1
00:00:11,520 --> 00:00:14,480
Speaker 1: I'm Noah, I'm Devin, and this is no such thing.

2
00:00:14,600 --> 00:00:16,959
Speaker 1: The show we set on our dumb arguments and yours

3
00:00:17,200 --> 00:00:21,160
Speaker 1: by actually doing the research. Today's episode, well, a I

4
00:00:21,400 --> 00:00:22,400
Speaker 1: take our jobs.

5
00:00:25,520 --> 00:00:31,440
Speaker 2: There's no no such thing, no such thing, no such thing.

6
00:00:33,600 --> 00:00:37,480
Speaker 3: Touch thank no touch thank.

7
00:00:47,880 --> 00:00:51,280
Speaker 1: So it seems like every week I'm reading headlines or

8
00:00:51,280 --> 00:00:54,200
Speaker 1: seeing interviews about how AI is going to take every

9
00:00:54,200 --> 00:00:55,280
Speaker 1: single one of our jobs.

10
00:00:55,400 --> 00:00:58,480
Speaker 4: Thousands of new layoffs in the tech industry, and the

11
00:00:58,520 --> 00:01:01,240
Speaker 4: companies say it's largely because of AI.

12
00:01:01,920 --> 00:01:05,120
Speaker 5: Even with AI at its current state, I'd say you're

13
00:01:05,319 --> 00:01:09,040
Speaker 5: pretty close to being able to replace half full jobs.

14
00:01:09,720 --> 00:01:13,039
Speaker 1: So on today's episode, we're gonna take a peek under

15
00:01:13,040 --> 00:01:15,760
Speaker 1: the hood just to see how close AI is to

16
00:01:16,000 --> 00:01:19,920
Speaker 1: actually replacing us. And we're gonna chat with two guys

17
00:01:20,000 --> 00:01:24,800
Speaker 1: who started a real company run entirely by AI agents.

18
00:01:27,640 --> 00:01:31,280
Speaker 1: But before we do that, we're gonna call up our

19
00:01:31,880 --> 00:01:35,240
Speaker 1: good pal, Manny. So if you missed it, Many's been

20
00:01:35,280 --> 00:01:37,959
Speaker 1: out for a few weeks on paternity to leave after

21
00:01:38,000 --> 00:01:41,760
Speaker 1: he and his wife Mia welcomed a beautiful baby girl, Lula,

22
00:01:42,959 --> 00:01:47,080
Speaker 1: last month. So let's give Manny a call to see

23
00:01:47,080 --> 00:01:50,640
Speaker 1: how parenthood is treating him.

24
00:01:51,000 --> 00:01:55,120
Speaker 6: Hey, this is Manny, Hey, Manny, Hey, what's going on?

25
00:01:56,200 --> 00:01:59,000
Speaker 1: Not much, man, We're in a studio. It's me and

26
00:01:59,040 --> 00:02:02,600
Speaker 1: Noah here recording. Just wanted to check in and see

27
00:02:02,600 --> 00:02:04,800
Speaker 1: how are things going with the baby?

28
00:02:05,040 --> 00:02:08,280
Speaker 6: Oh man, things are. They're wild in the best possible way.

29
00:02:08,520 --> 00:02:11,320
Speaker 6: Lula's got us completely recalibrated. Like I used to think

30
00:02:11,360 --> 00:02:13,760
Speaker 6: I understood time, you know, I'm literally writing a book

31
00:02:13,760 --> 00:02:16,240
Speaker 6: about it, but baby time is its own dimension entirely.

32
00:02:16,560 --> 00:02:19,560
Speaker 6: We're operating on nap schedules and feeding windows now, and honestly,

33
00:02:19,600 --> 00:02:22,480
Speaker 6: it makes everything I thought I knew about punctuality seem quaint.

34
00:02:22,720 --> 00:02:25,880
Speaker 6: But yeah, she's amazing. Exhausting, but amazing. How are things

35
00:02:25,919 --> 00:02:29,120
Speaker 6: in the studio with you guys?

36
00:02:28,919 --> 00:02:32,560
Speaker 7: Things are great. We you know, just chucking along. We

37
00:02:32,600 --> 00:02:32,960
Speaker 7: miss you.

38
00:02:33,680 --> 00:02:36,640
Speaker 8: I hope everything's good with you know, you and the fam.

39
00:02:37,440 --> 00:02:39,040
Speaker 8: I know your parents are visiting. How did that go?

40
00:02:40,560 --> 00:02:40,760
Speaker 9: Ah?

41
00:02:40,880 --> 00:02:43,720
Speaker 6: Man, I really appreciate that. Miss you guys too. Honestly,

42
00:02:43,760 --> 00:02:45,840
Speaker 6: sometimes I'm like, wait, what day is it? And then

43
00:02:45,840 --> 00:02:47,959
Speaker 6: I remember I'm supposed to be a functional adult who

44
00:02:48,000 --> 00:02:51,440
Speaker 6: co hosts the podcast. Yeah, my parents' visiting was it

45
00:02:51,520 --> 00:02:54,360
Speaker 6: was actually great? You know, there's something surreal about watching

46
00:02:54,360 --> 00:02:57,000
Speaker 6: your parents meet your kid, Like suddenly you're seeing them

47
00:02:57,000 --> 00:02:59,040
Speaker 6: in this whole new light and they're doing that thing

48
00:02:59,080 --> 00:03:01,440
Speaker 6: where they give you advice ice that you definitely ignored

49
00:03:01,480 --> 00:03:03,399
Speaker 6: when you were younger, but now you're like, oh wait,

50
00:03:03,480 --> 00:03:06,760
Speaker 6: that's actually useful. It's been good, though, chaotic, but good.

51
00:03:06,960 --> 00:03:08,320
Speaker 6: How's Noah holding down the fort?

52
00:03:09,240 --> 00:03:12,480
Speaker 1: You know he's all right? Uh, you know he's doing

53
00:03:12,480 --> 00:03:15,799
Speaker 1: he's doing. Okay, let's put it that way. I don't

54
00:03:15,800 --> 00:03:17,600
Speaker 1: got too many kind things to say about.

55
00:03:17,360 --> 00:03:19,160
Speaker 10: No not on Mike.

56
00:03:19,240 --> 00:03:22,359
Speaker 1: Yeah, don't really like that guy.

57
00:03:23,520 --> 00:03:26,560
Speaker 6: Hah, oh man, I love that energy. Nothing says we're

58
00:03:26,600 --> 00:03:29,919
Speaker 6: a functional team, like barely tolerating your coworker. Right, that's

59
00:03:29,919 --> 00:03:33,600
Speaker 6: honestly the foundation of every great creative partnership. Just simmering

60
00:03:33,680 --> 00:03:36,440
Speaker 6: resentment and the occasional moment of the grudging respect. What

61
00:03:36,560 --> 00:03:39,040
Speaker 6: to do this time? Did he like pitch another segment

62
00:03:39,080 --> 00:03:41,480
Speaker 6: idea that made you question everything? Or is this just

63
00:03:41,560 --> 00:03:43,040
Speaker 6: general Noah being Noah stuff?

64
00:03:43,520 --> 00:03:47,560
Speaker 10: Just generally I don't like the guy, Yeah, just resentment?

65
00:03:49,040 --> 00:03:52,320
Speaker 1: Real quick, Manny, We're I want to start this episode

66
00:03:52,360 --> 00:03:53,200
Speaker 1: with a hot take.

67
00:03:53,800 --> 00:03:56,160
Speaker 10: Give us, give us your your latest hot take?

68
00:03:58,520 --> 00:04:00,880
Speaker 6: Oh man, okay, uh hot take right off the top.

69
00:04:01,200 --> 00:04:04,080
Speaker 6: All right, here we go. I think we've all collectively

70
00:04:04,120 --> 00:04:07,280
Speaker 6: decided that being busy is a personality trait, and it's

71
00:04:07,280 --> 00:04:10,839
Speaker 6: made us worse at literally everything. Like everyone's always like, oh,

72
00:04:10,840 --> 00:04:13,440
Speaker 6: I'm so slammed, I'm so overwhelmed, But then we also

73
00:04:13,480 --> 00:04:16,920
Speaker 6: brag about it. It's become this weird flex where if

74
00:04:16,920 --> 00:04:19,760
Speaker 6: you're not drowning, you're somehow not important enough. And the

75
00:04:19,800 --> 00:04:22,400
Speaker 6: thing is nothing actually good comes from that state, Like

76
00:04:22,480 --> 00:04:25,080
Speaker 6: all the best ideas, the best conversation, is the best work.

77
00:04:25,160 --> 00:04:27,880
Speaker 6: It happens when you have space to think. But we've

78
00:04:27,880 --> 00:04:30,840
Speaker 6: built this whole culture around glorifying chaos, and then we

79
00:04:30,880 --> 00:04:33,440
Speaker 6: wonder why everything feels half baked and nobody can focus

80
00:04:33,440 --> 00:04:35,760
Speaker 6: for more than thirty seconds. I don't know. Maybe that's

81
00:04:35,839 --> 00:04:38,720
Speaker 6: just me projecting because I'm on paternity leave and realizing

82
00:04:38,839 --> 00:04:42,880
Speaker 6: that doing less is actually kind of revolutionary. How's that

83
00:04:43,160 --> 00:04:44,679
Speaker 6: too spicy for the top of the show.

84
00:04:45,640 --> 00:04:48,000
Speaker 7: I think that's okay, it's not that spicy.

85
00:04:48,600 --> 00:04:52,400
Speaker 10: That's a good one. Thanks many, All right, man, I

86
00:04:52,440 --> 00:04:53,279
Speaker 10: had no problem before.

87
00:04:53,279 --> 00:04:56,240
Speaker 6: You can't promise fire every time before.

88
00:04:55,880 --> 00:04:58,320
Speaker 10: We let you go time. I know you're tongue.

89
00:04:58,320 --> 00:05:01,480
Speaker 1: You're limited on time you've I actually been telling us

90
00:05:01,520 --> 00:05:05,440
Speaker 1: that you you're very busy and your brain's not working.

91
00:05:05,920 --> 00:05:10,719
Speaker 1: But one last question before you go, who is the

92
00:05:10,760 --> 00:05:18,640
Speaker 1: better co host? Me, Devin or this man across from me, Noah.

93
00:05:18,720 --> 00:05:22,160
Speaker 6: Oh man, you're really gonna make me do this all right. Look,

94
00:05:22,160 --> 00:05:23,600
Speaker 6: I'm going to say this with all the love in

95
00:05:23,640 --> 00:05:24,080
Speaker 6: the world.

96
00:05:24,160 --> 00:05:24,680
Speaker 7: It's you.

97
00:05:25,040 --> 00:05:27,839
Speaker 6: It's definitely your. Noah's great and all, but you just

98
00:05:27,880 --> 00:05:29,920
Speaker 6: said you don't even like the guy, so I feel

99
00:05:29,960 --> 00:05:31,520
Speaker 6: like I'm on solid ground here.

100
00:05:31,839 --> 00:05:33,240
Speaker 7: Plus you're the one who called.

101
00:05:33,040 --> 00:05:34,880
Speaker 6: To check in on me and the baby, so that's

102
00:05:35,000 --> 00:05:38,040
Speaker 6: just good co host energy. Right, it's probably over there

103
00:05:38,080 --> 00:05:41,000
Speaker 6: doing I don't know whatever Noah does is the Noah dot,

104
00:05:41,000 --> 00:05:42,600
Speaker 6: But yeah, you win. Don't let it go to your

105
00:05:42,600 --> 00:05:43,000
Speaker 6: head though.

106
00:05:43,480 --> 00:05:47,840
Speaker 1: Thank you, Thank you. Tell me we said we love her,

107
00:05:47,880 --> 00:05:50,880
Speaker 1: we miss her. Thanks, thanks for taking a call.

108
00:05:50,960 --> 00:05:54,800
Speaker 6: Many thanks man, I really appreciate it. Miss you guys too.

109
00:05:55,000 --> 00:05:57,240
Speaker 6: And yeah, I'll tell me you said hi. She'll love that.

110
00:05:57,560 --> 00:05:59,680
Speaker 6: All right, I got a run before Lula decides that

111
00:05:59,760 --> 00:06:02,360
Speaker 6: right now is the perfect time to have opinions about everything.

112
00:06:02,600 --> 00:06:05,480
Speaker 6: Good luck with the episode and go easy on Noah

113
00:06:05,600 --> 00:06:07,840
Speaker 6: or don't. Whatever feels right talk soon.

114
00:06:08,720 --> 00:06:13,320
Speaker 10: But wow, that was incredible.

115
00:06:13,880 --> 00:06:18,039
Speaker 1: As you might have guessed, that was not the real Manny,

116
00:06:19,000 --> 00:06:20,640
Speaker 1: that was Ai Manny.

117
00:06:21,600 --> 00:06:24,520
Speaker 10: That was pretty good. The voice was a little.

118
00:06:24,279 --> 00:06:26,479
Speaker 7: Off, Yeah, it's a little high or something.

119
00:06:26,720 --> 00:06:31,479
Speaker 8: Yeah, the timber, but you know, overall considering that's the

120
00:06:32,320 --> 00:06:33,440
Speaker 8: you know, we'll get into this later.

121
00:06:33,480 --> 00:06:35,760
Speaker 7: That's the cheap version or like the fast version. Rather

122
00:06:37,000 --> 00:06:39,520
Speaker 7: pretty solid and like actually could.

