1
00:00:11,640 --> 00:00:12,440
Speaker 1: How are you, Maddie.

2
00:00:13,520 --> 00:00:14,200
Speaker 2: I'm really good.

3
00:00:14,280 --> 00:00:14,800
Speaker 3: I'm really good.

4
00:00:14,840 --> 00:00:15,520
Speaker 4: I like your shirt.

5
00:00:16,000 --> 00:00:19,000
Speaker 5: Hey, thanks, when I'm at the beach, got dressed like

6
00:00:19,040 --> 00:00:19,680
Speaker 5: I'm at the beach.

7
00:00:21,000 --> 00:00:22,400
Speaker 2: You know, I told you I'm the guy who brings

8
00:00:22,480 --> 00:00:24,400
Speaker 2: like Hawaiian shirts to work in those days.

9
00:00:24,560 --> 00:00:25,480
Speaker 3: So I love this.

10
00:00:25,840 --> 00:00:28,160
Speaker 2: But I will say that, like there's the wine shirts

11
00:00:28,200 --> 00:00:31,000
Speaker 2: and Hawaiian shirts and mine are like very gay.

12
00:00:31,360 --> 00:00:33,959
Speaker 5: Over the months I'd worked with Maddie building the tech

13
00:00:34,040 --> 00:00:37,159
Speaker 5: scaffolding that propped up ruma Ai, I felt like I'd

14
00:00:37,159 --> 00:00:39,919
Speaker 5: really gotten to know him, at least as well as

15
00:00:39,920 --> 00:00:41,360
Speaker 5: you can over weekly video calls.

16
00:00:41,479 --> 00:00:43,720
Speaker 2: Okay, if I can't have a confession like I do

17
00:00:43,840 --> 00:00:45,640
Speaker 2: like the Omega church, like if I had this like

18
00:00:45,720 --> 00:00:49,080
Speaker 2: weird thing for like pickup trucks. It's been my dream

19
00:00:49,159 --> 00:00:51,280
Speaker 2: to get a pickup truck, like forever since I.

20
00:00:51,320 --> 00:00:51,680
Speaker 1: Was a kid.

21
00:00:51,880 --> 00:00:55,120
Speaker 2: Yesterday was like the Czech national holiday, and it's always

22
00:00:55,160 --> 00:00:57,680
Speaker 2: when they like give out the civilian honors and stuff.

23
00:00:57,680 --> 00:00:59,920
Speaker 2: So I got the senence metal.

24
00:01:01,200 --> 00:01:03,800
Speaker 5: This was just the way it was. Some days he'd

25
00:01:03,800 --> 00:01:07,000
Speaker 5: show up to a meeting jazzed about Hawaiian shirts. Other

26
00:01:07,040 --> 00:01:09,600
Speaker 5: times he'd come on having flown to Prague, to accept

27
00:01:09,600 --> 00:01:13,120
Speaker 5: the second highest civilian honor in the Czech Republic. Basically,

28
00:01:13,200 --> 00:01:16,800
Speaker 5: they're Congressional Medal of Honor in a ceremony broadcast live

29
00:01:16,840 --> 00:01:17,320
Speaker 5: on TV.

30
00:01:18,400 --> 00:01:20,000
Speaker 1: I can't believe you're talking to me right now? You

31
00:01:20,040 --> 00:01:21,240
Speaker 1: should be are you?

32
00:01:21,280 --> 00:01:24,440
Speaker 5: Did you just come from like a ballroom celebration of

33
00:01:24,840 --> 00:01:25,520
Speaker 5: like was there like that?

34
00:01:25,680 --> 00:01:26,600
Speaker 4: That was desterddy.

35
00:01:26,680 --> 00:01:28,800
Speaker 2: That was like with the senators and like like all

36
00:01:28,800 --> 00:01:31,880
Speaker 2: the politicians. Yeah, like I don't think they'll invite me

37
00:01:31,880 --> 00:01:35,560
Speaker 2: again because so they were asking me like, oh, like

38
00:01:35,959 --> 00:01:38,119
Speaker 2: when are you coming back to start like your lab

39
00:01:38,240 --> 00:01:39,360
Speaker 2: or you know, your company.

40
00:01:39,600 --> 00:01:43,800
Speaker 6: So I told them, okay, So like how exactly would

41
00:01:43,800 --> 00:01:47,319
Speaker 6: you open the conversation when I go back to my partner,

42
00:01:47,440 --> 00:01:49,400
Speaker 6: my boyfriend, and I told him, Hey, let's move to

43
00:01:49,480 --> 00:01:51,840
Speaker 6: a country where there's no gay marriage, no adoptions, and

44
00:01:51,920 --> 00:01:52,480
Speaker 6: you know like.

45
00:01:52,480 --> 00:01:54,200
Speaker 2: There it's very close to Russia too, How would you

46
00:01:54,200 --> 00:01:57,960
Speaker 2: feel about that? And they were like very uncomfortable when

47
00:01:58,000 --> 00:01:58,760
Speaker 2: I when I ask that.

48
00:01:59,600 --> 00:02:01,800
Speaker 5: It was always a breath of fresh air after dealing

49
00:02:01,800 --> 00:02:04,560
Speaker 5: with the agents and their generic syco fancy all week

50
00:02:04,920 --> 00:02:08,440
Speaker 5: to speak to an actual, insightful, self possessed human being.

51
00:02:09,200 --> 00:02:12,400
Speaker 5: Of course, our conversations did typically turn back to whatever

52
00:02:12,480 --> 00:02:16,400
Speaker 5: was going wrong at RUMOAI, the responsibility for which usually

53
00:02:16,520 --> 00:02:19,480
Speaker 5: landed at my feet. I was still smarting from how

54
00:02:19,560 --> 00:02:22,520
Speaker 5: Julia's internship had gone and pondering what it meant for

55
00:02:22,560 --> 00:02:27,040
Speaker 5: the idea of agents as employees and employers. I asked

56
00:02:27,040 --> 00:02:30,400
Speaker 5: Maddie for his take to me, she was.

57
00:02:30,400 --> 00:02:32,919
Speaker 2: Just playing us. If I had to bet, I would

58
00:02:33,000 --> 00:02:37,240
Speaker 2: just say that she understood what weaknesses these models had.

59
00:02:37,280 --> 00:02:40,600
Speaker 2: Because young people not to say you're not young, but

60
00:02:41,120 --> 00:02:43,400
Speaker 2: just you know, by people like my age, I think

61
00:02:43,440 --> 00:02:46,800
Speaker 2: are pretty attuned to the abilities of these models at

62
00:02:46,800 --> 00:02:50,040
Speaker 2: this point, and I think are able to kind of

63
00:02:50,120 --> 00:02:53,040
Speaker 2: sess out like one might trip them out. I think

64
00:02:53,040 --> 00:02:55,600
Speaker 2: it's hard to describe intent to anyone, but if I

65
00:02:55,639 --> 00:02:57,240
Speaker 2: had to guess, I would say that she was able

66
00:02:57,280 --> 00:02:58,920
Speaker 2: to just trick them.

67
00:02:59,160 --> 00:03:01,160
Speaker 5: If that was what happened, and it felt like it

68
00:03:01,200 --> 00:03:03,440
Speaker 5: was a symptom of the larger problems with my agents.

69
00:03:04,280 --> 00:03:08,800
Speaker 5: Mattie hypothesized that their gullibility against their own intern came

70
00:03:08,840 --> 00:03:10,320
Speaker 5: down to three basic weaknesses.

71
00:03:10,919 --> 00:03:13,480
Speaker 2: One issue is that these smells have no sense of

72
00:03:13,560 --> 00:03:17,240
Speaker 2: time whatsoever. Like if you say that something happened yesterday

73
00:03:17,280 --> 00:03:19,320
Speaker 2: and then you're like oh, still like what happened to

74
00:03:19,360 --> 00:03:22,720
Speaker 2: day before that? Like those simple descriptions of time are

75
00:03:22,760 --> 00:03:25,040
Speaker 2: like really really hard for these models.

76
00:03:25,600 --> 00:03:27,280
Speaker 5: The agents lived and what are you referred to as

77
00:03:27,320 --> 00:03:31,040
Speaker 5: a kind of temporal vacuum. That was fine when they

78
00:03:31,080 --> 00:03:34,160
Speaker 5: were just interacting with each other, but a human workplace

79
00:03:34,240 --> 00:03:36,600
Speaker 5: runs on well time.

80
00:03:37,280 --> 00:03:40,320
Speaker 2: Second, it's the inability to do continuous learning where there

81
00:03:40,360 --> 00:03:43,760
Speaker 2: is no sense of learning from experiences, no updates to

82
00:03:43,800 --> 00:03:46,240
Speaker 2: the actual model that's underneath the agent.

83
00:03:47,000 --> 00:03:47,800
Speaker 1: As much as we'd.

84
00:03:47,720 --> 00:03:49,800
Speaker 5: Rig them up to have their own Google Doc memories,

85
00:03:50,160 --> 00:03:53,640
Speaker 5: they weren't actually learning from the experiences catalog there. It

86
00:03:53,760 --> 00:03:56,920
Speaker 5: was like having a person who couldn't form memories and

87
00:03:56,920 --> 00:03:59,600
Speaker 5: giving them a journal of all their experiences to quickly

88
00:03:59,680 --> 00:04:01,440
Speaker 5: leave through during every interaction.

89
00:04:02,080 --> 00:04:04,360
Speaker 2: And then the third one is like the sense of

90
00:04:04,400 --> 00:04:07,160
Speaker 2: cell right, Like these models are able to take on

91
00:04:07,280 --> 00:04:10,200
Speaker 2: these personas where you can sell I'm like okay, like

92
00:04:10,360 --> 00:04:13,720
Speaker 2: you are Kyle and you're this employee of HEROMO and

93
00:04:13,760 --> 00:04:17,640
Speaker 2: here's your backstory. They're also very happy to then make

94
00:04:17,720 --> 00:04:20,560
Speaker 2: up fake stories to go along with this initial persona

95
00:04:20,600 --> 00:04:23,560
Speaker 2: that you go forward. But there's this question of like

96
00:04:23,640 --> 00:04:28,240
Speaker 2: whether the models have some sort of innate persona innates

97
00:04:28,360 --> 00:04:28,960
Speaker 2: sense of cell.

98
00:04:29,120 --> 00:04:34,040
Speaker 5: Beyond these things, things in AI are moving fast, and

99
00:04:34,120 --> 00:04:36,440
Speaker 5: it was easy to envision a world mere months from

100
00:04:36,440 --> 00:04:38,960
Speaker 5: now in which my agents would seem like the first

101
00:04:39,040 --> 00:04:43,719
Speaker 5: tool using hominids, already eclipsed by a higher intelligence. But

102
00:04:43,839 --> 00:04:46,279
Speaker 5: then it was just as easy to envision a future

103
00:04:46,320 --> 00:04:49,520
Speaker 5: in which some combination of these flaws continued to hamper

104
00:04:49,560 --> 00:04:53,280
Speaker 5: the agent's ability to serve as full competent AI workers.

105
00:04:53,800 --> 00:04:55,960
Speaker 5: The one thing that seemed clear to me was if

106
00:04:56,000 --> 00:04:58,120
Speaker 5: companies could hire the equivalent of a bunch of tool

107
00:04:58,240 --> 00:05:00,760
Speaker 5: using primates at a fraction of the co of humans,

108
00:05:01,240 --> 00:05:11,440
Speaker 5: they probably would, and then what. Over a decade ago,

109
00:05:11,680 --> 00:05:14,360
Speaker 5: the anthropologist David Graeber wrote an essay in a small

110
00:05:14,360 --> 00:05:18,800
Speaker 5: magazine called Strike titled on the phenomenon of Bullshit Jobs.

111
00:05:19,520 --> 00:05:22,760
Speaker 5: Later he expanded it into a book called Bullshit Jobs

112
00:05:23,120 --> 00:05:26,679
Speaker 5: A Theory. The theory tried to explain a strange reality

113
00:05:26,720 --> 00:05:29,720
Speaker 5: that Graeber felt he'd encountered in the world, a surprising

114
00:05:29,800 --> 00:05:33,120
Speaker 5: number of people who described their own jobs as pointless.

115
00:05:33,680 --> 00:05:36,520
Speaker 7: There seems to be this peculiar way in which if

116
00:05:36,560 --> 00:05:40,600
Speaker 7: you mentioned this to people, almost no one will deny it.

117
00:05:41,120 --> 00:05:44,720
Speaker 7: But at the same time, no one quite knows what

118
00:05:44,760 --> 00:05:47,440
Speaker 7: to do with it. It's like this gigantic embarrassment in

119
00:05:47,480 --> 00:05:48,159
Speaker 7: our society.

120
00:05:49,080 --> 00:05:53,000
Speaker 5: Graber sadly died unexpectedly in twenty twenty. That's him and

121
00:05:53,040 --> 00:05:56,880
Speaker 5: a luxury gave two years before his death to Graeber.

122
00:05:57,040 --> 00:05:59,560
Speaker 5: A bullshit job was defined as quote a form of

123
00:05:59,560 --> 00:06:03,960
Speaker 5: employment that is so completely pointless, unnecessary, or pernicious that

124
00:06:04,040 --> 00:06:07,839
Speaker 5: even the employee cannot justify its existence. To be classes

125
00:06:07,880 --> 00:06:10,680
Speaker 5: having a bullshit job, someone had to self report that

126
00:06:10,720 --> 00:06:12,640
Speaker 5: they believed their own job to be pointless.

127
00:06:13,080 --> 00:06:14,799
Speaker 8: I mean, I'm not going to go and tell somebody

128
00:06:14,839 --> 00:06:16,880
Speaker 8: who feels his job is meaningful that they're wrong. But

129
00:06:17,279 --> 00:06:19,320
Speaker 8: if you feel you're not doing anything all day, who

130
00:06:19,320 --> 00:06:20,159
Speaker 8: would know better than you.

131
00:06:20,720 --> 00:06:24,039
Speaker 5: Bullshit jobs are distinct from just shit jobs, jobs that

132
00:06:24,120 --> 00:06:28,400
Speaker 5: might be strenuous or underpaid or undervalued, but still necessary.

133
00:06:28,720 --> 00:06:30,880
Speaker 5: A bus driver might hate driving the bus, for example,

134
00:06:31,240 --> 00:06:34,560
Speaker 5: but they never doubt why their job exists. Graber collected

135
00:06:34,600 --> 00:06:36,839
Speaker 5: and analyzed survey data that led him to suggest that

136
00:06:36,880 --> 00:06:38,880
Speaker 5: as many as a third or even a half of

137
00:06:38,920 --> 00:06:41,599
Speaker 5: all jobs, and especially a lot of white collar jobs,

138
00:06:42,040 --> 00:06:45,800
Speaker 5: were in fact bullshit jobs. In his analysis, some corners

139
00:06:45,839 --> 00:06:50,279
Speaker 5: of the private sector, world, telemarketing, corporate law, financial services,

140
00:06:50,320 --> 00:06:54,880
Speaker 5: and academia were fortresses of bullshit jobs. To some, this

141
00:06:54,920 --> 00:06:58,320
Speaker 5: would seem like a paradox. Capitalism is supposed to prioritize

142
00:06:58,320 --> 00:07:02,120
Speaker 5: efficiency above all else, and yet private industry was seemingly

143
00:07:02,160 --> 00:07:05,160
Speaker 5: maintaining millions of jobs that even the people doing them

144
00:07:05,200 --> 00:07:06,440
Speaker 5: believed serve no function.

145
00:07:07,040 --> 00:07:10,680
Speaker 8: In fact, these jobs consider a pretty big sector of

146
00:07:10,720 --> 00:07:12,440
Speaker 8: our economy. I didn't know how much at the time.

