1
00:00:00,920 --> 00:00:03,360
Speaker 1: First Contact with Lori Siegel is a production of Dot

2
00:00:03,360 --> 00:00:10,800
Speaker 1: Dot Dot Media and iHeartRadio. Okay, we just got an

3
00:00:10,800 --> 00:00:14,800
Speaker 1: email from us who are interviewing. I'm interviewing tomorrow, and

4
00:00:14,840 --> 00:00:17,360
Speaker 1: he says, Hi, here's some data we got from the

5
00:00:17,400 --> 00:00:21,680
Speaker 1: conversation that's exciting. Some are screenshots of the app, which

6
00:00:21,680 --> 00:00:23,800
Speaker 1: I can give you a demo tomorrow, and other data

7
00:00:23,840 --> 00:00:25,599
Speaker 1: from our system we don't show in the app to

8
00:00:25,600 --> 00:00:27,920
Speaker 1: give you a sense of the underlying analytics. Are you

9
00:00:27,960 --> 00:00:31,280
Speaker 1: ready for the underlying analytics of our connection?

10
00:00:32,240 --> 00:00:32,440
Speaker 2: Yeah?

11
00:00:32,560 --> 00:00:35,040
Speaker 3: Because I honestly have no idea what type of insights

12
00:00:35,040 --> 00:00:35,840
Speaker 3: they're able to clean.

13
00:00:36,440 --> 00:00:38,880
Speaker 1: It's loading, it's taking a while.

14
00:00:39,880 --> 00:00:46,480
Speaker 2: Built in suspense, Okay, what.

15
00:00:53,920 --> 00:00:56,880
Speaker 1: Think about this for a second and don't judge too quickly.

16
00:00:57,640 --> 00:01:01,600
Speaker 1: Imagine an AI assistant that's on top of your text messages.

17
00:01:02,240 --> 00:01:05,880
Speaker 1: It reads your conversations, It turns your messages into a

18
00:01:05,959 --> 00:01:10,640
Speaker 1: novel of data points, basically a human psychology report, and

19
00:01:10,680 --> 00:01:14,600
Speaker 1: then it guides you. The assistant can say, Hey, the

20
00:01:14,640 --> 00:01:17,520
Speaker 1: person you're talking to is a bit more introverted. You

21
00:01:17,640 --> 00:01:19,280
Speaker 1: might want to be a little bit more delicate when

22
00:01:19,280 --> 00:01:22,440
Speaker 1: you message them. The assistant will give you a percentage

23
00:01:22,560 --> 00:01:25,840
Speaker 1: likelihood that the person you're texting with likes you. I'm

24
00:01:25,840 --> 00:01:30,200
Speaker 1: talking romantic feelings. It'll read hundreds of data points to

25
00:01:30,280 --> 00:01:34,200
Speaker 1: help you shape a better, more personalized conversation based on

26
00:01:34,280 --> 00:01:36,720
Speaker 1: what it picks up in the nuances of your language,

27
00:01:37,520 --> 00:01:40,840
Speaker 1: the pauses between texts, the emojis you use. Are you

28
00:01:41,000 --> 00:01:45,240
Speaker 1: using words that are positive or negative? So would you

29
00:01:45,440 --> 00:01:48,520
Speaker 1: use this type of technology? Well, if so, you're going

30
00:01:48,560 --> 00:01:50,560
Speaker 1: to have to give over a lot of your data

31
00:01:50,680 --> 00:01:54,280
Speaker 1: in exchange. I know what you're probably thinking, My text

32
00:01:54,280 --> 00:01:59,360
Speaker 1: messages are pretty personal. What about my privacy? The conversations

33
00:01:59,400 --> 00:02:02,280
Speaker 1: are anonymous and you can delete them whenever you want.

34
00:02:02,800 --> 00:02:06,120
Speaker 1: But it's twenty twenty. Leaks and hacks are the norm,

35
00:02:06,720 --> 00:02:10,160
Speaker 1: so of course they're ethical boundaries. The tech I'm talking

36
00:02:10,160 --> 00:02:13,520
Speaker 1: about was built by an entrepreneur named s Lee. It's

37
00:02:13,560 --> 00:02:16,560
Speaker 1: an app called may, an AI assistant that you can

38
00:02:16,600 --> 00:02:20,600
Speaker 1: download that, when unleashed, can tell a lot about us,

39
00:02:21,120 --> 00:02:24,760
Speaker 1: including our mental state. So in this episode, you're going

40
00:02:24,800 --> 00:02:28,320
Speaker 1: to hear me talk a lot about depression, love, all

41
00:02:28,360 --> 00:02:31,720
Speaker 1: the raw human elements that this tech can discover. So

42
00:02:32,080 --> 00:02:35,920
Speaker 1: what does an algorithm say about our messy human conversations

43
00:02:36,120 --> 00:02:39,119
Speaker 1: about us? Could it help us communicate better with each

44
00:02:39,160 --> 00:02:43,200
Speaker 1: other or are we crossing a line? The future is here,

45
00:02:43,560 --> 00:02:45,840
Speaker 1: we might as well take a deep dive. I'm Lorie

46
00:02:45,840 --> 00:02:54,320
Speaker 1: Siegel and this is First Contact. Well, welcome to First Contact.

47
00:02:54,560 --> 00:02:57,800
Speaker 1: This is a show that introduces you to people and

48
00:02:57,880 --> 00:03:00,160
Speaker 1: tech that changes what it means to be human. And

49
00:03:00,240 --> 00:03:03,399
Speaker 1: I think you're a perfect example of that. And as

50
00:03:03,480 --> 00:03:06,600
Speaker 1: normally I go into my first contact with founders, I

51
00:03:06,639 --> 00:03:09,480
Speaker 1: bring on the show and how we first met. But

52
00:03:09,520 --> 00:03:11,440
Speaker 1: we've never met in real life.

53
00:03:11,600 --> 00:03:12,600
Speaker 2: No, thank you for having me.

54
00:03:12,919 --> 00:03:15,680
Speaker 1: But that being said, you might not realize this, but

55
00:03:15,880 --> 00:03:19,880
Speaker 1: our first contact happened in this room. I interviewed a

56
00:03:19,880 --> 00:03:21,560
Speaker 1: guy named Shane Mac Do you know.

57
00:03:21,720 --> 00:03:23,720
Speaker 2: Shane, Yeah, yeah, I met him once.

58
00:03:23,919 --> 00:03:27,960
Speaker 1: Great guy, interesting guy, and he's building out bought technology

59
00:03:28,480 --> 00:03:32,040
Speaker 1: to date on the dating apps, and so he said,

60
00:03:32,080 --> 00:03:34,840
Speaker 1: in the middle of the interview, he brought up your

61
00:03:34,920 --> 00:03:38,520
Speaker 1: company and he said, they're building this really cool technology

62
00:03:38,520 --> 00:03:40,920
Speaker 1: that can analyze conversations and they can know if you're

63
00:03:40,960 --> 00:03:43,240
Speaker 1: interested in the other person by the way you're speaking.

64
00:03:43,760 --> 00:03:47,280
Speaker 1: I was like, WHOA, that's really cool. And so, you know,

65
00:03:47,320 --> 00:03:50,560
Speaker 1: I was doing research on you and something stuck out

66
00:03:50,640 --> 00:03:53,600
Speaker 1: that you said you talked about how your text messages

67
00:03:53,840 --> 00:03:55,080
Speaker 1: have body languages.

68
00:03:55,760 --> 00:03:55,960
Speaker 2: Yeah.

69
00:03:56,040 --> 00:03:58,640
Speaker 1: I love this idea that we could read our text

70
00:03:58,640 --> 00:04:01,720
Speaker 1: messages like body language. It's just so much about us

71
00:04:01,840 --> 00:04:03,480
Speaker 1: and I think that's like a good way to start

72
00:04:03,480 --> 00:04:04,360
Speaker 1: with what you're building.

73
00:04:05,000 --> 00:04:05,160
Speaker 2: Yeah.

74
00:04:05,240 --> 00:04:07,040
Speaker 4: Yeah, I think it's body language in a way in

75
00:04:07,040 --> 00:04:09,760
Speaker 4: that I guess there's kind of like a new language

76
00:04:09,840 --> 00:04:13,880
Speaker 4: evolving from text. So thirty years ago, nobody be able

77
00:04:13,960 --> 00:04:15,800
Speaker 4: to look at a phone and say, oh, this person

78
00:04:15,920 --> 00:04:18,920
Speaker 4: likes you based on the metadata in the text. Right,

79
00:04:19,160 --> 00:04:21,920
Speaker 4: So if somebody makes eye contact with you, if they smile,

80
00:04:22,560 --> 00:04:24,760
Speaker 4: these are all kinds of signs that we've learned over

81
00:04:24,800 --> 00:04:27,280
Speaker 4: time that this is an indication this person's interested in you.

82
00:04:28,000 --> 00:04:30,440
Speaker 4: And now in the form of text messages, it might

83
00:04:30,480 --> 00:04:32,760
Speaker 4: be a double text, it might be you know, a

84
00:04:32,760 --> 00:04:35,560
Speaker 4: lot of emoji usage or exclamation points. It might be

85
00:04:35,600 --> 00:04:38,479
Speaker 4: immediately replying to all your text messages.

86
00:04:39,040 --> 00:04:41,599
Speaker 1: And the idea is that with what you guys are building,

87
00:04:41,600 --> 00:04:45,680
Speaker 1: you could actually do analytics on it and really help

88
00:04:45,760 --> 00:04:50,280
Speaker 1: people understand relationships and if someone likes you, and much

89
00:04:50,320 --> 00:04:53,200
Speaker 1: more than just that. It started very basic like that,

90
00:04:53,400 --> 00:04:55,599
Speaker 1: but you can understand all sorts of things about people.

91
00:04:56,279 --> 00:05:00,839
Speaker 4: Yeah, Yeah, there's an immense number of data points in conversation,

92
00:05:01,320 --> 00:05:03,720
Speaker 4: so you can imagine even the conversation that we're having. Now,

93
00:05:04,320 --> 00:05:06,880
Speaker 4: there's been research out there that the number of times

94
00:05:06,880 --> 00:05:09,200
Speaker 4: a person says I versus the other person is an

95
00:05:09,200 --> 00:05:12,039
Speaker 4: indicator of the power dynamic between the two.

