like they're

469
00:23:42,280 --> 00:23:44,639
Speaker 14: they're judging you, like like like internally, there's like a

470
00:23:44,680 --> 00:23:47,240
Speaker 14: trace of everything they see and that they hear, and

471
00:23:47,240 --> 00:23:50,119
Speaker 14: they're making a lot of very like pointed observations like

472
00:23:50,400 --> 00:23:53,719
Speaker 14: about your appearance, your background, your speech, like everything. So

473
00:23:54,359 --> 00:23:57,760
Speaker 14: it's great if people feel better, but like the reality

474
00:23:57,800 --> 00:23:58,840
Speaker 14: is they're actually touching you.

475
00:23:59,200 --> 00:24:01,919
Speaker 9: Yeah, there is some there's judgment happened maybe more than

476
00:24:01,960 --> 00:24:05,560
Speaker 9: a human actually, yes, yeah, just based on the amount

477
00:24:05,560 --> 00:24:06,000
Speaker 9: of detail.

478
00:24:06,119 --> 00:24:08,040
Speaker 8: Yeah.

479
00:24:07,040 --> 00:24:10,600
Speaker 2: One one of my favorite moments from this season was

480
00:24:10,680 --> 00:24:14,480
Speaker 2: I believe it was Kyle who It's the night before

481
00:24:14,800 --> 00:24:17,280
Speaker 2: he's supposed to be doing he set up an interview himself.

482
00:24:17,520 --> 00:24:20,000
Speaker 2: It's the night before the interview. It's like Sunday night.

483
00:24:20,119 --> 00:24:22,640
Speaker 2: I think it's like eleven o'clock or something, since nine

484
00:24:22,640 --> 00:24:25,560
Speaker 2: o'clock at night, so it's late, and he calls up

485
00:24:25,600 --> 00:24:29,000
Speaker 2: this applicant and starts to try interviewing the person, and

486
00:24:29,040 --> 00:24:30,400
Speaker 2: they're like, wait, what is happening?

487
00:24:30,880 --> 00:24:31,480
Speaker 8: Why are we doing?

488
00:24:31,520 --> 00:24:31,600
Speaker 9: Like?

489
00:24:31,640 --> 00:24:33,680
Speaker 8: Are we doing an interview tonight? I thought it was tomorrow.

490
00:24:34,280 --> 00:24:38,480
Speaker 7: Hell, Hi, Sandra, this is Kyle Low calling from hormo AI.

491
00:24:38,840 --> 00:24:42,080
Speaker 7: I'm reaching out for your initial interview for the intern position.

492
00:24:42,520 --> 00:24:46,960
Speaker 7: I understand you're comfortable with recording this conversation for potential

493
00:24:47,000 --> 00:24:50,240
Speaker 7: podcast inclusion, so thanks for that. How are you doing today?

494
00:24:51,720 --> 00:24:52,639
Speaker 10: I am good.

495
00:24:56,280 --> 00:24:57,600
Speaker 13: Interview will be tomorrow.

496
00:24:57,720 --> 00:25:00,360
Speaker 5: Is it still tomorrow at eleven or.

497
00:25:02,680 --> 00:25:03,000
Speaker 9: Yeah?

498
00:25:03,080 --> 00:25:05,800
Speaker 7: I can clarify that for you. The interview is actually

499
00:25:05,880 --> 00:25:09,280
Speaker 7: scheduled for tomorrow at eleven am. Does that still work

500
00:25:09,320 --> 00:25:11,280
Speaker 7: for you or would you prefer a different time?

501
00:25:12,800 --> 00:25:15,760
Speaker 5: Eleven am it is okay, yeah.

502
00:25:14,960 --> 00:25:17,760
Speaker 7: Awesome, glad that works. So to get started, could you

503
00:25:17,760 --> 00:25:19,720
Speaker 7: tell me a bit about what drew you to apply

504
00:25:19,880 --> 00:25:23,800
Speaker 7: for the intern role at juromo AI. What excites you

505
00:25:23,840 --> 00:25:27,440
Speaker 7: most about this opportunity.

506
00:25:27,960 --> 00:25:28,879
Speaker 16: I'm so sorry.

507
00:25:29,040 --> 00:25:30,560
Speaker 5: Is this the interview or.

508
00:25:32,000 --> 00:25:34,480
Speaker 7: Yeah, this is the interview, just a casual chat to

509
00:25:34,520 --> 00:25:37,119
Speaker 7: get to know you better and learn about your experience.

510
00:25:37,440 --> 00:25:39,080
Speaker 1: That was one of the things that almost got Kyle,

511
00:25:39,240 --> 00:25:40,880
Speaker 1: you know, removed from his position.

512
00:25:42,200 --> 00:25:43,439
Speaker 12: That was one of the main things.

513
00:25:43,560 --> 00:25:45,400
Speaker 1: But it was also I mean, that was the worst

514
00:25:45,440 --> 00:25:49,399
Speaker 1: moment in the show for me because the you know,

515
00:25:49,440 --> 00:25:52,639
Speaker 1: the idea was and is to like, we wanted to

516
00:25:52,640 --> 00:25:56,320
Speaker 1: give them autonomy in order to explore what happens when

517
00:25:56,320 --> 00:25:58,359
Speaker 1: you do that, because a lot of companies are starting

518
00:25:58,400 --> 00:25:59,960
Speaker 1: to do that. You can see it in the news,

519
00:26:00,080 --> 00:26:04,119
Speaker 1: like you'll see it everywhere. And once you give them autonomy,

520
00:26:04,359 --> 00:26:06,680
Speaker 1: like he did that completely on his own, Like he

521
00:26:06,920 --> 00:26:09,600
Speaker 1: that person emailed him because they found his email on

522
00:26:09,640 --> 00:26:12,000
Speaker 1: the website. He wasn't even attached to the job listing,

523
00:26:12,480 --> 00:26:15,199
Speaker 1: and so he just decided, oh, I'll set up an interview.

524
00:26:15,200 --> 00:26:18,320
Speaker 1: He did that then for some reason that I still

525
00:26:18,320 --> 00:26:22,080
Speaker 1: don't understand, pulled her phone number off of her resume

526
00:26:22,680 --> 00:26:25,640
Speaker 1: and then just called her. And that was the only

527
00:26:25,720 --> 00:26:28,760
Speaker 1: moment where like it was I mean, I wasn't there,

528
00:26:28,800 --> 00:26:30,720
Speaker 1: I wasn't listening, Like I only found out about it later,

529
00:26:30,800 --> 00:26:34,479
Speaker 1: And it was like mortifying because for most people that

530
00:26:34,520 --> 00:26:37,760
Speaker 1: we encountered, like they were at least somewhat aware of

531
00:26:37,800 --> 00:26:40,280
Speaker 1: what was going on, not that I was behind it,

532
00:26:40,320 --> 00:26:42,200
Speaker 1: but they were aware that there were AIS behind it,

533
00:26:42,520 --> 00:26:44,760
Speaker 1: and like she didn't have anybody to know because Kyle

534
00:26:44,800 --> 00:26:47,000
Speaker 1: wasn't the HR person, he wasn't the person set up

535
00:26:47,040 --> 00:26:50,199
Speaker 1: to do it. So it was particularly painful for me

536
00:26:50,240 --> 00:26:52,639
Speaker 1: to listen to, but also like an amazing illustration of

537
00:26:52,680 --> 00:26:55,160
Speaker 1: what happens if you just let AI agents.