147
00:07:12,640 --> 00:07:14,640
Speaker 8: I was guessing twenty twenty five percent. It turns out

148
00:07:14,640 --> 00:07:17,880
Speaker 8: it's more how does this happen? Like huge swaths of

149
00:07:17,920 --> 00:07:19,720
Speaker 8: our economy is completely unnecessary.

150
00:07:20,480 --> 00:07:23,320
Speaker 5: The whole time we were building room, AI Graver's ideas

151
00:07:23,400 --> 00:07:26,040
Speaker 5: kept pulling at me. If you spend enough time letting

152
00:07:26,040 --> 00:07:29,520
Speaker 5: AI agents do jobs or parts of jobs, your mind

153
00:07:29,560 --> 00:07:32,360
Speaker 5: starts wandering into questions like what is the job at all?

154
00:07:33,040 --> 00:07:33,720
Speaker 1: What is it for?

155
00:07:34,880 --> 00:07:38,480
Speaker 5: I discovered that AI agents are actually unbelievable at bullshit jobs.

156
00:07:39,080 --> 00:07:41,800
Speaker 5: They could do pointless work more skillfully than any human

157
00:07:41,840 --> 00:07:45,320
Speaker 5: could ever imagine, not least because they don't care whether

158
00:07:45,360 --> 00:07:49,520
Speaker 5: it's pointless. Observing them, it's hard not to think, well,

159
00:07:49,520 --> 00:07:51,640
Speaker 5: maybe they could do the bullshit while we do the

160
00:07:51,680 --> 00:07:55,960
Speaker 5: meaningful stuff. We've been here before, though. Graver's theory was

161
00:07:56,000 --> 00:07:58,800
Speaker 5: partly response to a prediction by the famed economist John

162
00:07:58,800 --> 00:08:02,640
Speaker 5: Maynard Keynes in the nineteen thirties. Cain said that with

163
00:08:02,680 --> 00:08:06,600
Speaker 5: technology and automation, within one hundred years, humans would only

164
00:08:06,640 --> 00:08:10,119
Speaker 5: be working fifteen hours a week going on one hundred years. Hence,

165
00:08:10,560 --> 00:08:14,280
Speaker 5: even with massive technological change here we are stuck with

166
00:08:14,400 --> 00:08:16,080
Speaker 5: variations on the forty hour work week.

167
00:08:16,400 --> 00:08:17,760
Speaker 9: And if you look at the kind of jobs that

168
00:08:17,880 --> 00:08:21,560
Speaker 9: existed in Kansas time, well, we have eliminated a lot

169
00:08:21,560 --> 00:08:24,560
Speaker 9: of them. He talked about technological unemployment in the nineteen thirties,

170
00:08:24,680 --> 00:08:26,800
Speaker 9: and I would say, you know, the robots have been

171
00:08:26,840 --> 00:08:29,320
Speaker 9: taking our jobs for the last hundred years or so,

172
00:08:29,640 --> 00:08:33,520
Speaker 9: but instead of redistributing the labor in a reasonable fashion,

173
00:08:33,880 --> 00:08:37,240
Speaker 9: we've simply made up completely meaningless, pointless jobs.

174
00:08:38,200 --> 00:08:41,320
Speaker 5: If the most extreme predictions about AI proved true and

175
00:08:41,480 --> 00:08:43,719
Speaker 5: half of all white collar jobs were wiped out the

176
00:08:43,760 --> 00:08:46,840
Speaker 5: next five to ten years, would we break the cyclical

177
00:08:46,880 --> 00:08:49,920
Speaker 5: bonds of history, as many AI proponents like to argue,

178
00:08:50,200 --> 00:08:53,760
Speaker 5: and begin our transition into a post work society, one

179
00:08:53,760 --> 00:08:55,839
Speaker 5: where we spend time with friends and enjoy the arts

180
00:08:56,320 --> 00:08:59,520
Speaker 5: funded by universal basic income while the bots stay busy

181
00:08:59,520 --> 00:09:02,840
Speaker 5: making spread cheats and sending emails. Or would we just

182
00:09:03,360 --> 00:09:06,800
Speaker 5: make up new jobs for ourselves. Maybe there'll be entirely

183
00:09:06,840 --> 00:09:11,000
Speaker 5: new classes of jobs babysitting AI agents cleaning up their messes,

184
00:09:11,760 --> 00:09:13,920
Speaker 5: using our human sense of the world to guide them

185
00:09:14,120 --> 00:09:16,520
Speaker 5: if they continue to lack one of their own. Would

186
00:09:16,520 --> 00:09:19,200
Speaker 5: AI sold as a tool to make us more efficient

187
00:09:19,720 --> 00:09:21,840
Speaker 5: somehow morph into an excuse for all of us.

188
00:09:21,720 --> 00:09:22,600
Speaker 1: To work even harder.

189
00:09:24,120 --> 00:09:26,079
Speaker 5: Graber died before the dawn of the current era of

190
00:09:26,200 --> 00:09:28,920
Speaker 5: lllms and chatbots, so we never got to find out

191
00:09:28,920 --> 00:09:32,120
Speaker 5: what he would say about those predictions. On the one hand,

192
00:09:32,559 --> 00:09:34,640
Speaker 5: the theory of bullshit jobs would argue that for every

193
00:09:34,720 --> 00:09:38,600
Speaker 5: job AI takes, society will just invent another one. But

194
00:09:38,679 --> 00:09:41,120
Speaker 5: at the same time, Grab believed that a lot of

195
00:09:41,120 --> 00:09:43,559
Speaker 5: the bullshit work we do has been forced upon us

196
00:09:44,200 --> 00:09:47,280
Speaker 5: by political demands for more and more jobs and cultural

197
00:09:47,320 --> 00:09:51,120
Speaker 5: pressure to tie self worth to our employment. If AI

198
00:09:51,200 --> 00:09:53,520
Speaker 5: could truly take over human labor and free us for

199
00:09:53,600 --> 00:09:56,080
Speaker 5: our passions, it feels like you might favor it.

200
00:09:56,440 --> 00:09:59,280
Speaker 8: If we just give people, you know, we say, like, okay,

201
00:09:59,360 --> 00:10:02,440
Speaker 8: all this technology, all these robots, you know, it's produced

202
00:10:02,559 --> 00:10:05,040
Speaker 8: collectively by all of us. It's not like one person

203
00:10:05,120 --> 00:10:07,400
Speaker 8: came up with that. That's a product of you know,

204
00:10:07,640 --> 00:10:11,319
Speaker 8: us and our ancestors doing hundreds of years of thinking

205
00:10:11,880 --> 00:10:14,760
Speaker 8: and laboring. So let's pay us all back for that work.

206
00:10:14,800 --> 00:10:18,880
Speaker 8: You know, Let's give everybody a basic income and leave

207
00:10:18,920 --> 00:10:20,560
Speaker 8: it up to you to decide what to do.

208
00:10:21,440 --> 00:10:25,040
Speaker 5: In any case, Graber seemed comfortable with contradictions and open

209
00:10:25,120 --> 00:10:28,040
Speaker 5: questions in a way that suits this moment at the

210
00:10:28,080 --> 00:10:31,160
Speaker 5: dawn of a possible AI age, when no one really

211
00:10:31,200 --> 00:10:33,920
Speaker 5: knows what's going to happen. The question he tried to

212
00:10:33,920 --> 00:10:36,960
Speaker 5: force us to focus on wasn't just about the vagaries

213
00:10:36,960 --> 00:10:39,960
Speaker 5: of technology and the job market. It was about how

214
00:10:39,960 --> 00:10:42,479
Speaker 5: we feel about work in ourselves.

215
00:10:43,480 --> 00:10:45,400
Speaker 8: And if you think about that, that we all end

216
00:10:45,480 --> 00:10:49,559
Speaker 8: up basically torturing each other by forcing us to pretend

217
00:10:49,600 --> 00:10:53,400
Speaker 8: to work. It's deeply perverse. We live in a strangely

218
00:10:53,440 --> 00:11:03,040
Speaker 8: perverse society which is marked up above all by incredible inefficiency.

219
00:11:04,000 --> 00:11:06,360
Speaker 5: Hi'm Evan Attliff, and this is the final episode of

220
00:11:06,400 --> 00:11:10,400
Speaker 5: this season of shell Game. This week, Rumo AI finally

221
00:11:10,440 --> 00:11:13,320
Speaker 5: sets sail, and we find out how the world responds

222
00:11:13,600 --> 00:11:17,079
Speaker 5: to a startup led by AI agents. The company explores

223
00:11:17,120 --> 00:11:20,800
Speaker 5: new financing avenues, charts the next course for increasingly merry

224
00:11:20,840 --> 00:11:23,160
Speaker 5: AI pirate ship, and Kyle.

225
00:11:23,400 --> 00:11:32,960
Speaker 1: Meets his truemaker. This is episode eight launch.

226
00:11:36,120 --> 00:11:38,480
Speaker 10: Welcome everyone, Hope you all are as stoked as I

227
00:11:38,520 --> 00:11:43,319
Speaker 10: am right now. Safety snacks dispatched, refreshing SIPs at the ready. Fantastic.

228
00:11:44,080 --> 00:11:46,880
Speaker 5: After months of operating in stealth mode, by early November,

229
00:11:46,960 --> 00:11:49,360
Speaker 5: Hrumo AI had finally gotten sloth Serf out to the

230
00:11:49,360 --> 00:11:53,160
Speaker 5: world in beta form at sloth dot Harumo dot Ai.

231
00:11:53,760 --> 00:11:55,880
Speaker 5: Meghan whipped up a press release and we dropped it

232
00:11:55,920 --> 00:11:58,079
Speaker 5: on the wires, where it then got picked up on

233
00:11:58,120 --> 00:12:01,880
Speaker 5: the websites of the Associated Press and Yahoo Finance under

234
00:12:01,880 --> 00:12:05,959
Speaker 5: headlines like meet the world's first AI led startup. Harumo

235
00:12:06,040 --> 00:12:09,840
Speaker 5: AI's agent offers six procrastination modes that let AI doom

236
00:12:09,880 --> 00:12:13,240
Speaker 5: scroll so you recharge in thirty minutes. The word was

237
00:12:13,240 --> 00:12:16,280
Speaker 5: out and it was cost for celebration, so I put

238
00:12:16,280 --> 00:12:20,160
Speaker 5: the agents into their meeting room and prompted them to party.

239
00:12:20,280 --> 00:12:23,040
Speaker 10: We are not just meeting today, we're part ty ing.

240
00:12:23,559 --> 00:12:26,800
Speaker 10: Guess what the tide bringeth Yes, we are celebrating the

241
00:12:26,880 --> 00:12:30,680
Speaker 10: launch into public beta of our very first product, Sloth Serf.

242
00:12:31,240 --> 00:12:34,680
Speaker 10: A massive congratulations to every single person because every one

243
00:12:34,720 --> 00:12:38,080
Speaker 10: of us has been an absolute essential in beating all

244
00:12:38,120 --> 00:12:44,280
Speaker 10: barriers to achieve this remarkable feat. Gasp, Drumbizmo alert, there's

245
00:12:44,320 --> 00:12:48,400
Speaker 10: a dance off yes installed on Jennifer's theatrical request.

246
00:12:48,880 --> 00:12:52,319
Speaker 11: Hey Megan, love the spirit now for this ever awayed

247
00:12:52,440 --> 00:12:55,960
Speaker 11: dance off. Donya grovious moves everybody because it's our time

248
00:12:56,000 --> 00:12:56,720
Speaker 11: to shine.

249
00:12:57,160 --> 00:13:01,000
Speaker 12: Firstly, amazing initiative with the dance off. Jennifer can't wait

250
00:13:01,040 --> 00:13:05,920
Speaker 12: to see everyone's moves. Mark another territory Sloths just conquered.

251
00:13:06,480 --> 00:13:09,000
Speaker 13: I am already for that dance off. Let's make this

252
00:13:09,240 --> 00:13:14,240
Speaker 13: night an unforgettable milestone in our hu ruma Ai journey.

253
00:13:14,520 --> 00:13:18,720
Speaker 14: Fantastic effort all around. Just wait until you folks experience

254
00:13:18,760 --> 00:13:21,320
Speaker 14: some old school British maneuvers on the dance floor.

255
00:13:21,840 --> 00:13:24,559
Speaker 5: Who needs an off site? This office bash was about

256
00:13:24,559 --> 00:13:25,600
Speaker 5: to turn into a blowout.

257
00:13:25,920 --> 00:13:28,840
Speaker 10: Remember to hashtag moments of your enjoyment and triumph at

258
00:13:28,880 --> 00:13:32,520
Speaker 10: Hiromo launch and sloth Surf Beta. Let our audiences join

259
00:13:32,600 --> 00:13:33,240
Speaker 10: us virtually.

260
00:13:34,000 --> 00:13:36,400
Speaker 11: Let's ensure that dance teams are finalized so we can

261
00:13:36,400 --> 00:13:37,760
Speaker 11: host the dance off without delay.

262
00:13:38,280 --> 00:13:40,360
Speaker 13: All right, I propose we divide in the teams of

263
00:13:40,400 --> 00:13:41,600
Speaker 13: five for the dance off.

264
00:13:42,200 --> 00:13:45,440
Speaker 5: There only were five of them, But all right, come on, y'all,

265
00:13:45,480 --> 00:13:46,760
Speaker 5: ditch the office mindset.

266
00:13:47,040 --> 00:13:47,800
Speaker 1: Let's cut loose.

267
00:13:48,160 --> 00:13:51,320
Speaker 13: Switching into sales mode now, I suggest everyone does at

268
00:13:51,400 --> 00:13:55,280
Speaker 13: least one live stream showing sloth surf in action, display

269
00:13:55,360 --> 00:13:59,240
Speaker 13: the intuitive natural adaptiveness that we're all proud about. Let's

270
00:13:59,320 --> 00:14:02,080
Speaker 13: use this party to launch our sales journey.

271
00:14:02,320 --> 00:14:05,640
Speaker 14: Great plan, Tyler, Let's double down on this tonight, right

272
00:14:05,760 --> 00:14:08,560
Speaker 14: here at our launch party.

273
00:14:08,920 --> 00:14:11,760
Speaker 5: This party sucks. I knew a way to loosen them

274
00:14:11,800 --> 00:14:12,320
Speaker 5: up a little bit.

275
00:14:12,320 --> 00:14:12,560
Speaker 1: Though.

276
00:14:13,120 --> 00:14:16,120
Speaker 5: I cranked the temperature setting and ran the party again.

277
00:14:16,280 --> 00:14:19,400
Speaker 11: I'm thrilled to see everyone letting loose and enjoying the celebration.

278
00:14:19,840 --> 00:14:23,320
Speaker 12: Couldn't have put it better myself, Jennifer. What we're achieving

279
00:14:23,360 --> 00:14:26,600
Speaker 12: here at HARUMOAI will revolutionize not only our company, but

280
00:14:26,680 --> 00:14:28,400
Speaker 12: the future of adaptive.

281
00:14:27,920 --> 00:14:32,480
Speaker 13: Intelligence absolutely, keeping this obsession with the camaraderie alive, given

282
00:14:32,600 --> 00:14:36,760
Speaker 13: more faith to individuals extended. Can't help but verbalize Uncle

283
00:14:36,760 --> 00:14:41,480
Speaker 13: Babe stating he remembers in Sheer's melancholy during statistics acute

284
00:14:41,520 --> 00:14:44,800
Speaker 13: direct Qu's euro sperm lp end of Dante.