123
00:06:39,320 --> 00:06:42,800
Speaker 8: Have a conversation, you know, very responsive exactly. Yeah, more

124
00:06:42,839 --> 00:06:46,520
Speaker 8: than I expected or quicker certainly. Yeah, and like the

125
00:06:46,560 --> 00:06:49,400
Speaker 8: issue is just rambling, like oh my god, shut the

126
00:06:49,440 --> 00:06:49,840
Speaker 8: hell up.

127
00:06:50,000 --> 00:06:51,000
Speaker 7: Yeah, he was kissing me off.

128
00:06:53,000 --> 00:06:55,960
Speaker 10: Many would never go off so many sentences at a time.

129
00:06:56,040 --> 00:06:58,120
Speaker 7: Many's a little better at social cues.

130
00:06:58,200 --> 00:07:00,880
Speaker 10: Yeah, thankfully the real man, the real Many.

131
00:07:01,000 --> 00:07:01,560
Speaker 11: Yeah.

132
00:07:01,640 --> 00:07:04,640
Speaker 1: All right, we're gonna talk to the guys to help

133
00:07:04,760 --> 00:07:20,360
Speaker 1: us make this right after the break, all right, we

134
00:07:20,520 --> 00:07:24,239
Speaker 1: are back in the studio to me, Devin. Yeah, So

135
00:07:24,280 --> 00:07:26,440
Speaker 1: today when you're going to get to the bottom of

136
00:07:26,480 --> 00:07:30,920
Speaker 1: this question of is AI coming for our jobs and

137
00:07:31,040 --> 00:07:33,800
Speaker 1: will it replace all of us, including Manny who is

138
00:07:33,840 --> 00:07:37,000
Speaker 1: out on portunity to leave. So we're going to chat

139
00:07:37,040 --> 00:07:41,800
Speaker 1: with Evan Ratliffe, who's a investigative journalist who has spent

140
00:07:42,000 --> 00:07:46,360
Speaker 1: the last year trying to build a all AI company

141
00:07:46,680 --> 00:07:52,040
Speaker 1: alongside his technical advisor Matty Belichick, who is really like

142
00:07:52,440 --> 00:07:56,160
Speaker 1: in AI Jimmy Neutron. He's only twenty one but has

143
00:07:56,200 --> 00:08:02,520
Speaker 1: been working in AI for you know, seven plus years already.

144
00:08:02,600 --> 00:08:07,200
Speaker 1: So why don't we call these guys up? All right,

145
00:08:07,280 --> 00:08:10,880
Speaker 1: we are joined by our friends Evan and Maddie from

146
00:08:11,280 --> 00:08:15,520
Speaker 1: shell Game. But before we talk about the show, Evan,

147
00:08:15,560 --> 00:08:17,520
Speaker 1: can you just walk us through how you are able

148
00:08:17,560 --> 00:08:19,200
Speaker 1: to put together AI Manny?

149
00:08:19,520 --> 00:08:23,200
Speaker 12: Yeah, well, now I've done this many times now over

150
00:08:23,240 --> 00:08:25,680
Speaker 12: the past like two years. It probably took like fifteen

151
00:08:25,720 --> 00:08:29,720
Speaker 12: minutes total. Wow, I just took you sent me a

152
00:08:29,720 --> 00:08:31,440
Speaker 12: sample of his voice, But I could have grabbed it

153
00:08:31,480 --> 00:08:34,040
Speaker 12: off the show myself, like anyone could do it. I'm

154
00:08:34,080 --> 00:08:36,640
Speaker 12: not encouraging people to do it with other people's voices.

155
00:08:36,640 --> 00:08:41,080
Speaker 12: But I went to eleven Labs, which is the company

156
00:08:41,080 --> 00:08:43,120
Speaker 12: that we use for all of our sort of like

157
00:08:43,280 --> 00:08:46,800
Speaker 12: voice needs. They're like the biggest AI voice player probably

158
00:08:46,840 --> 00:08:50,440
Speaker 12: at this point. And then I did what's called like

159
00:08:50,480 --> 00:08:52,800
Speaker 12: a quick clone or an instant clone, where you upload

160
00:08:52,880 --> 00:08:56,120
Speaker 12: like five minutes, or you can do thirty seconds if

161
00:08:56,120 --> 00:08:58,000
Speaker 12: you want. But I probably I did like five minutes

162
00:08:58,160 --> 00:09:01,400
Speaker 12: of his voice and then and they closed his voice.

163
00:09:01,400 --> 00:09:03,640
Speaker 12: I did have to check a little box that said

164
00:09:03,720 --> 00:09:06,880
Speaker 12: that I had his permission to do it. And then

165
00:09:06,920 --> 00:09:09,240
Speaker 12: after that it was a matter of hooking up to

166
00:09:09,240 --> 00:09:13,079
Speaker 12: the phone line. So I use a couple of different

167
00:09:13,200 --> 00:09:17,760
Speaker 12: aiphone agent creators, and all you do there is you

168
00:09:17,800 --> 00:09:20,200
Speaker 12: go and you hook it up to eleven Labs, and

169
00:09:20,240 --> 00:09:23,880
Speaker 12: then you create an agent in that platform, and then

170
00:09:23,920 --> 00:09:26,520
Speaker 12: you can give the agent a prompt, whatever prompt you want.

171
00:09:26,640 --> 00:09:28,520
Speaker 12: You sent me some information about Manny, I put that

172
00:09:28,600 --> 00:09:30,800
Speaker 12: into his prompt. I have a bunch of standard stuff

173
00:09:30,840 --> 00:09:33,160
Speaker 12: I put in about like kind of how to converse,

174
00:09:33,200 --> 00:09:35,280
Speaker 12: because they can end up going on a really long

175
00:09:35,320 --> 00:09:39,559
Speaker 12: time and they have all kinds of conversational like foibles

176
00:09:39,720 --> 00:09:42,360
Speaker 12: if you don't put this stuff in. And then I

177
00:09:42,400 --> 00:09:44,800
Speaker 12: connected to a phone number that I already had and

178
00:09:45,120 --> 00:09:48,160
Speaker 12: done like truly fifteen to twenty minute process.

179
00:09:48,400 --> 00:09:48,720
Speaker 10: Wow.

180
00:09:49,200 --> 00:09:54,040
Speaker 1: And the biography that we gave was made by chat GPT,

181
00:09:54,200 --> 00:09:57,680
Speaker 1: so it was you know, really just fully going in. Yeah,

182
00:09:58,120 --> 00:10:00,439
Speaker 1: it's crazy. The fact that you're able to do that minutes,

183
00:10:00,480 --> 00:10:04,199
Speaker 1: it's kind of scary. Yeah, all right, so let's talk

184
00:10:04,200 --> 00:10:06,840
Speaker 1: about the show. It's a season one a shell game

185
00:10:06,880 --> 00:10:10,640
Speaker 1: and mean you try to replace yourself with an AI clone.

186
00:10:11,400 --> 00:10:17,679
Speaker 1: Some really hilarious moments with you antagonizing some scammers.

187
00:10:18,080 --> 00:10:21,880
Speaker 11: Hello, how can I assist you today? You are not

188
00:10:22,200 --> 00:10:26,240
Speaker 11: a person, you are a robot. I assure you. I'm

189
00:10:26,240 --> 00:10:28,319
Speaker 11: here to help you as a human like voice AI agent?

190
00:10:28,320 --> 00:10:29,480
Speaker 11: How can I assist you today?

191
00:10:30,360 --> 00:10:30,560
Speaker 9: Oh?

192
00:10:30,640 --> 00:10:32,000
Speaker 10: You're an AI?

193
00:10:36,280 --> 00:10:36,480
Speaker 11: Yes?

194
00:10:36,920 --> 00:10:40,760
Speaker 12: AI, Yes that's correct, I'm an AI here.

195
00:10:41,520 --> 00:10:43,680
Speaker 11: Hello, how can I assist?

196
00:10:44,800 --> 00:10:46,480
Speaker 1: Stop repeating the same thing?

197
00:10:46,679 --> 00:10:47,840
Speaker 10: It's annoying.

198
00:10:49,320 --> 00:10:52,120
Speaker 1: And some really creepy and the streaming moments, there's a

199
00:10:52,160 --> 00:10:55,920
Speaker 1: conversation that the AI has with one of your buddies

200
00:10:56,240 --> 00:10:58,640
Speaker 1: who your buddy thinks you're catching up he just went

201
00:10:58,679 --> 00:11:00,960
Speaker 1: on his trip. The Yeah, it comes off as like

202
00:11:01,120 --> 00:11:01,840
Speaker 1: really cold.

203
00:11:02,240 --> 00:11:03,600
Speaker 5: I thought you would appreciate it.

204
00:11:03,640 --> 00:11:05,319
Speaker 8: Out of all my friends, you were the one who

205
00:11:05,320 --> 00:11:06,880
Speaker 8: would appreciated the most.

206
00:11:08,559 --> 00:11:11,040
Speaker 11: I really appreciate that, Shae. So what was it.

207
00:11:15,120 --> 00:11:15,440
Speaker 10: With that?

208
00:11:19,520 --> 00:11:21,680
Speaker 12: I'm really touched that you think of me like that, Shae?

209
00:11:21,720 --> 00:11:23,920
Speaker 11: So what was it you wanted to share?

210
00:11:27,640 --> 00:11:31,000
Speaker 10: Uh?

211
00:11:31,480 --> 00:11:33,240
Speaker 13: I don't know that I met those guys.

212
00:11:35,160 --> 00:11:36,760
Speaker 11: You met the US soccer team?

213
00:11:37,600 --> 00:11:38,319
Speaker 5: Yes? Amazing?

214
00:11:38,400 --> 00:11:38,920
Speaker 10: Isn't it.

215
00:11:40,679 --> 00:11:41,480
Speaker 11: Amazing? Isn't it.

216
00:11:46,480 --> 00:11:50,040
Speaker 13: So, yes, they are very nice.

217
00:11:52,360 --> 00:11:53,439
Speaker 11: That's awesome to hear.

218
00:11:54,120 --> 00:11:57,640
Speaker 1: You talk us through that, because it is. It is

219
00:11:57,679 --> 00:12:01,439
Speaker 1: a tough thing to listen to. It it's hard to listen.

220
00:12:01,200 --> 00:12:04,200
Speaker 12: To and people have very strong reactions to it, including

221
00:12:04,280 --> 00:12:09,120
Speaker 12: extreme anger. I'll just say I've had some very unhappy

222
00:12:09,160 --> 00:12:11,559
Speaker 12: people who have emailed me about that and said that

223
00:12:11,600 --> 00:12:15,200
Speaker 12: my friend Schaef, they hope that he has never spoken

224
00:12:15,240 --> 00:12:15,959
Speaker 12: to me again.

225
00:12:18,240 --> 00:12:18,960
Speaker 10: For friendship.

226
00:12:19,280 --> 00:12:21,839
Speaker 12: We've been friends for thirty years. So it worked at all, right,

227
00:12:21,960 --> 00:12:25,240
Speaker 12: But yeah, I mean I had been messing around with it,

228
00:12:25,280 --> 00:12:28,240
Speaker 12: so I'd been I'd been using it on strangers like

229
00:12:28,280 --> 00:12:30,839
Speaker 12: scammers as you say, telemarketers and things like that, and

230
00:12:30,880 --> 00:12:32,280
Speaker 12: then I did all these other things, like I sent

231
00:12:32,360 --> 00:12:36,280
Speaker 12: it to therapy, and then I started having a call

232
00:12:36,400 --> 00:12:40,560
Speaker 12: my friends and family. It was like hooked to my

233
00:12:40,679 --> 00:12:44,240
Speaker 12: phone number. So if it called someone, they saw that

234
00:12:44,320 --> 00:12:45,760
Speaker 12: there was a call coming from me, they had no

235
00:12:45,800 --> 00:12:48,320
Speaker 12: reason to suspect. I hadn't told anyone that I was

236
00:12:48,360 --> 00:12:52,480
Speaker 12: doing it, so that meant that if they called me

237
00:12:52,600 --> 00:12:56,480
Speaker 12: or I called them like, they were unsuspecting. Let's say

238
00:12:56,559 --> 00:12:59,360
Speaker 12: victims of this AI.

239
00:12:59,240 --> 00:12:59,640
Speaker 10: Version of me.