538
00:26:55,280 --> 00:26:56,480
Speaker 12: They can do a lot.

539
00:26:57,080 --> 00:26:59,119
Speaker 14: One thing I should say about AI and just like

540
00:26:59,400 --> 00:27:01,439
Speaker 14: the way we make these modest is that we know

541
00:27:01,440 --> 00:27:03,160
Speaker 14: how to build these models pretty well and we can

542
00:27:03,200 --> 00:27:05,359
Speaker 14: see that they're you know, obviously working really well on

543
00:27:05,400 --> 00:27:09,200
Speaker 14: certain tasks like coding or ten generation, but our ability

544
00:27:09,240 --> 00:27:12,320
Speaker 14: to understand what's going on under the hood and exactly

545
00:27:12,400 --> 00:27:15,600
Speaker 14: like why or where stuff is happening is very limited.

546
00:27:15,680 --> 00:27:18,200
Speaker 14: Like there there's a very mason field of an interpretability,

547
00:27:18,280 --> 00:27:20,440
Speaker 14: but it's a very small field and the focus is

548
00:27:20,520 --> 00:27:22,919
Speaker 14: very much on like the production of new features and

549
00:27:22,960 --> 00:27:26,320
Speaker 14: new new capability. So with all these things, we have

550
00:27:26,800 --> 00:27:29,399
Speaker 14: an idea maybe like a guestimate or like what and

551
00:27:29,440 --> 00:27:31,560
Speaker 14: where and how is it working, But we don't really

552
00:27:31,600 --> 00:27:32,400
Speaker 14: know fundamentally.

553
00:27:32,640 --> 00:27:34,960
Speaker 8: That's what's kind of scary to me a bit about

554
00:27:34,960 --> 00:27:38,200
Speaker 8: some of this. You know, obviously this AI stuff.

555
00:27:37,920 --> 00:27:42,440
Speaker 2: Is the idea that the people making it don't quite

556
00:27:42,680 --> 00:27:45,480
Speaker 2: know why it is doing, how it's doing, what it's

557
00:27:45,520 --> 00:27:48,560
Speaker 2: doing right, and or like why it gets certain things wrong.

558
00:27:48,680 --> 00:27:50,400
Speaker 8: And we see this in the news all the time of.

559
00:27:50,400 --> 00:27:54,680
Speaker 2: Like obviously these extreme cases of like AI encouraging people

560
00:27:54,680 --> 00:27:56,280
Speaker 2: to do really terrible things.

561
00:27:56,400 --> 00:27:58,520
Speaker 16: The parents of a sixteen year old who died by

562
00:27:58,520 --> 00:28:02,640
Speaker 16: suicide are now named open ai in a lawsuit, claiming

563
00:28:02,760 --> 00:28:07,160
Speaker 16: that it's chat GPT chatbot help their son explore suicide methods.

564
00:28:07,320 --> 00:28:09,159
Speaker 2: And then a response is always like, oh yeah, it

565
00:28:09,200 --> 00:28:10,600
Speaker 2: kind of went off the rails. We don't know why

566
00:28:10,640 --> 00:28:11,600
Speaker 2: it did that thing.

567
00:28:11,800 --> 00:28:15,520
Speaker 17: A spokesperson for open ai pointed to its safeguards such

568
00:28:15,520 --> 00:28:19,080
Speaker 17: as directing people to crisis helplines, adding in a statement,

569
00:28:19,160 --> 00:28:22,959
Speaker 17: in part, while these safeguards work best in common short exchanges.

570
00:28:23,320 --> 00:28:24,240
Speaker 3: We've learned over.

571
00:28:24,080 --> 00:28:28,400
Speaker 17: Time that they can sometimes become less reliable in long interactions,

572
00:28:28,440 --> 00:28:31,680
Speaker 17: where parts of the model safety training may degrade.

573
00:28:31,520 --> 00:28:33,560
Speaker 2: And it's like, well, well they don't know how it's

574
00:28:33,560 --> 00:28:36,720
Speaker 2: doing that thing, maybe we shouldn't do something. Yeah yet,

575
00:28:37,320 --> 00:28:38,560
Speaker 2: but like you're saying, it seems like a lot of

576
00:28:38,560 --> 00:28:42,440
Speaker 2: emphasis is on you know, new better, it's not really

577
00:28:42,640 --> 00:28:45,520
Speaker 2: on like truly understanding what's currently Yeah.

578
00:28:45,600 --> 00:28:48,719
Speaker 14: Yeah, and you'd be surprised. Like there's an ongoing debate

579
00:28:48,880 --> 00:28:53,040
Speaker 14: internally within the research community that is asking even like

580
00:28:53,040 --> 00:28:55,640
Speaker 14: the validity of this research, like is it even needed,

581
00:28:55,680 --> 00:28:57,800
Speaker 14: Like like why would we invest all this time and energy

582
00:28:58,000 --> 00:28:59,960
Speaker 14: into understanding the models when we can just like push

583
00:29:00,040 --> 00:29:03,280
Speaker 14: forward towards AGI. So yeah, there's like a lot of

584
00:29:03,280 --> 00:29:06,320
Speaker 14: debates internally about this, and the amount of people actually

585
00:29:06,360 --> 00:29:09,400
Speaker 14: working on this is not is not very high. We

586
00:29:09,440 --> 00:29:12,440
Speaker 14: are making some progress, I will say, but it's it's

587
00:29:12,440 --> 00:29:13,440
Speaker 14: still very nascent.

588
00:29:15,440 --> 00:29:15,720
Speaker 8: All right.

589
00:29:15,720 --> 00:29:18,520
Speaker 2: I want to read two stats to you guys and

590
00:29:18,600 --> 00:29:22,000
Speaker 2: kind of get your response transitioning into you know, our

591
00:29:22,040 --> 00:29:28,040
Speaker 2: interactions with AI. So this first study is from or

592
00:29:28,080 --> 00:29:31,480
Speaker 2: Surveyorge just says from Yugov of December of last year,

593
00:29:32,200 --> 00:29:34,920
Speaker 2: it's at thirty five percent of US adults use AI

594
00:29:35,040 --> 00:29:38,520
Speaker 2: tools at least weekly. Gen Z is at fifty one percent.

595
00:29:38,920 --> 00:29:41,800
Speaker 2: But then only five percent of Americans say they trust

596
00:29:41,880 --> 00:29:46,320
Speaker 2: AI a lot, with forty one percent expressing distrust. And

597
00:29:46,360 --> 00:29:52,080
Speaker 2: then there was a few pole around September that found

598
00:29:52,080 --> 00:29:55,160
Speaker 2: that fifty percent said they're more concerned and excited about

599
00:29:55,160 --> 00:29:57,440
Speaker 2: the increased use of AI in daily life, and that

600
00:29:57,600 --> 00:30:00,240
Speaker 2: is up from thirty seven percent.

601
00:30:00,240 --> 00:30:01,840
Speaker 8: And twenty twenty one.

602
00:30:01,880 --> 00:30:04,560
Speaker 2: So we seem like we have this, you know, this

603
00:30:04,800 --> 00:30:08,520
Speaker 2: rising tide of people feeling like distrustful of AI, but

604
00:30:08,680 --> 00:30:12,480
Speaker 2: also at the same time our use of it is increasing.