96
00:05:12,200 --> 00:05:14,000
Speaker 2: So when you.

97
00:05:14,279 --> 00:05:18,040
Speaker 4: Add all the metadata along with the content of what

98
00:05:18,080 --> 00:05:20,560
Speaker 4: we're saying, there's a lot of things that get revealed

99
00:05:21,040 --> 00:05:22,839
Speaker 4: just from looking at the data and text messages.

100
00:05:23,320 --> 00:05:25,279
Speaker 1: So before we dig into the tech and you're building,

101
00:05:25,279 --> 00:05:26,920
Speaker 1: can we just take a giant step back because normally

102
00:05:26,960 --> 00:05:29,359
Speaker 1: when I interview folks, I generally have a history with

103
00:05:29,400 --> 00:05:31,640
Speaker 1: them or I know them, And this is, as you said,

104
00:05:31,680 --> 00:05:34,039
Speaker 1: my first contact with you. Where are you from?

105
00:05:35,040 --> 00:05:35,160
Speaker 3: So?

106
00:05:35,600 --> 00:05:37,040
Speaker 2: I was born in China.

107
00:05:37,640 --> 00:05:41,120
Speaker 4: I grew up in Boston, Okay, and I moved to

108
00:05:41,160 --> 00:05:43,720
Speaker 4: New York about twelve years ago. I always thought people

109
00:05:43,760 --> 00:05:46,920
Speaker 4: were really interesting to analyze. I did computer science and college,

110
00:05:46,960 --> 00:05:51,159
Speaker 4: and then after school I went into finance and covered

111
00:05:51,200 --> 00:05:52,160
Speaker 4: the insurance industry.

112
00:05:52,279 --> 00:05:54,200
Speaker 1: The insurance industry, Yeah, that's right.

113
00:05:54,760 --> 00:05:58,040
Speaker 4: I was working on an insurance startup and whenever I

114
00:05:58,040 --> 00:06:00,479
Speaker 4: would tell people what I was working on, it would

115
00:06:00,480 --> 00:06:03,799
Speaker 4: just be like crickets. And then I'd say, well, actually,

116
00:06:03,880 --> 00:06:05,680
Speaker 4: in my back pocket, I'm working on an algorithm that

117
00:06:05,720 --> 00:06:07,400
Speaker 4: can figure out how much people like you based on

118
00:06:07,440 --> 00:06:10,120
Speaker 4: your text messages. And you kind of see the eyeslight

119
00:06:10,200 --> 00:06:12,320
Speaker 4: of the jaws drop and then I'm like, Okay, that's

120
00:06:12,400 --> 00:06:14,680
Speaker 4: enough validation that I should probably be working on this.

121
00:06:15,920 --> 00:06:21,680
Speaker 1: In my experience, founders look to try to solve problems

122
00:06:21,680 --> 00:06:25,719
Speaker 1: that they have, So what problem did you have that

123
00:06:25,800 --> 00:06:26,880
Speaker 1: you're trying to solve?

124
00:06:27,920 --> 00:06:30,400
Speaker 2: So actually you kind of solve right through me.

125
00:06:31,160 --> 00:06:33,359
Speaker 4: I've been in New York for twelve years, and you know,

126
00:06:33,520 --> 00:06:36,880
Speaker 4: had my share of dating experiences. I think I'm okay

127
00:06:36,880 --> 00:06:39,200
Speaker 4: in real life, but then when it comes to text messages,

128
00:06:39,240 --> 00:06:39,800
Speaker 4: I'm terrible.

129
00:06:40,120 --> 00:06:40,400
Speaker 1: Why.

130
00:06:40,760 --> 00:06:42,960
Speaker 4: I guess it's a common affliction that actually a lot

131
00:06:42,960 --> 00:06:45,760
Speaker 4: of guys may have. They just don't know how to

132
00:06:45,800 --> 00:06:48,960
Speaker 4: text right. So I was always eager just to end

133
00:06:48,960 --> 00:06:51,800
Speaker 4: the text communication and get to real life, Whereas that

134
00:06:51,920 --> 00:06:54,320
Speaker 4: was probably interpreted as a you know, he doesn't like

135
00:06:54,360 --> 00:06:57,359
Speaker 4: me if he doesn't care, So I kind of wondered

136
00:06:57,400 --> 00:07:00,839
Speaker 4: just how many relationships could have been better or more

137
00:07:01,120 --> 00:07:02,760
Speaker 4: How to just see these things?

138
00:07:04,680 --> 00:07:08,920
Speaker 1: So you're texting with women and not having luck, and

139
00:07:08,960 --> 00:07:13,840
Speaker 1: you're frustrated and you have this background in computer science

140
00:07:14,040 --> 00:07:17,720
Speaker 1: and there's an algorithm that can help you communicate better.

141
00:07:18,080 --> 00:07:21,600
Speaker 4: Right, okay, so then what so Actually what kind of

142
00:07:21,640 --> 00:07:23,960
Speaker 4: brought this to fruition was a friend that moved into

143
00:07:24,040 --> 00:07:27,320
Speaker 4: the city and he said dating here sucks, and I

144
00:07:27,400 --> 00:07:28,640
Speaker 4: was like, why he goes?

145
00:07:29,360 --> 00:07:30,600
Speaker 2: So I went out with this girl a couple of

146
00:07:30,680 --> 00:07:31,040
Speaker 2: days ago.

147
00:07:31,440 --> 00:07:34,480
Speaker 4: We really hit it off, everything was great, and now

148
00:07:34,480 --> 00:07:36,720
Speaker 4: all of a sudden, she's not responding to my text messages.

149
00:07:37,400 --> 00:07:38,440
Speaker 2: Then he was confused.

150
00:07:38,840 --> 00:07:40,600
Speaker 4: So I take a look at it, and I kind

151
00:07:40,600 --> 00:07:42,520
Speaker 4: of flipped through all his text messages and I'm.

152
00:07:42,320 --> 00:07:45,120
Speaker 2: Like, she likes you, How did you know? So I

153
00:07:45,160 --> 00:07:46,440
Speaker 2: kind of looked at the body language.

154
00:07:46,680 --> 00:07:48,160
Speaker 1: I mean, it sounds like if you don't have luck

155
00:07:48,240 --> 00:07:50,840
Speaker 1: on text message, how are you able to analyze these

156
00:07:50,880 --> 00:07:53,640
Speaker 1: text messages and look at words as data points and

157
00:07:53,680 --> 00:07:54,120
Speaker 1: actually know?

158
00:07:54,600 --> 00:07:55,000
Speaker 2: Yeah.

159
00:07:55,040 --> 00:07:58,360
Speaker 4: The funny thing is actually I found over the years

160
00:07:58,440 --> 00:08:00,920
Speaker 4: that it is actually impossible to be objective about your

161
00:08:00,920 --> 00:08:01,800
Speaker 4: own relationships.

162
00:08:02,120 --> 00:08:02,320
Speaker 2: Right.

163
00:08:02,400 --> 00:08:04,440
Speaker 4: Often it it's a third party to kind of look

164
00:08:04,480 --> 00:08:06,560
Speaker 4: at it from you know, a bird's eye view and say, hey,

165
00:08:06,600 --> 00:08:08,320
Speaker 4: look here, we're all the warning signs.

166
00:08:08,320 --> 00:08:09,320
Speaker 2: How did you miss all that?

167
00:08:10,000 --> 00:08:12,000
Speaker 4: And I actually think That's how we deal with a

168
00:08:12,000 --> 00:08:14,960
Speaker 4: lot of our relationships, is we you know, we're not

169
00:08:15,000 --> 00:08:18,080
Speaker 4: able to see objectively, right despite knowing kind of all

170
00:08:18,080 --> 00:08:20,960
Speaker 4: the pitfalls of texting when it comes to my own relationship,

171
00:08:21,000 --> 00:08:23,560
Speaker 4: it helps me not at all, but I can kind

172
00:08:23,560 --> 00:08:25,920
Speaker 4: of see it in other people, and you know, I

173
00:08:25,960 --> 00:08:29,000
Speaker 4: can be more objective myself about it, and then with

174
00:08:29,080 --> 00:08:31,240
Speaker 4: the algorithm, I can actually give them data behind it.

175
00:08:31,960 --> 00:08:35,760
Speaker 1: And all of this is kind of background for an

176
00:08:35,880 --> 00:08:38,199
Speaker 1: algorithm that went in and this is this is all

177
00:08:38,240 --> 00:08:41,400
Speaker 1: before you created May it was Crush, right, the crush

178
00:08:41,480 --> 00:08:44,200
Speaker 1: app that you created, and so all of these data

179
00:08:44,200 --> 00:08:46,920
Speaker 1: points went into creating this crush app. So talk to

180
00:08:47,000 --> 00:08:47,559
Speaker 1: us about that.

181
00:08:47,880 --> 00:08:49,760
Speaker 4: Yeah, yeah, I would like to go on record to

182
00:08:49,880 --> 00:08:54,040
Speaker 4: publicly apologize for Crush. It was downloaded by one hundred

183
00:08:54,040 --> 00:08:56,760
Speaker 4: and fifty thousand people, Uh huh. And I joked with

184
00:08:56,800 --> 00:08:58,760
Speaker 4: my team that, like, we should put out a press

185
00:08:58,760 --> 00:09:01,600
Speaker 4: release to apologize for ruining one hundred and fifty thousand

186
00:09:01,600 --> 00:09:02,480
Speaker 4: people's days.

187
00:09:03,000 --> 00:09:04,880
Speaker 1: What was so terrible about it?

188
00:09:05,080 --> 00:09:09,239
Speaker 4: Because you know, you're probably if you're using an algorithm

189
00:09:09,280 --> 00:09:11,440
Speaker 4: to figure out kind of whether somebody likes you or not,

190
00:09:11,800 --> 00:09:13,760
Speaker 4: like chances are you're not going to be happy with

191
00:09:13,800 --> 00:09:14,920
Speaker 4: what you find.