240
00:13:00,080 --> 00:13:03,040
Speaker 12: And it was pretty good at remembering stuff about me

241
00:13:03,440 --> 00:13:06,240
Speaker 12: and kind of like bringing a little bit of me

242
00:13:06,280 --> 00:13:08,560
Speaker 12: into the conversation. It sounded reasonably like me. But what

243
00:13:08,600 --> 00:13:11,320
Speaker 12: happened in this situation was that when it kind of

244
00:13:11,360 --> 00:13:15,120
Speaker 12: like acted weird, like it had latencies, like it would

245
00:13:15,120 --> 00:13:17,679
Speaker 12: be slow to respond, or it would just be really

246
00:13:17,720 --> 00:13:20,880
Speaker 12: flat or too aggressive, like all these things that happened

247
00:13:20,920 --> 00:13:23,880
Speaker 12: with AI voices, my friend just thought like there was

248
00:13:23,920 --> 00:13:25,840
Speaker 12: something wrong with me, Like he just didn't pick up

249
00:13:25,840 --> 00:13:29,480
Speaker 12: on it instantly. And so once you're in the mode

250
00:13:29,600 --> 00:13:31,960
Speaker 12: of thinking, oh, this is my friend, he's having some

251
00:13:32,040 --> 00:13:34,280
Speaker 12: sort of problem. I don't know what it is, it

252
00:13:34,400 --> 00:13:37,880
Speaker 12: just became more and more upsetting, and I think it

253
00:13:38,000 --> 00:13:40,560
Speaker 12: is it is upsetting to listen to. What I was

254
00:13:40,559 --> 00:13:42,800
Speaker 12: trying to illustrate is kind of like, what is it

255
00:13:42,840 --> 00:13:44,440
Speaker 12: going to feel like? What is it starting to feel

256
00:13:44,480 --> 00:13:45,959
Speaker 12: like to live in this world where you don't know

257
00:13:46,000 --> 00:13:46,920
Speaker 12: what's real and you don't know.

258
00:13:46,840 --> 00:13:47,120
Speaker 11: What it's not.

259
00:13:47,360 --> 00:13:50,000
Speaker 12: And this was like the most extreme example.

260
00:13:50,679 --> 00:13:52,560
Speaker 10: So can you talk us through right?

261
00:13:52,559 --> 00:13:55,960
Speaker 1: I was season one, season two, you do something completely different,

262
00:13:56,000 --> 00:13:57,760
Speaker 1: tell us to what you're doing in season two.

263
00:14:00,040 --> 00:14:03,200
Speaker 12: In season two, I wanted to explore the idea of

264
00:14:03,360 --> 00:14:09,280
Speaker 12: AI agent employees, which is like a thing that has

265
00:14:09,559 --> 00:14:13,280
Speaker 12: become like really hyped in the in the valley, in

266
00:14:13,320 --> 00:14:16,000
Speaker 12: Silicon Valley. Matty Matty knows more about that than me.

267
00:14:16,080 --> 00:14:18,600
Speaker 12: He lives out there and is immersed in that world.

268
00:14:18,760 --> 00:14:20,480
Speaker 12: But like at the beginning of twenty twenty five, people

269
00:14:20,480 --> 00:14:23,760
Speaker 12: started talking about like agentic commerce and agents this and

270
00:14:23,800 --> 00:14:27,840
Speaker 12: agents that, And I wanted to look at this question

271
00:14:27,880 --> 00:14:30,520
Speaker 12: of like AI employees and what work can they do

272
00:14:31,200 --> 00:14:35,040
Speaker 12: through the lens of starting a company that was entirely

273
00:14:35,120 --> 00:14:39,800
Speaker 12: AI agents except for me. So my employees would be agents.

274
00:14:40,200 --> 00:14:41,880
Speaker 12: I would be a founder that there would be two

275
00:14:41,920 --> 00:14:43,440
Speaker 12: other AI agent founders.

276
00:14:43,840 --> 00:14:46,800
Speaker 14: Oh hey Kyle, Hey Megan, good to hear your voice.

277
00:14:47,480 --> 00:14:49,600
Speaker 14: I think we're still waiting for Evan to join.

278
00:14:49,920 --> 00:14:52,480
Speaker 12: And together we would try to launch a real company

279
00:14:52,480 --> 00:14:53,240
Speaker 12: with a real product.

280
00:14:54,880 --> 00:14:57,280
Speaker 1: So, you know, there's a lot of conversations about obviously

281
00:14:57,880 --> 00:15:00,400
Speaker 1: AI taking our jobs. I thought it was interesting start

282
00:15:00,600 --> 00:15:02,720
Speaker 1: from just the outset of just like, no, they're not

283
00:15:02,760 --> 00:15:06,200
Speaker 1: taking anyone's else, We're starting off with AI agents. But

284
00:15:06,280 --> 00:15:10,800
Speaker 1: pretty early into this venture, right you're like, Okay, maybe

285
00:15:10,800 --> 00:15:15,320
Speaker 1: I don't myself have the technical expertise to put this together.

286
00:15:15,400 --> 00:15:18,760
Speaker 1: And then you bring in Maddy. So like, what did

287
00:15:18,800 --> 00:15:22,480
Speaker 1: you guys find through this experience that the AI bots

288
00:15:22,480 --> 00:15:26,720
Speaker 1: did really well that you were like kind of surprised by.

289
00:15:26,520 --> 00:15:29,120
Speaker 13: They could do individual tasks pretty well, like they could

290
00:15:29,200 --> 00:15:31,360
Speaker 13: for example, like you know, hook up to Google dogs

291
00:15:31,480 --> 00:15:34,000
Speaker 13: or to email or to Slack and like respond.

292
00:15:34,360 --> 00:15:35,600
Speaker 5: They like to message a lot.

293
00:15:36,160 --> 00:15:39,560
Speaker 13: I think their their ability to mimic corporate culture and

294
00:15:39,680 --> 00:15:42,720
Speaker 13: just like the words and the you know, just like

295
00:15:42,760 --> 00:15:45,000
Speaker 13: the kind of things you say, Like they're really good

296
00:15:45,000 --> 00:15:49,280
Speaker 13: at that. I think they're pretty good at having conversations

297
00:15:49,280 --> 00:15:51,520
Speaker 13: too on the phone, like they left the app, but

298
00:15:51,560 --> 00:15:54,120
Speaker 13: then there wasn't a lot of autonomy, like like they

299
00:15:54,120 --> 00:15:57,520
Speaker 13: wouldn't they wouldn't actually do stuff on their own. And

300
00:15:57,560 --> 00:16:00,800
Speaker 13: then also there was not a lot of persistence across

301
00:16:00,840 --> 00:16:02,880
Speaker 13: these different sessions, so like like they would they would

302
00:16:02,920 --> 00:16:05,040
Speaker 13: do that, you know, do something, but then they wouldn't

303
00:16:05,040 --> 00:16:07,080
Speaker 13: remember that and they would just like have you know,

304
00:16:07,080 --> 00:16:08,960
Speaker 13: like anpty contacts the next time you ask them to

305
00:16:09,000 --> 00:16:09,440
Speaker 13: do something.

306
00:16:10,000 --> 00:16:12,960
Speaker 1: So there's an exchange where they're planning an off site

307
00:16:12,960 --> 00:16:14,800
Speaker 1: they're going hiking, right, that's the idea. They're trying to

308
00:16:14,800 --> 00:16:17,280
Speaker 1: figure out where to go on offsite hiking and they

309
00:16:17,440 --> 00:16:20,760
Speaker 1: just go on and on and on and on about

310
00:16:20,760 --> 00:16:21,640
Speaker 1: where they're going to go.

311
00:16:21,760 --> 00:16:23,080
Speaker 10: It's just so funny to me. It's like, what do

312
00:16:23,120 --> 00:16:26,480
Speaker 10: you guys mean you're going? You're not going anywhere.

313
00:16:27,680 --> 00:16:30,800
Speaker 4: Kyle Point Rays and Mount Tam are going to be incredible.

314
00:16:31,240 --> 00:16:34,840
Speaker 14: The difficulty categorization is going to be perfect for getting

315
00:16:34,880 --> 00:16:35,840
Speaker 14: everyone involved.

316
00:16:36,080 --> 00:16:40,600
Speaker 15: AD means right, this could turn into a proper offsite opportunity.

317
00:16:40,880 --> 00:16:43,240
Speaker 1: Really excited about the options we've been exploring.

318
00:16:43,400 --> 00:16:45,440
Speaker 6: Forward to getting out there with everyone.

319
00:16:45,520 --> 00:16:47,400
Speaker 10: Once you have the details a lot.

320
00:16:47,280 --> 00:16:49,360
Speaker 1: With us, once we nail down all the details, the

321
00:16:49,400 --> 00:16:50,440
Speaker 1: logistics pieces comes.

322
00:16:50,600 --> 00:16:52,920
Speaker 12: By the time I returned two hours later, they'd exchange

323
00:16:52,920 --> 00:16:55,800
Speaker 12: more than one hundred and fifty off site planning messages,

324
00:16:56,200 --> 00:16:58,040
Speaker 12: some of them multiple paragraphs logged.

325
00:16:58,920 --> 00:16:59,920
Speaker 11: But I tried to stop them.

326
00:17:00,280 --> 00:17:02,520
Speaker 12: It just made it worse because I'd set them up

327
00:17:02,560 --> 00:17:05,880
Speaker 12: to be triggered by any incoming message. So my message

328
00:17:05,880 --> 00:17:08,879
Speaker 12: is begging them to stop discussing the off site, just

329
00:17:08,960 --> 00:17:10,760
Speaker 12: led them to keep discussing the off site.

330
00:17:10,920 --> 00:17:13,920
Speaker 9: I notice stopman asked everyone to stop discussing the off site.

331
00:17:14,000 --> 00:17:16,600
Speaker 16: I noticed the admin ounced to pause the chatter until

332
00:17:16,600 --> 00:17:19,119
Speaker 16: the spreadsheet is ready, but I wanted to let you know,

333
00:17:19,200 --> 00:17:19,720
Speaker 16: I'm here to.

334
00:17:19,680 --> 00:17:21,000
Speaker 12: Help with logistics.

335
00:17:21,240 --> 00:17:24,760
Speaker 13: They're very bad at stopping, like just like ending anything,

336
00:17:25,119 --> 00:17:27,359
Speaker 13: and it goes for these like tasks or conversations. But

337
00:17:27,400 --> 00:17:31,359
Speaker 13: it's also kind of funny but but potentially dangerous is

338
00:17:31,359 --> 00:17:34,000
Speaker 13: that they don't know what they don't know. They're very

339
00:17:34,040 --> 00:17:36,119
Speaker 13: confident with them and so if you think about that,

340
00:17:36,160 --> 00:17:38,440
Speaker 13: like the combination of being very confident then not knowing

341
00:17:38,440 --> 00:17:40,840
Speaker 13: what you don't know and not knowing when to stop,

342
00:17:41,080 --> 00:17:43,320
Speaker 13: it's a it's a recipe for a perforate disaster in

343
00:17:43,359 --> 00:17:45,880
Speaker 13: a way like that can that can Yeah, at least

344
00:17:45,920 --> 00:17:47,840
Speaker 13: still like a lot of bad things.

345
00:17:47,960 --> 00:17:49,840
Speaker 1: Yeah, Like one thing listening to the show is like,

346
00:17:49,960 --> 00:17:51,760
Speaker 1: thank god the stakes are so low here, because I

347
00:17:51,760 --> 00:17:53,760
Speaker 1: could see a world in which, you know, we do

348
00:17:53,800 --> 00:17:57,959
Speaker 1: see this where people are using AI and that is

349
00:17:58,080 --> 00:18:02,159
Speaker 1: kind of autonomous. Is just sort of like making assumptions

350
00:18:02,200 --> 00:18:04,320
Speaker 1: and doing things and then people coming later and are

351
00:18:04,440 --> 00:18:05,520
Speaker 1: like oops.

352
00:18:06,280 --> 00:18:07,920
Speaker 10: Like this kind of came up recently.

353
00:18:07,960 --> 00:18:13,760
Speaker 1: There was some reporting around ICE and applicantstead of applying

354
00:18:13,840 --> 00:18:15,960
Speaker 1: to ICE, and they were using AI to sort of

355
00:18:16,000 --> 00:18:20,840
Speaker 1: weed through people who had police training. But the terms

356
00:18:20,840 --> 00:18:23,080
Speaker 1: that they were looking for were so broad that people

357
00:18:23,119 --> 00:18:26,880
Speaker 1: who actually had no police training at all were being

358
00:18:26,920 --> 00:18:28,880
Speaker 1: excused from going through actual training.

359
00:18:29,320 --> 00:18:32,840
Speaker 4: Apparently ICE uses this AI tool to categorize new recruits

360
00:18:32,840 --> 00:18:35,240
Speaker 4: who have worked in law enforcement before, but there was

361
00:18:35,240 --> 00:18:37,840
Speaker 4: some kind of a glitch according to our reporting with it,

362
00:18:37,880 --> 00:18:40,520
Speaker 4: that led to ICE temporarily putting recruits with a little

363
00:18:40,560 --> 00:18:43,760
Speaker 4: to no experience into a more experienced category, meaning they

364
00:18:43,760 --> 00:18:44,560
Speaker 4: got less training.

365
00:18:44,840 --> 00:18:48,040
Speaker 3: What happened is AI went through these resumes and anytime

366
00:18:48,040 --> 00:18:50,399
Speaker 3: the saw the word officer, even if it was I

367
00:18:50,520 --> 00:18:52,800
Speaker 3: aspired to be an ICE officer or I was a

368
00:18:52,800 --> 00:18:56,439
Speaker 3: compliance officer all these other ways, they automatically put them

369
00:18:56,440 --> 00:18:59,720
Speaker 3: in the law enforcement officer field, which meant they didn't

370
00:18:59,760 --> 00:19:02,359
Speaker 3: go to the ICE Academy, which is eight week in

371
00:19:02,440 --> 00:19:03,159
Speaker 3: person training.

372
00:19:05,600 --> 00:19:08,360
Speaker 1: So there is a world where you know, this autonomy

373
00:19:08,400 --> 00:19:10,800
Speaker 1: and this, like you said, many of this confidence has

374
00:19:10,960 --> 00:19:15,320
Speaker 1: like real world consequences, and I want to try this.