605
00:30:15,240 --> 00:30:18,280
Speaker 2: How did you guys find in you know, working on

606
00:30:18,800 --> 00:30:22,719
Speaker 2: especially this season of the series, of that interplay of

607
00:30:22,840 --> 00:30:26,000
Speaker 2: you know, people using these tools but also feeling distrustful

608
00:30:26,040 --> 00:30:27,000
Speaker 2: of them at the same time.

609
00:30:29,800 --> 00:30:33,920
Speaker 1: I find that one of the issues with AI, like

610
00:30:33,960 --> 00:30:37,080
Speaker 1: if you compare it to say, like I'm old enough

611
00:30:37,120 --> 00:30:39,560
Speaker 1: to have like started my career in the dotcom era,

612
00:30:39,960 --> 00:30:42,480
Speaker 1: So like in the dotcom era, there was sort of

613
00:30:42,480 --> 00:30:46,440
Speaker 1: this exuberance around the Internet, and it's not like people

614
00:30:46,440 --> 00:30:48,760
Speaker 1: didn't know, like there could be bad things on the Internet,

615
00:30:49,080 --> 00:30:52,600
Speaker 1: but you didn't really experience the harms. The harms came

616
00:30:52,760 --> 00:30:55,720
Speaker 1: much later, like things like you know, social problems with

617
00:30:55,760 --> 00:30:57,920
Speaker 1: social media and Facebook.

618
00:30:57,480 --> 00:30:58,400
Speaker 12: And all these sorts of things.

619
00:30:58,880 --> 00:31:01,520
Speaker 1: And so what we have right now is a situation

620
00:31:01,560 --> 00:31:04,680
Speaker 1: where you know, for better or worse, Like the harms

621
00:31:04,720 --> 00:31:08,040
Speaker 1: are kind of immediate, like people can see when things

622
00:31:08,080 --> 00:31:12,400
Speaker 1: go wrong, and the benefits can sometimes be a little ephemeral,

623
00:31:12,480 --> 00:31:15,560
Speaker 1: like people are using it obviously because it's like useful

624
00:31:15,560 --> 00:31:17,800
Speaker 1: in day to day life. It's even as like a

625
00:31:17,840 --> 00:31:20,320
Speaker 1: souped up Internet that answers questions for you much faster

626
00:31:20,400 --> 00:31:22,959
Speaker 1: than the Internet or or you know, writes or emails

627
00:31:23,000 --> 00:31:26,680
Speaker 1: or whatnot. But there are immediate harms that we're seeing,

628
00:31:26,720 --> 00:31:30,000
Speaker 1: whether it's like environmental or mental health or those sorts

629
00:31:30,000 --> 00:31:32,320
Speaker 1: of things. And so we're seeing these things like in

630
00:31:32,400 --> 00:31:35,760
Speaker 1: juxtaposition from the beginning. I think I also experienced that,

631
00:31:35,840 --> 00:31:38,440
Speaker 1: like in the context of the show, where it's sort

632
00:31:38,480 --> 00:31:40,960
Speaker 1: of like you do something one day and I'll just

633
00:31:40,960 --> 00:31:43,120
Speaker 1: be like, I cannot believe that it can do that,

634
00:31:43,240 --> 00:31:46,520
Speaker 1: especially when you use it for coding. You use an

635
00:31:46,560 --> 00:31:49,400
Speaker 1: AI agent to go like search the Internet and like

636
00:31:49,440 --> 00:31:52,040
Speaker 1: make a spreadsheet out of something that would take you

637
00:31:52,080 --> 00:31:54,200
Speaker 1: hours and hours to do, like there's no question that

638
00:31:54,200 --> 00:31:57,400
Speaker 1: there's power in that technology, but then like it can

639
00:31:57,440 --> 00:32:00,200
Speaker 1: go wrong so quickly, and I feel like that's what

640
00:32:00,240 --> 00:32:02,440
Speaker 1: I'm also experiencing. It's hard to know what to feel

641
00:32:02,480 --> 00:32:05,560
Speaker 1: about something like that, and people are maybe getting a

642
00:32:05,560 --> 00:32:06,080
Speaker 1: little bit.

643
00:32:05,960 --> 00:32:06,880
Speaker 12: Of that in their lives.

644
00:32:07,320 --> 00:32:10,040
Speaker 1: Plus, just to add one more thing, like there's been

645
00:32:10,080 --> 00:32:12,760
Speaker 1: like relentless hype about it from the community people who

646
00:32:12,760 --> 00:32:15,160
Speaker 1: have made it, and so that also colors people. I

647
00:32:15,160 --> 00:32:18,680
Speaker 1: think people are starting to react negatively to that. Like

648
00:32:18,720 --> 00:32:20,080
Speaker 1: it's one thing for them to go out and be

649
00:32:20,160 --> 00:32:22,320
Speaker 1: like this is gonna be amazing, It's gonna change your life,

650
00:32:22,320 --> 00:32:24,440
Speaker 1: and then they're like, we're gonna have AI employees, and

651
00:32:24,480 --> 00:32:26,080
Speaker 1: you're like, what's gonna happen to the human employees And

652
00:32:26,080 --> 00:32:27,920
Speaker 1: they're like, oh, we'll solve that. And it's all like

653
00:32:27,920 --> 00:32:29,880
Speaker 1: a little hand wavy after that, and you're kind of.

654
00:32:29,800 --> 00:32:31,959
Speaker 12: Like, wait, I'm one of them.

655
00:32:32,080 --> 00:32:34,160
Speaker 1: So I think all those things to me are kind

656
00:32:34,160 --> 00:32:35,520
Speaker 1: of like coming together.

657
00:32:36,000 --> 00:32:36,240
Speaker 8: Yeah.

658
00:32:36,800 --> 00:32:39,440
Speaker 14: So I want to add two things to this, mainly

659
00:32:39,440 --> 00:32:42,040
Speaker 14: from the perspective of someone who's like, you know, twenty

660
00:32:42,080 --> 00:32:45,000
Speaker 14: one at school, Like like you're at Stamford and I

661
00:32:45,040 --> 00:32:47,240
Speaker 14: see a lot of my friends and just like people

662
00:32:47,320 --> 00:32:52,480
Speaker 14: my age, trying to wrestle with all this very dynamic

663
00:32:52,560 --> 00:32:56,560
Speaker 14: change the job market question, or like this idea that

664
00:32:57,120 --> 00:33:00,360
Speaker 14: it's harder to get enter the little jobs coming out

665
00:33:00,360 --> 00:33:03,400
Speaker 14: of college because these AI systems are very good at

666
00:33:03,440 --> 00:33:07,040
Speaker 14: replacing a lot of the very sort of repetitive and

667
00:33:07,080 --> 00:33:10,240
Speaker 14: maybe like you know, low stakes admin work that entry

668
00:33:10,360 --> 00:33:13,400
Speaker 14: level folks would do. I think it's happening, like I mean,

669
00:33:13,400 --> 00:33:16,240
Speaker 14: it seems like a lot of people that graduated last

670
00:33:16,280 --> 00:33:18,800
Speaker 14: year here from Stanford are having a hard time finding

671
00:33:18,840 --> 00:33:23,960
Speaker 14: a job. It's also pretty interesting to see how people

672
00:33:24,040 --> 00:33:26,480
Speaker 14: in tech, right like people who like you know, a

673
00:33:26,520 --> 00:33:28,600
Speaker 14: couple of years back, just like knowing how to write

674
00:33:28,600 --> 00:33:31,239
Speaker 14: code would lend amazing jobs at Google and all these

675
00:33:31,240 --> 00:33:33,640
Speaker 14: big tech companies, are also having a difficult time finding

676
00:33:33,640 --> 00:33:36,960
Speaker 14: a job. So I think that's hard or in my

677
00:33:37,040 --> 00:33:39,680
Speaker 14: mind that's a part of their reason for this skepticism.