192
00:09:15,120 --> 00:09:17,160
Speaker 1: And what did the algorithm take into account? Can you

193
00:09:17,200 --> 00:09:18,959
Speaker 1: talk to us about the technology behind and then we'll

194
00:09:18,960 --> 00:09:21,160
Speaker 1: get into the technology you're using now, which is much

195
00:09:21,440 --> 00:09:24,760
Speaker 1: beyond just crush and does someone like you or not?

196
00:09:24,880 --> 00:09:26,520
Speaker 1: What did it actually take into account?

197
00:09:27,080 --> 00:09:30,480
Speaker 4: It took into account the average response times, the length

198
00:09:30,520 --> 00:09:34,079
Speaker 4: of the text message counted, an emoji usage like exclamation

199
00:09:34,160 --> 00:09:36,600
Speaker 4: point usage, whether they sent you a link or a picture,

200
00:09:37,520 --> 00:09:41,160
Speaker 4: and then the number of conversations each side initiated.

201
00:09:42,160 --> 00:09:44,520
Speaker 1: I mean, it is pretty extraordinary if you think about

202
00:09:44,640 --> 00:09:46,920
Speaker 1: the amount of data we give each other on a

203
00:09:47,000 --> 00:09:49,800
Speaker 1: daily basis, like the time of day we text, the

204
00:09:49,840 --> 00:09:52,520
Speaker 1: types of words we use, the types of emojis, like,

205
00:09:53,000 --> 00:09:56,040
Speaker 1: all of this is an insane data set that creates

206
00:09:56,040 --> 00:10:00,200
Speaker 1: a personality profile that I mean, it's just an extraordinary

207
00:10:00,240 --> 00:10:03,240
Speaker 1: amount of information about us that you're utilizing and that

208
00:10:03,480 --> 00:10:08,080
Speaker 1: advertisers are utilizing that when put to use, can be used.

209
00:10:08,120 --> 00:10:09,600
Speaker 1: And I guess you're trying to use it to kind

210
00:10:09,600 --> 00:10:12,640
Speaker 1: of help us better ourselves and have better conversations. But

211
00:10:13,120 --> 00:10:16,040
Speaker 1: it is an extraordinary amount of data that I don't

212
00:10:16,080 --> 00:10:19,280
Speaker 1: think people quite understand, like how as human beings we

213
00:10:19,440 --> 00:10:23,120
Speaker 1: decode pretty easily, right, yeah, yeah, so I.

214
00:10:23,040 --> 00:10:24,440
Speaker 2: Try to back up my text messages.

215
00:10:24,520 --> 00:10:26,079
Speaker 4: Right, so when I get a new phone, you get

216
00:10:26,080 --> 00:10:28,280
Speaker 4: a new app and it kind of loads back your

217
00:10:28,320 --> 00:10:31,439
Speaker 4: old text messages. But the program I was using took

218
00:10:31,440 --> 00:10:34,439
Speaker 4: a really long time and actually shload every text message.

219
00:10:34,720 --> 00:10:38,040
Speaker 4: And I actually spent three hours looking at all my

220
00:10:38,160 --> 00:10:41,160
Speaker 4: text messages flash before me at half a second each.

221
00:10:42,040 --> 00:10:45,120
Speaker 4: It's an incredible amount of data. Sure, and the average

222
00:10:45,160 --> 00:10:49,960
Speaker 4: person actually texts about a novel's worth of their thoughts

223
00:10:50,200 --> 00:10:52,199
Speaker 4: a year with your thumbs.

224
00:10:52,800 --> 00:10:55,480
Speaker 1: Wow, I mean, And we'll get into this a little

225
00:10:55,480 --> 00:10:58,080
Speaker 1: bit later, but I gave you text messages with my

226
00:10:58,200 --> 00:11:02,760
Speaker 1: co founder and dear friend, and as we were giving

227
00:11:02,760 --> 00:11:04,200
Speaker 1: you the text message is, by the way, we only

228
00:11:04,200 --> 00:11:05,520
Speaker 1: give you like a year or a year and a

229
00:11:05,520 --> 00:11:08,400
Speaker 1: half for it, but we have like twelve years worth

230
00:11:08,520 --> 00:11:10,680
Speaker 1: or something. And it was maybe one of the most

231
00:11:10,720 --> 00:11:15,520
Speaker 1: like horrifying interesting experiences watching as they were downloading or

232
00:11:15,559 --> 00:11:18,520
Speaker 1: looking at them. It was just like you know, looking

233
00:11:18,559 --> 00:11:21,920
Speaker 1: at all those words and such like a personal experience

234
00:11:22,000 --> 00:11:25,360
Speaker 1: through technology to do. So I want to get into

235
00:11:25,400 --> 00:11:29,440
Speaker 1: what you're doing now with may. So you have moved past,

236
00:11:29,559 --> 00:11:31,839
Speaker 1: so we're not just doing love anymore where we've moved

237
00:11:31,840 --> 00:11:35,000
Speaker 1: past just an app to help people find out if

238
00:11:35,040 --> 00:11:36,560
Speaker 1: other people are interested in them.

239
00:11:36,760 --> 00:11:38,240
Speaker 2: Yeah, yeah, we have moved on from that.

240
00:11:38,400 --> 00:11:40,480
Speaker 1: And so when did we make the pivot? Talk us

241
00:11:40,520 --> 00:11:40,880
Speaker 1: through that?

242
00:11:41,480 --> 00:11:44,480
Speaker 4: So my goal had always been to be like the

243
00:11:44,640 --> 00:11:48,800
Speaker 4: default texting app on a phone. So the most used

244
00:11:49,040 --> 00:11:52,400
Speaker 4: application on our phone is for messaging. Yeah, we had

245
00:11:52,440 --> 00:11:55,719
Speaker 4: instant messaging twenty years ago and there's almost been no

246
00:11:55,800 --> 00:11:59,480
Speaker 4: innovation whatsoever on it, right, And so I go, hey,

247
00:11:59,520 --> 00:12:02,760
Speaker 4: what if I am able to build a messaging app

248
00:12:03,000 --> 00:12:05,480
Speaker 4: that actually has kind of an AI that's just sitting

249
00:12:05,559 --> 00:12:08,840
Speaker 4: right on it kind of analyzing in conversations maybe like

250
00:12:09,000 --> 00:12:11,360
Speaker 4: it's almost like a guardian the Angel, you know, looking

251
00:12:11,360 --> 00:12:14,320
Speaker 4: over your shoulders being like, oh wait, you're coming across

252
00:12:14,320 --> 00:12:17,600
Speaker 4: as being rude or you know you're missing this about

253
00:12:17,600 --> 00:12:20,640
Speaker 4: this person. So I always thought there was the capability

254
00:12:20,679 --> 00:12:23,720
Speaker 4: for that. So Crush was basically the very first step

255
00:12:23,760 --> 00:12:26,559
Speaker 4: in doing that. You know, I know, with a lot

256
00:12:26,600 --> 00:12:30,280
Speaker 4: of data, a lot of amazing AI is possible, and

257
00:12:30,640 --> 00:12:33,240
Speaker 4: that wasn't possible without first having the data that we

258
00:12:33,320 --> 00:12:34,040
Speaker 4: got from Crush.

259
00:12:34,520 --> 00:12:36,760
Speaker 1: First, Well, may is, did I read that you named

260
00:12:36,760 --> 00:12:39,800
Speaker 1: that after your mom? I did that's interesting.

261
00:12:40,600 --> 00:12:43,079
Speaker 4: It was also coincidental in a way because I was

262
00:12:43,120 --> 00:12:48,040
Speaker 4: looking for a name for the AI and I thought

263
00:12:48,120 --> 00:12:50,840
Speaker 4: messaging could be done so much better. So in a way,

264
00:12:51,600 --> 00:12:55,600
Speaker 4: MAY also stands for messaging Improved, and maybe one day

265
00:12:55,720 --> 00:13:00,120
Speaker 4: it'll be it'll stand for me improved, because you know,

266
00:13:00,120 --> 00:13:02,240
Speaker 4: we're going to take this data and discover what we

267
00:13:02,320 --> 00:13:07,280
Speaker 4: can to give you the best information for you to

268
00:13:07,760 --> 00:13:10,079
Speaker 4: operate as an improve version of yourself.

269
00:13:10,840 --> 00:13:13,400
Speaker 1: And it means beautiful in Chinese.

270
00:13:13,480 --> 00:13:14,880
Speaker 2: Yes, it means beautiful in Chinese.

271
00:13:15,080 --> 00:13:17,160
Speaker 1: You know, I want to get into the algorithm and

272
00:13:17,200 --> 00:13:18,679
Speaker 1: how it works. So I think we're talking a lot

273
00:13:18,760 --> 00:13:21,160
Speaker 1: of like in theory about how this thing works and

274
00:13:21,160 --> 00:13:23,880
Speaker 1: all this. So let's like go hardcore into the technology

275
00:13:23,880 --> 00:13:27,520
Speaker 1: behind it. So what exactly does MAY do, like like

276
00:13:28,000 --> 00:13:30,480
Speaker 1: break it down for us and you know, the most

277
00:13:30,559 --> 00:13:31,320
Speaker 1: human way you can.

278
00:13:31,640 --> 00:13:35,600
Speaker 4: Okay, we're basically a replacement of the default texting app

279
00:13:35,640 --> 00:13:38,960
Speaker 4: on the phone. So if you and I texted after this,

280
00:13:39,520 --> 00:13:42,360
Speaker 4: with enough messages, the AI will be able to pick

281
00:13:42,440 --> 00:13:44,880
Speaker 4: up certain things about your personality and start giving me

282
00:13:45,520 --> 00:13:48,920
Speaker 4: advice or kind of little insights that it picks up

283
00:13:48,920 --> 00:13:52,560
Speaker 4: about you, Like what like it might pick up that

284
00:13:53,080 --> 00:13:55,360
Speaker 4: you know, maybe you're more organized than I am.