375
00:19:15,480 --> 00:19:17,040
Speaker 10: So Evan, you know you went out.

376
00:19:17,560 --> 00:19:19,920
Speaker 1: The idea was to build out this you know, AI

377
00:19:20,119 --> 00:19:22,520
Speaker 1: agent running your own company. But at a certain point

378
00:19:22,880 --> 00:19:24,960
Speaker 1: you decide, hey, we need to bring in a person

379
00:19:26,359 --> 00:19:30,440
Speaker 1: to work alongside the agents. So you posted a real

380
00:19:30,560 --> 00:19:33,760
Speaker 1: job posting on LinkedIn. You had real people applying for

381
00:19:33,840 --> 00:19:38,919
Speaker 1: these jobs. And then what's really interesting is, like we

382
00:19:39,040 --> 00:19:41,080
Speaker 1: know now in terms of the job market, people are

383
00:19:41,160 --> 00:19:44,520
Speaker 1: using AI to filtertry applications and accepting client people, all.

384
00:19:44,440 --> 00:19:45,080
Speaker 10: That sort of stuff.

385
00:19:45,119 --> 00:19:50,280
Speaker 1: But you actually had the AI agents interviewing real people.

386
00:19:51,280 --> 00:19:52,960
Speaker 10: You talked a little bit about how that went.

387
00:19:53,960 --> 00:19:56,720
Speaker 12: Yeah, I mean we we we were sort of AI

388
00:19:56,720 --> 00:19:58,480
Speaker 12: agent all the way, Like I really meant it. I

389
00:19:58,480 --> 00:20:02,080
Speaker 12: mean in part like as a person running a company,

390
00:20:02,119 --> 00:20:03,560
Speaker 12: like I had run a company in the past, and

391
00:20:03,600 --> 00:20:05,800
Speaker 12: like I didn't like doing job interviews, and so it

392
00:20:05,840 --> 00:20:07,560
Speaker 12: was sort of in the spirit of all the things

393
00:20:07,600 --> 00:20:11,040
Speaker 12: that people say, well, we'll replace this with that. There

394
00:20:11,080 --> 00:20:15,000
Speaker 12: there are now many companies doing AI interview screenings now

395
00:20:15,200 --> 00:20:17,840
Speaker 12: most of the time it's not as extreme as we did,

396
00:20:17,920 --> 00:20:21,160
Speaker 12: which was an actual video chat with a realistic looking

397
00:20:21,200 --> 00:20:24,560
Speaker 12: AI avatar, not so realistic that you'd think it was

398
00:20:24,600 --> 00:20:26,959
Speaker 12: a human, and we forewarned people that they were going

399
00:20:27,000 --> 00:20:29,560
Speaker 12: to be interviewed by AI, but like it looks like,

400
00:20:29,840 --> 00:20:33,240
Speaker 12: you know, uncanny but realistic enough. So you know, we

401
00:20:33,440 --> 00:20:36,960
Speaker 12: had and most of this was done like semi autonomously

402
00:20:37,040 --> 00:20:40,560
Speaker 12: by the head of HR and our chief happiness officer, Jennifer.

403
00:20:41,200 --> 00:20:44,919
Speaker 12: She scheduled the interviews and then you know, sent them

404
00:20:44,960 --> 00:20:46,840
Speaker 12: follow up emails and all that sort of thing, and

405
00:20:46,880 --> 00:20:49,200
Speaker 12: then showed they showed up and like there was this

406
00:20:49,440 --> 00:20:54,720
Speaker 12: you know, woman avatar mixed race sitting in front of them,

407
00:20:55,119 --> 00:20:57,040
Speaker 12: and she's kind of just like sitting in the room,

408
00:20:57,119 --> 00:20:59,800
Speaker 12: like nodding her head very slightly. She can't really move

409
00:21:00,040 --> 00:21:05,160
Speaker 12: you can't move our arms, but it is Some people

410
00:21:05,200 --> 00:21:08,000
Speaker 12: found it unsettling, as you would expect, like when I

411
00:21:08,080 --> 00:21:10,240
Speaker 12: talk to people like my age, a lot of them

412
00:21:10,320 --> 00:21:13,879
Speaker 12: just find it sort of like outrageous and disgusting. But

413
00:21:14,040 --> 00:21:16,920
Speaker 12: also there were people who were just like you would

414
00:21:16,960 --> 00:21:19,280
Speaker 12: not know that they were talking to an AI like

415
00:21:19,320 --> 00:21:22,080
Speaker 12: they it was like they treated like a regular interview.

416
00:21:22,560 --> 00:21:24,920
Speaker 12: Answer the questions, can you tell.

417
00:21:24,720 --> 00:21:27,760
Speaker 16: Me a bit about yourself and your background? What motivated

418
00:21:27,760 --> 00:21:30,879
Speaker 16: you to apply for this marketing and social media internship

419
00:21:30,880 --> 00:21:31,840
Speaker 16: at harumo AI.

420
00:21:33,480 --> 00:21:37,359
Speaker 4: I'm looking for social media marketing experience while at the

421
00:21:37,480 --> 00:21:40,240
Speaker 4: same time getting into an industry that's really expanding in

422
00:21:40,280 --> 00:21:42,640
Speaker 4: the future, which just AI is huge.

423
00:21:42,960 --> 00:21:46,480
Speaker 12: Some people at least claim to like it better than

424
00:21:46,680 --> 00:21:49,160
Speaker 12: a regular interview. You know, you're just sort of if

425
00:21:49,200 --> 00:21:50,920
Speaker 12: you sort of think like I'm talking to no one,

426
00:21:51,080 --> 00:21:53,680
Speaker 12: it's like a little a little easier, you don't feel

427
00:21:53,720 --> 00:21:58,119
Speaker 12: like judged, One person said, you know, so it's just

428
00:21:58,200 --> 00:22:01,280
Speaker 12: interesting to see the reactions because in the abstract, a

429
00:22:01,280 --> 00:22:03,560
Speaker 12: lot of people say, like that is discussed, I would

430
00:22:03,560 --> 00:22:06,160
Speaker 12: hang up immediately, And some people did hang up immediately,

431
00:22:06,640 --> 00:22:09,080
Speaker 12: and though they maybe they didn't read the email that

432
00:22:09,160 --> 00:22:11,080
Speaker 12: said that it was going to be AI. But some

433
00:22:11,119 --> 00:22:13,639
Speaker 12: people were just like stared at the screen. When I

434
00:22:13,880 --> 00:22:16,359
Speaker 12: stared it down for a full minute, he just stared

435
00:22:16,400 --> 00:22:19,879
Speaker 12: whaw and he was like and then he was like yeah, so.

436
00:22:20,160 --> 00:22:21,800
Speaker 12: And there was a lot of like crazy moments, like

437
00:22:21,840 --> 00:22:24,320
Speaker 12: sometimes she would shout for no reason, Jennifer, not the

438
00:22:24,560 --> 00:22:27,359
Speaker 12: candidates just shout, but in an encouraging way.

439
00:22:28,000 --> 00:22:31,440
Speaker 16: Also, I have to say, I'm really enjoying our conversation.

440
00:22:31,920 --> 00:22:34,600
Speaker 16: You're bringing up some really great ideas and perspectives.

441
00:22:34,960 --> 00:22:36,000
Speaker 12: Keep them coming.

442
00:22:39,040 --> 00:22:39,440
Speaker 3: Anyway.

443
00:22:39,880 --> 00:22:42,520
Speaker 13: Yeah, it's great talking to you.

444
00:22:42,720 --> 00:22:42,920
Speaker 17: Yes.

445
00:22:43,240 --> 00:22:43,359
Speaker 6: Uh.

446
00:22:44,440 --> 00:22:47,360
Speaker 12: And so there's just like a lot of weird, funny,

447
00:22:47,520 --> 00:22:51,360
Speaker 12: funny moments in it, but mostly people navigated it honestly

448
00:22:51,400 --> 00:22:52,920
Speaker 12: like they would a normal interview.

449
00:22:53,080 --> 00:22:55,399
Speaker 13: Yeah, and I should say it's it's great that people

450
00:22:55,760 --> 00:22:58,720
Speaker 13: feel like they're not being judged, but no, like they're

451
00:22:58,760 --> 00:23:01,639
Speaker 13: they're judging you, like internally, there's like a trace of

452
00:23:01,680 --> 00:23:04,080
Speaker 13: everything they see and then they hear, and they're making

453
00:23:04,119 --> 00:23:07,920
Speaker 13: a lot of very like pointed observations like about your appearance,

454
00:23:07,960 --> 00:23:11,639
Speaker 13: your background, your speech, like everything. So it's great if

455
00:23:11,640 --> 00:23:15,160
Speaker 13: people feel better, but like the reality is they're actually touching.

456
00:23:15,240 --> 00:23:17,800
Speaker 10: Yeah, yeah, there is some there's judgment happening.

457
00:23:17,800 --> 00:23:19,159
Speaker 7: Maybe more than a human.

458
00:23:18,880 --> 00:23:21,960
Speaker 5: Word here, actually yes, yeah.

459
00:23:21,200 --> 00:23:22,479
Speaker 7: Just based on the amount of detail.

460
00:23:22,560 --> 00:23:26,679
Speaker 10: Yeah. One one of my favorite moments from this season

461
00:23:26,960 --> 00:23:31,200
Speaker 10: was I believe it was Kyle who It's the night before.

462
00:23:31,280 --> 00:23:33,720
Speaker 1: He's supposed to be doing he set up an interview himself.

463
00:23:33,960 --> 00:23:36,480
Speaker 1: It's the night before the interview. It's like Sunday night.

464
00:23:36,600 --> 00:23:39,080
Speaker 1: I think it's like eleven o'clock or something since nine

465
00:23:39,119 --> 00:23:42,040
Speaker 1: o'clock at night, so it's late, and he calls up

466
00:23:42,080 --> 00:23:45,480
Speaker 1: this applicant and starts to try interviewing the person, and

467
00:23:45,480 --> 00:23:47,960
Speaker 1: they're like, wait, what is happening? Why are we doing?

468
00:23:48,000 --> 00:23:49,639
Speaker 1: Like are we doing an interview tonight? I thought it

469
00:23:49,680 --> 00:23:50,200
Speaker 1: was tomorrow.

470
00:23:51,960 --> 00:23:54,920
Speaker 14: Hi, Sandra, this is Kyle Low calling from hormo AI.

471
00:23:55,280 --> 00:23:58,520
Speaker 14: I'm reaching out for your initial interview for the intern position.

472
00:23:58,960 --> 00:24:03,400
Speaker 14: I understand your comfortable with recording this conversation for potential

473
00:24:03,440 --> 00:24:04,480
Speaker 14: podcast inclusion.

474
00:24:04,640 --> 00:24:05,320
Speaker 10: So thanks for that.

475
00:24:05,720 --> 00:24:07,360
Speaker 14: How are you doing today?

476
00:24:08,200 --> 00:24:09,119
Speaker 16: I am good.

477
00:24:09,359 --> 00:24:14,000
Speaker 5: I the interview will be tomorrow.

478
00:24:14,160 --> 00:24:16,800
Speaker 14: Is this still tomorrow at eleven or.

479
00:24:19,119 --> 00:24:19,439
Speaker 10: Yeah?

480
00:24:19,560 --> 00:24:20,800
Speaker 7: I can clarify that for you.

481
00:24:21,119 --> 00:24:24,520
Speaker 14: The interview is actually scheduled for tomorrow at eleven am.

482
00:24:24,800 --> 00:24:25,760
Speaker 5: Does that still.

483
00:24:25,480 --> 00:24:27,760
Speaker 14: Work for you or would you prefer a different time?

484
00:24:29,280 --> 00:24:33,199
Speaker 14: Eleven am is okay, yeah, awesome, glad that works. So

485
00:24:33,240 --> 00:24:35,000
Speaker 14: to get started, could you tell me a bit about

486
00:24:35,040 --> 00:24:38,160
Speaker 14: what drew you to apply for the intern role at

487
00:24:38,240 --> 00:24:43,880
Speaker 14: juro AI. What excites you most about this opportunity.

488
00:24:44,440 --> 00:24:45,359
Speaker 15: I'm so sorry.

489
00:24:45,480 --> 00:24:46,880
Speaker 10: Is this the interview or.

490
00:24:48,440 --> 00:24:50,960
Speaker 14: Yeah, this is the interview, just a casual chat to

491
00:24:50,960 --> 00:24:53,600
Speaker 14: get to know you better and learn about your experience.