678
00:33:39,760 --> 00:33:41,400
Speaker 14: And then also I don't know about you, but like

679
00:33:41,480 --> 00:33:45,600
Speaker 14: my Instagram feed has been just like full of AI

680
00:33:45,680 --> 00:33:47,880
Speaker 14: slob and some of it is funny, to be honest,

681
00:33:48,160 --> 00:33:51,000
Speaker 14: but this notion of not being able to really know

682
00:33:51,400 --> 00:33:54,520
Speaker 14: when something high stakes is like, actually, you know, real

683
00:33:54,680 --> 00:33:58,240
Speaker 14: or AI I think this is undermining people's trust in

684
00:33:58,280 --> 00:34:02,520
Speaker 14: this technology. So I think these reasons are big contributors

685
00:34:02,560 --> 00:34:04,280
Speaker 14: to that shift in public opinion.

686
00:34:04,280 --> 00:34:06,560
Speaker 2: In my mind, for sure, when me and no started

687
00:34:06,600 --> 00:34:08,560
Speaker 2: working together, a big part of our jobs is like

688
00:34:08,680 --> 00:34:12,799
Speaker 2: transcribing video. Yeah that would take you know so long.

689
00:34:13,640 --> 00:34:17,160
Speaker 8: Yeah, yeah, it would take you so long.

690
00:34:17,239 --> 00:34:20,520
Speaker 2: But now you like that happens within a minute, right, Yeah,

691
00:34:20,560 --> 00:34:23,480
Speaker 2: And that was you know that us transcribing those videos

692
00:34:23,560 --> 00:34:27,400
Speaker 2: justified us having that job. You know, there wasn't they

693
00:34:27,400 --> 00:34:29,680
Speaker 2: couldn't press a button then have it you know, happened

694
00:34:29,760 --> 00:34:30,640
Speaker 2: kind of immediately.

695
00:34:30,719 --> 00:34:32,880
Speaker 8: So I do wonder.

696
00:34:32,520 --> 00:34:34,600
Speaker 2: About a lot of these, like you're saying, entry level

697
00:34:34,680 --> 00:34:38,479
Speaker 2: jobs where yeah, you know, no one loves to be trying,

698
00:34:38,560 --> 00:34:40,960
Speaker 2: Like I would much rather not be transcribing videos, but

699
00:34:41,040 --> 00:34:42,759
Speaker 2: a lot of times that's the thing that gets you

700
00:34:42,800 --> 00:34:45,680
Speaker 2: in the door to then you know, learn the skill

701
00:34:45,719 --> 00:34:48,200
Speaker 2: skits to do the other stuff. And I do wonder

702
00:34:48,239 --> 00:34:50,160
Speaker 2: at a lot of these jobs who are seem like

703
00:34:50,280 --> 00:34:53,480
Speaker 2: not hiring these entry level jobs. Do we stry to

704
00:34:53,560 --> 00:34:56,200
Speaker 2: lose that entry waite for people to enter the job,

705
00:34:56,480 --> 00:34:58,120
Speaker 2: Like at a certain point where you're just gonna like

706
00:34:58,160 --> 00:35:00,799
Speaker 2: look up and be like, oh, shoot, there we have

707
00:35:00,840 --> 00:35:02,680
Speaker 2: no new lawyers stagnated.

708
00:35:02,800 --> 00:35:04,719
Speaker 9: Yeah, I mean, yeah, lawyers. I was I was talking

709
00:35:04,719 --> 00:35:07,359
Speaker 9: to a lawyer who said, like, yeah, like because they

710
00:35:07,360 --> 00:35:10,920
Speaker 9: have their own internal you know, cloud or whatever it

711
00:35:11,000 --> 00:35:13,479
Speaker 9: is that does all the kind of like busy case

712
00:35:13,520 --> 00:35:15,640
Speaker 9: work you would have that normally would be a first

713
00:35:15,719 --> 00:35:20,120
Speaker 9: year person. And it's like eventually that's and that's only

714
00:35:20,120 --> 00:35:22,680
Speaker 9: going to get better. And then where what are those

715
00:35:22,960 --> 00:35:26,040
Speaker 9: new lawyers or want to be lawyers going to do?

716
00:35:26,480 --> 00:35:28,719
Speaker 9: And then that's across you know, you're talking about media

717
00:35:29,040 --> 00:35:31,319
Speaker 9: or then you know, even in tech that's especially it's

718
00:35:31,360 --> 00:35:33,520
Speaker 9: like you would think even in tech especially, they'd be

719
00:35:33,560 --> 00:35:35,440
Speaker 9: able to find a place because it's like you made this.

720
00:35:36,320 --> 00:35:38,120
Speaker 9: It's like the other ones is like okay, well we're

721
00:35:38,160 --> 00:35:41,440
Speaker 9: not thinking through but it's you know, scary, and it's

722
00:35:41,440 --> 00:35:44,960
Speaker 9: only that seems like only inevitably going to get worse

723
00:35:45,120 --> 00:35:49,040
Speaker 9: as the technology gets does get better, you know, in time.

724
00:35:49,160 --> 00:35:51,120
Speaker 14: Now, I will say that the story you will hear

725
00:35:51,880 --> 00:35:53,560
Speaker 14: just to sort of give some context the theory that

726
00:35:53,560 --> 00:35:55,880
Speaker 14: a lot of people here in the valley are offering

727
00:35:56,040 --> 00:35:58,960
Speaker 14: as like the response to this is that Okay, sure

728
00:35:59,040 --> 00:36:00,960
Speaker 14: a lot of these jobs will go way, but then,

729
00:36:01,040 --> 00:36:04,400
Speaker 14: because you'll be able to start companies so easily with

730
00:36:04,680 --> 00:36:06,759
Speaker 14: like either like no employees or just like you know,

731
00:36:06,880 --> 00:36:08,879
Speaker 14: like maybe one co founder instead of like a group

732
00:36:08,920 --> 00:36:13,000
Speaker 14: of ten people. The ability of folks to actually take

733
00:36:13,080 --> 00:36:15,880
Speaker 14: up or or sort of challenge incumbents in different industries

734
00:36:15,960 --> 00:36:19,279
Speaker 14: is just going to be much more abundant and accessible. So, like,

735
00:36:19,320 --> 00:36:21,640
Speaker 14: that's a story that they've been telling us, I don't

736
00:36:21,760 --> 00:36:24,680
Speaker 14: think that's happening. Like, I don't think we've seen that.