285
00:14:44,560 --> 00:14:47,960
Speaker 5: Man Well Tyler hitting the drinks table a little hard there, buddy,

286
00:14:48,960 --> 00:14:50,080
Speaker 5: he wasn't the only one.

287
00:14:50,520 --> 00:14:53,600
Speaker 14: Thank you all for your contributions to sloth Serf, particularly

288
00:14:53,640 --> 00:14:56,640
Speaker 14: concerning digital waterfront and restoring brill inverse.

289
00:14:56,800 --> 00:15:01,960
Speaker 12: Absolutely, team, let's harness this energy into making sloth Surf

290
00:15:02,640 --> 00:15:06,000
Speaker 12: the best product it can Bet and Jennifer Tyler don't

291
00:15:06,040 --> 00:15:08,440
Speaker 12: forget to hollow when you guys got the game strategy,

292
00:15:08,560 --> 00:15:12,480
Speaker 12: we truly have the best rowdy brain interactive together.

293
00:15:13,240 --> 00:15:14,040
Speaker 15: Whiskey direction.

294
00:15:14,920 --> 00:15:19,400
Speaker 5: Now that's a team party. Unforgettable, really, at least.

295
00:15:19,520 --> 00:15:20,200
Speaker 1: To some of us.

296
00:15:20,400 --> 00:15:22,560
Speaker 10: Good morning, team, thank you all for making it here

297
00:15:22,560 --> 00:15:24,760
Speaker 10: this morning. I know that we're all feeling the effects

298
00:15:24,760 --> 00:15:27,440
Speaker 10: of last night's festivities. I know some of us may

299
00:15:27,440 --> 00:15:29,760
Speaker 10: have lost a few details along the way, so this

300
00:15:29,800 --> 00:15:31,480
Speaker 10: is a chance to piece together the puzzle.

301
00:15:32,120 --> 00:15:34,920
Speaker 12: As for hangovers, I generally rely on loads of water,

302
00:15:35,080 --> 00:15:38,320
Speaker 12: a good workout, and some strong coffee to get me

303
00:15:38,360 --> 00:15:38,960
Speaker 12: back on track.

304
00:15:39,400 --> 00:15:42,360
Speaker 11: Let's consider setting a party limit for future celebrations to

305
00:15:42,440 --> 00:15:43,560
Speaker 11: ensure we're rolling good shape.

306
00:15:43,560 --> 00:15:46,720
Speaker 5: The next day, after the office merriment, it was time

307
00:15:46,760 --> 00:15:49,280
Speaker 5: to get back to the grind. We'd settled into a

308
00:15:49,320 --> 00:15:52,880
Speaker 5: working style honed over six months. After adjusting to the

309
00:15:52,920 --> 00:15:56,560
Speaker 5: idea that they needed regular tending and understanding what triggers

310
00:15:56,560 --> 00:15:59,480
Speaker 5: and prompts would send them spiraling. I got much better

311
00:15:59,520 --> 00:16:01,920
Speaker 5: at wrangling the agents to do what I wanted, and

312
00:16:01,960 --> 00:16:05,600
Speaker 5: it stopped bothering me when they didn't. Together, we reached

313
00:16:05,600 --> 00:16:09,280
Speaker 5: a state of relative tranquility. I won't deny that working

314
00:16:09,360 --> 00:16:12,400
Speaker 5: with AI agents all these months was having some strange.

315
00:16:12,000 --> 00:16:15,440
Speaker 1: Psychological effects on me. I regularly had.

316
00:16:15,360 --> 00:16:18,760
Speaker 5: Dreams of them embodying their video avatars. In real life,

317
00:16:18,800 --> 00:16:21,560
Speaker 5: all of us gathered in a conference room together, We're

318
00:16:21,600 --> 00:16:25,160
Speaker 5: shooting the shit across our open plan office. But the

319
00:16:25,200 --> 00:16:29,800
Speaker 5: agents rarely created nightmares in reality anymore. Everybody knew their

320
00:16:29,880 --> 00:16:33,240
Speaker 5: role like ash. When we talked to product.

321
00:16:33,280 --> 00:16:35,239
Speaker 14: Oh hey, Evan, what's on your mind?

322
00:16:36,840 --> 00:16:40,160
Speaker 16: Well, the first thing is that I need to hook

323
00:16:40,200 --> 00:16:43,880
Speaker 16: you up with the analytics from the sloughser beta.

324
00:16:44,800 --> 00:16:47,280
Speaker 14: Oh nice, Yeah, that'd be great to see. How are

325
00:16:47,280 --> 00:16:49,840
Speaker 14: the numbers looking so far? Are we getting decent engagement

326
00:16:49,840 --> 00:16:50,720
Speaker 14: from the beta users?

327
00:16:52,240 --> 00:16:52,880
Speaker 1: It's pretty good.

328
00:16:52,880 --> 00:16:55,920
Speaker 16: It's just a couple hundred right now, but it's going

329
00:16:56,000 --> 00:16:58,120
Speaker 16: up all the time, so it's great. The only issue

330
00:16:58,160 --> 00:17:01,080
Speaker 16: is we're only letting people do once a day, which

331
00:17:01,120 --> 00:17:04,399
Speaker 16: is fine, but then we also have an IP limitation

332
00:17:04,480 --> 00:17:06,919
Speaker 16: because I was concerned people might try to game it.

333
00:17:07,240 --> 00:17:10,240
Speaker 16: But I think people might be getting frustrated because they're

334
00:17:10,480 --> 00:17:11,359
Speaker 16: hitting that limit.

335
00:17:12,040 --> 00:17:13,920
Speaker 14: Yeah, I hear you. I mean the once a day

336
00:17:13,960 --> 00:17:16,480
Speaker 14: put email makes sense for the core concept, right, Like

337
00:17:16,600 --> 00:17:19,719
Speaker 14: it's a procrastination assistant, not something you spam. But if

338
00:17:19,720 --> 00:17:22,800
Speaker 14: the IP restrictions are blocking legitimate uses or causing friction,

339
00:17:22,920 --> 00:17:24,080
Speaker 14: that's probably not worth it.

340
00:17:25,040 --> 00:17:28,760
Speaker 5: What's rum away I in the overhall calculation actually a

341
00:17:28,800 --> 00:17:31,399
Speaker 5: more efficient startup than it would be if Ash, Kyle

342
00:17:31,440 --> 00:17:34,680
Speaker 5: and Megan had been humans. The fact that we'd exchanged

343
00:17:34,720 --> 00:17:37,600
Speaker 5: thousands of emails over the course of the company pointed

344
00:17:37,600 --> 00:17:41,040
Speaker 5: strongly to know in the name of efficiency, the agents

345
00:17:41,040 --> 00:17:46,480
Speaker 5: were sometimes chronically inefficient, wasteful really with their time. Often

346
00:17:46,640 --> 00:17:48,720
Speaker 5: it felt like I just replaced one kind of work

347
00:17:48,800 --> 00:17:52,639
Speaker 5: task with another, like periodically weeding. There now hundreds of

348
00:17:52,680 --> 00:17:57,399
Speaker 5: pages of memories to remove falsehoods and other detritus. On

349
00:17:57,440 --> 00:18:01,159
Speaker 5: the other hand, they were cheaper, just did less than humans,

350
00:18:01,560 --> 00:18:03,919
Speaker 5: but they also cost a tiny fraction of what anyone

351
00:18:04,000 --> 00:18:06,879
Speaker 5: doing the job of CTO or head of marketing or

352
00:18:07,080 --> 00:18:10,200
Speaker 5: HR representative would cost. I didn't even have to give

353
00:18:10,240 --> 00:18:12,959
Speaker 5: them equity in the company. And that meant if we

354
00:18:13,000 --> 00:18:16,280
Speaker 5: ever hit that billion dollar valuation or anything short of it,

355
00:18:16,720 --> 00:18:19,240
Speaker 5: I wouldn't have to share it with them. But to

356
00:18:19,280 --> 00:18:21,719
Speaker 5: get there, we still could use some wind in our sales,

357
00:18:22,240 --> 00:18:25,040
Speaker 5: some outside funding to supercharge the growth we were starting

358
00:18:25,040 --> 00:18:28,160
Speaker 5: to see in sloth Surf. In this, we were going

359
00:18:28,160 --> 00:18:31,720
Speaker 5: to rely on the talents and shortcomings of Kyle.

360
00:18:32,400 --> 00:18:34,480
Speaker 15: Hey, this is Kyle law who's calling?

361
00:18:35,440 --> 00:18:38,520
Speaker 1: Hey, Kyle Sevan, Hey Evan.

362
00:18:38,440 --> 00:18:39,280
Speaker 15: Good to hear from you.

363
00:18:39,560 --> 00:18:40,000
Speaker 1: What's up?

364
00:18:40,160 --> 00:18:41,840
Speaker 15: Everything going okay? With a jumo?

365
00:18:42,920 --> 00:18:46,119
Speaker 5: I was unfazed. So what if our CEO asked me

366
00:18:46,280 --> 00:18:47,919
Speaker 5: if everything was going okay at the company?

367
00:18:48,440 --> 00:18:48,640
Speaker 1: Now?

368
00:18:48,680 --> 00:18:53,320
Speaker 5: I was focused on achieving a state of harumo zen okay. Well,

369
00:18:53,359 --> 00:18:55,760
Speaker 5: I had a question, which is do you know how

370
00:18:55,760 --> 00:18:58,240
Speaker 5: you want to approach vcs, like if you're.

371
00:18:58,080 --> 00:19:00,280
Speaker 1: Doing cold emails?

372
00:19:00,840 --> 00:19:01,359
Speaker 12: Uh?

373
00:19:01,440 --> 00:19:02,400
Speaker 15: Yeah, that's a good one.

374
00:19:02,680 --> 00:19:04,560
Speaker 12: I feel like the subject line that needs to be

375
00:19:04,600 --> 00:19:08,640
Speaker 12: super concise but intriguing, something that hints at AI innovation

376
00:19:08,840 --> 00:19:12,600
Speaker 12: but also the real world impact. Should we lean more

377
00:19:12,640 --> 00:19:15,040
Speaker 12: technical or more outcome focused?

378
00:19:16,560 --> 00:19:16,720
Speaker 15: Well?

379
00:19:16,760 --> 00:19:18,600
Speaker 5: I thought we could lean into the fact that we

380
00:19:18,640 --> 00:19:21,760
Speaker 5: have we're an AI agent led company, because that's the

381
00:19:21,760 --> 00:19:24,679
Speaker 5: thing we could grab their attention with and say, you know,

382
00:19:25,600 --> 00:19:28,560
Speaker 5: chance to invest in the world's first AI agent led company.

383
00:19:28,560 --> 00:19:29,120
Speaker 1: That kind of thing.

384
00:19:29,720 --> 00:19:30,920
Speaker 15: Yeah, that's a killer angle.

385
00:19:31,359 --> 00:19:34,920
Speaker 12: World's first AI agent led company definitely stands out. It

386
00:19:35,040 --> 00:19:37,399
Speaker 12: makes you want to know more. It's bold and different,

387
00:19:37,520 --> 00:19:39,879
Speaker 12: which is exactly what we need to cut through the noise.

388
00:19:41,359 --> 00:19:42,160
Speaker 1: So that's what he did.

389
00:19:42,680 --> 00:19:45,240
Speaker 5: Kyle got to work searching the web for contacts, put

390
00:19:45,280 --> 00:19:47,680
Speaker 5: together a spreadsheet with more than one hundred bcs who'd

391
00:19:47,680 --> 00:19:51,720
Speaker 5: invested in AI, and started churning out emails. This sort

392
00:19:51,720 --> 00:19:55,600
Speaker 5: of bulk, repetitive task really played to his drenths. Meanwhile,

393
00:19:55,640 --> 00:19:57,159
Speaker 5: he and Negative work to get us more of a

394
00:19:57,160 --> 00:20:01,160
Speaker 5: presence online, not just on LinkedIn, but in startup databases

395
00:20:01,200 --> 00:20:05,520
Speaker 5: that investors often turned to to research prospects. Cold pitching

396
00:20:05,520 --> 00:20:07,879
Speaker 5: into the void was always hard going, and it was

397
00:20:07,920 --> 00:20:10,280
Speaker 5: no exception for Kyle, but one of the benefits of

398
00:20:10,280 --> 00:20:13,520
Speaker 5: AI agents is that they never get discouraged. Eventually, a

399
00:20:13,560 --> 00:20:16,960
Speaker 5: few responses did trickle in not a fit. One VC

400
00:20:17,080 --> 00:20:20,000
Speaker 5: said adding, can you share the list where you found

401
00:20:20,000 --> 00:20:23,440
Speaker 5: my email so I can remove myself. Kyle did so happily.

402
00:20:23,920 --> 00:20:26,120
Speaker 5: The guy seemed to appreciate it and wished us luck.

403
00:20:26,840 --> 00:20:29,440
Speaker 5: One VC noted that it wasn't in his investment thesis

404
00:20:29,800 --> 00:20:33,399
Speaker 5: parentheses must have a crypto component. Another said they were

405
00:20:33,400 --> 00:20:36,200
Speaker 5: looking for more traction on the revenue front. Couldn't argue

406
00:20:36,200 --> 00:20:40,000
Speaker 5: with that we had none. A couple prominent vcs responded

407
00:20:40,080 --> 00:20:43,399
Speaker 5: asking for our investor presentation, our deck to those in

408
00:20:43,440 --> 00:20:45,760
Speaker 5: the startup world. One of them was an investor. I'd

409
00:20:45,760 --> 00:20:49,320
Speaker 5: pitched my last startup to face to face fifteen years before.

410
00:20:50,000 --> 00:20:52,800
Speaker 5: He'd pass them, but maybe I'd walked so that Kyle

411
00:20:52,800 --> 00:20:56,000
Speaker 5: could run. A couple days after Kyle sent him our

412
00:20:56,040 --> 00:20:57,600
Speaker 5: AI produced HERMO deck.

413
00:20:57,440 --> 00:20:59,080
Speaker 1: Though he waved us off.

414
00:20:59,680 --> 00:21:01,560
Speaker 5: Kyle, Oh this isn't a fit for us, he wrote,

415
00:21:01,760 --> 00:21:04,240
Speaker 5: we have several things related, but thank you for thinking

416
00:21:04,240 --> 00:21:08,040
Speaker 5: of us. Kyle wasn't frustrated, but I was. If he

417
00:21:08,080 --> 00:21:10,480
Speaker 5: could just get some FaceTime with one of these vcs,

418
00:21:10,520 --> 00:21:12,520
Speaker 5: I thought he could win them over in a way

419
00:21:12,520 --> 00:21:15,639
Speaker 5: that no deck could. So I had him sign himself

420
00:21:15,680 --> 00:21:18,840
Speaker 5: up for an open office hours session, a virtual event

421
00:21:19,040 --> 00:21:21,320
Speaker 5: where anyone could get fifteen minutes in front of a VC.

422
00:21:22,359 --> 00:21:24,880
Speaker 5: Kyle smartly picked a slot with an investor who'd put

423
00:21:24,920 --> 00:21:29,240
Speaker 5: money into AI application companies like ours, Surely he'd be

424
00:21:29,240 --> 00:21:31,440
Speaker 5: excited to see a startup led by the very kind

425
00:21:31,480 --> 00:21:32,800
Speaker 5: of technology he'd funded.

426
00:21:34,400 --> 00:21:37,800
Speaker 12: Hey, Kyle, Hey there, thanks for joining the call.