285
00:13:55,520 --> 00:13:58,360
Speaker 1: Well that's certainly not true. So everyone in this really

286
00:13:58,400 --> 00:14:00,280
Speaker 1: shaking their heads. Why don't we think.

287
00:14:00,440 --> 00:14:05,080
Speaker 2: One or more more you know, altruistic or empathetic than

288
00:14:05,120 --> 00:14:05,360
Speaker 2: I am.

289
00:14:05,520 --> 00:14:07,080
Speaker 1: Okay, we could go with that way.

290
00:14:07,160 --> 00:14:07,839
Speaker 2: Okay.

291
00:14:08,920 --> 00:14:12,080
Speaker 4: So the idea is, you know, as we're having these conversations,

292
00:14:12,400 --> 00:14:15,000
Speaker 4: and with enough data that goes through the system, the

293
00:14:15,040 --> 00:14:17,560
Speaker 4: AA has actually been able to figure certain things out

294
00:14:17,600 --> 00:14:21,560
Speaker 4: about yourself and the other person. So, you know, for example,

295
00:14:21,600 --> 00:14:24,880
Speaker 4: it might try to predict what the age or gender

296
00:14:24,920 --> 00:14:27,640
Speaker 4: of the other person is. You know, if there's a

297
00:14:27,880 --> 00:14:30,680
Speaker 4: history of conversation and it figures out that the other

298
00:14:30,720 --> 00:14:32,720
Speaker 4: person is using a lot more negative words than they

299
00:14:32,800 --> 00:14:36,080
Speaker 4: used to and might say, hey, Laurie seems like she's

300
00:14:36,560 --> 00:14:37,280
Speaker 4: a little different.

301
00:14:37,560 --> 00:14:38,760
Speaker 2: You might want to check in with her.

302
00:14:39,680 --> 00:14:43,400
Speaker 4: So we kind of went through as many algorithms as

303
00:14:43,800 --> 00:14:47,080
Speaker 4: possible to figure out, Hey, how can we be like

304
00:14:47,120 --> 00:14:51,080
Speaker 4: that little assistant, your little personal relationship assistant as you

305
00:14:51,120 --> 00:14:53,400
Speaker 4: go about, you know, through your day and texting.

306
00:14:53,760 --> 00:14:55,760
Speaker 1: What do you mean when it's like negative words? I

307
00:14:55,800 --> 00:14:59,480
Speaker 1: don't know if folks really truly understand what that actually means.

308
00:15:00,120 --> 00:15:02,800
Speaker 4: Yeah, a lot of like natural language work has been

309
00:15:02,960 --> 00:15:05,800
Speaker 4: based on like sentiment working amount sentiment, like hey, if

310
00:15:05,880 --> 00:15:10,280
Speaker 4: you're saying sad, angry, or you know, I'm happy, I'm

311
00:15:10,320 --> 00:15:13,400
Speaker 4: glad to see you. These are negative and positive words.

312
00:15:14,320 --> 00:15:17,360
Speaker 4: So if you take the dumbest algorithm and just say, hey,

313
00:15:17,440 --> 00:15:19,160
Speaker 4: let me just look at the number of times you

314
00:15:19,200 --> 00:15:23,200
Speaker 4: say something negative, use a negative word today, and say, now,

315
00:15:23,400 --> 00:15:26,160
Speaker 4: on average you use twenty two out of five thousand words,

316
00:15:26,480 --> 00:15:29,040
Speaker 4: but now you're at forty four, and so you're twice

317
00:15:29,040 --> 00:15:30,240
Speaker 4: as negative as you.

318
00:15:30,480 --> 00:15:30,920
Speaker 2: Tend to be.

319
00:15:31,440 --> 00:15:33,800
Speaker 4: And so we have algorithms that will look at statistically

320
00:15:33,960 --> 00:15:37,800
Speaker 4: like these numbers and find when there's an anomaly, and

321
00:15:37,880 --> 00:15:39,440
Speaker 4: you know, try to point that out to the user.

322
00:15:40,600 --> 00:15:42,240
Speaker 1: We've got to take a quick break to hear from

323
00:15:42,280 --> 00:15:45,600
Speaker 1: our sponsors, but when we come back, we'll talk May

324
00:15:45,760 --> 00:15:48,680
Speaker 1: and mental health. What if an algorithm identified that you

325
00:15:48,680 --> 00:15:52,640
Speaker 1: were falling into depression and then using data, it suggested

326
00:15:52,680 --> 00:15:55,240
Speaker 1: the best person in your contact list to reach out to.

327
00:15:55,960 --> 00:15:59,320
Speaker 1: And here's what's surprising, it's not the person you normally

328
00:15:59,360 --> 00:16:25,280
Speaker 1: would think. More after the break and what other types

329
00:16:25,320 --> 00:16:28,280
Speaker 1: of things does it analyze? It still analyzes if someone

330
00:16:28,640 --> 00:16:31,760
Speaker 1: is into you or the ideas that it can analyze

331
00:16:31,760 --> 00:16:36,160
Speaker 1: work relationships all sorts of different types of relationships, right, yeah, yeah,

332
00:16:36,320 --> 00:16:39,040
Speaker 1: and how and how? Yes, just give it the specifics

333
00:16:39,080 --> 00:16:41,120
Speaker 1: I think are really interesting to us because I don't

334
00:16:41,120 --> 00:16:44,000
Speaker 1: think people quite understand the data points that go into

335
00:16:44,080 --> 00:16:44,920
Speaker 1: a lot of this stuff.

336
00:16:45,320 --> 00:16:45,560
Speaker 2: Right.

337
00:16:46,240 --> 00:16:48,880
Speaker 4: So with Crush, one of the first things that we

338
00:16:48,880 --> 00:16:51,520
Speaker 4: would ask is, hey, what type of relationship is this?

339
00:16:52,360 --> 00:16:54,800
Speaker 4: Because you know you're going to be texting somebody you'd

340
00:16:54,840 --> 00:16:57,200
Speaker 4: like a little differently from your family and from your friends.

341
00:16:57,200 --> 00:16:59,240
Speaker 4: So we wanted to kind of separate out our data set.

342
00:17:00,200 --> 00:17:02,400
Speaker 4: So we actually kind of had a lot of labels

343
00:17:02,720 --> 00:17:05,320
Speaker 4: to do AI. There's only two pieces of things that

344
00:17:05,359 --> 00:17:09,159
Speaker 4: you really need. It's a ton of data which we

345
00:17:09,320 --> 00:17:14,320
Speaker 4: had and labels. So basically, let's just abstract it and say, okay,

346
00:17:14,560 --> 00:17:17,680
Speaker 4: let's talk about pictures of dogs and cats.

347
00:17:18,000 --> 00:17:18,200
Speaker 2: Right.

348
00:17:18,600 --> 00:17:21,160
Speaker 4: If you have hundreds of thousands of pictures of dogs

349
00:17:21,200 --> 00:17:23,440
Speaker 4: and cats and you go, okay, well that's a dog,

350
00:17:23,480 --> 00:17:25,240
Speaker 4: that's a cat, that's a dog, that's a cat. Right,

351
00:17:25,840 --> 00:17:28,560
Speaker 4: there's ways to turn that picture into kind of math

352
00:17:29,359 --> 00:17:33,200
Speaker 4: and then have the computer find patterns amongst those pictures

353
00:17:33,560 --> 00:17:35,560
Speaker 4: and then start predicting whether it was a dog or

354
00:17:35,600 --> 00:17:39,080
Speaker 4: a cat picture that was seeing. So in the same way,

355
00:17:39,600 --> 00:17:43,440
Speaker 4: if you had enough text message conversations when where one

356
00:17:43,480 --> 00:17:45,960
Speaker 4: party says, yeah, that was you know, I had a

357
00:17:46,000 --> 00:17:49,919
Speaker 4: question on that person, and that person that's a family

358
00:17:49,960 --> 00:17:52,800
Speaker 4: member or that's my buddy. If you're able to kind

359
00:17:52,800 --> 00:17:56,439
Speaker 4: of do the same thing with enough data, it starts

360
00:17:56,440 --> 00:17:59,880
Speaker 4: actually doing this better than people can. There's no way

361
00:18:00,119 --> 00:18:04,159
Speaker 4: person could have actually listened to two hundred and fifty

362
00:18:04,480 --> 00:18:09,560
Speaker 4: thousand conversations, right, or conversations between you know, five hundred

363
00:18:09,600 --> 00:18:12,240
Speaker 4: thousand people. But a machine is able to do that

364
00:18:12,320 --> 00:18:15,040
Speaker 4: and take every single one of those data points and say, hey,

365
00:18:15,600 --> 00:18:19,760
Speaker 4: if this person, you know, within like the first twenty

366
00:18:19,800 --> 00:18:24,359
Speaker 4: words say something like handsome and ninety nine out of

367
00:18:24,400 --> 00:18:27,239
Speaker 4: one hundred or those that person had marked that they

368
00:18:27,280 --> 00:18:31,640
Speaker 4: were romantically interested in that person. That machine is able

369
00:18:31,680 --> 00:18:34,160
Speaker 4: to do things and recognize patterns in ways that people can't.

370
00:18:35,160 --> 00:18:38,080
Speaker 4: And if you think about how a person actually works,

371
00:18:38,600 --> 00:18:41,359
Speaker 4: like if you're young and this is your first romantic relationship,

372
00:18:41,560 --> 00:18:43,040
Speaker 4: you might not be able to read any of the

373
00:18:43,080 --> 00:18:45,239
Speaker 4: signs that this person likes you, right, because you've never

374
00:18:45,280 --> 00:18:48,000
Speaker 4: seen it before. But you take somebody who's you know,

375
00:18:48,920 --> 00:18:53,080
Speaker 4: seen enough relationships and you know, your grandma might actually

376
00:18:53,080 --> 00:18:55,879
Speaker 4: be your the closest thing to an AI that we know,

377
00:18:56,440 --> 00:18:59,919
Speaker 4: because she's just seen enough information that she can go,

378
00:19:00,200 --> 00:19:02,720
Speaker 4: oh yeah, I heard him say that. So yeah, in

379
00:19:02,760 --> 00:19:05,760
Speaker 4: my experience, like, yeah, that has not turned out well.