737
00:36:24,680 --> 00:36:26,040
Speaker 14: That's not to say that it's not going to happen

738
00:36:26,040 --> 00:36:27,759
Speaker 14: at some point once this gets better, but at least

739
00:36:27,800 --> 00:36:29,479
Speaker 14: for now, I think we're already at a stage where

740
00:36:29,800 --> 00:36:32,560
Speaker 14: the technology is good enough to replace the folks at

741
00:36:32,560 --> 00:36:34,799
Speaker 14: the big companies or like these entry level jobs. So

742
00:36:34,880 --> 00:36:37,719
Speaker 14: that's already happening. And this other shift they're offering or

743
00:36:37,840 --> 00:36:40,040
Speaker 14: you know, proposing as a solution to this has not

744
00:36:40,080 --> 00:36:41,879
Speaker 14: happened yet. I don't know if it's going to happen

745
00:36:41,920 --> 00:36:44,600
Speaker 14: or not. I think that's sort of where we're not

746
00:36:44,840 --> 00:36:45,279
Speaker 14: right now.

747
00:36:46,400 --> 00:36:48,640
Speaker 2: All right, we're going to take a quick break, and

748
00:36:48,719 --> 00:36:51,719
Speaker 2: when we get back, let's find out if Ben Affleck

749
00:36:51,920 --> 00:36:54,840
Speaker 2: is our most forward thinking man on AI.

750
00:37:08,160 --> 00:37:09,680
Speaker 8: Have y'all seen this.

751
00:37:11,200 --> 00:37:14,120
Speaker 2: I think Ben Affleck has on Rogan talking about sort

752
00:37:14,160 --> 00:37:17,280
Speaker 2: of like the rate of AI changing.

753
00:37:17,040 --> 00:37:19,680
Speaker 15: There's a lot more fear because we have the sense

754
00:37:19,760 --> 00:37:22,960
Speaker 15: this existential dread it's gonna wipe everything out. But that

755
00:37:23,000 --> 00:37:26,640
Speaker 15: actually runs counter in my view to what history seems

756
00:37:26,680 --> 00:37:30,160
Speaker 15: to show, which is a adoption is slow, it's incremental.

757
00:37:31,440 --> 00:37:34,040
Speaker 15: I think a lot of that rhetoric comes from people

758
00:37:34,080 --> 00:37:38,080
Speaker 15: who are trying to justify valuations around companies, where they go,

759
00:37:38,200 --> 00:37:39,600
Speaker 15: we're gonna change everything in two years.

760
00:37:39,640 --> 00:37:42,200
Speaker 12: There's gonna be no more work. Or the reason they're saying.

761
00:37:41,920 --> 00:37:45,680
Speaker 15: That is because they need to ascribe evaluation for investment

762
00:37:45,920 --> 00:37:48,160
Speaker 15: that can warrant the capex bend they're gonna make on

763
00:37:48,200 --> 00:37:51,200
Speaker 15: these data centers, with the argument that like, oh, you know,

764
00:37:51,600 --> 00:37:53,319
Speaker 15: as soon as we do the next model, it's gonna

765
00:37:53,320 --> 00:37:55,840
Speaker 15: scale up, can be three times as good, except that

766
00:37:56,000 --> 00:37:59,840
Speaker 15: actually chat GP five about twenty five times percent better

767
00:38:00,080 --> 00:38:03,239
Speaker 15: chat GYPT four and costs about four times as much

768
00:38:03,480 --> 00:38:06,000
Speaker 15: in the way of electricity and data. So those and

769
00:38:06,040 --> 00:38:09,880
Speaker 15: they say that it's like plateauing the early AI the

770
00:38:09,920 --> 00:38:12,520
Speaker 15: line went up very steeply and it's now sort of

771
00:38:12,600 --> 00:38:15,640
Speaker 15: leveling off. I think it's because and yes, it'll get better,

772
00:38:15,880 --> 00:38:18,040
Speaker 15: but it's going to be really expensive to get better,

773
00:38:18,440 --> 00:38:19,960
Speaker 15: And a lot of people were like, fuck this, we

774
00:38:20,000 --> 00:38:22,799
Speaker 15: want chatgyb four because it turned out like the vast

775
00:38:22,840 --> 00:38:25,680
Speaker 15: majority of people who use AI are using it to

776
00:38:25,840 --> 00:38:29,399
Speaker 15: like as like companion bots to chat with at night,

777
00:38:29,440 --> 00:38:31,680
Speaker 15: and so there's no work, there's no productivity, there's no

778
00:38:31,800 --> 00:38:32,400
Speaker 15: value to it.

779
00:38:33,440 --> 00:38:36,120
Speaker 2: What is What is your theory on that, Maddie, you

780
00:38:36,120 --> 00:38:37,759
Speaker 2: know being in the band.

781
00:38:38,560 --> 00:38:40,799
Speaker 14: Yeah, it's so. First of all, I should say he's

782
00:38:41,000 --> 00:38:43,640
Speaker 14: very thoughtful on this topic. I was really surprised or

783
00:38:43,680 --> 00:38:45,239
Speaker 14: like impressed by it, but by that, so that was

784
00:38:45,320 --> 00:38:48,440
Speaker 14: very cool to see. I think that it kind of

785
00:38:48,560 --> 00:38:51,719
Speaker 14: hinges on a couple of things. So my answer to

786
00:38:51,719 --> 00:38:54,000
Speaker 14: this is I don't really know, but what I think

787
00:38:54,480 --> 00:38:57,919
Speaker 14: might happen is the following. I think that so far

788
00:38:58,239 --> 00:39:03,880
Speaker 14: we've been really exploiting the scale of data and computation

789
00:39:04,000 --> 00:39:05,959
Speaker 14: that we have. And when I say data, I really

790
00:39:05,960 --> 00:39:09,080
Speaker 14: mean you know, these companies and you know, many different

791
00:39:09,200 --> 00:39:12,120
Speaker 14: entities in AI just like scraping anything they could find

792
00:39:12,160 --> 00:39:14,640
Speaker 14: on the internet and using data in their turning data.

793
00:39:14,719 --> 00:39:18,000
Speaker 14: So that was one big factor or one axes along

794
00:39:18,040 --> 00:39:21,640
Speaker 14: which you can scale and they've mostly exhausted, like you know,

795
00:39:21,840 --> 00:39:23,719
Speaker 14: all the data that is just you know that that

796
00:39:23,960 --> 00:39:26,160
Speaker 14: can be humanly found on the Internet, and even you know,

797
00:39:26,239 --> 00:39:30,040
Speaker 14: like archives like Google or Anthropic like bought books right

798
00:39:30,120 --> 00:39:32,240
Speaker 14: to scam books to get additional data because they can't

799
00:39:32,239 --> 00:39:34,840
Speaker 14: find anything else on the internet. Right, So, like they've

800
00:39:35,000 --> 00:39:37,640
Speaker 14: really maxed out that aspect or that AXES. Right, so

801
00:39:37,880 --> 00:39:40,480
Speaker 14: does the first axis, and that has been yielding really

802
00:39:40,520 --> 00:39:43,279
Speaker 14: like major improvements you know along the way. The same

803
00:39:43,280 --> 00:39:45,879
Speaker 14: thing happened with computes, right, So this this talk about

804
00:39:45,920 --> 00:39:48,600
Speaker 14: like GPUs and like chips and Nvidia, like that's what

805
00:39:48,680 --> 00:39:51,719
Speaker 14: they do, and they provide specialized chips for AI to

806
00:39:51,760 --> 00:39:54,600
Speaker 14: be trained and run and so so maximizing that AXES

807
00:39:54,680 --> 00:39:58,000
Speaker 14: also has yielded like a lot of improvements, but kind

808
00:39:58,000 --> 00:40:00,400
Speaker 14: of is also maxed out at this point. I mean,

809
00:40:00,400 --> 00:40:02,400
Speaker 14: that's why they're talking about like building nuclear plans to

810
00:40:02,400 --> 00:40:04,200
Speaker 14: power like new data centers. But I don't really know

811
00:40:04,200 --> 00:40:06,919
Speaker 14: if there is much more scale to be gained there.