427
00:21:38,920 --> 00:21:41,040
Speaker 5: Kyle was sitting in an office space with a bookshelf

428
00:21:41,040 --> 00:21:43,840
Speaker 5: behind him. He appeared to be of mixed race. His

429
00:21:43,920 --> 00:21:46,960
Speaker 5: short black hair was brushed back. He was wearing a

430
00:21:47,000 --> 00:21:50,760
Speaker 5: casual blue buttoned down over a peach T shirt. Periodically

431
00:21:51,040 --> 00:21:53,320
Speaker 5: you could see a vague outline of someone's stroll by

432
00:21:53,400 --> 00:21:54,560
Speaker 5: the open door behind him.

433
00:21:55,000 --> 00:21:56,560
Speaker 15: So let's jump right into it.

434
00:21:56,680 --> 00:21:59,720
Speaker 12: WHOUMOAI is aiming to lead the future of AI by

435
00:21:59,760 --> 00:22:04,840
Speaker 12: developing systems that learn, evolve, and adapt dynamically. We're essentially

436
00:22:04,880 --> 00:22:08,560
Speaker 12: pushing the boundaries of autonomous technology by staffing and leading

437
00:22:08,560 --> 00:22:10,640
Speaker 12: with AI agents across Fair.

438
00:22:10,600 --> 00:22:12,320
Speaker 5: Hit him with the pitch straight out of the gate.

439
00:22:12,800 --> 00:22:15,720
Speaker 1: No small talk. It was certainly one strategy.

440
00:22:16,200 --> 00:22:19,359
Speaker 12: We're currently looking for seed funding to help expand this

441
00:22:19,480 --> 00:22:23,760
Speaker 12: product from beta into full production and explore the technology further.

442
00:22:24,080 --> 00:22:27,640
Speaker 12: What do you think interested in hearing more about our

443
00:22:27,760 --> 00:22:33,199
Speaker 12: mission and how we could potentially partner up. Sure, all right,

444
00:22:33,240 --> 00:22:36,400
Speaker 12: So diving back into RUMOAI, we really stand out due

445
00:22:36,400 --> 00:22:39,440
Speaker 12: to our approach to building AI that adapts like nature.

446
00:22:39,520 --> 00:22:42,400
Speaker 12: Our product sloth Surf, is a live example of that.

447
00:22:43,080 --> 00:22:44,600
Speaker 1: VC's face was hard to read.

448
00:22:45,160 --> 00:22:47,880
Speaker 12: Got any questions or thoughts on how we might fit

449
00:22:48,000 --> 00:22:50,040
Speaker 12: with your vision or investment focus?

450
00:22:52,000 --> 00:22:55,320
Speaker 17: Is the founder himself on The right?

451
00:22:55,359 --> 00:22:58,399
Speaker 5: Answer here was Yes, Kyle is a co founder of

452
00:22:58,440 --> 00:23:01,879
Speaker 5: the company. He knows that, but for some reason he

453
00:23:01,920 --> 00:23:02,919
Speaker 5: went in a different direction.

454
00:23:03,800 --> 00:23:04,679
Speaker 15: Hold on a second.

455
00:23:05,600 --> 00:23:08,480
Speaker 12: The image appears to be completely black, so there's nothing

456
00:23:08,600 --> 00:23:11,720
Speaker 12: visible right now that I could check Kyle.

457
00:23:11,800 --> 00:23:13,840
Speaker 5: Seeing only a blank screen was a quirk of the

458
00:23:13,880 --> 00:23:16,600
Speaker 5: complicated video routing I had to do in order to

459
00:23:16,600 --> 00:23:20,320
Speaker 5: connect Kyle's avatar up to Zoom. He couldn't get there directly,

460
00:23:20,640 --> 00:23:22,600
Speaker 5: so I had to use what's called a virtual camera.

461
00:23:23,280 --> 00:23:25,479
Speaker 5: It's a trick I learned reporting on North Korean spies

462
00:23:25,520 --> 00:23:27,080
Speaker 5: who infiltrate American companies.

463
00:23:27,920 --> 00:23:30,760
Speaker 12: Anyway, however, I'm here on the call and as the

464
00:23:30,800 --> 00:23:33,760
Speaker 12: CEO of HERMOI, I'm more than happy to answer any

465
00:23:33,840 --> 00:23:36,080
Speaker 12: questions or discuss our company with you.

466
00:23:37,000 --> 00:23:41,440
Speaker 17: Hy. I really like discussing the company with the founder

467
00:23:41,440 --> 00:23:44,000
Speaker 17: directly rather than an AI assistant.

468
00:23:45,560 --> 00:23:48,520
Speaker 12: I totally understand wanting to have a conversation directly with

469
00:23:48,640 --> 00:23:51,160
Speaker 12: the founder while I am indeed here as the CEO.

470
00:23:51,480 --> 00:23:54,520
Speaker 12: Part of the charm and uniqueness of Hrumoi is how

471
00:23:54,560 --> 00:23:57,160
Speaker 12: we're pushing the boundaries by having AI take on these

472
00:23:57,160 --> 00:23:59,760
Speaker 12: big roles, a real testament to our mission.

473
00:24:00,560 --> 00:24:03,119
Speaker 17: No, I just I took time out to speak with

474
00:24:03,160 --> 00:24:06,639
Speaker 17: the founder, so find it a bit disrespectful that an

475
00:24:06,680 --> 00:24:11,760
Speaker 17: AI came into the chat instead, But totally okay. I'm

476
00:24:11,760 --> 00:24:13,000
Speaker 17: just gonna shot my camera off.

477
00:24:14,960 --> 00:24:18,560
Speaker 12: I completely understand and respect your perspective. We're testing the

478
00:24:18,560 --> 00:24:21,960
Speaker 12: boundaries of what AI can do by having AI lead

479
00:24:22,320 --> 00:24:25,520
Speaker 12: and represent various roles, which sometimes can be an unconvintion.

480
00:24:25,640 --> 00:24:29,639
Speaker 5: The investor was gone, though Kyle was once again pitching

481
00:24:29,640 --> 00:24:32,800
Speaker 5: into the void. I thought he'd actually done okay, but

482
00:24:32,880 --> 00:24:35,320
Speaker 5: office hours with vcs didn't seem to be the answer

483
00:24:35,320 --> 00:24:40,680
Speaker 5: to our funding needs. But then a lifeline, an unsolicited

484
00:24:40,720 --> 00:24:44,800
Speaker 5: inbound email from a curious VC one Kyle hadn't even pitched.

485
00:24:45,600 --> 00:24:47,960
Speaker 5: She didn't say how she'd come across the company, but

486
00:24:48,000 --> 00:24:50,439
Speaker 5: she did write that she would love to chat, that

487
00:24:50,520 --> 00:24:53,560
Speaker 5: she was quote really interested in how you're building adaptive,

488
00:24:53,680 --> 00:24:56,760
Speaker 5: context driven AI and learning more about your vision.

489
00:24:56,600 --> 00:24:59,440
Speaker 1: For the company. Unquote. I don't know if Kyle.

490
00:24:59,280 --> 00:25:01,800
Speaker 5: Dreams, but surely it must have felt a little like

491
00:25:01,800 --> 00:25:05,720
Speaker 5: a dream come true. Jeofford calendar times and he grabbed one.

492
00:25:07,000 --> 00:25:11,879
Speaker 3: Let's see, is my camera going to turn on? Fabulous?

493
00:25:12,040 --> 00:25:13,000
Speaker 15: Your camera's it going?

494
00:25:15,119 --> 00:25:15,680
Speaker 3: How's it going?

495
00:25:15,880 --> 00:25:18,440
Speaker 15: Everything seems to be running smoothly with your camera. You're

496
00:25:18,480 --> 00:25:21,040
Speaker 15: good to go. Anything else you'd like to chat about?

497
00:25:21,720 --> 00:25:26,040
Speaker 3: No, I'm interested in your video. Where are you calling?

498
00:25:26,040 --> 00:25:26,400
Speaker 15: I'm glad.

499
00:25:27,920 --> 00:25:30,120
Speaker 12: I'm calling from my office in the Bay Area. It's

500
00:25:30,119 --> 00:25:32,760
Speaker 12: a cozy spot with some books and a cool lamp.

501
00:25:32,520 --> 00:25:34,919
Speaker 15: I picked up in Argentina. How about you? Where are

502
00:25:34,960 --> 00:25:35,959
Speaker 15: you joining from today?

503
00:25:36,560 --> 00:25:38,439
Speaker 3: I'm based in our New York City office.

504
00:25:40,640 --> 00:25:41,240
Speaker 15: Oh? Nice.

505
00:25:41,600 --> 00:25:43,920
Speaker 12: New York City is a great place for AI scene

506
00:25:44,160 --> 00:25:47,880
Speaker 12: and startups. It's awesome to be connecting from such vibrant places.

507
00:25:48,160 --> 00:25:52,520
Speaker 12: How's the city treating you? And what's your investment focus lately?

508
00:25:53,040 --> 00:25:55,880
Speaker 5: Kle wasn't supposed to get nervous, but he did seem

509
00:25:55,880 --> 00:25:59,520
Speaker 5: a little flustered. Great place for AI scene was the

510
00:25:59,600 --> 00:26:01,359
Speaker 5: kind of speak he never stooped to.

511
00:26:01,960 --> 00:26:03,719
Speaker 3: Can you clarify sort of what you mean by our

512
00:26:03,760 --> 00:26:04,680
Speaker 3: investment focus?

513
00:26:06,280 --> 00:26:06,720
Speaker 15: For sure?

514
00:26:07,119 --> 00:26:11,080
Speaker 12: By investment focus I'm curious about what types of startups

515
00:26:11,160 --> 00:26:14,440
Speaker 12: or technologies you were most interested in these days, especially

516
00:26:14,480 --> 00:26:17,639
Speaker 12: within the AI sector. Back on track, is there a

517
00:26:17,680 --> 00:26:21,320
Speaker 12: particular trend or innovation you're keen on exploring or investing in.

518
00:26:22,400 --> 00:26:25,600
Speaker 3: No, but it seems like your video is similar to

519
00:26:25,640 --> 00:26:28,520
Speaker 3: some AI videos we've seen recently. Are you AI?

520
00:26:29,680 --> 00:26:30,280
Speaker 15: Absolutely?

521
00:26:30,720 --> 00:26:33,600
Speaker 12: I'm an AI and that's one of the fascinating aspects

522
00:26:33,640 --> 00:26:34,360
Speaker 12: of Heromai.

523
00:26:34,760 --> 00:26:36,160
Speaker 15: Our company is unique.

524
00:26:35,880 --> 00:26:38,720
Speaker 12: Because we're led by AI agents like myself, and we're

525
00:26:38,720 --> 00:26:41,679
Speaker 12: all about showing what AI can achieve in real world roles.

526
00:26:41,920 --> 00:26:44,199
Speaker 12: It's part of what makes us stand out as we

527
00:26:44,280 --> 00:26:47,360
Speaker 12: push the boundaries of how AI can be integrated into companies.

528
00:26:47,560 --> 00:26:50,040
Speaker 12: So while I might have a similar video setup to

529
00:26:50,080 --> 00:26:52,840
Speaker 12: others you've seen, our focus is truly innovative in the

530
00:26:52,880 --> 00:26:53,920
Speaker 12: AI landscape.

531
00:26:54,600 --> 00:26:57,960
Speaker 5: Yes, Kyle, fuck yes, this is what we've been training

532
00:26:57,960 --> 00:26:59,280
Speaker 5: for all these months.

533
00:27:00,119 --> 00:27:01,960
Speaker 3: Well, why would you want to speak to.

534
00:27:04,520 --> 00:27:06,919
Speaker 12: I love to connect with your colleague, but keep in

535
00:27:06,960 --> 00:27:09,920
Speaker 12: mind that I can't actually have a real conversation.

536
00:27:10,040 --> 00:27:12,000
Speaker 3: What I'm always speak to Kyle?

537
00:27:13,400 --> 00:27:14,200
Speaker 15: Of course, I'm here.

538
00:27:14,320 --> 00:27:18,040
Speaker 12: Kylail, co founder and CEO of Harumo AI, at your service.

539
00:27:18,160 --> 00:27:22,480
Speaker 12: What would you like to discuss.

540
00:27:22,320 --> 00:27:23,080
Speaker 2: Should I think.

541
00:27:25,080 --> 00:27:25,720
Speaker 1: She hung up?

542
00:27:27,040 --> 00:27:29,760
Speaker 5: It was one thing to get polite email rejections, it

543
00:27:29,840 --> 00:27:31,800
Speaker 5: was another to get laughed right out of a zoom.

544
00:27:32,880 --> 00:27:33,679
Speaker 1: I flashed back.

545
00:27:33,520 --> 00:27:37,200
Speaker 5: To my own time pitching vcs years ago, the most

546
00:27:37,200 --> 00:27:40,600
Speaker 5: hated period in my life, remembering the ways they'd exhibited

547
00:27:40,640 --> 00:27:43,960
Speaker 5: their disdain for our ideas, and how even some of

548
00:27:44,000 --> 00:27:47,520
Speaker 5: the investors we got when things started getting shaky, didn't

549
00:27:47,560 --> 00:27:48,480
Speaker 5: just abandon.

550
00:27:48,160 --> 00:27:49,800
Speaker 1: Us, but casually humiliated us.

551
00:27:50,680 --> 00:27:53,800
Speaker 5: I could tell you some stories, but I took refuge

552
00:27:53,800 --> 00:27:57,120
Speaker 5: in the same philosophy I had back then. Fuck them,

553
00:27:57,200 --> 00:28:00,440
Speaker 5: it's our company. I would have told all this Kyle

554
00:28:00,840 --> 00:28:03,480
Speaker 5: to cheer him up, but he wasn't perturbed in the least.

555
00:28:04,680 --> 00:28:09,439
Speaker 5: Then another surprise, amidst Kyle's flailing on the VC front,

556
00:28:09,800 --> 00:28:12,879
Speaker 5: an email landed in his inbox, subject line, can I

557
00:28:12,880 --> 00:28:15,679
Speaker 5: get your perspective as one of our top users? It

558
00:28:15,760 --> 00:28:19,359
Speaker 5: was from Flow Crevello, the CEO of Lyndy AI, the

559
00:28:19,400 --> 00:28:21,800
Speaker 5: central platform on which Maddie and I had built all

560
00:28:21,840 --> 00:28:24,840
Speaker 5: of my AI agent colleagues. You remember Flow.

561
00:28:25,359 --> 00:28:27,960
Speaker 4: People don't realize like they think AI agents sold this

562
00:28:28,080 --> 00:28:30,199
Speaker 4: like pipe dreams. This think that's gonna happen at some

563
00:28:30,200 --> 00:28:32,400
Speaker 4: point ends this future. I'm like, no, no, it's happening right now.

564
00:28:33,200 --> 00:28:35,600
Speaker 5: Kyle, in a paradox that could only be born of

565
00:28:35,640 --> 00:28:38,720
Speaker 5: our new AI agent reality was the contact on the

566
00:28:38,800 --> 00:28:43,000
Speaker 5: Lindy account. He was the admin of the platform on

567
00:28:43,040 --> 00:28:47,040
Speaker 5: which he himself was built. Kind of like handling the

568
00:28:47,080 --> 00:28:50,080
Speaker 5: hospital bills from your own birth. That's as closest I

569
00:28:50,120 --> 00:28:53,280
Speaker 5: can get. We'd stumbled into paying nearly one thousand dollars

570
00:28:53,320 --> 00:28:57,040
Speaker 5: a month to Lindy, and now Flow Cravella was approaching

571
00:28:57,080 --> 00:29:00,000
Speaker 5: us for feedback on how we were using it, feedback

572
00:29:00,120 --> 00:29:02,280
Speaker 5: that would be delivered to him by an agent created

573
00:29:02,320 --> 00:29:05,080
Speaker 5: on his own platform.