380
00:19:05,800 --> 00:19:07,880
Speaker 1: And your grandma might be as close to an AI.

381
00:19:08,080 --> 00:19:10,680
Speaker 1: It's really interesting. So I mean, it's like it's pattern

382
00:19:10,720 --> 00:19:13,680
Speaker 1: recognition and also based on kind of personality types that

383
00:19:13,760 --> 00:19:16,600
Speaker 1: you guys define. Is that how you guys do it?

384
00:19:16,600 --> 00:19:18,320
Speaker 2: It's the big five person rightfile?

385
00:19:18,400 --> 00:19:22,440
Speaker 1: Okay, And so I guess, like, you know what, what's

386
00:19:22,480 --> 00:19:25,640
Speaker 1: the The point of it is to help us communicate better.

387
00:19:26,119 --> 00:19:28,240
Speaker 1: The point of it is to to give us the

388
00:19:28,280 --> 00:19:30,439
Speaker 1: words to say to each other because we're messy and

389
00:19:30,520 --> 00:19:33,760
Speaker 1: human and sometimes we just don't know. It's an iteration

390
00:19:33,800 --> 00:19:35,520
Speaker 1: of you not being able to talk to women, I mean,

391
00:19:35,680 --> 00:19:37,240
Speaker 1: and I don't mean that in a bad way, like

392
00:19:37,440 --> 00:19:39,720
Speaker 1: I mean that in the way that you know I

393
00:19:39,960 --> 00:19:44,080
Speaker 1: sometimes have trouble talking, you know, to a family member

394
00:19:44,160 --> 00:19:47,320
Speaker 1: or something. It just gives us a pattern or a

395
00:19:47,359 --> 00:19:51,639
Speaker 1: blueprint through AI by which to speak to people. Is

396
00:19:51,640 --> 00:19:53,400
Speaker 1: that that's if I'm reading it right.

397
00:19:53,760 --> 00:19:55,120
Speaker 2: Well, well, think about it this way.

398
00:19:56,080 --> 00:19:58,560
Speaker 4: How often when you compose a text message, do you

399
00:19:58,640 --> 00:20:02,359
Speaker 4: think about am I coming across across as like rude

400
00:20:02,480 --> 00:20:02,879
Speaker 4: or well?

401
00:20:02,920 --> 00:20:05,919
Speaker 1: I think my issue sometimes I don't think before I text.

402
00:20:06,960 --> 00:20:09,679
Speaker 4: But it's amazing I think when there's For most people,

403
00:20:09,720 --> 00:20:14,199
Speaker 4: I think when they text something, they scrutinize it, thinking

404
00:20:14,400 --> 00:20:16,000
Speaker 4: what is the other person going to think when they

405
00:20:16,000 --> 00:20:16,760
Speaker 4: get this message?

406
00:20:17,040 --> 00:20:17,240
Speaker 2: Right?

407
00:20:17,800 --> 00:20:20,960
Speaker 4: And I think most of us kind of think the worst. Right,

408
00:20:21,160 --> 00:20:23,280
Speaker 4: But just think about every time you texted somebody. If

409
00:20:23,280 --> 00:20:26,080
Speaker 4: you didn't like somebody, you wouldn't text them. So for example,

410
00:20:26,280 --> 00:20:28,639
Speaker 4: you know, if the algorithm starts picking out that the

411
00:20:28,680 --> 00:20:32,480
Speaker 4: person on the other side is a lot more reserved,

412
00:20:32,720 --> 00:20:35,880
Speaker 4: it'll say, oh, you know, this person seems like they're

413
00:20:35,880 --> 00:20:37,720
Speaker 4: a lot more reserved than you. You might want to

414
00:20:37,720 --> 00:20:40,800
Speaker 4: try a little harder to relate to them.

415
00:20:40,840 --> 00:20:43,720
Speaker 1: What other kind of soft nudges will the AI give us?

416
00:20:43,920 --> 00:20:46,240
Speaker 4: Well, on the flip side, if you know you're reserved

417
00:20:46,240 --> 00:20:48,720
Speaker 4: and the other person is a lot more outgoing, and

418
00:20:48,840 --> 00:20:50,959
Speaker 4: might say, hey, this person is a lot more outgoing

419
00:20:50,960 --> 00:20:52,520
Speaker 4: than you, you might want to try to keep it light,

420
00:20:52,840 --> 00:20:53,639
Speaker 4: you know, keep things.

421
00:20:53,560 --> 00:20:54,360
Speaker 2: Light with them.

422
00:20:54,600 --> 00:20:57,080
Speaker 1: Does the AI? I mean as always AI is so

423
00:20:57,240 --> 00:21:00,919
Speaker 1: flawed right like and can get it really wrong sometimes,

424
00:21:00,920 --> 00:21:03,040
Speaker 1: So like, what if the AI is giving me advice,

425
00:21:03,119 --> 00:21:04,800
Speaker 1: that's just terrible.

426
00:21:04,520 --> 00:21:08,720
Speaker 4: Right, Yeah, that's fair, It's not infallible, but I would

427
00:21:08,800 --> 00:21:11,119
Speaker 4: argue that it is probably better than your friends.

428
00:21:11,400 --> 00:21:13,560
Speaker 1: I want to talk a little bit about the mental

429
00:21:13,600 --> 00:21:15,680
Speaker 1: health aspect of this, because it's putting a lot of

430
00:21:15,720 --> 00:21:17,760
Speaker 1: weight on May. Like if May is looking through a

431
00:21:17,760 --> 00:21:20,680
Speaker 1: lot of these messages and seeing people at these very

432
00:21:20,760 --> 00:21:24,640
Speaker 1: vulnerable moments, and I'm sure you must see some crazy stuff,

433
00:21:24,760 --> 00:21:28,399
Speaker 1: like you must really see some people in pain, seeing

434
00:21:28,480 --> 00:21:33,639
Speaker 1: some pretty sad things about depression or suicide, what's the

435
00:21:33,680 --> 00:21:35,320
Speaker 1: responsibility of the AI.

436
00:21:37,040 --> 00:21:39,280
Speaker 4: Yeah, I mean that's a good question. I don't think

437
00:21:39,320 --> 00:21:43,320
Speaker 4: anybody really has a good answer for that. Look, we

438
00:21:43,359 --> 00:21:46,320
Speaker 4: actually try to steer away from giving people advice for

439
00:21:46,359 --> 00:21:49,159
Speaker 4: as much as possible. But all the feedback that we

440
00:21:49,200 --> 00:21:52,040
Speaker 4: got from people were like, hey, okay, so you told

441
00:21:52,040 --> 00:21:54,600
Speaker 4: me this about it, and shoot, what am I supposed

442
00:21:54,600 --> 00:21:55,119
Speaker 4: to do about it?

443
00:21:55,520 --> 00:21:55,720
Speaker 2: Right?

444
00:21:55,920 --> 00:21:59,359
Speaker 4: People saw that there was application here potentially outside dating.

445
00:22:00,280 --> 00:22:01,960
Speaker 4: That was one of the first things that we turned

446
00:22:01,960 --> 00:22:04,480
Speaker 4: to when we go, Hey, if the AI is able

447
00:22:04,520 --> 00:22:08,080
Speaker 4: to look at your conversations with everybody, your your parents,

448
00:22:08,160 --> 00:22:11,560
Speaker 4: your girlfriend, your best friend, it's able to kind of

449
00:22:11,640 --> 00:22:15,760
Speaker 4: understand things about you that like, really nobody has the perspective,

450
00:22:16,000 --> 00:22:19,160
Speaker 4: right you only get to see somebody in a certain dimension,

451
00:22:20,000 --> 00:22:22,160
Speaker 4: whereas we're, you know, we have a three sixty view,

452
00:22:22,600 --> 00:22:25,160
Speaker 4: and we kind of theorize that, hey, maybe we can actually,

453
00:22:25,320 --> 00:22:29,320
Speaker 4: you know, figure out patterns. And so we started looking

454
00:22:29,359 --> 00:22:33,280
Speaker 4: into conversations, and you know, for the longest time, I

455
00:22:33,359 --> 00:22:36,560
Speaker 4: was looking for another startup in mental health or some

456
00:22:36,680 --> 00:22:39,760
Speaker 4: data set where it's kind of morb to say, but

457
00:22:40,240 --> 00:22:41,200
Speaker 4: could I get the text.

458
00:22:41,040 --> 00:22:43,439
Speaker 2: Messaging history of somebody who killed themselves?

459
00:22:43,840 --> 00:22:47,120
Speaker 4: Because we have the analytics tools to kind of you know,

460
00:22:47,960 --> 00:22:51,359
Speaker 4: comb through the history, and what if we were able

461
00:22:51,359 --> 00:22:54,600
Speaker 4: to use that to figure out patterns that preceded the

462
00:22:54,640 --> 00:22:59,440
Speaker 4: actual act. And so I talked to many many psychologists

463
00:23:00,200 --> 00:23:04,600
Speaker 4: startup companies and the answer was always we don't know,

464
00:23:05,359 --> 00:23:07,679
Speaker 4: or we don't have the data. And even if we

465
00:23:07,720 --> 00:23:10,800
Speaker 4: did have the data, privacy and regulation is keeping us

466
00:23:10,840 --> 00:23:15,080
Speaker 4: from doing anything. And then it dawned on me, well,

467
00:23:15,480 --> 00:23:18,320
Speaker 4: we have conversations of one hundred and fifty thousand people.

468
00:23:19,359 --> 00:23:21,480
Speaker 4: What did we just search through these messages and try

469
00:23:21,480 --> 00:23:25,080
Speaker 4: to look for their phrase tried to kill myself? We

470
00:23:25,160 --> 00:23:29,720
Speaker 4: actually found about ten occurrences of that where it was

471
00:23:29,720 --> 00:23:31,640
Speaker 4: from the user and not of contact, so we had

472
00:23:31,640 --> 00:23:36,560
Speaker 4: a lot more information on them, and they actually in

473
00:23:36,600 --> 00:23:40,000
Speaker 4: those ten they actually included a date. So one of

474
00:23:40,000 --> 00:23:44,600
Speaker 4: them was, Hey, sorry, I've been MIA, but you know,

475
00:23:44,960 --> 00:23:46,919
Speaker 4: I've been going through a rough patch and I actually

476
00:23:46,920 --> 00:23:49,959
Speaker 4: tried to kill myself two weekends ago, but all I

477
00:23:50,119 --> 00:23:52,520
Speaker 4: managed to do was down a bottle of pills and

478
00:23:52,520 --> 00:23:55,200
Speaker 4: send myself to the hospital, So you know that person tried.