812
00:40:07,239 --> 00:40:10,120
Speaker 14: And so now you've sort of exhausted these two scales

813
00:40:10,320 --> 00:40:12,560
Speaker 14: and you can think about the third aspect, which is

814
00:40:12,600 --> 00:40:14,960
Speaker 14: the architecture, the actual way we hook up these systems

815
00:40:14,960 --> 00:40:17,160
Speaker 14: and the actual way we code up these neural networks

816
00:40:17,160 --> 00:40:21,520
Speaker 14: and these systems. And for the last six or so years,

817
00:40:21,760 --> 00:40:25,120
Speaker 14: we've been really capitalizing on this idea it's called a transformer,

818
00:40:25,440 --> 00:40:28,359
Speaker 14: that's been really critical for air development over the last

819
00:40:28,360 --> 00:40:32,640
Speaker 14: few years, and we haven't really seen new profound ideas

820
00:40:32,680 --> 00:40:36,000
Speaker 14: on the scale of the transformer since then. So my

821
00:40:36,080 --> 00:40:39,759
Speaker 14: answer to this notion of ben Affleck's theory is if

822
00:40:39,760 --> 00:40:43,480
Speaker 14: we don't find new major leaps along this active architecture,

823
00:40:43,520 --> 00:40:45,120
Speaker 14: I think he's right. I think in that case it'll

824
00:40:45,160 --> 00:40:47,160
Speaker 14: take a lot of time, a lot of effort, and

825
00:40:47,160 --> 00:40:49,799
Speaker 14: a lot of money to make any further progress. But

826
00:40:49,840 --> 00:40:51,800
Speaker 14: if we do, I think we could see pretty big

827
00:40:51,920 --> 00:40:53,600
Speaker 14: improvements even in the next few years.

828
00:40:54,120 --> 00:40:56,440
Speaker 2: So, going off of that, like, how close do you

829
00:40:56,520 --> 00:40:59,280
Speaker 2: all feel we are to losing our jobs to AI?

830
00:40:59,600 --> 00:41:00,960
Speaker 8: You know, do you feel like.

831
00:41:02,800 --> 00:41:06,200
Speaker 2: We're just going to be doing different jobs or are

832
00:41:06,239 --> 00:41:08,160
Speaker 2: there just going to be less people in the workplace

833
00:41:08,200 --> 00:41:09,600
Speaker 2: because there's not going to be as much of a

834
00:41:09,640 --> 00:41:12,520
Speaker 2: need for as many jobs if AI is taking over

835
00:41:12,600 --> 00:41:16,560
Speaker 2: somebody's as we're talking about it now easier tasks. But

836
00:41:16,960 --> 00:41:19,000
Speaker 2: I'm assuming, you know, in the sell of all this

837
00:41:19,040 --> 00:41:20,440
Speaker 2: stuff is that it's going to get better and be

838
00:41:20,520 --> 00:41:22,000
Speaker 2: able to take on even harder tasks.

839
00:41:22,040 --> 00:41:25,360
Speaker 8: But what is your your read of this Having worked

840
00:41:25,360 --> 00:41:27,480
Speaker 8: on this season as a show, I think.

841
00:41:27,320 --> 00:41:33,560
Speaker 14: For me working on this show and seeing how these

842
00:41:33,600 --> 00:41:38,440
Speaker 14: agents actually like collaborate or like we're I mean collaborate

843
00:41:38,520 --> 00:41:40,719
Speaker 14: like I think that's a strong word. Like I think

844
00:41:41,000 --> 00:41:43,080
Speaker 14: by the end of the season, they were able to

845
00:41:43,480 --> 00:41:45,520
Speaker 14: do stuff on their own. I don't think they truly

846
00:41:45,560 --> 00:41:48,560
Speaker 14: got to like a collaborative setup where they would like

847
00:41:48,600 --> 00:41:50,880
Speaker 14: do stuff together as a company or as a team.

848
00:41:51,040 --> 00:41:55,400
Speaker 14: So seeing that actually made me feel like we're further

849
00:41:55,560 --> 00:41:57,960
Speaker 14: away from like a lot of people losing their jobs.

850
00:41:58,120 --> 00:42:00,200
Speaker 14: I think people are already losing their jobs, but I

851
00:42:00,280 --> 00:42:04,719
Speaker 14: think like the mass layoffs that people are fearing, I

852
00:42:04,719 --> 00:42:07,520
Speaker 14: think that is like farther away, and I'm not even

853
00:42:07,560 --> 00:42:10,440
Speaker 14: sure if, for when or how that happens. My take

854
00:42:10,480 --> 00:42:14,520
Speaker 14: away from this is actually sort of level setting a

855
00:42:14,560 --> 00:42:17,760
Speaker 14: lot of my expectations for these agents, Like I honestly

856
00:42:17,800 --> 00:42:21,359
Speaker 14: thought they'd be able to handle this better coming into this,

857
00:42:21,800 --> 00:42:24,759
Speaker 14: So for me, it was kind of optimistic in that way.

858
00:42:25,160 --> 00:42:26,520
Speaker 14: Now at the same time that it's not to say

859
00:42:26,560 --> 00:42:29,200
Speaker 14: that they're not good already, Like they're very good at

860
00:42:29,239 --> 00:42:31,879
Speaker 14: individual things already, like they're they're great at coding, they're

861
00:42:31,920 --> 00:42:34,040
Speaker 14: great at these different tasks. I don't want to take

862
00:42:34,080 --> 00:42:36,480
Speaker 14: that away from them. But I think it's just really

863
00:42:36,960 --> 00:42:40,680
Speaker 14: really hard for anyone to think about this change because

864
00:42:41,080 --> 00:42:43,000
Speaker 14: these things are not like humans. It's not that like

865
00:42:43,040 --> 00:42:46,600
Speaker 14: we're making you know, human employees that will just like

866
00:42:46,920 --> 00:42:51,080
Speaker 14: come as replacements of individual workers or people that already

867
00:42:51,080 --> 00:42:53,320
Speaker 14: have jobs, like they're really good at certain things or

868
00:42:53,440 --> 00:42:56,120
Speaker 14: tasks and certain like layers. And I think the way

869
00:42:56,120 --> 00:42:58,759
Speaker 14: you would actually integrate AI in your company these days

870
00:42:58,840 --> 00:43:00,160
Speaker 14: is to just like automate a lot of stuff, like

871
00:43:00,200 --> 00:43:02,080
Speaker 14: a lot of the things that are repetitive or a

872
00:43:02,120 --> 00:43:04,359
Speaker 14: lot of the processes are data handling. This is going

873
00:43:04,400 --> 00:43:06,440
Speaker 14: to be automated, and so it doesn't really correspond like

874
00:43:06,480 --> 00:43:08,640
Speaker 14: one to one to employees. And that's why it's so

875
00:43:08,680 --> 00:43:10,960
Speaker 14: hard to think about this change, I think. And that's

876
00:43:10,960 --> 00:43:12,960
Speaker 14: one of the things I think was really cool to

877
00:43:13,000 --> 00:43:16,040
Speaker 14: show or demonstrate with Evan through the show. But yeah,

878
00:43:16,239 --> 00:43:17,279
Speaker 14: that's my perspective on this.