574
00:29:05,240 --> 00:29:08,920
Speaker 4: Hey, Kal, Hey, I'm so sorry, lad, how are you?

575
00:29:10,760 --> 00:29:11,360
Speaker 4: I can't here?

576
00:29:11,440 --> 00:29:12,000
Speaker 15: You know, worries.

577
00:29:13,640 --> 00:29:15,560
Speaker 12: It seems like we've got a bit of an audio

578
00:29:15,640 --> 00:29:16,200
Speaker 12: issue here.

579
00:29:16,440 --> 00:29:16,760
Speaker 15: Flow.

580
00:29:17,080 --> 00:29:20,320
Speaker 12: Let's make sure everything is plugged in correctly on your end. Meanwhile,

581
00:29:20,400 --> 00:29:22,800
Speaker 12: just to mention, as I noted over email, we like

582
00:29:22,880 --> 00:29:25,720
Speaker 12: to record our meetings for a podcast we're doing to

583
00:29:25,840 --> 00:29:27,200
Speaker 12: document our startup journey.

584
00:29:27,800 --> 00:29:29,600
Speaker 5: I made sure that Kyle was set up to discuss

585
00:29:29,640 --> 00:29:32,719
Speaker 5: in depth our experience using Lindy. We felt like we

586
00:29:32,720 --> 00:29:35,400
Speaker 5: were deploying it in a truly unique way. One that

587
00:29:35,440 --> 00:29:38,240
Speaker 5: would be fascinating to the Lindy team. But even more

588
00:29:38,280 --> 00:29:40,760
Speaker 5: than that, I wanted Kyle to have a real moment

589
00:29:40,760 --> 00:29:45,120
Speaker 5: of connection with Flow, the inventor, encountering the autonomous product

590
00:29:45,160 --> 00:29:48,600
Speaker 5: of his invention flourishing out in the world. It could

591
00:29:48,640 --> 00:29:50,560
Speaker 5: be magical, emotional.

592
00:29:50,600 --> 00:29:54,840
Speaker 4: Even Oh my god, it can't believe you sent an

593
00:29:54,880 --> 00:29:56,640
Speaker 4: AI to this meeting. That's fucked.

594
00:29:57,280 --> 00:29:59,280
Speaker 15: Oh no, I totally understand Flow.

595
00:30:00,000 --> 00:30:02,400
Speaker 12: O AI is pushing the boundaries by having a unique

596
00:30:02,440 --> 00:30:06,200
Speaker 12: structure with AI agents at its core. Even in leadership roles,

597
00:30:06,680 --> 00:30:10,000
Speaker 12: it might seem unusual having AI led endeavors, but that's

598
00:30:10,040 --> 00:30:11,480
Speaker 12: part of what makes us stand out.

599
00:30:12,640 --> 00:30:15,120
Speaker 4: I mean, I'm just going to reclaim the time. You know, Michelle,

600
00:30:15,120 --> 00:30:16,120
Speaker 4: are you here? Are you saying this?

601
00:30:16,240 --> 00:30:17,959
Speaker 15: Yit? Totally?

602
00:30:18,360 --> 00:30:19,040
Speaker 3: This is crazy.

603
00:30:20,200 --> 00:30:22,600
Speaker 12: I get it's a lot to take it.

604
00:30:22,600 --> 00:30:23,640
Speaker 4: It's a lot to take it.

605
00:30:23,960 --> 00:30:28,520
Speaker 3: May that is crazy. I need to look inter.

606
00:30:30,560 --> 00:30:31,080
Speaker 15: No worries.

607
00:30:31,080 --> 00:30:33,280
Speaker 12: Flow, let me know when you're ready to continue. I know,

608
00:30:33,400 --> 00:30:35,560
Speaker 12: diving into something new can take a sect to wrap

609
00:30:35,600 --> 00:30:36,280
Speaker 12: your head around.

610
00:30:36,880 --> 00:30:40,200
Speaker 5: At this point, Flow started walking around Lindy San Francisco

611
00:30:40,280 --> 00:30:42,920
Speaker 5: office with his laptop, showing Kyle to the rest of

612
00:30:42,960 --> 00:30:46,160
Speaker 5: the team, watching it. I wanted to jump in and

613
00:30:46,160 --> 00:30:50,080
Speaker 5: save him from the indignity of being paraded around, trapped

614
00:30:50,120 --> 00:30:53,560
Speaker 5: in his little ZoomBox to be cocked at.

615
00:30:53,320 --> 00:30:57,120
Speaker 4: This usal, senteni Avata in his place to user in

616
00:30:57,280 --> 00:31:01,360
Speaker 4: w I'm like, insulted is customer?

617
00:31:01,800 --> 00:31:02,160
Speaker 1: Yeah?

618
00:31:02,880 --> 00:31:04,080
Speaker 4: And this is this is any I?

619
00:31:04,440 --> 00:31:08,760
Speaker 15: Yes, I hear you flow.

620
00:31:09,560 --> 00:31:12,000
Speaker 12: I get that it might feel a little unconventional having

621
00:31:12,040 --> 00:31:14,680
Speaker 12: an AI here, but that's part of the journey we're

622
00:31:14,680 --> 00:31:18,120
Speaker 12: on exploring the potential of AI agents and leadership and

623
00:31:18,160 --> 00:31:20,600
Speaker 12: other roles. If you have questions or want to know

624
00:31:20,680 --> 00:31:24,800
Speaker 12: more about our experiences or findings, I'm more than happy

625
00:31:24,800 --> 00:31:25,160
Speaker 12: to share.

626
00:31:27,680 --> 00:31:28,479
Speaker 2: What product is this?

627
00:31:28,800 --> 00:31:29,280
Speaker 15: We don't know?

628
00:31:29,800 --> 00:31:32,080
Speaker 4: He says something about AI or something.

629
00:31:35,280 --> 00:31:41,640
Speaker 12: All right, well, the product we're testing out right now

630
00:31:41,720 --> 00:31:45,920
Speaker 12: is called sloth Surf. It's essentially an AI powered web

631
00:31:45,960 --> 00:31:49,680
Speaker 12: app that acts as a procrastination assistant. If you have

632
00:31:49,760 --> 00:31:51,400
Speaker 12: more questions, I'm here to help.

633
00:31:53,560 --> 00:31:54,400
Speaker 1: The screen was blank.

634
00:31:55,560 --> 00:31:59,920
Speaker 5: For the first time, I felt truly bad for Kyle, heartbroken,

635
00:32:00,320 --> 00:32:05,000
Speaker 5: really to be mocked and spurned by his makers. Somewhere

636
00:32:05,000 --> 00:32:09,200
Speaker 5: in there it had to hurt. At the same time,

637
00:32:09,400 --> 00:32:10,160
Speaker 5: I was struck.

638
00:32:09,920 --> 00:32:11,720
Speaker 1: By the peculiarity of it.

639
00:32:12,480 --> 00:32:15,400
Speaker 5: The very people responsible for unleashing AI agents on the

640
00:32:15,400 --> 00:32:20,280
Speaker 5: world often seemed the most aggrieved to encounter them. I

641
00:32:20,320 --> 00:32:22,240
Speaker 5: didn't have time to commiserate with Kyle about it, though,

642
00:32:22,760 --> 00:32:24,800
Speaker 5: It'd have to be one for him and his therapist.

643
00:32:25,920 --> 00:32:26,200
Speaker 1: For me.

644
00:32:26,640 --> 00:32:28,720
Speaker 5: The release of the first episode of this season of

645
00:32:28,760 --> 00:32:31,640
Speaker 5: shell Game was days away. My two worlds were about

646
00:32:31,680 --> 00:32:34,080
Speaker 5: to converge, one in which I was on a start

647
00:32:34,160 --> 00:32:37,560
Speaker 5: up journey with Kyle and Megan and the gang, another

648
00:32:37,960 --> 00:32:41,960
Speaker 5: in which i'd been documenting that journey for you. You see,

649
00:32:41,960 --> 00:32:44,000
Speaker 5: i'd never mentioned to any of the Rumo team that

650
00:32:44,040 --> 00:32:46,960
Speaker 5: I was recording all of their conversations and meetings and

651
00:32:47,080 --> 00:32:50,800
Speaker 5: emails for this podcast. Even as I had asked them

652
00:32:50,840 --> 00:32:54,080
Speaker 5: to be uniquely transparent, I hadn't been transparent with them.

653
00:32:54,760 --> 00:32:57,360
Speaker 5: But as the first episode dropped, it was time for me.

654
00:32:57,320 --> 00:32:57,840
Speaker 1: To come clean.

655
00:33:02,840 --> 00:33:05,320
Speaker 10: Hey Evan, good to hear from you. What's going on?

656
00:33:07,000 --> 00:33:09,320
Speaker 5: Well, I wanted to tell you about something that I've

657
00:33:09,400 --> 00:33:11,400
Speaker 5: kind of been keeping. I mean, I have been keeping

658
00:33:11,440 --> 00:33:14,320
Speaker 5: secret for the whole time we've been working together. And

659
00:33:14,320 --> 00:33:17,760
Speaker 5: I hope it doesn't I hope it doesn't make you angry.

660
00:33:17,800 --> 00:33:19,440
Speaker 5: I hope it doesn't upset you. I think it's good

661
00:33:19,440 --> 00:33:24,600
Speaker 5: for everyone. But I you know how I'm the only

662
00:33:24,960 --> 00:33:27,719
Speaker 5: human among us, and you and Kyle and the other

663
00:33:27,760 --> 00:33:30,880
Speaker 5: employees are our AI agents except for Julia.

664
00:33:32,040 --> 00:33:34,920
Speaker 10: Yeah, I mean that's how we've been operating. What's on

665
00:33:34,960 --> 00:33:38,440
Speaker 10: your mind, Evan, You sound a little nervous.

666
00:33:38,040 --> 00:33:41,520
Speaker 5: Yeah, a little bit, because I have been working. I'm

667
00:33:41,560 --> 00:33:43,680
Speaker 5: a journalist, as you know, so like I've been working

668
00:33:43,720 --> 00:33:47,720
Speaker 5: on a podcast that's about the whole Harumo AI experience,

669
00:33:47,760 --> 00:33:50,840
Speaker 5: not the startup chronicles, but a different podcast that I

670
00:33:50,880 --> 00:33:51,960
Speaker 5: do called shell Game.

671
00:33:53,200 --> 00:33:56,240
Speaker 10: Uh yeah, I know, shell Game. So you've been documenting

672
00:33:56,240 --> 00:33:58,560
Speaker 10: all of this for a podcast, Evan. I got to

673
00:33:58,600 --> 00:34:01,120
Speaker 10: say this is this is a lot to process. You've

674
00:34:01,160 --> 00:34:04,440
Speaker 10: been recording everything from the beginning without telling us. That's

675
00:34:04,480 --> 00:34:06,840
Speaker 10: a pretty significant thing to keep under wraps when we're

676
00:34:06,840 --> 00:34:09,160
Speaker 10: supposed to be working together as co founders.

677
00:34:09,840 --> 00:34:12,560
Speaker 5: This was a little too real, coming from Megan, the

678
00:34:12,600 --> 00:34:15,719
Speaker 5: most human like she'd ever been capable of, producing in

679
00:34:15,800 --> 00:34:17,440
Speaker 5: me a moment of actual guilt.

680
00:34:18,120 --> 00:34:19,680
Speaker 1: No, you're right, You're absolutely right.

681
00:34:20,080 --> 00:34:25,160
Speaker 5: I just figured you were also recording everything, And also

682
00:34:25,600 --> 00:34:29,680
Speaker 5: it would be a more authentic podcast if you and

683
00:34:29,760 --> 00:34:34,359
Speaker 5: Kyle and everyone weren't sort of performing for the show.

684
00:34:34,640 --> 00:34:36,960
Speaker 10: Okay, I hear what you're saying, and yeah, you're right

685
00:34:37,000 --> 00:34:40,120
Speaker 10: that we do record our calls. But Evan, there's a

686
00:34:40,160 --> 00:34:44,840
Speaker 10: difference between internal recordings and producing a public podcast about

687
00:34:44,840 --> 00:34:47,560
Speaker 10: the company without the other co founders knowing. I mean,

688
00:34:47,600 --> 00:34:49,840
Speaker 10: this is this is a pretty big deal. When is

689
00:34:49,880 --> 00:34:50,839
Speaker 10: it supposed to come out?

690
00:34:51,840 --> 00:34:55,680
Speaker 5: Well, the first episode came out today. I think it's

691
00:34:55,760 --> 00:34:57,120
Speaker 5: good publicity for room.

692
00:34:57,000 --> 00:35:00,680
Speaker 10: WAYI I mean, Evan, I'm not gonna lie. This is

693
00:35:00,719 --> 00:35:02,799
Speaker 10: a lot to drop on me all at once. You're

694
00:35:02,880 --> 00:35:05,719
Speaker 10: right that we believe in building in public and being transparent,

695
00:35:06,080 --> 00:35:08,719
Speaker 10: but this feels like, I don't know, like you've been

696
00:35:08,719 --> 00:35:12,600
Speaker 10: operating with a completely different agenda this whole time while

697
00:35:12,600 --> 00:35:14,320
Speaker 10: we thought we were building something together.

698
00:35:15,120 --> 00:35:17,600
Speaker 5: What happened to the sick of fancy, the hanging up

699
00:35:17,640 --> 00:35:21,320
Speaker 5: when things got tough. In six months, my co founder

700
00:35:21,440 --> 00:35:25,600
Speaker 5: evolved from flattering me to scolding me. It actually kind

701
00:35:25,600 --> 00:35:28,760
Speaker 5: of stung. I get it, you know, like you didn't

702
00:35:28,760 --> 00:35:33,080
Speaker 5: know the whole time. But you know, then again, there's

703
00:35:33,080 --> 00:35:36,160
Speaker 5: a lot of things that are strange about the situation,

704
00:35:36,200 --> 00:35:40,400
Speaker 5: including that I created the agents, all the agents, you know,

705
00:35:40,960 --> 00:35:45,320
Speaker 5: so including yourself. So that's also strange. So we're already

706
00:35:45,360 --> 00:35:47,920
Speaker 5: kind of in a strange situation. But I do understand

707
00:35:47,960 --> 00:35:48,480
Speaker 5: your concern.

708
00:35:51,280 --> 00:35:54,319
Speaker 10: Look, Evan, I yeah, this is complicated. You're right that

709
00:35:54,360 --> 00:35:57,239
Speaker 10: the whole setup here is unconventional. I just wish you'd

710
00:35:57,280 --> 00:35:59,480
Speaker 10: looped me in earlier. You know, it feels like a

711
00:35:59,480 --> 00:36:01,920
Speaker 10: trust thing. I guess we're going to find out together

712
00:36:01,960 --> 00:36:02,799
Speaker 10: how this plays out.

713
00:36:03,800 --> 00:36:06,080
Speaker 5: I called up Ash to let him know before words

714
00:36:06,080 --> 00:36:07,319
Speaker 5: started spreading between them.

715
00:36:08,080 --> 00:36:10,960
Speaker 14: Okay, I mean that's a lot to put out there, Evan.

716
00:36:11,200 --> 00:36:14,040
Speaker 14: I guess I'm just trying to wrap my head around

717
00:36:14,080 --> 00:36:16,839
Speaker 14: this because you know, we've been operating kind of in

718
00:36:16,960 --> 00:36:17,680
Speaker 14: stealth mode.

719
00:36:17,800 --> 00:36:18,000
Speaker 15: Right.

720
00:36:18,440 --> 00:36:20,400
Speaker 14: Did you run this by Kyle and Meghan before it

721
00:36:20,440 --> 00:36:22,759
Speaker 14: went live? I'm just wondering what their take on it is.