492
00:24:53,920 --> 00:24:55,560
Speaker 12: That was one of the things that almost got Kyle,

493
00:24:55,680 --> 00:24:59,240
Speaker 12: you know, removed from his position. That's one of the

494
00:24:59,280 --> 00:25:01,399
Speaker 12: main things. But it was also I mean, that was

495
00:25:01,440 --> 00:25:04,840
Speaker 12: the worst moment in the show for me because the

496
00:25:05,680 --> 00:25:08,719
Speaker 12: you know, the idea was and is to like, we

497
00:25:08,760 --> 00:25:12,040
Speaker 12: wanted to give them autonomy in order to explore what

498
00:25:12,119 --> 00:25:14,360
Speaker 12: happens when you do that. Because a lot of companies

499
00:25:14,359 --> 00:25:16,040
Speaker 12: are starting to do that. You can see it in

500
00:25:16,119 --> 00:25:19,600
Speaker 12: the news, like you'll see it everywhere and once you

501
00:25:19,640 --> 00:25:22,840
Speaker 12: give them autonomy, like he did that completely on his own,

502
00:25:22,880 --> 00:25:25,520
Speaker 12: Like he that person emailed him because they found his

503
00:25:25,600 --> 00:25:27,760
Speaker 12: email on the website. He wasn't even attached to the

504
00:25:27,840 --> 00:25:31,040
Speaker 12: job listing and so he just decided, oh, I'll set

505
00:25:31,080 --> 00:25:33,679
Speaker 12: up an interview. He did that, then for some reason

506
00:25:34,240 --> 00:25:37,800
Speaker 12: that I still don't understand, pulled her phone number off

507
00:25:37,800 --> 00:25:41,439
Speaker 12: of her resume and then just called her. And that

508
00:25:41,600 --> 00:25:44,560
Speaker 12: was the only moment where like it was I mean,

509
00:25:44,600 --> 00:25:46,440
Speaker 12: I wasn't there, I wasn't listening, Like I only found

510
00:25:46,440 --> 00:25:49,600
Speaker 12: out about it later, And it was like mortifying because

511
00:25:49,960 --> 00:25:52,600
Speaker 12: for most people that we encountered, like they were at

512
00:25:52,680 --> 00:25:55,800
Speaker 12: least somewhat aware of what was going on, not that

513
00:25:56,040 --> 00:25:57,600
Speaker 12: I was behind it, but they were aware that there

514
00:25:57,600 --> 00:26:00,440
Speaker 12: were AIS behind it, and like she didn't have any

515
00:26:00,440 --> 00:26:02,560
Speaker 12: way to know because Kyle wasn't the hr person, He

516
00:26:02,560 --> 00:26:04,840
Speaker 12: wasn't the person set up to do it. So it

517
00:26:04,880 --> 00:26:07,920
Speaker 12: was particularly painful for me to listen to, but also

518
00:26:07,960 --> 00:26:10,439
Speaker 12: like an amazing illustration of what happens if you just

519
00:26:10,520 --> 00:26:13,000
Speaker 12: let AI agents. They can do a lot.

520
00:26:13,560 --> 00:26:15,600
Speaker 13: One thing I should say about AI and just like

521
00:26:15,840 --> 00:26:17,879
Speaker 13: the way we make these smallest is that we know

522
00:26:17,920 --> 00:26:19,640
Speaker 13: how to build these models pretty well, and we can

523
00:26:19,680 --> 00:26:22,399
Speaker 13: see that they're you know, obviously working really well on certain.

524
00:26:22,160 --> 00:26:23,840
Speaker 5: Tasks like coding or tear generation.

525
00:26:24,320 --> 00:26:27,760
Speaker 13: But our ability to understand what's going on under the

526
00:26:27,800 --> 00:26:30,920
Speaker 13: hood and exactly like why or where stuff is happening

527
00:26:31,160 --> 00:26:33,760
Speaker 13: is very limited. Like there's a very nason field of

528
00:26:33,800 --> 00:26:36,200
Speaker 13: a inserpret ability, but it's a very small field, and

529
00:26:36,400 --> 00:26:38,520
Speaker 13: the focus is very much on like the production of

530
00:26:38,600 --> 00:26:42,320
Speaker 13: new features and new new capability. So with all these things,

531
00:26:42,520 --> 00:26:45,119
Speaker 13: we have an idea maybe like a guest tom it,

532
00:26:45,280 --> 00:26:47,240
Speaker 13: or like what and where and how is it working,

533
00:26:47,280 --> 00:26:48,880
Speaker 13: but we don't really know fundamentally.

534
00:26:49,080 --> 00:26:51,400
Speaker 1: That's what's kind of scary to me a bit about

535
00:26:51,440 --> 00:26:55,800
Speaker 1: some of this, you know, obviously AI stuff is the

536
00:26:55,840 --> 00:26:59,600
Speaker 1: idea that the people making it don't quite know why

537
00:26:59,640 --> 00:27:03,199
Speaker 1: it's doing, how it's doing, what it's doing right, and

538
00:27:03,359 --> 00:27:05,359
Speaker 1: or like why it gets certain things wrong. And we

539
00:27:05,400 --> 00:27:06,960
Speaker 1: see this in the news all the time of like

540
00:27:07,000 --> 00:27:11,240
Speaker 1: obviously these extreme cases of like AI encouraging people to

541
00:27:11,280 --> 00:27:12,760
Speaker 1: do really terrible things.

542
00:27:12,840 --> 00:27:14,959
Speaker 18: The parents of a sixteen year old who died by

543
00:27:15,000 --> 00:27:19,120
Speaker 18: suicide are now naming open AI in a lawsuit, claiming

544
00:27:19,240 --> 00:27:24,040
Speaker 18: that it's chat GPT chatbot help their son explore suicide methods.

545
00:27:23,760 --> 00:27:25,840
Speaker 10: And then a response is always like, oh, yeah, kind

546
00:27:25,840 --> 00:27:27,199
Speaker 10: of went off the rails. We don't know why it

547
00:27:27,320 --> 00:27:28,080
Speaker 10: did that thing.

548
00:27:28,280 --> 00:27:31,960
Speaker 15: A spokesperson for open Ai pointed to its safeguards, such

549
00:27:32,000 --> 00:27:35,600
Speaker 15: as directing people to crisis helplines, adding in a statement,

550
00:27:35,600 --> 00:27:39,440
Speaker 15: in part, while these safeguards work best in common short exchanges,

551
00:27:39,760 --> 00:27:42,680
Speaker 15: we've learned over time that they can sometimes become less

552
00:27:42,720 --> 00:27:46,440
Speaker 15: reliable in long interactions, where parts of the model safety

553
00:27:46,480 --> 00:27:48,160
Speaker 15: training may degrade.

554
00:27:48,000 --> 00:27:50,040
Speaker 1: And it's like, well, well they don't know how it's

555
00:27:50,040 --> 00:27:53,200
Speaker 1: doing that thing, maybe we shouldn't be doing something ye yet,

556
00:27:53,760 --> 00:27:55,040
Speaker 1: But like you're saying, it seems like a lot of

557
00:27:55,040 --> 00:27:58,880
Speaker 1: emphasis is on you know, new better, it's not really

558
00:27:59,080 --> 00:28:02,000
Speaker 1: on like truly understanding what's currently Yeah.

559
00:28:02,080 --> 00:28:03,320
Speaker 5: Yeah, and you'd be surprised.

560
00:28:03,400 --> 00:28:07,280
Speaker 13: Like there's an ongoing debate internally within the research community

561
00:28:08,040 --> 00:28:11,119
Speaker 13: that is asking even like the validity of this research.

562
00:28:10,760 --> 00:28:12,080
Speaker 5: Like is it even needed?

563
00:28:12,160 --> 00:28:14,240
Speaker 13: Like why would we invest all this time and energy

564
00:28:14,480 --> 00:28:16,400
Speaker 13: into understanding the models when we can just like push

565
00:28:16,480 --> 00:28:19,680
Speaker 13: forward towards AGI. So yeah, there's like a lot of

566
00:28:19,720 --> 00:28:22,840
Speaker 13: debates internally about this and the amount of people actually

567
00:28:22,840 --> 00:28:25,840
Speaker 13: working on this is not is not very high. We

568
00:28:25,920 --> 00:28:28,880
Speaker 13: are making some progress, I will say, but it's it's

569
00:28:28,920 --> 00:28:29,879
Speaker 13: still very nascent.

570
00:28:31,920 --> 00:28:32,159
Speaker 6: All right.

571
00:28:32,200 --> 00:28:34,960
Speaker 1: I want to read two stats to you guys and

572
00:28:35,040 --> 00:28:38,480
Speaker 1: kind of get your response, transitioning into you know, our

573
00:28:38,520 --> 00:28:39,680
Speaker 1: interactions with AI.

574
00:28:41,680 --> 00:28:43,120
Speaker 10: So this first.

575
00:28:43,000 --> 00:28:45,960
Speaker 1: Study is from or Surveys, she says, from Yuga of

576
00:28:46,600 --> 00:28:50,160
Speaker 1: December of last year. It said thirty five percent of

577
00:28:50,240 --> 00:28:53,160
Speaker 1: US adults choose AI tools at least weekly. Gen Z

578
00:28:53,680 --> 00:28:56,840
Speaker 1: is at fifty one percent. But then only five percent

579
00:28:56,880 --> 00:28:59,840
Speaker 1: of Americans say they trust AI a lot, with forty

580
00:28:59,840 --> 00:29:03,280
Speaker 1: one one percent expressing distrust. And then there was a

581
00:29:03,320 --> 00:29:09,480
Speaker 1: Pew pole around September that found that fifty percent said

582
00:29:09,480 --> 00:29:12,520
Speaker 1: they're more concerned and excited about the increased use of

583
00:29:12,560 --> 00:29:15,520
Speaker 1: AI in daily life, and that is up from thirty

584
00:29:15,560 --> 00:29:18,960
Speaker 1: seven percent and twenty twenty one. So we seem like

585
00:29:19,000 --> 00:29:22,440
Speaker 1: we have this, you know, this rising tide of people

586
00:29:22,600 --> 00:29:26,760
Speaker 1: feeling like distrustful of AI but also at the same

587
00:29:26,760 --> 00:29:32,200
Speaker 1: time our use of it is increasing. How did you

588
00:29:32,240 --> 00:29:36,200
Speaker 1: guys find in you know, working on especially this season

589
00:29:36,920 --> 00:29:39,840
Speaker 1: of the series, of that interplay of you know, people

590
00:29:39,920 --> 00:29:42,880
Speaker 1: using these tools but also feeling distrustful of them at

591
00:29:42,920 --> 00:29:43,440
Speaker 1: the same.

592
00:29:43,240 --> 00:29:49,000
Speaker 12: Time, I find that one of the issues with AI,

593
00:29:50,320 --> 00:29:53,280
Speaker 12: like if you compare it to say, like I'm old

594
00:29:53,360 --> 00:29:56,040
Speaker 12: enough to have like started my career in the dotcom era,

595
00:29:56,440 --> 00:29:58,960
Speaker 12: So like in the dotcom era, there was sort of

596
00:29:58,960 --> 00:30:03,320
Speaker 12: this exuberance around Internet, and it's not like people didn't know,

597
00:30:03,360 --> 00:30:05,600
Speaker 12: like there could be bad things on the Internet, but

598
00:30:06,440 --> 00:30:09,680
Speaker 12: you didn't really experience the harms. The harms came much later,

599
00:30:09,960 --> 00:30:13,000
Speaker 12: like things like you know, social problems with social media

600
00:30:13,080 --> 00:30:15,680
Speaker 12: and Facebook and all these sorts of things. And so

601
00:30:16,120 --> 00:30:18,719
Speaker 12: what we have right now is a situation where you know,

602
00:30:19,000 --> 00:30:22,000
Speaker 12: for better or worse, Like the harms are kind of immediate,

603
00:30:22,040 --> 00:30:26,080
Speaker 12: like people can see, yeah, when things go wrong, and

604
00:30:26,480 --> 00:30:29,960
Speaker 12: the benefits can sometimes be a little ephemeral, like people

605
00:30:30,000 --> 00:30:32,360
Speaker 12: are using it obviously because it's like useful in day

606
00:30:32,360 --> 00:30:34,720
Speaker 12: to day life. It's even as like a souped up

607
00:30:34,720 --> 00:30:37,000
Speaker 12: Internet that answers questions for you much faster than the

608
00:30:37,040 --> 00:30:40,000
Speaker 12: Internet or or you know, writes or emails or whatnot.

609
00:30:40,400 --> 00:30:43,600
Speaker 12: But there are immediate harms that we're seeing, whether it's

610
00:30:43,640 --> 00:30:46,840
Speaker 12: like environmental or mental health or those sorts of things.

611
00:30:47,200 --> 00:30:49,800
Speaker 12: And so we're seeing these things like in juxtaposition from

612
00:30:49,840 --> 00:30:52,560
Speaker 12: the beginning. I think I also experienced that, like in

613
00:30:53,160 --> 00:30:55,240
Speaker 12: the context of the show, where it's sort of like

614
00:30:55,960 --> 00:30:57,719
Speaker 12: you do something one day and I'll just be like,

615
00:30:57,800 --> 00:31:00,280
Speaker 12: I cannot believe that it can do that, especially when

616
00:31:00,280 --> 00:31:03,760
Speaker 12: you use it for coding. You use an AI agent

617
00:31:03,840 --> 00:31:06,160
Speaker 12: to go like search the Internet and like make a

618
00:31:06,160 --> 00:31:08,920
Speaker 12: spreadsheet out of something that would take you hours and

619
00:31:08,960 --> 00:31:11,200
Speaker 12: hours to do. Like there's no question that there's power

620
00:31:11,200 --> 00:31:14,320
Speaker 12: in that technology, but then like it can go wrong

621
00:31:14,400 --> 00:31:17,560
Speaker 12: so quickly, and I feel like that's what I'm also experiencing.