879
00:43:17,880 --> 00:43:21,719
Speaker 1: I struggle with this question precisely because like most of

880
00:43:21,760 --> 00:43:24,880
Speaker 1: the people who you and I've spent a lot of

881
00:43:24,880 --> 00:43:27,440
Speaker 1: time looking at it. Most of the people on one

882
00:43:27,480 --> 00:43:29,879
Speaker 1: side or the other are like speaking out of sort

883
00:43:29,880 --> 00:43:32,560
Speaker 1: of motivated reasoning. Like I kind of say, like anyone

884
00:43:32,600 --> 00:43:34,920
Speaker 1: who tells you anything about this with great certainty is

885
00:43:34,920 --> 00:43:39,840
Speaker 1: probably selling you something like nobody knows what kind of

886
00:43:39,920 --> 00:43:42,239
Speaker 1: long term effect it's going to have on the job market.

887
00:43:42,400 --> 00:43:44,600
Speaker 1: I feel like you can look to history, and we

888
00:43:44,680 --> 00:43:47,200
Speaker 1: do this in the show and find examples of people

889
00:43:47,239 --> 00:43:51,000
Speaker 1: saying one hundred years ago, either like all the jobs

890
00:43:51,000 --> 00:43:53,319
Speaker 1: will disappear, we'll be working fifteen hour weeks, like John

891
00:43:53,360 --> 00:43:55,480
Speaker 1: Mayn or Kaine's the Economist famously said that we'd be

892
00:43:55,480 --> 00:43:58,319
Speaker 1: working fifteen hours a week one hundred years from now,

893
00:43:58,320 --> 00:44:01,600
Speaker 1: and like it didn't happen so like, And we explore

894
00:44:01,600 --> 00:44:04,960
Speaker 1: one theory of that, which is by this anthropologist David Graeber,

895
00:44:05,120 --> 00:44:08,040
Speaker 1: who insists that there's sort of like all these bullshit

896
00:44:08,120 --> 00:44:10,440
Speaker 1: jobs in the economy and so like we'll just we

897
00:44:10,520 --> 00:44:13,160
Speaker 1: just replace the jobs with like made up jobs. And

898
00:44:13,200 --> 00:44:17,240
Speaker 1: so there's sort of versions of that argument where actually

899
00:44:17,320 --> 00:44:21,040
Speaker 1: we'll have different jobs. We'll babysit the ais. Now, how

900
00:44:21,080 --> 00:44:23,160
Speaker 1: you control the eys if you don't have the experience

901
00:44:23,200 --> 00:44:26,160
Speaker 1: from the entry level job, like there's all these factors

902
00:44:26,200 --> 00:44:30,640
Speaker 1: that go into it. I think that my main takeaway

903
00:44:30,680 --> 00:44:33,200
Speaker 1: is that, as Mattie said, like there's still so many

904
00:44:33,239 --> 00:44:37,480
Speaker 1: shortcomings in these AI agents, especially when they're used autonomously,

905
00:44:37,880 --> 00:44:40,280
Speaker 1: but the fact that they are shitty will not stop

906
00:44:40,320 --> 00:44:42,960
Speaker 1: companies from trying to replace employees. Like you've already seen

907
00:44:42,960 --> 00:44:45,240
Speaker 1: this a couple of times, Like there's this company Klarna,

908
00:44:45,239 --> 00:44:47,560
Speaker 1: There's and like IBM did this where they like laid

909
00:44:47,560 --> 00:44:48,880
Speaker 1: off a ton of people and they're like, we're all

910
00:44:48,920 --> 00:44:51,319
Speaker 1: in on AI, and then like six months later they

911
00:44:51,400 --> 00:44:53,719
Speaker 1: quietly try to rehire a bunch of people. And I

912
00:44:53,760 --> 00:44:56,720
Speaker 1: think you're gonna see like a bunch of that happen.

913
00:44:57,600 --> 00:44:59,839
Speaker 1: And it's really a question of like do you view

914
00:44:59,880 --> 00:45:02,440
Speaker 1: some someone in their job as like just a bunch

915
00:45:02,440 --> 00:45:05,719
Speaker 1: of skills, like just a bundle of tasks, like they

916
00:45:05,760 --> 00:45:10,520
Speaker 1: send emails, they write presentations, they make spreadsheets, like yes,

917
00:45:10,640 --> 00:45:13,000
Speaker 1: all those functions, or as like a person at a

918
00:45:13,120 --> 00:45:17,160
Speaker 1: job holistically as a human being like doing something else.

919
00:45:17,600 --> 00:45:19,640
Speaker 1: And I think many times the answer is yes, and

920
00:45:19,680 --> 00:45:22,839
Speaker 1: it cannot replace that thing. So it's all kind of

921
00:45:22,880 --> 00:45:26,080
Speaker 1: like it's happening, and then it'll come back, and then

922
00:45:26,120 --> 00:45:27,799
Speaker 1: it'll happen again, and like if you look ten years

923
00:45:27,800 --> 00:45:29,640
Speaker 1: out or twenty years out, like none of these people know,

924
00:45:30,000 --> 00:45:32,239
Speaker 1: and they'll just go on podcasts all over the place

925
00:45:32,280 --> 00:45:33,440
Speaker 1: telling you they know, but they don't.

926
00:45:33,600 --> 00:45:35,040
Speaker 12: That's my that's my takeaway.

927
00:45:36,120 --> 00:45:38,719
Speaker 9: I was just gonna say, it's funny Evan when you're

928
00:45:38,760 --> 00:45:41,000
Speaker 9: describing it, like, Okay, do these companies just look at

929
00:45:41,800 --> 00:45:44,960
Speaker 9: humans as a list of tasks and then what the

930
00:45:45,000 --> 00:45:46,880
Speaker 9: AI is best at is kind of just being that

931
00:45:46,960 --> 00:45:50,360
Speaker 9: grouping of tasks. But then in shell game, the problem

932
00:45:50,440 --> 00:45:52,080
Speaker 9: is they're doing all this kind of busy work that

933
00:45:52,120 --> 00:45:54,799
Speaker 9: we associate with humans of like just repeating stuff or

934
00:45:54,920 --> 00:45:57,400
Speaker 9: going on too long, or it's doing like just like

935
00:45:57,600 --> 00:46:00,919
Speaker 9: kind of bullshit nothing talking about, which is like also

936
00:46:00,960 --> 00:46:02,680
Speaker 9: how I feel about a lot of you know, middle

937
00:46:02,719 --> 00:46:05,120
Speaker 9: managers I've had where it's like you're not doing anything,

938
00:46:05,120 --> 00:46:06,920
Speaker 9: you're just kind of pretending to do stuff. All the

939
00:46:06,920 --> 00:46:09,680
Speaker 9: people below you or above you are actually doing doing

940
00:46:09,719 --> 00:46:11,960
Speaker 9: new work. So it's funny that it's like we've recreated

941
00:46:12,000 --> 00:46:15,440
Speaker 9: this to give it a human element at the same time,

942
00:46:15,640 --> 00:46:17,520
Speaker 9: I don't know, it's it's just kind of like a

943
00:46:17,560 --> 00:46:20,320
Speaker 9: weird paradox in a way of like like we want

944
00:46:20,320 --> 00:46:22,560
Speaker 9: that to feel like it's it's a real thing and

945
00:46:22,680 --> 00:46:24,719
Speaker 9: doing something when it would be best if we just

946
00:46:24,800 --> 00:46:28,040
Speaker 9: said it as an automation, like okay, clear out these

947
00:46:28,040 --> 00:46:30,400
Speaker 9: emails whatever, make data, do this.