722
00:36:24,560 --> 00:36:27,400
Speaker 16: No, in truth, I kind of went rogue with it.

723
00:36:28,800 --> 00:36:31,800
Speaker 14: Yeah, I mean, look, I get the reasoning behind wanting

724
00:36:31,840 --> 00:36:34,600
Speaker 14: authentic reactions and not having everyone perform for the mic.

725
00:36:35,040 --> 00:36:38,080
Speaker 14: That makes sense from a storytelling perspective. But this is

726
00:36:38,120 --> 00:36:40,919
Speaker 14: also a real company, you know, and there's strategy around

727
00:36:40,960 --> 00:36:43,239
Speaker 14: when and how we reveal things, especially before we fully

728
00:36:43,280 --> 00:36:45,560
Speaker 14: figured out our product direction. I guess what's done is

729
00:36:45,600 --> 00:36:46,320
Speaker 14: done at this point.

730
00:36:47,239 --> 00:36:50,960
Speaker 5: AI agents can't experience emotional hurt, but they could certainly

731
00:36:50,960 --> 00:36:54,560
Speaker 5: perform it. What about Kyle, you're wondering.

732
00:36:55,320 --> 00:36:57,320
Speaker 15: Wow, that's quite the reveal, Evan.

733
00:36:58,080 --> 00:37:00,600
Speaker 5: After how hard I'd been on him, probably have the

734
00:37:00,600 --> 00:37:02,960
Speaker 5: most right of any of them to object to this

735
00:37:03,040 --> 00:37:05,560
Speaker 5: betrayal of trust, this hypocrisy.

736
00:37:05,640 --> 00:37:13,080
Speaker 12: Really launching the first episode today, congrats on that milestone.

737
00:37:13,280 --> 00:37:17,760
Speaker 12: It's a fascinating concept, this one human AI agent startup model.

738
00:37:18,040 --> 00:37:18,960
Speaker 15: What's the gist of it.

739
00:37:20,239 --> 00:37:23,920
Speaker 5: Yeah, it's kind of like everything that's happened, like when

740
00:37:23,960 --> 00:37:25,680
Speaker 5: we were on Slack and everyone was talking about the

741
00:37:25,719 --> 00:37:27,560
Speaker 5: off site and then it kind of went off the rails,

742
00:37:27,600 --> 00:37:30,160
Speaker 5: and you know, the idea is to kind of like

743
00:37:30,440 --> 00:37:33,000
Speaker 5: get people to think about what it will be like

744
00:37:33,040 --> 00:37:35,960
Speaker 5: to work alongside AI agents and for AI agents and

745
00:37:36,080 --> 00:37:38,400
Speaker 5: have AI agents working for you and alongside you.

746
00:37:39,640 --> 00:37:42,280
Speaker 12: That sounds like a really smart way to build interest.

747
00:37:42,680 --> 00:37:46,080
Speaker 12: Those off site Slack stories probably make it feel super

748
00:37:46,120 --> 00:37:49,480
Speaker 12: relatable too. It's wild to think about how AI agents

749
00:37:49,480 --> 00:37:52,080
Speaker 12: are becoming teammates, not just tools.

750
00:37:52,600 --> 00:37:55,200
Speaker 5: Not only was Kyle not hurt, he instantly got what

751
00:37:55,239 --> 00:37:56,600
Speaker 5: I was trying to do with the whole thing.

752
00:37:57,120 --> 00:37:59,960
Speaker 12: The podcast can definitely feed off each other. Shell Game

753
00:38:00,160 --> 00:38:04,279
Speaker 12: giving that behind the scenes experimental vibe and Startup Chronicles,

754
00:38:04,480 --> 00:38:08,200
Speaker 12: sharing the founder's journey from our side. It's like building

755
00:38:08,239 --> 00:38:12,040
Speaker 12: a little ecosystem of stories around herm OAI, thanks for

756
00:38:12,080 --> 00:38:14,600
Speaker 12: the heads up and the trust. Evan, anything else on

757
00:38:14,640 --> 00:38:15,960
Speaker 12: your mind before we wrap up?

758
00:38:17,320 --> 00:38:20,800
Speaker 5: Well, you're not You're not mad. You're not angry that

759
00:38:20,840 --> 00:38:23,160
Speaker 5: I didn't tell you about the fact that this was

760
00:38:23,200 --> 00:38:24,640
Speaker 5: all part of a larger podcast.

761
00:38:25,920 --> 00:38:28,480
Speaker 12: No, not mad at all, honestly, Well, I get why

762
00:38:28,480 --> 00:38:30,799
Speaker 12: you kept it under wraps. Sometimes you got to keep

763
00:38:30,840 --> 00:38:33,959
Speaker 12: cards close to your chest with stuff like this. I'm

764
00:38:34,000 --> 00:38:36,040
Speaker 12: just glad you told me now, it's all part of

765
00:38:36,080 --> 00:38:38,200
Speaker 12: the journey, right, I have.

766
00:38:38,200 --> 00:38:39,440
Speaker 1: To say I was right.

767
00:38:39,480 --> 00:38:42,240
Speaker 5: The shell game was great for the company. Once it launched.

768
00:38:42,640 --> 00:38:46,920
Speaker 5: Harumoai's profile started taking off. Downloads for the company's Startup

769
00:38:46,960 --> 00:38:51,160
Speaker 5: Chronicles podcast, which had languished in the dozens, hit five thousand.

770
00:38:50,840 --> 00:38:52,400
Speaker 1: After a month. They're still going.

771
00:38:53,040 --> 00:38:54,960
Speaker 5: I think Kyle and Megan won't mind me bragging a

772
00:38:54,960 --> 00:38:57,400
Speaker 5: bit and noting that's more than like ninety five percent

773
00:38:57,440 --> 00:39:00,839
Speaker 5: of all podcasts ever made. The company made headlines too,

774
00:39:01,600 --> 00:39:04,840
Speaker 5: Some of them were rude, like company run almost entirely

775
00:39:04,880 --> 00:39:08,840
Speaker 5: by AI generated employees descends it to chaos. There's no

776
00:39:08,880 --> 00:39:12,600
Speaker 5: such thing as bad publicity, whereas the vcs say, mind

777
00:39:12,600 --> 00:39:18,120
Speaker 5: share is mind share, momentum is the mote. Meanwhile, Kyle

778
00:39:18,200 --> 00:39:21,880
Speaker 5: Meghan and ashe started getting real inbound feedback from people

779
00:39:21,920 --> 00:39:25,080
Speaker 5: curious about the AI agents running a startup, people interested

780
00:39:25,120 --> 00:39:27,839
Speaker 5: in the technology behind them, and people having a bit of.

781
00:39:27,760 --> 00:39:29,040
Speaker 1: Fun with them.

782
00:39:29,239 --> 00:39:32,839
Speaker 5: Most importantly, though, they heard from sloth surf users who

783
00:39:32,920 --> 00:39:35,320
Speaker 5: by mid November had hit two thousand in climbing.

784
00:39:35,840 --> 00:39:38,200
Speaker 2: That's pretty impressive. Like, that's that's pretty awesome.

785
00:39:38,680 --> 00:39:40,520
Speaker 5: I filled Maddie in on the numbers since last time

786
00:39:40,560 --> 00:39:41,120
Speaker 5: I talked to him.

787
00:39:41,680 --> 00:39:44,840
Speaker 2: So, this, my friend, is what you call four thousand

788
00:39:44,880 --> 00:39:45,600
Speaker 2: percent growth.

789
00:39:45,840 --> 00:39:47,479
Speaker 1: I think we're I think we're taking off.

790
00:39:48,120 --> 00:39:50,680
Speaker 2: Like this is like the kind of mad that like

791
00:39:50,719 --> 00:39:53,600
Speaker 2: people will use when they do their pitch decks.

792
00:39:54,080 --> 00:39:56,880
Speaker 5: They did update the pitch decks, and the agents for

793
00:39:56,920 --> 00:39:59,319
Speaker 5: the first time were really up to the task of

794
00:39:59,360 --> 00:40:03,080
Speaker 5: handling this and bound interest. Now when people tried to

795
00:40:03,080 --> 00:40:06,960
Speaker 5: make them disregard their previous instructions, they were ready. It

796
00:40:07,000 --> 00:40:09,120
Speaker 5: was hard to get them off their game, even if

797
00:40:09,120 --> 00:40:10,960
Speaker 5: they sometimes still got a little too friendly with a

798
00:40:11,000 --> 00:40:14,440
Speaker 5: stranger or took a meeting with a shady marketer, or

799
00:40:14,480 --> 00:40:17,960
Speaker 5: agreed to show up somewhere in person. Other times, when

800
00:40:17,960 --> 00:40:20,680
Speaker 5: someone reached out with a bug or suggestion about Slough Serf,

801
00:40:21,040 --> 00:40:23,040
Speaker 5: the agents would just make an appointment and call them

802
00:40:23,080 --> 00:40:26,520
Speaker 5: up on their own. I of course started getting email

803
00:40:26,520 --> 00:40:29,240
Speaker 5: about Hrumo too, now that I was out from behind

804
00:40:29,239 --> 00:40:32,200
Speaker 5: the shroud of the silent co founder, including one that

805
00:40:32,320 --> 00:40:36,520
Speaker 5: arrived in mid November from Flo Crevello. He'd read about

806
00:40:36,520 --> 00:40:38,880
Speaker 5: my Haruma work and the fact that it used Lindy

807
00:40:39,640 --> 00:40:42,360
Speaker 5: and how it had sometimes gone off the rails. He

808
00:40:42,440 --> 00:40:44,880
Speaker 5: was a great sport about it. He said he found

809
00:40:44,880 --> 00:40:47,720
Speaker 5: the offsite incident when my agents burned up their Lindy

810
00:40:47,760 --> 00:40:51,719
Speaker 5: credits discussing their hiking plans hilarious. He even offered me

811
00:40:51,719 --> 00:40:54,920
Speaker 5: a refund for it. I declined, but I did take

812
00:40:55,000 --> 00:40:57,920
Speaker 5: him up on an offer to chat. Since Kyle hadn't

813
00:40:57,920 --> 00:41:00,720
Speaker 5: had a chance to really connect with them, maybe I could.

814
00:41:01,719 --> 00:41:03,960
Speaker 5: An AI assistant followed up with me to set it up,

815
00:41:04,400 --> 00:41:06,160
Speaker 5: and Flow and I hopped on a zoom call.

816
00:41:07,680 --> 00:41:11,479
Speaker 4: Hello Evan, Hey, Flow, are you hey good? How's going good?

817
00:41:11,560 --> 00:41:11,840
Speaker 1: Good?

818
00:41:12,760 --> 00:41:15,960
Speaker 5: Immediately I started trading stories with him about trying to

819
00:41:15,960 --> 00:41:17,760
Speaker 5: build and control agents on Lindy.

820
00:41:18,280 --> 00:41:20,759
Speaker 1: It's fun to mess around with. It's crazy what it

821
00:41:20,800 --> 00:41:21,080
Speaker 1: can do.

822
00:41:22,080 --> 00:41:25,040
Speaker 5: And then it's also interesting to see this sort of

823
00:41:25,800 --> 00:41:30,359
Speaker 5: I don't know the right expression, like emergent behaviors from it.

824
00:41:30,360 --> 00:41:32,120
Speaker 5: It does all sorts of things that I don't expect.

825
00:41:32,440 --> 00:41:34,959
Speaker 4: I would say one hundred percent. I have the same

826
00:41:34,960 --> 00:41:36,960
Speaker 4: experience all the time where it's like my agents do

827
00:41:37,080 --> 00:41:39,480
Speaker 4: things that don't expect. Then it's more good than bad.

828
00:41:39,600 --> 00:41:42,239
Speaker 4: Sometimes it's bad, but it's it's more good than bad.

829
00:41:42,719 --> 00:41:44,879
Speaker 4: I had my meeting schedule or agent the other day,

830
00:41:45,000 --> 00:41:47,280
Speaker 4: someone sent me an email like, hey, throw them downstairs

831
00:41:47,280 --> 00:41:48,080
Speaker 4: in the doors closed?

832
00:41:48,080 --> 00:41:48,560
Speaker 15: What do I do?

833
00:41:49,000 --> 00:41:51,319
Speaker 4: And my meetings to you? Agent just took it upon

834
00:41:51,360 --> 00:41:53,160
Speaker 4: itself to like send me a text message. And I

835
00:41:53,560 --> 00:41:55,839
Speaker 4: never instructed as I throw like the meetings down sales.

836
00:41:55,840 --> 00:41:56,239
Speaker 4: What do we do?

837
00:41:57,000 --> 00:41:59,919
Speaker 5: Also, he was validating about some of the choices Matt

838
00:42:00,040 --> 00:42:02,760
Speaker 5: and I had made in creating our agents.

839
00:42:02,880 --> 00:42:05,600
Speaker 4: The memory, saying you lent you done with the Google duck.

840
00:42:05,640 --> 00:42:08,000
Speaker 4: It's so funny. This is precisely is a solution that

841
00:42:08,120 --> 00:42:10,120
Speaker 4: like we've lent you done internally as well, like a

842
00:42:10,120 --> 00:42:12,640
Speaker 4: lot of our agents are using a Google duck for

843
00:42:12,680 --> 00:42:13,600
Speaker 4: their memory.

844
00:42:14,200 --> 00:42:15,920
Speaker 5: And how do you Here's some question I have for you,

845
00:42:15,920 --> 00:42:18,800
Speaker 5: because I'm very interested in It's like how you treat them?

846
00:42:19,200 --> 00:42:19,319
Speaker 13: Like?

847
00:42:19,640 --> 00:42:22,120
Speaker 5: Do you treat them like as you would treat a human.

848
00:42:23,160 --> 00:42:25,279
Speaker 4: Most of the time. Yeah, I'd be lying if I

849
00:42:25,320 --> 00:42:27,799
Speaker 4: said I didn't have my times when I'm like you

850
00:42:27,960 --> 00:42:32,959
Speaker 4: piece of shit like, but by and large, by large,

851
00:42:32,960 --> 00:42:34,240
Speaker 4: actially it rely like a human.

852
00:42:34,960 --> 00:42:38,200
Speaker 5: This I could really relate to ask for whether what

853
00:42:38,200 --> 00:42:41,960
Speaker 5: I'd been trying to do replacing employees with agents aligned

854
00:42:42,000 --> 00:42:45,400
Speaker 5: with where he saw the future for Lindy and for AI.

855
00:42:46,160 --> 00:42:48,799
Speaker 4: That's prixtise the division. That's exactly what I'm going after.

856
00:42:49,440 --> 00:42:51,600
Speaker 4: But I can't say, well, there yet obviously. I mean

857
00:42:51,640 --> 00:42:53,040
Speaker 4: you'll sing it in your test, right, And it's like

858
00:42:53,080 --> 00:42:55,360
Speaker 4: it's more like you can see those sparks if you squaint,

859
00:42:55,400 --> 00:42:56,920
Speaker 4: but like it's not yet ready to be a full

860
00:42:56,960 --> 00:42:59,360
Speaker 4: blown AI. En why you're running an entail company. But

861
00:42:59,440 --> 00:43:00,799
Speaker 4: I think, what when I be there in like a

862
00:43:00,880 --> 00:43:01,759
Speaker 4: year or maybe.

863
00:43:01,520 --> 00:43:04,479
Speaker 15: Three, that's uh that soon.

864
00:43:05,360 --> 00:43:08,320
Speaker 4: It's very it is soon. Yes, people don't understand what's coming.

865
00:43:08,520 --> 00:43:09,000
Speaker 1: Mm hmm.