479
00:23:55,800 --> 00:23:58,040
Speaker 4: So with that data, we're able to now look at

480
00:23:58,080 --> 00:24:00,840
Speaker 4: the patterns and some of the things that found were

481
00:24:00,880 --> 00:24:02,200
Speaker 4: actually really really interesting.

482
00:24:03,800 --> 00:24:06,480
Speaker 1: How do you feel when you're looking at messages this

483
00:24:06,560 --> 00:24:08,520
Speaker 1: is your product, right, Like, how do you feel when

484
00:24:08,560 --> 00:24:12,760
Speaker 1: you're looking at messages and someone is saying, on this date,

485
00:24:12,840 --> 00:24:16,879
Speaker 1: I tried to kill myself. Like, I don't know really

486
00:24:17,040 --> 00:24:19,920
Speaker 1: what the question is other than like, how does that

487
00:24:20,040 --> 00:24:23,840
Speaker 1: as a founder and with your responsibility and your technology,

488
00:24:23,840 --> 00:24:24,879
Speaker 1: how does that make you feel?

489
00:24:27,119 --> 00:24:30,760
Speaker 4: I mean, I think going through some of these messages,

490
00:24:30,800 --> 00:24:32,600
Speaker 4: and by the way, we don't know the identity of

491
00:24:32,640 --> 00:24:35,320
Speaker 4: the people because we never asked them for their name, right,

492
00:24:36,280 --> 00:24:40,399
Speaker 4: I think it was an exercise in empathy that if

493
00:24:40,720 --> 00:24:45,760
Speaker 4: everybody had the opportunity to go through I would suggest

494
00:24:45,800 --> 00:24:47,840
Speaker 4: it because.

495
00:24:48,160 --> 00:24:49,080
Speaker 2: Going through.

496
00:24:51,520 --> 00:24:54,639
Speaker 4: The interactions of some of these people with everybody in

497
00:24:54,680 --> 00:24:59,640
Speaker 4: their life, it felt like a soap opera. I did

498
00:24:59,640 --> 00:25:02,200
Speaker 4: it with one where I looked at the two months

499
00:25:02,200 --> 00:25:05,720
Speaker 4: of conversations before that and I tried to jot down

500
00:25:05,800 --> 00:25:08,760
Speaker 4: every person, Oh yeah, this number three, seven oh four,

501
00:25:09,160 --> 00:25:10,720
Speaker 4: that must have been her friend, or that must have

502
00:25:10,720 --> 00:25:16,760
Speaker 4: been a girlfriend. It was It was exhausting because imagine

503
00:25:16,800 --> 00:25:18,600
Speaker 4: you know that this person was going to go through

504
00:25:18,640 --> 00:25:21,400
Speaker 4: hardship a month from now, and you're reading the conversations

505
00:25:21,680 --> 00:25:23,359
Speaker 4: of them crying out for help to people.

506
00:25:24,200 --> 00:25:25,240
Speaker 2: It was kind of heartbreaking.

507
00:25:27,200 --> 00:25:31,159
Speaker 4: And there were some conversations where you know, this person

508
00:25:31,200 --> 00:25:33,919
Speaker 4: came in and just started listening to them and checking

509
00:25:33,920 --> 00:25:35,680
Speaker 4: in with them and saying, hey, how are you doing.

510
00:25:36,720 --> 00:25:39,240
Speaker 4: I just remember mentally like cheering at that point, like

511
00:25:39,280 --> 00:25:44,400
Speaker 4: thank God for you. And so, you know, we went

512
00:25:44,480 --> 00:25:49,639
Speaker 4: through this because I knew that an algorithm might be

513
00:25:49,680 --> 00:25:52,120
Speaker 4: able to find that person. And I think when people

514
00:25:52,119 --> 00:25:54,719
Speaker 4: go through tough times, there really is no solution.

515
00:25:54,840 --> 00:25:56,360
Speaker 2: There's no easy solution, right.

516
00:25:56,520 --> 00:25:59,560
Speaker 4: People might say, hey, if somebody is depressed and you

517
00:25:59,600 --> 00:26:02,359
Speaker 4: find that out, what you should do is you know

518
00:26:02,400 --> 00:26:04,280
Speaker 4: there's a lot of bots out there or a mental

519
00:26:04,359 --> 00:26:07,479
Speaker 4: health hotline. I think all these people know that there

520
00:26:07,480 --> 00:26:10,360
Speaker 4: are bots of mental health hotlines, but they don't reach out.

521
00:26:10,840 --> 00:26:13,600
Speaker 4: And I actually think what we worked on is a

522
00:26:13,640 --> 00:26:16,359
Speaker 4: lot more complete solution than anything that's out there, because

523
00:26:16,400 --> 00:26:18,320
Speaker 4: the only people that can actually help you are the

524
00:26:18,359 --> 00:26:20,480
Speaker 4: people that are important to you, that you talk to,

525
00:26:20,840 --> 00:26:24,200
Speaker 4: that you care about in your life. If we could

526
00:26:24,280 --> 00:26:28,240
Speaker 4: use these algorithms to find that person that we could

527
00:26:28,320 --> 00:26:34,400
Speaker 4: recognize cared about you the most, but also was depressive

528
00:26:34,480 --> 00:26:37,720
Speaker 4: or could be empathetic to you. And it doesn't matter

529
00:26:37,760 --> 00:26:41,159
Speaker 4: if you have mother Teresa in your phone. If she

530
00:26:41,200 --> 00:26:43,560
Speaker 4: doesn't reply to your text messages, she's not going to

531
00:26:43,560 --> 00:26:45,919
Speaker 4: be able to help you. So we needed the algorithms

532
00:26:45,920 --> 00:26:48,960
Speaker 4: to find, you know, somebody who tries a lot harder

533
00:26:48,960 --> 00:26:51,120
Speaker 4: than you do to communicate than you do with them.

534
00:26:52,200 --> 00:26:54,840
Speaker 4: And so the algorithm goes through all that and finds

535
00:26:55,000 --> 00:26:58,560
Speaker 4: the person that might be right for you and suggests

536
00:26:58,560 --> 00:27:00,520
Speaker 4: that you reach out to the person.

537
00:27:01,400 --> 00:27:03,600
Speaker 1: That's what you guys do. If you notice that someone

538
00:27:03,720 --> 00:27:06,720
Speaker 1: is getting depressive or has threatened their own life, may

539
00:27:06,920 --> 00:27:09,520
Speaker 1: will suggest that AI will suggest you reach out to

540
00:27:09,520 --> 00:27:13,240
Speaker 1: someone and based on your AI, you identify a person

541
00:27:13,840 --> 00:27:16,240
Speaker 1: based on these messages who you think would be a

542
00:27:16,240 --> 00:27:17,240
Speaker 1: good person to talk to.

543
00:27:17,640 --> 00:27:17,920
Speaker 2: Yes.

544
00:27:18,200 --> 00:27:20,440
Speaker 4: Yes, And the reason we had to go through kind

545
00:27:20,440 --> 00:27:22,800
Speaker 4: of reading these messages was we had to back test

546
00:27:22,840 --> 00:27:25,520
Speaker 4: the algorithm. So we would go through these messages and

547
00:27:25,560 --> 00:27:27,520
Speaker 4: there might be five or six people that they're texting,

548
00:27:28,359 --> 00:27:30,920
Speaker 4: and we go, I think this is the best person

549
00:27:30,960 --> 00:27:34,200
Speaker 4: to reach out to. So we went through multiple iterations

550
00:27:34,200 --> 00:27:38,000
Speaker 4: of the algorithm, and then we found that the one

551
00:27:38,040 --> 00:27:41,480
Speaker 4: iteration where that algorithm was going to pick that exact

552
00:27:41,600 --> 00:27:42,960
Speaker 4: person that we were going to pick.

553
00:27:45,080 --> 00:27:49,040
Speaker 1: But, like, doesn't that feel like plan God rolling the dice.

554
00:27:49,680 --> 00:27:50,520
Speaker 2: I think.

555
00:27:52,040 --> 00:27:55,600
Speaker 4: It's normal to see a person in need and say, hey,

556
00:27:56,240 --> 00:27:58,720
Speaker 4: I think I could help that person. I feel like

557
00:27:59,000 --> 00:28:03,560
Speaker 4: it's a responsibility in a way then, right, because I

558
00:28:03,600 --> 00:28:10,199
Speaker 4: think in society, when you know there are others, there's

559
00:28:10,320 --> 00:28:13,159
Speaker 4: a part of a responsibility that grows out of that

560
00:28:13,720 --> 00:28:16,679
Speaker 4: for the other person to help you in need, you know.

561
00:28:17,040 --> 00:28:21,600
Speaker 4: So I actually do think that it is the responsibility

562
00:28:21,600 --> 00:28:24,080
Speaker 4: of the people who have access to.

563
00:28:24,040 --> 00:28:24,560
Speaker 2: Something like this.

564
00:28:25,640 --> 00:28:28,560
Speaker 1: Yeah, I mean, I think it's a really interesting and

565
00:28:28,600 --> 00:28:31,080
Speaker 1: ethical conversation of what you guys are sitting on and

566
00:28:31,119 --> 00:28:33,120
Speaker 1: I want to get into the privacy. But also think

567
00:28:33,119 --> 00:28:36,160
Speaker 1: about Facebook, right, Like all these social networks can tell

568
00:28:36,240 --> 00:28:38,920
Speaker 1: if people are becoming depressive or threatening their own lives,

569
00:28:38,920 --> 00:28:42,320
Speaker 1: and it is you know, do they contact authorities, do

570
00:28:42,360 --> 00:28:45,280
Speaker 1: they put the suicide hotline number? And so it is

571
00:28:45,280 --> 00:28:48,080
Speaker 1: pretty unique that you guys are identifying and suggesting a

572
00:28:48,120 --> 00:28:52,840
Speaker 1: person based on artificial intelligence and text message data that

573
00:28:52,880 --> 00:28:55,960
Speaker 1: you have. It. It's a choice, right, and no one

574
00:28:56,000 --> 00:28:57,240
Speaker 1: knows what the right choice is.