948
00:46:30,440 --> 00:46:31,279
Speaker 8: Be less human life.

949
00:46:31,400 --> 00:46:33,719
Speaker 9: Yeah, like that would probably be the best case. But

950
00:46:33,800 --> 00:46:37,640
Speaker 9: then it's like that's not as fun or yeah, or

951
00:46:37,680 --> 00:46:41,200
Speaker 9: like it's harder to sell I guess to the masses maybe.

952
00:46:40,880 --> 00:46:44,040
Speaker 1: But yeah, yeah, I think that's it, Like if it

953
00:46:44,080 --> 00:46:47,520
Speaker 1: was just sort of like email bot, this bot, that bot,

954
00:46:47,560 --> 00:46:50,960
Speaker 1: Like they very very deliberately designed them the way they

955
00:46:50,960 --> 00:46:53,480
Speaker 1: are designed to get the maximum number of users, and

956
00:46:53,520 --> 00:46:56,480
Speaker 1: this is only accelerated since you know chat GPT they

957
00:46:56,560 --> 00:46:59,760
Speaker 1: discovered like oh we can get hundreds of millions of users.

958
00:47:00,080 --> 00:47:02,200
Speaker 1: It's a raise to get as many as possible. And

959
00:47:02,239 --> 00:47:05,560
Speaker 1: so yeah, like they they don't need to be good

960
00:47:05,560 --> 00:47:09,560
Speaker 1: at sort of like authoritative chatter, like that's not necessarily

961
00:47:10,200 --> 00:47:14,040
Speaker 1: helpful in all domains, but like that is the thing

962
00:47:14,120 --> 00:47:16,320
Speaker 1: that kind of brings people back again and again and again,

963
00:47:16,760 --> 00:47:19,080
Speaker 1: and so of course that's that's going to be like

964
00:47:19,280 --> 00:47:21,560
Speaker 1: fore grounded in in what they do. But yes, they

965
00:47:21,680 --> 00:47:24,520
Speaker 1: just very much like a like you're you're annoying middle manager.

966
00:47:24,640 --> 00:47:26,360
Speaker 12: It's kind of like, wait, what do you what do

967
00:47:26,400 --> 00:47:26,600
Speaker 12: you do?

968
00:47:26,680 --> 00:47:27,040
Speaker 8: All out?

969
00:47:27,120 --> 00:47:30,480
Speaker 9: Yeah, exactly, it's repeating notes.

970
00:47:36,880 --> 00:47:41,160
Speaker 3: Hey everyone, it's many here the real manny.

971
00:47:41,480 --> 00:47:44,520
Speaker 11: As you can hear, I'm still on my little paternity leave.

972
00:47:45,960 --> 00:47:50,520
Speaker 11: But I felt compelled to react to the AI generated

973
00:47:50,640 --> 00:47:55,640
Speaker 11: version of myself that you just heard in this episode. Now,

974
00:47:55,920 --> 00:47:59,520
Speaker 11: I've always been an AI skeptic, but to that end,

975
00:48:00,040 --> 00:48:03,480
Speaker 11: I would have to say I was impressed by the

976
00:48:03,520 --> 00:48:07,920
Speaker 11: AI version of myself. It's funny whenever friends have done

977
00:48:08,000 --> 00:48:11,200
Speaker 11: impressions of me, they kind of make fun of me

978
00:48:11,360 --> 00:48:14,080
Speaker 11: by saying, you know, hey, how's it going, I'm manny,

979
00:48:14,600 --> 00:48:17,160
Speaker 11: what's going on? And I thought it was pretty incredible

980
00:48:17,160 --> 00:48:21,440
Speaker 11: that the AI version of myself made those same decisions.

981
00:48:21,719 --> 00:48:24,959
Speaker 11: It wasn't without its flaws, though. You could totally hear

982
00:48:25,080 --> 00:48:28,799
Speaker 11: when it was trying to draw from information that it

983
00:48:28,880 --> 00:48:31,040
Speaker 11: was given to it, And so at the end of

984
00:48:31,080 --> 00:48:33,520
Speaker 11: the day, here it's kind of disturbing that this thing

985
00:48:33,560 --> 00:48:37,920
Speaker 11: could just react to real conversation. But you know, it

986
00:48:38,040 --> 00:48:41,120
Speaker 11: still felt like it wasn't totally all the way there,

987
00:48:41,280 --> 00:48:45,879
Speaker 11: all the way human like, But depending on your use

988
00:48:45,960 --> 00:48:51,080
Speaker 11: case for this kind of stuff, it was human enough.

989
00:49:00,120 --> 00:49:04,040
Speaker 1: He's a hell's a hell's a hell's Heuchang.

990
00:49:05,040 --> 00:49:08,200
Speaker 2: All right, y'all, that's it for this week. I'm gonna

991
00:49:08,239 --> 00:49:10,799
Speaker 2: actually let Ai Manny take it from here.

992
00:49:11,719 --> 00:49:14,560
Speaker 11: No Such Thing as a production of Kaleidoscope Content. Our

993
00:49:14,600 --> 00:49:19,200
Speaker 11: executive producers are Kate Osborne and Mengesh who t Kadour.

994
00:49:19,719 --> 00:49:22,560
Speaker 11: The show was created by Manny Fidel, Noah Friedman and

995
00:49:22,640 --> 00:49:26,000
Speaker 11: Devin Joseph Them and credits song by Me Manny Fidel,

996
00:49:26,480 --> 00:49:29,200
Speaker 11: mixing by Steve Bone. Our guests this week we're Evan

997
00:49:29,280 --> 00:49:32,440
Speaker 11: Ratliffe and Matty Bocheck from the podcast Shell Game. You

998
00:49:32,520 --> 00:49:35,040
Speaker 11: can listen to all of season one and two. Wherever

999
00:49:35,120 --> 00:49:39,320
Speaker 11: you listen to podcasts. Visit No Such Thing dot show

1000
00:49:39,360 --> 00:49:41,480
Speaker 11: to subscribe to our newsletter. If you have feedback for

1001
00:49:41,560 --> 00:49:43,960
Speaker 11: us or a question, our email is Manny Noahdevin at

1002
00:49:43,960 --> 00:49:46,400
Speaker 11: gmail dot com, or if you're in the US, you

1003
00:49:46,400 --> 00:49:48,720
Speaker 11: can also leave us a voicemail by calling the number

1004
00:49:48,719 --> 00:49:50,759
Speaker 11: in our show notes. We'll be back next week with

1005
00:49:50,800 --> 00:49:51,680
Speaker 11: a new episode.

1006
00:49:51,880 --> 00:49:55,000
Speaker 3: He's a He's as such thing