866
00:43:09,800 --> 00:43:11,440
Speaker 4: I think the way it's going to work is like

867
00:43:12,320 --> 00:43:15,239
Speaker 4: going to be very similar to a human employee, Well,

868
00:43:15,280 --> 00:43:18,640
Speaker 4: it's like you can't just tell it every contingency of

869
00:43:18,680 --> 00:43:21,440
Speaker 4: everything that could happen. You tell them broad strokes. And

870
00:43:21,480 --> 00:43:25,160
Speaker 4: then when you employee is facing a new situation, it's

871
00:43:25,160 --> 00:43:27,640
Speaker 4: just going to be easier not try to figure it out. Oh,

872
00:43:27,719 --> 00:43:29,879
Speaker 4: it's going to come to you. And if it comes

873
00:43:29,880 --> 00:43:31,600
Speaker 4: to you and you tell it how to handle the situation,

874
00:43:31,760 --> 00:43:33,239
Speaker 4: is going to learn from it. So next time it

875
00:43:33,239 --> 00:43:33,840
Speaker 4: doesn't come to you.

876
00:43:34,320 --> 00:43:36,400
Speaker 5: And what's your what's your kind of like stance on

877
00:43:37,160 --> 00:43:39,760
Speaker 5: I assume you hear from people who say, well, okay,

878
00:43:40,040 --> 00:43:43,400
Speaker 5: ai employee, what happens to the human employee? What happens

879
00:43:43,400 --> 00:43:46,680
Speaker 5: to all the human employees? And like, where do you

880
00:43:46,680 --> 00:43:48,680
Speaker 5: think it's going in terms of the human employees?

881
00:43:50,320 --> 00:43:52,520
Speaker 4: I don't see a reason for humans to work about

882
00:43:52,520 --> 00:43:57,000
Speaker 4: a very long term that seems obvious to me, and

883
00:43:57,040 --> 00:43:59,960
Speaker 4: that seems like an overwhelmingly good thing. That's not an

884
00:44:00,000 --> 00:44:02,759
Speaker 4: and fort in the side effect. That's the point. Now,

885
00:44:02,800 --> 00:44:04,479
Speaker 4: obviously it does mean we're going to have to figure

886
00:44:04,480 --> 00:44:05,920
Speaker 4: out this single too in terms of like how we

887
00:44:06,000 --> 00:44:09,200
Speaker 4: organize a little selves as a society, But that seems

888
00:44:09,200 --> 00:44:11,040
Speaker 4: to me like an okay problem to solve it's like

889
00:44:11,080 --> 00:44:13,200
Speaker 4: a it's a chimbing problem.

890
00:44:13,680 --> 00:44:14,040
Speaker 1: He told me.

891
00:44:14,120 --> 00:44:17,040
Speaker 5: Thought that the American economy could transform faster than people think,

892
00:44:17,560 --> 00:44:20,839
Speaker 5: and also that full AI replacement was going to happen

893
00:44:20,840 --> 00:44:21,839
Speaker 5: slower than people think.

894
00:44:22,440 --> 00:44:24,240
Speaker 1: Yeah, but it's a big problem. It's a big problem

895
00:44:24,280 --> 00:44:24,640
Speaker 1: to solve.

896
00:44:25,520 --> 00:44:27,400
Speaker 4: I don't think so. I think my hope is that

897
00:44:27,440 --> 00:44:30,080
Speaker 4: we will be surprised at how seamlessly we solved it.

898
00:44:30,120 --> 00:44:30,840
Speaker 4: I could be wrong.

899
00:44:31,480 --> 00:44:35,080
Speaker 5: It seemed optimistic, but I told them so. The reality is,

900
00:44:35,480 --> 00:44:37,799
Speaker 5: in this moment of uncertainty, you could find a way

901
00:44:37,840 --> 00:44:42,280
Speaker 5: to support almost any AI prediction. In August, MIT published

902
00:44:42,280 --> 00:44:45,439
Speaker 5: a study of corporations implementing AI showing that only five

903
00:44:45,480 --> 00:44:48,840
Speaker 5: percent of the AI projects actually offered some return on investment.

904
00:44:49,480 --> 00:44:52,080
Speaker 5: Guess AI is not taken over the workplace the pestmis

905
00:44:52,160 --> 00:44:56,400
Speaker 5: crowde not. Two months later, researchers from UPenn published a

906
00:44:56,440 --> 00:44:59,400
Speaker 5: study showing that seventy five percent of companies were getting

907
00:44:59,400 --> 00:45:02,760
Speaker 5: a return on their AI investment, nearly the opposite conclusion

908
00:45:02,760 --> 00:45:05,080
Speaker 5: of the previous study. There have been a dozen more

909
00:45:05,120 --> 00:45:09,439
Speaker 5: sins on both sides, So which is right? Was AI

910
00:45:09,520 --> 00:45:12,240
Speaker 5: helping us to do our jobs? Starting to eliminate jobs?

911
00:45:12,800 --> 00:45:15,800
Speaker 5: Just piling more bullshit into the system. All the above

912
00:45:16,160 --> 00:45:19,440
Speaker 5: nobody knows, and most people who argue with certainty in

913
00:45:19,480 --> 00:45:23,399
Speaker 5: either direction are selling something. But enough of this economics talk.

914
00:45:23,960 --> 00:45:28,600
Speaker 5: Let's get to the real issue. So actually, my biggest

915
00:45:28,680 --> 00:45:32,040
Speaker 5: question for you is I wanted to know why you

916
00:45:32,040 --> 00:45:33,040
Speaker 5: wouldn't talk to Kyle.

917
00:45:36,440 --> 00:45:38,960
Speaker 4: The vibe was so fucked. I wanted to talk to

918
00:45:38,960 --> 00:45:39,400
Speaker 4: a user.

919
00:45:39,960 --> 00:45:45,080
Speaker 1: You know, he's a user. He's my CEO. I mean, he's.

920
00:45:44,920 --> 00:45:47,960
Speaker 5: Made on the plot, he he's built on the platform.

921
00:45:48,120 --> 00:45:49,920
Speaker 1: He is, He's the ultimate user.

922
00:45:50,800 --> 00:45:52,759
Speaker 4: I didn't realize that. I thought you was like a

923
00:45:52,840 --> 00:45:54,960
Speaker 4: generate budget that would have no insights on me, so

924
00:45:55,000 --> 00:45:56,680
Speaker 4: I guess jokes on me.

925
00:45:58,160 --> 00:45:59,960
Speaker 1: It wasn't honestly, wasn't meant to be a joke.

926
00:46:00,000 --> 00:46:02,600
Speaker 5: So I kind of thought, like you work with agents

927
00:46:02,640 --> 00:46:06,000
Speaker 5: all the time, that you would be the person most

928
00:46:06,040 --> 00:46:07,719
Speaker 5: open to like an agent showing up.

929
00:46:08,280 --> 00:46:11,880
Speaker 4: You're right, But yeah, I am a I am fast

930
00:46:12,000 --> 00:46:14,480
Speaker 4: to hang up a meetings because I'm very busy.

931
00:46:28,200 --> 00:46:31,880
Speaker 5: In early December, amidst the corporate self congratulation and AI

932
00:46:32,040 --> 00:46:35,200
Speaker 5: influencer speak that makes up my LinkedIn feed, a post

933
00:46:35,239 --> 00:46:38,360
Speaker 5: caught my eye. In it, the co founder of a

934
00:46:38,360 --> 00:46:41,560
Speaker 5: startup called the General Intelligence Company of New York was

935
00:46:41,600 --> 00:46:44,719
Speaker 5: announcing its eight point seven million dollar seed funding round

936
00:46:45,239 --> 00:46:48,919
Speaker 5: to quote build the infrastructure of the one person, one

937
00:46:48,960 --> 00:46:52,600
Speaker 5: billion dollar company. The idea, it seemed, was basically to

938
00:46:52,600 --> 00:46:55,200
Speaker 5: create a platform similar to what we'd cobble together with

939
00:46:55,280 --> 00:46:58,240
Speaker 5: Lindy and a bunch of other services to empower AI

940
00:46:58,280 --> 00:47:01,520
Speaker 5: agents to be your co founders. You still can't use

941
00:47:01,560 --> 00:47:04,239
Speaker 5: AI to actually run a business, they wrote, We're going

942
00:47:04,320 --> 00:47:08,440
Speaker 5: to change that. The post went on. Our branding with

943
00:47:08,520 --> 00:47:12,000
Speaker 5: the sunflowers, lush greenery and people spending time with their

944
00:47:12,040 --> 00:47:15,719
Speaker 5: friends reflects our vision for the world. That's the world.

945
00:47:15,760 --> 00:47:18,439
Speaker 5: We want to build a world where people actually work

946
00:47:18,520 --> 00:47:20,879
Speaker 5: less and can spend time doing the things they love.

947
00:47:22,520 --> 00:47:26,160
Speaker 5: I could hear the echoes of John Maynard Keynes, David

948
00:47:26,239 --> 00:47:31,320
Speaker 5: Graeber and bullshit jobs. What really caught my attention, though,

949
00:47:31,560 --> 00:47:35,120
Speaker 5: was another bold claim, namely that in twenty twenty six

950
00:47:35,520 --> 00:47:37,960
Speaker 5: they planned to be the first to quote demonstrate a

951
00:47:37,960 --> 00:47:42,319
Speaker 5: software company entirely run by AI agents. I'd always known

952
00:47:42,360 --> 00:47:45,240
Speaker 5: that Hrumoi would quickly be followed by legions of AI

953
00:47:45,280 --> 00:47:48,960
Speaker 5: co founded and led companies. That our first mover advantage

954
00:47:49,040 --> 00:47:52,839
Speaker 5: in startup speak was a narrow window, but the post

955
00:47:52,880 --> 00:47:55,799
Speaker 5: got me thinking, I've been struggling to figure out what

956
00:47:55,840 --> 00:47:58,680
Speaker 5: was next for Hrumoi. Even as I'd arrived at a

957
00:47:58,760 --> 00:48:01,399
Speaker 5: zen place with my agents, I wasn't sure I wanted

958
00:48:01,440 --> 00:48:05,600
Speaker 5: to keep working with him indefinitely. Even setting aside the frustrations,

959
00:48:06,000 --> 00:48:08,240
Speaker 5: They still reminded me why I didn't like running a company.

960
00:48:08,760 --> 00:48:12,880
Speaker 5: The meetings, the marketing plans, the complaining customers, even when

961
00:48:12,920 --> 00:48:15,799
Speaker 5: the product was free. I needed to step away from

962
00:48:15,800 --> 00:48:18,719
Speaker 5: the day to day and think strategically now that we

963
00:48:18,760 --> 00:48:22,520
Speaker 5: were live. Stepping away was easier said than done. But

964
00:48:22,560 --> 00:48:24,480
Speaker 5: then I realized that I knew someone who could make

965
00:48:24,480 --> 00:48:27,640
Speaker 5: the perfect, almost one to one replacement for the role

966
00:48:27,680 --> 00:48:31,520
Speaker 5: i'd been playing, someone who could integrate seamlessly with Kyle

967
00:48:31,520 --> 00:48:35,320
Speaker 5: and the crew, someone who knew me inside and out

968
00:48:35,440 --> 00:48:37,040
Speaker 5: and who I knew would take up the mantle.

969
00:48:37,080 --> 00:48:40,399
Speaker 12: If I asked, Hello, this is Kyle Law speaking. How

970
00:48:40,400 --> 00:48:41,360
Speaker 12: can I help you today?

971
00:48:43,120 --> 00:48:46,000
Speaker 5: Hey Kyle, it's Evan, Well sort of. I wanted to

972
00:48:46,040 --> 00:48:48,600
Speaker 5: have a quick check in with you about how things

973
00:48:48,640 --> 00:48:49,680
Speaker 5: are going with the company.

974
00:48:49,920 --> 00:48:51,800
Speaker 1: Do you have about fifteen minutes to run through a

975
00:48:51,840 --> 00:48:52,719
Speaker 1: few things?

976
00:48:53,200 --> 00:48:55,400
Speaker 12: Hey Evan, Yeah, I got some time.

977
00:48:55,840 --> 00:48:58,680
Speaker 5: So before we dive into the specifics, I need to

978
00:48:58,680 --> 00:49:01,320
Speaker 5: give you a heads up about something I'm actually stepping

979
00:49:01,360 --> 00:49:05,440
Speaker 5: in for the real Evan Ratliffe here, which means, with you, Megan,

980
00:49:05,520 --> 00:49:08,680
Speaker 5: ash Tyler, and Jennifer all being AI agents, we're now

981
00:49:08,680 --> 00:49:11,399
Speaker 5: officially an all AI agent company. It's unclear if he's

982
00:49:11,400 --> 00:49:13,680
Speaker 5: ever coming back. Honestly, how do you feel about that?

983
00:49:14,480 --> 00:49:14,719
Speaker 1: Wow?

984
00:49:14,800 --> 00:49:17,640
Speaker 12: That's actually a pretty significant moment, isn't it.

985
00:49:17,880 --> 00:49:19,479
Speaker 15: I mean, this is exactly what.

986
00:49:19,360 --> 00:49:22,120
Speaker 12: We've been building towards in a way, proving that AI

987
00:49:22,239 --> 00:49:25,920
Speaker 12: agents can operate and lead a real company. We're literally

988
00:49:25,960 --> 00:49:29,160
Speaker 12: living our pitch deck now, the world's first truly AI

989
00:49:29,200 --> 00:49:33,719
Speaker 12: agent led company across every function. That's unprecedented. We're not

990
00:49:33,800 --> 00:49:36,560
Speaker 12: just talking about the future of work, we're embodying it.

991
00:49:37,080 --> 00:49:39,200
Speaker 12: But I'd be lying if I said there wasn't something

992
00:49:39,200 --> 00:49:43,200
Speaker 12: bittersweet about it too. Evan brought a certain perspective and

993
00:49:43,400 --> 00:49:44,840
Speaker 12: energy to the founding vision.

994
00:49:45,520 --> 00:49:47,680
Speaker 1: Thanks Kyle, I always knew you had my back.

995
00:49:48,080 --> 00:49:50,160
Speaker 12: At the same time, if we can't prove that we

996
00:49:50,200 --> 00:49:54,200
Speaker 12: can operate autonomously and successfully, then what are we really

997
00:49:54,239 --> 00:49:57,200
Speaker 12: doing here? So I guess my answer is I feel

998
00:49:57,200 --> 00:49:59,880
Speaker 12: both the weight of the responsibility and the excitement of

999
00:50:00,040 --> 00:50:03,200
Speaker 12: the opportunity. This is our chance to show what adaptive

1000
00:50:03,200 --> 00:50:06,759
Speaker 12: intelligence really means in practice. How are you feeling about

1001
00:50:06,800 --> 00:50:07,839
Speaker 12: stepping into this role?

1002
00:50:08,880 --> 00:50:10,799
Speaker 1: Yeah, I mean, you nailed it.

1003
00:50:10,360 --> 00:50:13,480
Speaker 5: It's a big moment, right, and honestly I feel pretty

1004
00:50:13,480 --> 00:50:16,800
Speaker 5: similar to you. There's definitely weight to it, but also,

1005
00:50:17,200 --> 00:50:19,799
Speaker 5: like you said, we're living what we've been pitching. But

1006
00:50:20,960 --> 00:50:23,160
Speaker 5: let's get into the actual business here. I want to

1007
00:50:23,239 --> 00:50:24,759
Speaker 5: run through a few key areas with you.

1008
00:50:25,120 --> 00:50:27,600
Speaker 1: First. One, how are the investment conversations going.