575
00:28:58,120 --> 00:29:00,800
Speaker 4: So I think we've seen an example. I think of

576
00:29:00,840 --> 00:29:03,800
Speaker 4: the big tech firms that actually went and did something right.

577
00:29:04,320 --> 00:29:05,760
Speaker 4: I don't think they have a right to do that.

578
00:29:06,400 --> 00:29:08,440
Speaker 4: It's just logically, I don't think they have a right

579
00:29:08,480 --> 00:29:10,920
Speaker 4: to do that. Whereas what we do is we put

580
00:29:10,920 --> 00:29:13,600
Speaker 4: the power back into that person and we say, hey,

581
00:29:13,640 --> 00:29:15,680
Speaker 4: you might not have noticed this, but this person that

582
00:29:16,160 --> 00:29:18,240
Speaker 4: you know checks in with you, a friend that you

583
00:29:18,240 --> 00:29:21,320
Speaker 4: don't talk to that frequently. We figured out that this

584
00:29:21,440 --> 00:29:24,680
Speaker 4: was a person that might really understand what you're going through, right,

585
00:29:24,960 --> 00:29:27,160
Speaker 4: and then it's your choice to reach out to that person.

586
00:29:27,600 --> 00:29:29,080
Speaker 4: We don't do anything on your behalf.

587
00:29:29,600 --> 00:29:31,240
Speaker 1: I mean, and I think one of the most interesting

588
00:29:31,280 --> 00:29:34,760
Speaker 1: things you said there is like you pick a person

589
00:29:34,880 --> 00:29:38,080
Speaker 1: for someone who's threatening their own life to reach out to.

590
00:29:38,600 --> 00:29:40,400
Speaker 1: One of the tidbits that you kind of slipped in

591
00:29:40,400 --> 00:29:42,760
Speaker 1: there is the aau guys try to pick out someone

592
00:29:42,800 --> 00:29:46,240
Speaker 1: who's also depressive, which is certainly very interesting to me.

593
00:29:47,040 --> 00:29:50,440
Speaker 4: Now, preface this, Ten people is a small amount of data, right,

594
00:29:51,280 --> 00:29:53,640
Speaker 4: But when I was going through that exercise and I

595
00:29:53,680 --> 00:29:55,920
Speaker 4: plotted these things on a graph, I think people will

596
00:29:55,960 --> 00:29:58,120
Speaker 4: say like, hey, I don't know what the patterns of

597
00:29:58,400 --> 00:30:01,760
Speaker 4: suicidal tendencies look like. It must be really hard to

598
00:30:01,760 --> 00:30:04,880
Speaker 4: figure out. When you see some of these graphs, you're like,

599
00:30:05,040 --> 00:30:07,760
Speaker 4: oh my god, this is perfectly obvious.

600
00:30:08,320 --> 00:30:10,240
Speaker 2: So let's just say, you know, what do you mean?

601
00:30:10,480 --> 00:30:11,280
Speaker 2: So the usage of.

602
00:30:11,240 --> 00:30:15,040
Speaker 4: Negative words and absolutest words, there's research out there that's say,

603
00:30:15,040 --> 00:30:17,600
Speaker 4: when people are depressive, they tend to use these words more.

604
00:30:18,240 --> 00:30:20,280
Speaker 4: When we plotted these ten people out on a graph,

605
00:30:20,680 --> 00:30:24,680
Speaker 4: we plot in the negative word usage of them versus

606
00:30:24,880 --> 00:30:26,200
Speaker 4: everybody that they're talking to.

607
00:30:26,600 --> 00:30:29,520
Speaker 1: What words just asking for us?

608
00:30:30,760 --> 00:30:31,560
Speaker 2: It's a lot of words.

609
00:30:31,640 --> 00:30:37,160
Speaker 4: It's just negativity, negativity, you know, no, hate, sad, angry.

610
00:30:37,840 --> 00:30:41,280
Speaker 4: When we plotted out the word usage, negative word usage

611
00:30:41,360 --> 00:30:46,320
Speaker 4: on a graph between the person and everybody collectively, that

612
00:30:46,360 --> 00:30:49,480
Speaker 4: they talked to, it tends to show a pretty regular wave.

613
00:30:49,720 --> 00:30:53,560
Speaker 4: And we found that when the user was more negative

614
00:30:53,760 --> 00:30:57,239
Speaker 4: than they ever were, if the other people that they

615
00:30:57,280 --> 00:31:02,400
Speaker 4: talked to were just as negative, everything was fine. It's

616
00:31:02,440 --> 00:31:05,480
Speaker 4: when they were the most negative and there was the

617
00:31:05,480 --> 00:31:08,720
Speaker 4: biggest gap between how negative they were and the other

618
00:31:08,800 --> 00:31:14,560
Speaker 4: people that you started seeing behaviors. And I think when

619
00:31:14,560 --> 00:31:16,640
Speaker 4: I went through that exercise, by the time I came

620
00:31:16,720 --> 00:31:20,240
Speaker 4: to like the six or seventh graph, I didn't even

621
00:31:20,320 --> 00:31:20,720
Speaker 4: need to.

622
00:31:20,720 --> 00:31:22,000
Speaker 2: Look at the text messages.

623
00:31:22,160 --> 00:31:23,840
Speaker 4: I just looked at the graph and said, I'm going

624
00:31:23,920 --> 00:31:25,959
Speaker 4: to look at May twenty six because I think that

625
00:31:26,040 --> 00:31:28,680
Speaker 4: day they might have tried something wow, and I was

626
00:31:28,720 --> 00:31:30,080
Speaker 4: pretty close to most of them.

627
00:31:30,240 --> 00:31:32,560
Speaker 1: It's crazy, I mean even crazy. You think about the future,

628
00:31:32,600 --> 00:31:34,920
Speaker 1: like if you could predict I don't know, and this

629
00:31:34,960 --> 00:31:37,360
Speaker 1: gets into like minority report and a lot of stuff

630
00:31:37,400 --> 00:31:40,160
Speaker 1: like you know, this is used for good, but this

631
00:31:40,240 --> 00:31:42,600
Speaker 1: could also be used in a pretty scary way, like

632
00:31:42,640 --> 00:31:45,000
Speaker 1: could you start making some of these predictions if someone's

633
00:31:45,040 --> 00:31:50,640
Speaker 1: gonna rub a bank, do something bad, harm themselves, hurt someone.

634
00:31:50,880 --> 00:31:52,800
Speaker 1: You know, you could start making some of these really

635
00:31:52,800 --> 00:31:55,040
Speaker 1: interesting predictions based off of some of the data that

636
00:31:55,080 --> 00:31:57,240
Speaker 1: you guys are getting Although we'll get into the privacy

637
00:31:57,280 --> 00:31:59,360
Speaker 1: because it's a huge part of this, but it's anonymized

638
00:31:59,400 --> 00:32:01,840
Speaker 1: data at the moment. But I'm sure you can make

639
00:32:01,880 --> 00:32:04,000
Speaker 1: some insane predictions, right.

640
00:32:04,200 --> 00:32:09,200
Speaker 4: Yeah, yeah, right, like what, Well, we've never tried, but

641
00:32:09,280 --> 00:32:13,800
Speaker 4: you could theoretically, yes, but yes, I think there's the

642
00:32:13,800 --> 00:32:17,200
Speaker 4: possibility for that. And we've actually been in the four

643
00:32:17,280 --> 00:32:19,520
Speaker 4: years i've been working on this, really been open to

644
00:32:19,520 --> 00:32:22,240
Speaker 4: anybody coming along saying, hey, like I'm an expert in this,

645
00:32:22,320 --> 00:32:25,360
Speaker 4: I would love to look into this. I think this

646
00:32:25,520 --> 00:32:28,360
Speaker 4: data is very powerful, and so long as it's being

647
00:32:28,480 --> 00:32:30,840
Speaker 4: used by the people who kind of intend to use

648
00:32:30,880 --> 00:32:36,440
Speaker 4: it for public good and for education, I feel fine

649
00:32:36,440 --> 00:32:39,480
Speaker 4: with that. Right because none of us really know how

650
00:32:39,600 --> 00:32:44,720
Speaker 4: this technology will be used. It's almost like a hammer

651
00:32:44,720 --> 00:32:48,360
Speaker 4: could be used to build a house or tear something down, right,

652
00:32:48,480 --> 00:32:51,280
Speaker 4: It's up to the people how they choose to use it.

653
00:32:51,480 --> 00:32:53,760
Speaker 1: Well, you guys are a small startup, but big companies

654
00:32:53,880 --> 00:32:56,000
Speaker 1: use big data in all sorts of different types of ways.

655
00:32:56,000 --> 00:33:00,080
Speaker 1: This is the national conversation around our data and our privacy.

656
00:33:01,520 --> 00:33:03,760
Speaker 1: We've got to take another break to hear from our sponsors.

657
00:33:03,800 --> 00:33:06,080
Speaker 1: But when we come back I know what you're thinking.

658
00:33:06,520 --> 00:33:09,200
Speaker 1: Tech that sees all my text messages and figures out

659
00:33:09,200 --> 00:33:14,160
Speaker 1: these intimate things about my personality and relationships. What about privacy?

660
00:33:14,720 --> 00:33:34,080
Speaker 1: We'll dig in after the break. I want to get

661
00:33:34,080 --> 00:33:36,760
Speaker 1: into privacy a little bit because we've been speaking around it.

662
00:33:36,800 --> 00:33:38,800
Speaker 1: But it's probably, like I would say, the most important

663
00:33:38,840 --> 00:33:41,600
Speaker 1: part of what you guys are doing. You guys are

664
00:33:41,640 --> 00:33:44,800
Speaker 1: an AI that sits on top of messages and reads messages.