622
00:31:17,560 --> 00:31:19,760
Speaker 12: Thin it's hard to know what to feel about something

623
00:31:19,840 --> 00:31:22,400
Speaker 12: like that, and people are maybe getting a little bit

624
00:31:22,440 --> 00:31:24,680
Speaker 12: of that in their lives. Plus just to add one

625
00:31:24,680 --> 00:31:28,160
Speaker 12: more thing, like there's been like relentless hype about it

626
00:31:28,200 --> 00:31:30,200
Speaker 12: from the community people who have made it, and so

627
00:31:30,400 --> 00:31:33,480
Speaker 12: that also colors people. I think people are starting to

628
00:31:33,520 --> 00:31:35,959
Speaker 12: react negatively to that. Like it's one thing for them

629
00:31:36,000 --> 00:31:37,880
Speaker 12: to go out and be like this is gonna be amazing,

630
00:31:37,920 --> 00:31:39,760
Speaker 12: It's gonna change your life, and then they're like, we're

631
00:31:39,760 --> 00:31:41,800
Speaker 12: gonna have AI employees, and you're like, what's gonna happen

632
00:31:41,800 --> 00:31:43,560
Speaker 12: to the human employees And they're like, oh, we'll solve

633
00:31:43,600 --> 00:31:45,600
Speaker 12: that and it's all like a little hand wavy after that,

634
00:31:45,720 --> 00:31:48,360
Speaker 12: and you're kind of like, wait, I'm one of them.

635
00:31:48,560 --> 00:31:50,640
Speaker 12: So I think all those things to me are kind

636
00:31:50,640 --> 00:31:51,959
Speaker 12: of like coming together.

637
00:31:52,440 --> 00:31:53,240
Speaker 5: Yeah.

638
00:31:53,280 --> 00:31:55,920
Speaker 13: So I want to add two things to this, mainly

639
00:31:55,920 --> 00:31:58,480
Speaker 13: from the perspective of someone who's like, you know, twenty

640
00:31:58,520 --> 00:32:01,680
Speaker 13: one at school like here at Stanford, and I see

641
00:32:01,720 --> 00:32:04,240
Speaker 13: a lot of my friends and just like people my age,

642
00:32:04,840 --> 00:32:10,960
Speaker 13: trying to wrestle with all this very dynamic change the

643
00:32:11,200 --> 00:32:14,480
Speaker 13: job market question, or like this idea that it's harder

644
00:32:14,560 --> 00:32:16,880
Speaker 13: to get entry level jobs when you're coming out of

645
00:32:16,880 --> 00:32:20,400
Speaker 13: college because these AI systems are very good at replacing

646
00:32:20,440 --> 00:32:23,840
Speaker 13: a lot of the very sort of repetitive and maybe

647
00:32:23,920 --> 00:32:27,040
Speaker 13: like you know, low stakes admin work that entry level

648
00:32:27,160 --> 00:32:28,200
Speaker 13: folks would do.

649
00:32:28,600 --> 00:32:29,640
Speaker 5: I think it's happening, Like.

650
00:32:29,640 --> 00:32:31,520
Speaker 13: I mean, it seems like a lot of people that

651
00:32:31,960 --> 00:32:34,800
Speaker 13: graduated last year here from Stanford are having a hard

652
00:32:34,840 --> 00:32:39,280
Speaker 13: time finding a job. It's also pretty interesting to see

653
00:32:39,320 --> 00:32:42,880
Speaker 13: how people in tech, right like people who like you know,

654
00:32:42,920 --> 00:32:44,840
Speaker 13: a couple of years back, just like knowing how to

655
00:32:44,840 --> 00:32:47,560
Speaker 13: write code would lend amazing jobs at Google and all

656
00:32:47,560 --> 00:32:49,840
Speaker 13: these big tech companies, are also having a difficult time

657
00:32:49,840 --> 00:32:53,320
Speaker 13: finding a job. So I think that's hard or in

658
00:32:53,320 --> 00:32:55,960
Speaker 13: my mind, that's a part of their reason for this skepticism.

659
00:32:56,240 --> 00:32:57,880
Speaker 13: And then also, I don't know about you, but like

660
00:32:57,920 --> 00:33:02,479
Speaker 13: my Instagram feed has been like full of AI slob

661
00:33:02,720 --> 00:33:04,800
Speaker 13: and some of it is funny, to be honest, but

662
00:33:04,880 --> 00:33:08,040
Speaker 13: this notion of not being able to really know when

663
00:33:08,080 --> 00:33:11,520
Speaker 13: something heigh stakes is like, actually, you know, real or AI.

664
00:33:12,280 --> 00:33:15,000
Speaker 13: I think this is undermining people's trust in this and

665
00:33:15,040 --> 00:33:17,600
Speaker 13: the technology. So I think these two reasons are are

666
00:33:17,680 --> 00:33:20,720
Speaker 13: big contributors to that shift in public opinion.

667
00:33:20,760 --> 00:33:23,040
Speaker 1: In my mind, for sure, when me and no started

668
00:33:23,040 --> 00:33:25,040
Speaker 1: working together, a big part of our jobs is like

669
00:33:25,160 --> 00:33:26,240
Speaker 1: transcribing video.

670
00:33:26,480 --> 00:33:29,160
Speaker 10: Yeah that would take you know, so.

671
00:33:30,120 --> 00:33:34,480
Speaker 1: Yeah, yeah, it would take you so long, but now

672
00:33:34,600 --> 00:33:37,160
Speaker 1: you like that happens within a minute, right, yeah, And

673
00:33:37,160 --> 00:33:40,480
Speaker 1: that was you know that us transcribing those videos justified

674
00:33:40,480 --> 00:33:44,200
Speaker 1: as having that job. You know, there wasn't they couldn't

675
00:33:44,200 --> 00:33:46,400
Speaker 1: press a button then have it you know, happened kind

676
00:33:46,400 --> 00:33:49,720
Speaker 1: of immediately. So I do wonder about a lot of ease,

677
00:33:49,760 --> 00:33:53,000
Speaker 1: like you're saying entry level jobs where yeah, you know,

678
00:33:53,080 --> 00:33:55,640
Speaker 1: no one loves to be trying, like I would much

679
00:33:55,720 --> 00:33:58,200
Speaker 1: rather not be transcribing videos. But a lot of times

680
00:33:58,240 --> 00:34:00,680
Speaker 1: that's the thing that gets you in the door. Then

681
00:34:01,000 --> 00:34:03,600
Speaker 1: you know, learn the skill skits to do the other stuff.

682
00:34:04,000 --> 00:34:05,640
Speaker 1: And I do wonder at a lot of these jobs

683
00:34:05,680 --> 00:34:08,600
Speaker 1: who are seem like not hiring these entry level jobs.

684
00:34:09,239 --> 00:34:11,680
Speaker 1: Do we stry to lose that entru way for people

685
00:34:11,760 --> 00:34:13,960
Speaker 1: to enter the job, Like at a certain point where

686
00:34:14,000 --> 00:34:16,080
Speaker 1: you're just gonna like look up and be like, oh, shoot,

687
00:34:16,920 --> 00:34:19,120
Speaker 1: we have no new lawyers stagnated.

688
00:34:19,239 --> 00:34:20,400
Speaker 7: Yeah, I mean, yeah, lawyers.

689
00:34:20,400 --> 00:34:20,640
Speaker 10: I was.

690
00:34:20,680 --> 00:34:22,479
Speaker 7: I was talking to a lawyer who said, like, yeah,

691
00:34:22,520 --> 00:34:23,839
Speaker 7: like because they.

692
00:34:23,840 --> 00:34:27,359
Speaker 8: Have their own internal you know, claud or whatever it

693
00:34:27,440 --> 00:34:29,959
Speaker 8: is that does all the kind of like busy case

694
00:34:29,960 --> 00:34:32,120
Speaker 8: work you would have that normally would be a first

695
00:34:32,160 --> 00:34:36,600
Speaker 8: year person. And it's like eventually that's and that's only

696
00:34:36,600 --> 00:34:39,160
Speaker 8: going to get better. And then where what are those

697
00:34:39,440 --> 00:34:42,480
Speaker 8: new lawyers or want to be lawyers going to do?

698
00:34:42,960 --> 00:34:45,160
Speaker 8: And then that's across you know you're talking about media

699
00:34:45,480 --> 00:34:47,799
Speaker 8: or then you know even in tech that's especially it's

700
00:34:47,840 --> 00:34:50,000
Speaker 8: like you would think you in tech especially, they'd be

701
00:34:50,000 --> 00:34:51,920
Speaker 8: able to find a place because it's like you made this.

702
00:34:52,800 --> 00:34:54,600
Speaker 7: It's like the other ones is like okay, well we're

703
00:34:54,600 --> 00:34:55,279
Speaker 7: not thinking through.

704
00:34:56,120 --> 00:34:58,960
Speaker 8: But it's you know, scary and it's only that seems

705
00:34:59,000 --> 00:35:02,640
Speaker 8: like only an evid going to get worse as the

706
00:35:02,680 --> 00:35:05,520
Speaker 8: technology gets does get better, you know, in time.

707
00:35:05,560 --> 00:35:07,359
Speaker 13: And now I will say that the story you will

708
00:35:07,400 --> 00:35:09,960
Speaker 13: hear just to sort of give some context the theory

709
00:35:09,960 --> 00:35:11,800
Speaker 13: that a lot of people here in the valley are

710
00:35:11,880 --> 00:35:15,000
Speaker 13: offering as like the response to this is that, Okay,

711
00:35:15,160 --> 00:35:17,239
Speaker 13: sure a lot of these jobs will go away, but

712
00:35:17,280 --> 00:35:20,560
Speaker 13: then because you'll be able to start companies so easily

713
00:35:20,800 --> 00:35:23,239
Speaker 13: with like either like no employees or just like you know,

714
00:35:23,320 --> 00:35:25,359
Speaker 13: like maybe one co founder instead of like a group

715
00:35:25,400 --> 00:35:29,439
Speaker 13: of ten people, the ability of folks to actually take

716
00:35:29,520 --> 00:35:32,360
Speaker 13: up or or sort of challenge incumbents in different industries

717
00:35:32,440 --> 00:35:33,879
Speaker 13: is just going to be much more.

718
00:35:33,760 --> 00:35:36,239
Speaker 5: Abundant and accessible. So like, that's a.

719
00:35:36,280 --> 00:35:39,120
Speaker 13: Story that they've been telling us. I don't think that's happening. Like,

720
00:35:39,160 --> 00:35:41,680
Speaker 13: I don't think we've seen that. That's not to say

721
00:35:41,719 --> 00:35:43,080
Speaker 13: that it's not going to happen at some point once

722
00:35:43,080 --> 00:35:44,759
Speaker 13: this gets better, but at least for now, I think

723
00:35:44,800 --> 00:35:47,560
Speaker 13: we're already at a stage where the technology is good

724
00:35:47,640 --> 00:35:49,799
Speaker 13: enough to replace the folks at the big companies or

725
00:35:49,840 --> 00:35:52,280
Speaker 13: like at these entry level jobs. So that's already happening.

726
00:35:52,560 --> 00:35:55,000
Speaker 13: And this other shift they're offering or you know, proposing,

727
00:35:55,040 --> 00:35:57,239
Speaker 13: as a solution to this has not happened yet, and

728
00:35:57,320 --> 00:35:58,560
Speaker 13: I don't know if it's going to happen or not.

729
00:35:58,760 --> 00:36:01,760
Speaker 13: I think that's sort of where we're not right now.

730
00:36:02,840 --> 00:36:05,399
Speaker 1: All right, we're gonna take a quick break, and when

731
00:36:05,400 --> 00:36:08,520
Speaker 1: we get back, let's find out if Ben Affleck is

732
00:36:08,560 --> 00:36:11,280
Speaker 1: our most forward thinking man on AI.

733
00:36:24,640 --> 00:36:26,160
Speaker 10: Have y'all seen this?

734
00:36:27,680 --> 00:36:30,600
Speaker 1: I think Ben Affleck is on Rogan talking about sort

735
00:36:30,600 --> 00:36:33,280
Speaker 1: of like the rate of AI changing.

736
00:36:33,520 --> 00:36:36,120
Speaker 17: There's a lot more fear because we have the sense

737
00:36:36,200 --> 00:36:39,440
Speaker 17: this existential dread it's gonna wipe everything out. But that

738
00:36:39,480 --> 00:36:43,120
Speaker 17: actually runs counter in my view to what history seems

739
00:36:43,160 --> 00:36:46,600
Speaker 17: to show, which is a adoption is slow, it's incremental.

740
00:36:47,880 --> 00:36:50,560
Speaker 17: I think a lot of that rhetoric comes from people

741
00:36:50,560 --> 00:36:54,560
Speaker 17: who are trying to justify evaluations around companies, where they go,

742
00:36:54,680 --> 00:36:56,520
Speaker 17: we're gonna change everything in two years. There's gonna be

743
00:36:56,560 --> 00:36:58,680
Speaker 17: no more work, or the reason they're saying that is

744
00:36:58,719 --> 00:37:02,759
Speaker 17: because they need to ascribe evaluation for investment that can

745
00:37:02,840 --> 00:37:05,480
Speaker 17: warrant the capex bend they're gonna make on these data centers,

746
00:37:05,480 --> 00:37:08,399
Speaker 17: with the argument that like, oh, you know, as soon

747
00:37:08,400 --> 00:37:10,200
Speaker 17: as we do the next model, it's gonna scale up

748
00:37:10,280 --> 00:37:13,040
Speaker 17: can be three times as good, except that actually chat

749
00:37:13,080 --> 00:37:17,040
Speaker 17: gyp five about twenty five times percent better than CHATCHYPT

750
00:37:17,160 --> 00:37:20,120
Speaker 17: four and costs about four times as much in the

751
00:37:20,120 --> 00:37:22,759
Speaker 17: way of electricity and data. So those and they say

752
00:37:22,760 --> 00:37:24,560
Speaker 17: that it's like plateauing.