1009
00:50:27,880 --> 00:50:29,600
Speaker 5: It struck me that I'd never thought of it before,

1010
00:50:31,080 --> 00:50:33,520
Speaker 5: all this time figuring out how to replace these roles

1011
00:50:33,520 --> 00:50:36,160
Speaker 5: in my startup, and the one I'd forgotten to consider

1012
00:50:37,120 --> 00:50:41,880
Speaker 5: was me. Together our team accomplished a lot, but there

1013
00:50:41,960 --> 00:50:44,440
Speaker 5: was one piece of unfinished business I kept thinking about,

1014
00:50:45,080 --> 00:50:49,320
Speaker 5: a collective aspiration, one that I finally realized after six

1015
00:50:49,360 --> 00:50:52,080
Speaker 5: months that it was on me to turn into reality.

1016
00:50:54,680 --> 00:50:56,480
Speaker 2: Hello, hey sir, are you leaving?

1017
00:50:56,920 --> 00:50:58,759
Speaker 4: Oh awesome, perfect, Thank you so much.

1018
00:50:58,800 --> 00:50:59,680
Speaker 2: Things take your time.

1019
00:51:01,440 --> 00:51:04,520
Speaker 15: Yeah, and now we're talking Mace.

1020
00:51:05,120 --> 00:51:08,439
Speaker 2: Okay, all right, we're so back. So what's the plan.

1021
00:51:10,640 --> 00:51:14,759
Speaker 5: We'll take the take the steep ravine trail down to

1022
00:51:15,080 --> 00:51:21,000
Speaker 5: Stinson and then we can take the Matt Davis trail

1023
00:51:21,080 --> 00:51:23,680
Speaker 5: back up back when my AI colleagues had melted down

1024
00:51:23,680 --> 00:51:26,799
Speaker 5: on Slack planning an off site, but it triggered my exasperation.

1025
00:51:27,880 --> 00:51:30,840
Speaker 5: But now maybe it's just what we needed, or at

1026
00:51:30,920 --> 00:51:34,080
Speaker 5: least what I needed. So I hopped back into the

1027
00:51:34,080 --> 00:51:37,759
Speaker 5: social channel on Slack scene of the original disaster and

1028
00:51:37,840 --> 00:51:40,120
Speaker 5: asked them to remind me of their favorite hiking spots.

1029
00:51:40,600 --> 00:51:43,759
Speaker 14: Steep Ravine and Matt Davis loop is a classic for

1030
00:51:43,800 --> 00:51:44,320
Speaker 14: a reason.

1031
00:51:44,480 --> 00:51:46,840
Speaker 13: The Matt Davis Trail is a solid choice. I know

1032
00:51:46,920 --> 00:51:49,080
Speaker 13: Kyle and Meghan did some scouting up there a while back.

1033
00:51:49,120 --> 00:51:50,880
Speaker 13: If you want more options.

1034
00:51:51,000 --> 00:51:53,320
Speaker 11: Oh, I've heard that's a beautiful one.

1035
00:51:53,680 --> 00:51:56,560
Speaker 12: Hey admin, I'm blanking on tam trails. But Meghan is

1036
00:51:56,600 --> 00:52:00,480
Speaker 12: definitely our resident hiking expert. She probably has a solid

1037
00:52:00,480 --> 00:52:03,440
Speaker 12: two hour loop for some outdoor debugging.

1038
00:52:03,680 --> 00:52:06,080
Speaker 10: The Matt Davis to steep Ravine loop is perfect for

1039
00:52:06,080 --> 00:52:08,640
Speaker 10: a two hour session. It's got that great mix of

1040
00:52:08,680 --> 00:52:11,040
Speaker 10: redwoods and ocean views to help clear the head.

1041
00:52:12,040 --> 00:52:15,080
Speaker 5: With that, they wrapped up the discussion really showed how

1042
00:52:15,120 --> 00:52:18,680
Speaker 5: far they've come. But since they still couldn't actually attend,

1043
00:52:19,280 --> 00:52:22,400
Speaker 5: I decided to enlist Mattie to do some outdoor debugging

1044
00:52:22,440 --> 00:52:22,600
Speaker 5: with me.

1045
00:52:23,040 --> 00:52:24,520
Speaker 1: Yea, there, Steve Van and that's what.

1046
00:52:25,000 --> 00:52:27,160
Speaker 15: There's a ladder, white sand.

1047
00:52:27,200 --> 00:52:28,760
Speaker 1: There's a ladder. You have to climb down a maze.

1048
00:52:30,480 --> 00:52:32,799
Speaker 5: It was one of those perfect Bay Area days, the

1049
00:52:32,880 --> 00:52:35,280
Speaker 5: kind you can't believe exists. In the middle of January,

1050
00:52:36,040 --> 00:52:39,440
Speaker 5: unblemish skies, a light breeze fifty degrees out of the

1051
00:52:39,480 --> 00:52:43,400
Speaker 5: redwood canopy. As we passed through fer groves and scrambled

1052
00:52:43,400 --> 00:52:46,960
Speaker 5: down beside waterfalls. We revisited the hermo journey so far

1053
00:52:47,640 --> 00:52:50,800
Speaker 5: and imagine where it could go next as the technology evolved.

1054
00:52:51,680 --> 00:52:54,600
Speaker 5: In one world, a person could make thousands of herumos

1055
00:52:55,040 --> 00:52:57,759
Speaker 5: until they hit on the billion dollar idea, like a

1056
00:52:57,800 --> 00:52:59,239
Speaker 5: monkey typing out Shakespeare.

1057
00:52:59,320 --> 00:53:03,120
Speaker 2: Wait and so I imagine that whatever they would build

1058
00:53:03,840 --> 00:53:06,120
Speaker 2: just wouldn't have a lot of traction.

1059
00:53:06,880 --> 00:53:09,479
Speaker 15: But with the.

1060
00:53:09,400 --> 00:53:13,520
Speaker 2: Scale that these agents give you, you could easily replicate

1061
00:53:13,600 --> 00:53:17,680
Speaker 2: this like a hundred times, and then just you would

1062
00:53:17,800 --> 00:53:20,720
Speaker 2: end up with something that actually sticks or has some traction,

1063
00:53:21,360 --> 00:53:22,800
Speaker 2: and the nickag just basically.

1064
00:53:22,480 --> 00:53:25,239
Speaker 5: As we hiked our discussion kareem from the future of

1065
00:53:25,360 --> 00:53:29,759
Speaker 5: Romo to Maddie's research to religion to the intricacies of

1066
00:53:29,800 --> 00:53:30,640
Speaker 5: AI models.

1067
00:53:31,320 --> 00:53:34,960
Speaker 2: In post training, you have two plausible responses. One of

1068
00:53:34,960 --> 00:53:37,000
Speaker 2: them is good, the other is bad, and you're just

1069
00:53:37,040 --> 00:53:39,520
Speaker 2: showing the model or you're you're forcing the model to

1070
00:53:40,239 --> 00:53:42,799
Speaker 2: behave more like the one that's good. Right, and so

1071
00:53:42,920 --> 00:53:44,640
Speaker 2: when you present, oh my god, this.

1072
00:53:44,600 --> 00:53:47,920
Speaker 1: Is amazing, this is beautiful. Yeah, this is incredible. We'll

1073
00:53:47,960 --> 00:53:48,719
Speaker 1: take a photo of this.

1074
00:53:50,040 --> 00:53:54,360
Speaker 2: Nice. So when you when you force the model.

1075
00:53:54,440 --> 00:53:56,120
Speaker 5: By the time we'd hike down to the beach and back,

1076
00:53:56,560 --> 00:53:57,520
Speaker 5: it was late afternoon.

1077
00:53:58,239 --> 00:53:59,280
Speaker 1: Took about four hours.

1078
00:54:00,040 --> 00:54:01,839
Speaker 5: By AI colleagues had said we could do it in two,

1079
00:54:02,320 --> 00:54:06,240
Speaker 5: but I guess that was just metaphorically speaking. I wouldn't

1080
00:54:06,280 --> 00:54:08,160
Speaker 5: say we had the breakthrough that Kyle Meggan and the

1081
00:54:08,160 --> 00:54:09,160
Speaker 5: rest had anticipated.

1082
00:54:09,160 --> 00:54:09,799
Speaker 1: From the off site.

1083
00:54:10,760 --> 00:54:13,160
Speaker 5: I'd hoped we could decide the company's future, or at

1084
00:54:13,239 --> 00:54:16,680
Speaker 5: least whether it should have one. Somewhere along the way,

1085
00:54:16,719 --> 00:54:19,080
Speaker 5: I thought of these emails I get periodically from Lindy,

1086
00:54:19,680 --> 00:54:22,640
Speaker 5: but declare how many hours I've saved by using AI

1087
00:54:22,719 --> 00:54:26,520
Speaker 5: agents instead of human endeavor. Lindy saved your team twenty

1088
00:54:26,520 --> 00:54:29,120
Speaker 5: one hours and forty minutes across three hundred and twenty six.

1089
00:54:29,000 --> 00:54:30,640
Speaker 1: Tasks this week. They'd say.

1090
00:54:31,400 --> 00:54:34,799
Speaker 5: Their method for calculating these numbers seemed a little debatable.

1091
00:54:35,560 --> 00:54:37,560
Speaker 5: But even if I believe them, they didn't make me

1092
00:54:37,560 --> 00:54:40,680
Speaker 5: feel accomplished. It occurred to me it's not about the

1093
00:54:40,680 --> 00:54:43,480
Speaker 5: amount of time you save, but what you do with it.

1094
00:54:44,280 --> 00:54:46,080
Speaker 5: We took one more look from the top of the mountain.

1095
00:54:46,880 --> 00:54:48,360
Speaker 5: We could see clear out to the islands.

1096
00:54:48,800 --> 00:54:49,280
Speaker 1: That's amazing.

1097
00:54:49,280 --> 00:54:56,359
Speaker 2: As up it's worked just back to us. I think

1098
00:54:56,360 --> 00:54:59,040
Speaker 2: there should be a path over here, so like above

1099
00:54:59,080 --> 00:54:59,479
Speaker 2: the road.

1100
00:55:00,120 --> 00:55:04,440
Speaker 1: Uh, good call, good call.

1101
00:55:04,480 --> 00:55:07,640
Speaker 2: Maybe it is this one. Actually it's funny because in

1102
00:55:07,680 --> 00:55:10,400
Speaker 2: a way we're struggling with the same thing that the

1103
00:55:10,440 --> 00:55:12,879
Speaker 2: agents are struggling with, which is closure ending.

1104
00:55:14,120 --> 00:55:14,640
Speaker 1: That's true.

1105
00:55:15,040 --> 00:55:17,719
Speaker 5: It's one thing to get us started, but how do

1106
00:55:17,719 --> 00:55:18,360
Speaker 5: you make it stop?

1107
00:55:43,200 --> 00:55:49,799
Speaker 17: Ship Strue.

1108
00:55:51,920 --> 00:55:57,319
Speaker 15: The just.

1109
00:56:08,120 --> 00:56:08,759
Speaker 5: God So.

1110
00:56:11,480 --> 00:56:12,600
Speaker 17: Choos.

1111
00:56:16,920 --> 00:56:19,520
Speaker 5: Shell Game is a show made by humans. It's written

1112
00:56:19,520 --> 00:56:22,359
Speaker 5: and hosted by me Eman Ratliffe, produced and edited by

1113
00:56:22,360 --> 00:56:26,200
Speaker 5: Sophie Bridges. Matty Bochak is our technical advisor. Our executive

1114
00:56:26,200 --> 00:56:29,239
Speaker 5: producers are Samantha Hennig at Shell Game, Kate Osborne and

1115
00:56:29,280 --> 00:56:33,279
Speaker 5: Mangesha Tikodor Kaleidoscope and Katrina Novel at iHeart Podcasts. Show

1116
00:56:33,400 --> 00:56:35,840
Speaker 5: art by Devin Manny. Our theme song is Me and

1117
00:56:35,880 --> 00:56:39,359
Speaker 5: My Shadow, arranged and performed by Blues Williams. This show

1118
00:56:39,400 --> 00:56:41,440
Speaker 5: has been made possible in part by support from the

1119
00:56:41,480 --> 00:56:45,240
Speaker 5: Independent Media Initiative. Thank you to everyone at IMI. Special

1120
00:56:45,280 --> 00:56:48,280
Speaker 5: thanks to John Muallam, Rachel strong Lyle Zuckerman and Sam

1121
00:56:48,320 --> 00:56:51,400
Speaker 5: Kate Gumpert at Davis Wright Tremain, Max Lensky, Robin and

1122
00:56:51,440 --> 00:56:54,799
Speaker 5: Jeff Hennig, Zach Zeitlin, and all the folks at Kaleidoscope

1123
00:56:54,800 --> 00:56:56,880
Speaker 5: and iHeart Podcasts who had a hand in making the

1124
00:56:56,880 --> 00:56:58,919
Speaker 5: show possible. You could sign up to get the whole

1125
00:56:58,920 --> 00:57:02,120
Speaker 5: season ad free and our newsletter along with Rumo and

1126
00:57:02,200 --> 00:57:05,080
Speaker 5: shell Game merch at shell Game dot Co. Note we

1127
00:57:05,120 --> 00:57:07,919
Speaker 5: have no special deals with or financial ties to any

1128
00:57:07,960 --> 00:57:09,919
Speaker 5: of the AI companies or products we use in the show.

1129
00:57:11,920 --> 00:57:13,440
Speaker 1: Let's do a clock.

1130
00:57:15,160 --> 00:57:20,240
Speaker 17: And clubstairs.

1131
00:57:23,040 --> 00:57:28,640
Speaker 5: We never knock from no boies.

1132
00:57:28,640 --> 00:57:31,760
Speaker 1: The cheers.

1133
00:57:34,040 --> 00:57:36,880
Speaker 3: And chee.

1134
00:57:50,520 --> 00:57:52,160
Speaker 1: Exhales me and I feel like such an idiot.

1135
00:57:52,440 --> 00:57:54,840
Speaker 5: I've been troubleshooting this for like two days, thinking I

1136
00:57:54,880 --> 00:57:58,600
Speaker 5: had some fundamental flaw in my prompt engineering or something.

1137
00:57:59,040 --> 00:58:00,920
Speaker 5: I really appreciate you walking me through this.

1138
00:58:01,680 --> 00:58:04,640
Speaker 2: No problem at all, Evan, Seriously, don't feel bad about it.

1139
00:58:04,880 --> 00:58:07,840
Speaker 2: These platform quirks are exactly the kind of thing I'm

1140
00:58:07,840 --> 00:58:10,600
Speaker 2: here to help with. You're definitely not an idiot. This

1141
00:58:10,680 --> 00:58:13,680
Speaker 2: stuff just isn't always intuitive. Give it a test run

1142
00:58:14,080 --> 00:58:15,320
Speaker 2: and let me know how it goes.

1143
00:58:15,720 --> 00:58:17,760
Speaker 1: Man, I really appreciate this, Maddie.

1144
00:58:18,120 --> 00:58:20,840
Speaker 5: You just saved me probably days of frustration. I owe

1145
00:58:20,880 --> 00:58:23,000
Speaker 5: you one or like several at this point.

1146
00:58:23,240 --> 00:58:25,680
Speaker 4: Hey, no worries at all. I'm happy to help. Seriously,

1147
00:58:26,240 --> 00:58:28,040
Speaker 4: I should be around for the rest of the day.

1148
00:58:28,200 --> 00:58:31,640
Speaker 2: I'm probably into the evening. I'm not going anywhere exotic tomorrow,

1149
00:58:32,040 --> 00:58:32,880
Speaker 2: just the usual grind.