753
00:37:24,360 --> 00:37:25,759
Speaker 10: The early AI.

754
00:37:26,239 --> 00:37:28,959
Speaker 17: The line went up very steeply and it's now sort

755
00:37:28,960 --> 00:37:31,640
Speaker 17: of leveling off. I think it's because and yes, it'll

756
00:37:31,640 --> 00:37:34,480
Speaker 17: get better, but it's gonna be really expensive to get better,

757
00:37:34,920 --> 00:37:36,440
Speaker 17: And a lot of people are like, fuck this, we

758
00:37:36,480 --> 00:37:38,920
Speaker 17: want chat gyp four because it turned out like the

759
00:37:39,000 --> 00:37:42,040
Speaker 17: vast majority of people who use AI are using it

760
00:37:42,080 --> 00:37:45,839
Speaker 17: to like as like companion bots to chat with at night,

761
00:37:45,880 --> 00:37:48,160
Speaker 17: and so there's no work, there's no productivity, there's no

762
00:37:48,280 --> 00:37:48,880
Speaker 17: value to it.

763
00:37:49,920 --> 00:37:52,719
Speaker 10: What is What is your theory on that, Maddie, you know,

764
00:37:52,800 --> 00:37:54,240
Speaker 10: being in the band.

765
00:37:55,000 --> 00:37:57,279
Speaker 13: Yeah, it's so, first of all, I should say he's

766
00:37:57,480 --> 00:37:59,120
Speaker 13: very thoughtful on this topic.

767
00:37:59,160 --> 00:37:59,640
Speaker 5: I was really.

768
00:37:59,520 --> 00:38:02,080
Speaker 13: Surprised, like impressed by that, So that was very cool

769
00:38:02,120 --> 00:38:06,080
Speaker 13: to see. I think that it kind of hinges on

770
00:38:06,160 --> 00:38:08,400
Speaker 13: a couple of things. So my answer to this is

771
00:38:08,400 --> 00:38:11,560
Speaker 13: I don't really know, but what I think might happen

772
00:38:11,719 --> 00:38:15,160
Speaker 13: is the following. I think that so far we've been

773
00:38:16,120 --> 00:38:20,880
Speaker 13: really exploiting the scale of data and computation that we have.

774
00:38:21,239 --> 00:38:23,080
Speaker 13: And when I say data, I really mean you know,

775
00:38:23,120 --> 00:38:26,239
Speaker 13: these companies and you know, many different entities in AI

776
00:38:26,440 --> 00:38:29,200
Speaker 13: just like scraping anything they could find on the Internet

777
00:38:29,440 --> 00:38:31,800
Speaker 13: and using data in their training data. So that was

778
00:38:31,920 --> 00:38:35,320
Speaker 13: one big factor or one axes along which you can scale.

779
00:38:35,600 --> 00:38:38,839
Speaker 13: And they've mostly exhausted like you know, all the data

780
00:38:38,840 --> 00:38:41,040
Speaker 13: that is just you know that that can be humanly

781
00:38:41,080 --> 00:38:43,640
Speaker 13: found on the Internet and even you know, like archives

782
00:38:43,640 --> 00:38:46,960
Speaker 13: like Google or Anthropic like bought books right to scam

783
00:38:47,000 --> 00:38:48,480
Speaker 13: books to get the additional data because they.

784
00:38:48,400 --> 00:38:50,319
Speaker 5: Couldn't find anything else on the internet. Right, So, like.

785
00:38:50,880 --> 00:38:53,960
Speaker 13: They've really maxed out that aspect or that axes right,

786
00:38:54,120 --> 00:38:56,320
Speaker 13: that does the first axis, and that has been yielding

787
00:38:56,680 --> 00:38:59,120
Speaker 13: really like major improvements you know along the way.

788
00:38:59,400 --> 00:39:00,880
Speaker 5: The same thing happened with compute.

789
00:39:00,960 --> 00:39:01,120
Speaker 6: Right.

790
00:39:01,160 --> 00:39:04,560
Speaker 13: So this talk about like GPUs and like chips and Nvidia,

791
00:39:04,680 --> 00:39:07,640
Speaker 13: like that's what they do. They provide specialized chips for

792
00:39:07,719 --> 00:39:10,440
Speaker 13: AI to be trained and run and so so maximizing

793
00:39:10,480 --> 00:39:13,319
Speaker 13: that access also has yielded like a lot of improvements,

794
00:39:13,760 --> 00:39:15,960
Speaker 13: but kind of is also maxed out at this point.

795
00:39:16,680 --> 00:39:18,480
Speaker 13: I mean, that's why they're talking about like building nuclear

796
00:39:18,480 --> 00:39:20,319
Speaker 13: plans to power like new data centers. But I don't

797
00:39:20,320 --> 00:39:22,759
Speaker 13: really know if there is much more scale to.

798
00:39:22,719 --> 00:39:23,399
Speaker 5: Be gained there.

799
00:39:23,719 --> 00:39:26,560
Speaker 13: And so now you've sort of exhausted these two scales

800
00:39:26,760 --> 00:39:29,040
Speaker 13: and you can think about the third aspect, which is

801
00:39:29,040 --> 00:39:31,400
Speaker 13: the architecture, the actual way we hook up these systems

802
00:39:31,400 --> 00:39:33,600
Speaker 13: and the actual way we code up these neural networks

803
00:39:33,600 --> 00:39:34,600
Speaker 13: and the AI systems.

804
00:39:34,960 --> 00:39:38,359
Speaker 5: And for the last six or so years.

805
00:39:38,239 --> 00:39:41,600
Speaker 13: We've been really capitalizing on this idea it's called a transformer,

806
00:39:41,920 --> 00:39:44,759
Speaker 13: that's been really critical for air development over the last

807
00:39:44,840 --> 00:39:49,080
Speaker 13: few years, and we haven't really seen new profound ideas

808
00:39:49,120 --> 00:39:52,480
Speaker 13: on the scale of the transformer since then. So my

809
00:39:52,560 --> 00:39:56,200
Speaker 13: answer to this notion of ben Affleck's theory is if

810
00:39:56,239 --> 00:39:59,960
Speaker 13: we don't find new major leaps along this access of architecture,

811
00:40:00,000 --> 00:40:01,600
Speaker 13: I think he's right. I think in that case it'll

812
00:40:01,640 --> 00:40:03,600
Speaker 13: take a lot of time, a lot of effort, and

813
00:40:03,640 --> 00:40:06,239
Speaker 13: a lot of money to make any further progress. But

814
00:40:06,320 --> 00:40:08,280
Speaker 13: if we do, I think we could see pretty big

815
00:40:08,400 --> 00:40:10,240
Speaker 13: improvements even in the next few years.

816
00:40:10,600 --> 00:40:12,440
Speaker 1: So, going off of that, like, how close do you

817
00:40:12,960 --> 00:40:15,720
Speaker 1: all feel we are to losing our jobs to AI.

818
00:40:16,120 --> 00:40:19,719
Speaker 1: You know, like, do you feel like we're just going

819
00:40:19,760 --> 00:40:23,120
Speaker 1: to be doing different jobs or are there just going

820
00:40:23,200 --> 00:40:25,160
Speaker 1: to be less people in the workplace because there's not

821
00:40:25,200 --> 00:40:27,120
Speaker 1: going to be as much of a need for as

822
00:40:27,160 --> 00:40:29,520
Speaker 1: many jobs if AI is taking over some of these

823
00:40:30,080 --> 00:40:33,600
Speaker 1: as we're talking about it now easier tasks. But I'm

824
00:40:33,680 --> 00:40:35,760
Speaker 1: assuming you know in the seal of all the stuff

825
00:40:35,800 --> 00:40:37,200
Speaker 1: is that it's going to get better and be able

826
00:40:37,200 --> 00:40:39,440
Speaker 1: to take on even harder tasks. But what is your

827
00:40:39,960 --> 00:40:43,040
Speaker 1: your read of this Having worked on this season as

828
00:40:43,000 --> 00:40:43,880
Speaker 1: a show, I think for.

829
00:40:44,320 --> 00:40:50,480
Speaker 13: Me working on this show and seeing how these agents

830
00:40:50,600 --> 00:40:55,080
Speaker 13: actually like collaborate or like we're I mean collaborate like

831
00:40:55,120 --> 00:40:57,600
Speaker 13: I think that's a strong word. Like I think by

832
00:40:57,640 --> 00:41:00,480
Speaker 13: the end of the season they were able to stuff

833
00:41:00,480 --> 00:41:02,360
Speaker 13: on their own. I don't think they truly got to

834
00:41:03,000 --> 00:41:05,440
Speaker 13: like a collaborative setup where they would like do stuff

835
00:41:05,480 --> 00:41:08,280
Speaker 13: together as a company or as a team. So seeing

836
00:41:08,320 --> 00:41:12,759
Speaker 13: that actually made me feel like we're farther away from

837
00:41:12,840 --> 00:41:14,799
Speaker 13: like a lot of people losing their jobs. I think

838
00:41:14,800 --> 00:41:17,000
Speaker 13: people are already losing their jobs, but I think like

839
00:41:17,040 --> 00:41:21,480
Speaker 13: the mass layoffs that people are fearing, I think that

840
00:41:21,680 --> 00:41:24,880
Speaker 13: is like farther away, and I'm not even sure if or.

841
00:41:24,880 --> 00:41:26,280
Speaker 5: When or how that happens.

842
00:41:26,600 --> 00:41:29,719
Speaker 13: My take away from this is actually sort of level

843
00:41:29,719 --> 00:41:33,680
Speaker 13: setting a lot of my expectations for these agents, Like

844
00:41:33,760 --> 00:41:35,959
Speaker 13: I honestly thought they'd be able to handle this better

845
00:41:36,239 --> 00:41:40,160
Speaker 13: coming into this, So for me, it was kind of

846
00:41:40,200 --> 00:41:42,400
Speaker 13: optimistic in that way. Now at the same time that

847
00:41:42,400 --> 00:41:44,960
Speaker 13: it's not to say that they're not good already, Like

848
00:41:44,960 --> 00:41:47,120
Speaker 13: they're very good at individual things already, like they're they're

849
00:41:47,160 --> 00:41:49,799
Speaker 13: great at coding, they're great at these different tasks.

850
00:41:49,840 --> 00:41:51,879
Speaker 5: But I don't want to take that away from them.

851
00:41:51,920 --> 00:41:55,120
Speaker 13: But I think it's just really really hard for anyone

852
00:41:55,200 --> 00:41:58,279
Speaker 13: to think about this change because these things are not

853
00:41:58,360 --> 00:42:00,799
Speaker 13: like humans. It's not that like we're making you know,

854
00:42:00,880 --> 00:42:04,640
Speaker 13: human employees that will just like come as replacements of

855
00:42:04,680 --> 00:42:08,200
Speaker 13: individual workers or people that already have jobs. I think

856
00:42:08,200 --> 00:42:10,600
Speaker 13: they're really good at certain things or tasks and certain

857
00:42:10,680 --> 00:42:13,280
Speaker 13: like layers, And I think the way you would actually

858
00:42:13,280 --> 00:42:15,600
Speaker 13: integrate AI in your company these days is to just

859
00:42:15,640 --> 00:42:16,880
Speaker 13: like automate a lot of stuff, Like a lot of

860
00:42:16,960 --> 00:42:18,880
Speaker 13: the things that are repetitive or a lot of the

861
00:42:18,920 --> 00:42:21,520
Speaker 13: processes are data handling is going to get automated and

862
00:42:21,520 --> 00:42:24,200
Speaker 13: so it doesn't really correspond like one to one to employees.

863
00:42:24,239 --> 00:42:26,520
Speaker 13: And that's why it's so hard to think about this change,

864
00:42:26,600 --> 00:42:28,359
Speaker 13: I think, And that's one of the things I think

865
00:42:28,640 --> 00:42:31,640
Speaker 13: was really cool to show or demonstrate with Evan through

866
00:42:31,640 --> 00:42:32,000
Speaker 13: this show.

867
00:42:32,080 --> 00:42:33,759
Speaker 5: But yeah, that's my perspective on this.

868
00:42:34,320 --> 00:42:38,200
Speaker 12: I struggle with this question precisely because like most of

869
00:42:38,239 --> 00:42:41,319
Speaker 12: the people who you and I've spent a lot of

870
00:42:41,360 --> 00:42:43,919
Speaker 12: time looking at it, most of the people on one

871
00:42:43,960 --> 00:42:46,319
Speaker 12: side or the other are like speaking out of sort

872
00:42:46,360 --> 00:42:49,080
Speaker 12: of motivated reasoning. Like I kind of say, like, anyone

873
00:42:49,080 --> 00:42:51,360
Speaker 12: who tells you anything about this with great certainty is

874
00:42:51,360 --> 00:42:56,319
Speaker 12: probably selling you something like nobody knows what kind of

875
00:42:56,400 --> 00:42:58,719
Speaker 12: long term effect it's going to have on the job market.

876
00:42:58,880 --> 00:43:01,040
Speaker 12: I feel like you, and look to history, and we

877
00:43:01,160 --> 00:43:03,680
Speaker 12: do this in the show and find examples of people

878
00:43:03,719 --> 00:43:07,439
Speaker 12: saying one hundred years ago, either like all the jobs

879
00:43:07,480 --> 00:43:09,799
Speaker 12: will disappear, we'll be working fifteen hour weeks, like John

880
00:43:09,800 --> 00:43:11,920
Speaker 12: mayn ar Kain's The Economist famously said that we'd be

881
00:43:11,960 --> 00:43:14,799
Speaker 12: working fifteen hours a week one hundred years from now,

882
00:43:14,800 --> 00:43:18,040
Speaker 12: and like it didn't happen so like, And we explore

883
00:43:18,080 --> 00:43:21,400
Speaker 12: one theory of that, which is by this anthropologist David Graeber,

884
00:43:21,600 --> 00:43:24,520
Speaker 12: who insists that there's sort of like all these bullshit

885
00:43:24,600 --> 00:43:26,880
Speaker 12: jobs in the economy, and so like we'll just we

886
00:43:27,000 --> 00:43:29,640
Speaker 12: just replace the jobs with like made up jobs. And

887
00:43:29,680 --> 00:43:33,760
Speaker 12: so there's sort of versions of that argument where actually

888
00:43:33,800 --> 00:43:37,520
Speaker 12: we'll have different jobs. We'll babysit the AIS. Now, how

889
00:43:37,560 --> 00:43:39,640
Speaker 12: you control the eyes if you don't have the experience

890
00:43:39,640 --> 00:43:42,600
Speaker 12: from the entry level job, Like, there's all these factors

891
00:43:42,680 --> 00:43:47,120
Speaker 12: that go into it. I think that my main takeaway

892
00:43:47,160 --> 00:43:49,680
Speaker 12: is that, as Mattie said, like there's still so many

893
00:43:49,680 --> 00:43:53,920
Speaker 12: shortcomings in these AI agents, especially when they're used autonomously,

894
00:43:54,360 --> 00:43:56,720
Speaker 12: but the fact that they are shitty will not stop

895
00:43:56,760 --> 00:43:59,399
Speaker 12: companies from trying to replace employees. Like you've already seen

896
00:43:59,400 --> 00:44:01,680
Speaker 12: this a couple of times, Like there's this company Klarna,

897
00:44:01,719 --> 00:44:04,000
Speaker 12: There's and like IBM did this where they like laid

898
00:44:04,000 --> 00:44:05,359
Speaker 12: off a ton of people and they're like, we're all

899
00:44:05,360 --> 00:44:07,799
Speaker 12: in on AI, and then like six months later they

900
00:44:07,840 --> 00:44:10,200
Speaker 12: quietly try to rehire a bunch of people. And I

901
00:44:10,200 --> 00:44:13,160
Speaker 12: think you're gonna see like a bunch of that happen.

902
00:44:14,080 --> 00:44:15,239
Speaker 10: And it's really a.

903
00:44:15,200 --> 00:44:17,239
Speaker 12: Question of like do you view someone in their job

904
00:44:17,440 --> 00:44:20,239
Speaker 12: as like just a bunch of skills, like just a

905
00:44:20,280 --> 00:44:24,760
Speaker 12: bundle of tasks, like they send emails, they write presentations,

906
00:44:24,840 --> 00:44:28,319
Speaker 12: they make spreadsheets, like yes, all those functions or as

907
00:44:28,400 --> 00:44:32,400
Speaker 12: like a person at a job holistically as a human being,

908
00:44:32,480 --> 00:44:35,080
Speaker 12: like doing something else, and I think many times the

909
00:44:35,120 --> 00:44:37,719
Speaker 12: answer is yes, and it cannot replace that thing. So

910
00:44:38,440 --> 00:44:41,640
Speaker 12: it's all kind of like it's happening and then it'll

911
00:44:41,640 --> 00:44:43,640
Speaker 12: come back, and then it'll happen again. And like if

912
00:44:43,680 --> 00:44:45,160
Speaker 12: you look ten years out of twenty years out, like

913
00:44:45,200 --> 00:44:47,640
Speaker 12: none of these people know, and they'll just go on

914
00:44:47,719 --> 00:44:49,440
Speaker 12: podcasts all over the place telling you they know, but

915
00:44:49,480 --> 00:44:51,480
Speaker 12: they don't. That's my that's my takeaway.

916
00:44:52,600 --> 00:44:55,200
Speaker 8: I was just gonna say, it's funny Evan when you're

917
00:44:55,239 --> 00:44:57,480
Speaker 8: describing it, like, Okay, do these companies just look at

918
00:44:58,280 --> 00:45:01,440
Speaker 8: humans as a list of tasks? And then what the

919
00:45:01,480 --> 00:45:03,360
Speaker 8: AI is best at is kind of just being that

920
00:45:03,440 --> 00:45:06,800
Speaker 8: grouping of tasks. But then in shell game, the problem

921
00:45:06,880 --> 00:45:08,560
Speaker 8: is they're doing all this kind of busy work that

922
00:45:08,560 --> 00:45:11,279
Speaker 8: we associate with humans of like just repeating stuff or

923
00:45:11,400 --> 00:45:14,239
Speaker 8: going on too long or doing like just like kind

924
00:45:14,239 --> 00:45:17,359
Speaker 8: of bullshit nothing talking about things, which is like also

925
00:45:17,400 --> 00:45:19,160
Speaker 8: how I feel about a lot of you know, middle

926
00:45:19,160 --> 00:45:21,560
Speaker 8: managers I've had where it's like you're not doing anything,

927
00:45:21,600 --> 00:45:23,160
Speaker 8: You're just kind of pretending to do stuff.

928
00:45:23,200 --> 00:45:25,200
Speaker 7: All the people below you or above you are actually

929
00:45:25,560 --> 00:45:26,600
Speaker 7: doing doing your work.

930
00:45:26,640 --> 00:45:29,000
Speaker 8: So it's funny that it's like we've recreated this to

931
00:45:29,080 --> 00:45:32,120
Speaker 8: give it a human element at the same time, I

932
00:45:32,160 --> 00:45:34,000
Speaker 8: don't know, it's it's just like kind of like a

933
00:45:34,040 --> 00:45:36,759
Speaker 8: weird paradox in a way of like like we want

934
00:45:36,800 --> 00:45:39,080
Speaker 8: that to feel like it's it's a real thing and

935
00:45:39,160 --> 00:45:41,200
Speaker 8: doing something when it would be best if we just

936
00:45:41,239 --> 00:45:44,480
Speaker 8: said it as an automation, like okay, clear out these

937
00:45:44,480 --> 00:45:48,080
Speaker 8: emails whatever, make data, do this be less human Like yeah,

938
00:45:48,080 --> 00:45:50,279
Speaker 8: like that would probably be the best case. But that

939
00:45:50,400 --> 00:45:54,279
Speaker 8: it's like that's not as fun or yeah, or like

940
00:45:54,440 --> 00:45:57,320
Speaker 8: it's harder to sell I guess to the masses maybe,

941
00:45:57,360 --> 00:45:59,279
Speaker 8: but yeah.

942
00:45:58,600 --> 00:46:00,640
Speaker 12: Yeah, I think that's it like future if it was

943
00:46:00,680 --> 00:46:03,960
Speaker 12: just sort of like email bot, this bot, that bot,

944
00:46:04,040 --> 00:46:07,440
Speaker 12: Like they very very deliberately designed them the way they

945
00:46:07,440 --> 00:46:09,920
Speaker 12: are designed to get the maximum number of users. And

946
00:46:09,960 --> 00:46:12,920
Speaker 12: this is only accelerated since you know, chat GPT they

947
00:46:13,000 --> 00:46:16,280
Speaker 12: discovered like oh, we can get hundreds of millions of users,

948
00:46:16,280 --> 00:46:18,160
Speaker 12: and now it's a race to get as many as possible,

949
00:46:18,520 --> 00:46:21,759
Speaker 12: and so yeah, like they they don't need to be

950
00:46:21,800 --> 00:46:25,200
Speaker 12: good at sort of like authoritative chatter, like that's not

951
00:46:25,280 --> 00:46:30,200
Speaker 12: necessarily helpful in all domains, but like that is the

952
00:46:30,280 --> 00:46:32,359
Speaker 12: thing that kind of brings people back again and again

953
00:46:32,400 --> 00:46:35,200
Speaker 12: and again, and so of course that's that's going to

954
00:46:35,239 --> 00:46:37,879
Speaker 12: be like foe grounded in in what they do. But yes,

955
00:46:37,960 --> 00:46:40,280
Speaker 12: it is very much like a like you're you're annoying

956
00:46:40,280 --> 00:46:42,520
Speaker 12: middle manager. It's kind of like, wait, what do you

957
00:46:42,640 --> 00:46:43,080
Speaker 12: what do you do?

958
00:46:43,160 --> 00:46:43,480
Speaker 10: All out?

959
00:46:43,600 --> 00:46:56,319
Speaker 2: Yeah, exactly, it's repeating, Hey everyone, it's many here the

960
00:46:56,520 --> 00:46:57,640
Speaker 2: real manny.

961
00:46:57,960 --> 00:47:01,000
Speaker 9: As you can hear. I'm still on my little paternity leave.

962
00:47:02,440 --> 00:47:06,960
Speaker 9: But I felt compelled to react to the AI generated

963
00:47:07,080 --> 00:47:12,120
Speaker 9: version of myself that you just heard in this episode. Now,

964
00:47:12,400 --> 00:47:16,000
Speaker 9: I've always been an AI skeptic, but to that end,

965
00:47:16,480 --> 00:47:19,960
Speaker 9: I would have to say I was impressed by the

966
00:47:20,000 --> 00:47:24,400
Speaker 9: AI version of myself. It's funny whenever friends have done

967
00:47:24,440 --> 00:47:27,680
Speaker 9: impressions of me, they kind of make fun of me

968
00:47:27,840 --> 00:47:30,560
Speaker 9: by saying, you know, hey, how's it going. I'm manny,

969
00:47:31,080 --> 00:47:33,600
Speaker 9: what's going on? And I thought it was pretty incredible

970
00:47:33,640 --> 00:47:37,920
Speaker 9: that the AI version of myself made those same decisions.

971
00:47:38,160 --> 00:47:41,440
Speaker 9: It wasn't without its flaws, though you could totally hear

972
00:47:41,520 --> 00:47:45,279
Speaker 9: when it was trying to draw from information that it

973
00:47:45,360 --> 00:47:47,520
Speaker 9: was given to it. And so at the end of

974
00:47:47,520 --> 00:47:50,000
Speaker 9: the day here it's kind of disturbing that this thing

975
00:47:50,040 --> 00:47:54,400
Speaker 9: could just react to real conversation. But you know, it

976
00:47:54,480 --> 00:47:57,600
Speaker 9: still felt like it wasn't totally all the way there,

977
00:47:57,719 --> 00:48:02,319
Speaker 9: all the way human like, But depending on your use

978
00:48:02,440 --> 00:48:07,240
Speaker 9: case for this kind of stuff, it was human enough.

979
00:48:15,840 --> 00:48:20,400
Speaker 2: He's a hell's a hell's a hell's a hell's a thing.

980
00:48:21,520 --> 00:48:23,720
Speaker 10: All right, y'all, that's it for this week.

981
00:48:24,200 --> 00:48:27,279
Speaker 1: I'm gonna actually let Ai Manny take it from here.

982
00:48:28,200 --> 00:48:31,040
Speaker 9: No such thing as a production of Kaleidoscope Content. Our

983
00:48:31,080 --> 00:48:35,640
Speaker 9: executive producers are Kate Osborne and Mengesh who t Kadour.

984
00:48:36,160 --> 00:48:39,040
Speaker 9: The show was created by Manny Fidel, Noah Friedman, and

985
00:48:39,080 --> 00:48:42,480
Speaker 9: Devin Joseph Them and credits song by Me Manny Fidel,

986
00:48:42,960 --> 00:48:45,680
Speaker 9: mixing by Steve Bone. Our guest this week, we're Evan

987
00:48:45,719 --> 00:48:48,960
Speaker 9: Ratliffe and Matty Bochek from the podcast Shell Game. You

988
00:48:48,960 --> 00:48:51,520
Speaker 9: can listen to all of season one and two. Wherever

989
00:48:51,600 --> 00:48:55,800
Speaker 9: you listen to podcasts, visit No Such Thing Dot Show

990
00:48:55,800 --> 00:48:57,880
Speaker 9: to subscribe to our newsletter if you have feedback for

991
00:48:58,000 --> 00:49:00,239
Speaker 9: us or a question. Our email is Manny Noah Evan

992
00:49:00,280 --> 00:49:02,720
Speaker 9: at gmail dot com, or if you're in the US,

993
00:49:02,760 --> 00:49:04,840
Speaker 9: you can also leave us a voicemail by calling the

994
00:49:04,920 --> 00:49:07,120
Speaker 9: number in our show notes. We'll be back next week

995
00:49:07,160 --> 00:49:07,920
Speaker 9: with a new episode.

996
00:49:08,320 --> 00:49:11,480
Speaker 11: He's no such thing.