665
00:33:44,880 --> 00:33:47,680
Speaker 1: So like the elephant in the room is like, whoa,

666
00:33:47,880 --> 00:33:51,320
Speaker 1: what about our privacy? So what is your answer to

667
00:33:51,360 --> 00:33:54,080
Speaker 1: the question of you know, how do you protect user data?

668
00:33:54,800 --> 00:33:56,880
Speaker 2: Yeah, I think it's like security.

669
00:33:57,120 --> 00:34:00,840
Speaker 4: Nothing is fool proof, but you want to try to

670
00:34:00,840 --> 00:34:03,760
Speaker 4: do your best at every step. So we knew text

671
00:34:03,760 --> 00:34:09,080
Speaker 4: messaging data is probably some of the most intimate data, right,

672
00:34:10,280 --> 00:34:13,120
Speaker 4: so everything within may everything is off by default. We

673
00:34:13,160 --> 00:34:16,160
Speaker 4: don't collect anything. If you want to turn on the AI,

674
00:34:16,760 --> 00:34:19,040
Speaker 4: we pop up this pop up that says, hey, we're

675
00:34:19,040 --> 00:34:22,040
Speaker 4: collecting these messages. If you're not comfortable with this, don't

676
00:34:22,040 --> 00:34:24,319
Speaker 4: go any further. We did that with Crush and we

677
00:34:24,400 --> 00:34:27,080
Speaker 4: probably turned away half of the people, which is fine.

678
00:34:27,840 --> 00:34:30,160
Speaker 4: We want people to know exactly what they're getting into.

679
00:34:30,600 --> 00:34:32,440
Speaker 4: But then when we collect the data, we don't collect

680
00:34:32,480 --> 00:34:35,400
Speaker 4: anything more than we have to. So, for example, we

681
00:34:35,400 --> 00:34:38,239
Speaker 4: don't collect the pictures because we're not at a stage

682
00:34:38,239 --> 00:34:41,720
Speaker 4: where we can analyze pictures, so we only take.

683
00:34:41,560 --> 00:34:44,520
Speaker 2: What we need. We don't need your name to help you.

684
00:34:45,000 --> 00:34:48,480
Speaker 4: The only personally identifiable data that we have we get

685
00:34:48,600 --> 00:34:50,719
Speaker 4: is the telephone number, and we have to verify that

686
00:34:50,760 --> 00:34:52,680
Speaker 4: you own the account. There has to be a way

687
00:34:52,719 --> 00:34:56,160
Speaker 4: to authenticate you. We take that and actually hash it

688
00:34:56,320 --> 00:35:00,840
Speaker 4: to an ID that means nothing to anybody. And actually

689
00:35:00,840 --> 00:35:03,560
Speaker 4: that was a way to protect us against ourselves. What

690
00:35:03,600 --> 00:35:06,799
Speaker 4: do you mean meaning we can't look in our database,

691
00:35:07,000 --> 00:35:11,239
Speaker 4: find something, find a user, and then take that ide

692
00:35:11,400 --> 00:35:13,160
Speaker 4: and try to figure out what telephone.

693
00:35:12,760 --> 00:35:13,359
Speaker 2: Number it is.

694
00:35:14,280 --> 00:35:16,359
Speaker 1: So none of your employees could actually be like, oh,

695
00:35:16,400 --> 00:35:19,360
Speaker 1: this is a pretty interesting conversation with so and so,

696
00:35:19,440 --> 00:35:20,480
Speaker 1: I'd like to go look it up.

697
00:35:20,520 --> 00:35:24,040
Speaker 4: They couldn't do that, Yeah, right, And then we go, okay,

698
00:35:24,080 --> 00:35:26,320
Speaker 4: if you want to delete your data from our database,

699
00:35:26,840 --> 00:35:28,920
Speaker 4: we make it really easy to I know there's a

700
00:35:28,960 --> 00:35:30,799
Speaker 4: lot of apps out there that make it impossible too.

701
00:35:31,480 --> 00:35:33,680
Speaker 4: We just put the button in the app and just say, okay,

702
00:35:33,760 --> 00:35:34,319
Speaker 4: now you're going.

703
00:35:34,239 --> 00:35:34,880
Speaker 2: To be deleted.

704
00:35:35,719 --> 00:35:37,759
Speaker 4: So every step along the way we thought, okay, well,

705
00:35:37,840 --> 00:35:40,759
Speaker 4: you know, we're users of apps too, we're at a

706
00:35:40,760 --> 00:35:43,360
Speaker 4: cross roads. Which choice do we make? And we always

707
00:35:43,400 --> 00:35:45,520
Speaker 4: made the one that we thought was for the sake

708
00:35:45,560 --> 00:35:46,640
Speaker 4: of preserving privacy.

709
00:35:46,880 --> 00:35:50,560
Speaker 1: What's interesting though, is when we were thinking about Okay, well,

710
00:35:50,880 --> 00:35:53,440
Speaker 1: I wanted you to analyze some of my messages and

711
00:35:53,480 --> 00:35:55,480
Speaker 1: I don't have an Android. So right now it's available

712
00:35:55,480 --> 00:35:58,760
Speaker 1: on Android and the iOS app, you have to upload

713
00:35:58,800 --> 00:36:00,960
Speaker 1: your WhatsApp Yeah, messages.

714
00:36:01,000 --> 00:36:04,760
Speaker 2: Why is that so on iOS?

715
00:36:05,000 --> 00:36:08,239
Speaker 4: There's no third party apps that can access your text messages?

716
00:36:08,800 --> 00:36:09,120
Speaker 2: Got it?

717
00:36:09,200 --> 00:36:11,000
Speaker 1: Do you think at any point that would change. I'm

718
00:36:11,040 --> 00:36:14,560
Speaker 1: assuming that that's a hindrance to the business to a degree, right,

719
00:36:14,960 --> 00:36:15,560
Speaker 1: to a degree.

720
00:36:15,760 --> 00:36:19,960
Speaker 4: Yeah, I think Apple would say this is for the

721
00:36:20,000 --> 00:36:23,080
Speaker 4: sake of user privacy. Yeah, but I actually think the

722
00:36:23,080 --> 00:36:27,040
Speaker 4: better answers give them a choice. Right, if you, as

723
00:36:27,040 --> 00:36:29,480
Speaker 4: a user, and I'm telling you exactly what I'm using

724
00:36:29,520 --> 00:36:32,280
Speaker 4: this data for, if you choose to do so, shouldn't

725
00:36:32,320 --> 00:36:35,000
Speaker 4: you have the right to it's your messages?

726
00:36:35,640 --> 00:36:38,239
Speaker 1: Well, what's interesting though, I mean, and we had this

727
00:36:38,280 --> 00:36:40,360
Speaker 1: whole debate when we were thinking about what text messages

728
00:36:40,400 --> 00:36:42,200
Speaker 1: to send over to you. I was like, well, I

729
00:36:42,200 --> 00:36:44,319
Speaker 1: guess I have to ask so and so permission. I'm like, no,

730
00:36:44,440 --> 00:36:46,400
Speaker 1: I don't like it goes into this question do you

731
00:36:46,560 --> 00:36:50,760
Speaker 1: own your text messages? And there's legal precedent that ss

732
00:36:50,800 --> 00:36:55,200
Speaker 1: like if US send me a message, it's public domain,

733
00:36:55,280 --> 00:36:57,600
Speaker 1: right or I can you know, it doesn't matter, like

734
00:36:57,640 --> 00:37:00,200
Speaker 1: you don't have any right right, so like you can

735
00:37:00,239 --> 00:37:03,799
Speaker 1: do this analysis of anyone. I could send over text

736
00:37:03,840 --> 00:37:05,880
Speaker 1: messages of me and any friend and they don't have

737
00:37:05,920 --> 00:37:06,759
Speaker 1: to consent.

738
00:37:06,840 --> 00:37:09,880
Speaker 2: Right right. In the US, that is legal president.

739
00:37:09,960 --> 00:37:12,400
Speaker 1: Right, And then overseas, you guys comply.

740
00:37:12,719 --> 00:37:15,920
Speaker 4: Yeah, Yeah, GDPR is the big one in Europe. Yeah,

741
00:37:16,080 --> 00:37:19,520
Speaker 4: and we're compliant with that. In Canada they have a

742
00:37:19,560 --> 00:37:23,520
Speaker 4: little more stringent I guess property rights as it applies

743
00:37:23,600 --> 00:37:24,640
Speaker 4: to text messages.

744
00:37:24,960 --> 00:37:26,880
Speaker 1: And I saw you said something about you guys are

745
00:37:26,880 --> 00:37:29,120
Speaker 1: a small startup and you've never made a cent from

746
00:37:29,200 --> 00:37:31,399
Speaker 1: data and all this kind of stuff, and you've never

747
00:37:31,440 --> 00:37:33,759
Speaker 1: sold data. I mean, and the time that I've covered tech,

748
00:37:33,800 --> 00:37:36,319
Speaker 1: I've heard that before and I believe you, right, I

749
00:37:36,360 --> 00:37:38,759
Speaker 1: believe it. I think the issue, I think is that

750
00:37:38,760 --> 00:37:41,400
Speaker 1: that could change right in a heartbeat, and we've seen

751
00:37:41,400 --> 00:37:44,080
Speaker 1: that change and so there's less I think there's less

752
00:37:44,080 --> 00:37:46,480
Speaker 1: trust now, So how do you ensure to your users?

753
00:37:46,480 --> 00:37:48,680
Speaker 1: And maybe the problem isn't just small startups, it's really

754
00:37:48,719 --> 00:37:50,760
Speaker 1: the big companies that we need to be talking about

755
00:37:51,239 --> 00:37:54,040
Speaker 1: that won't change, right, that their data will continue to

756
00:37:54,040 --> 00:37:55,560
Speaker 1: be protected as you guys grow.

757
00:37:56,040 --> 00:37:59,839
Speaker 4: Right, Well, you are correct. We have never sold any

758
00:38:00,840 --> 00:38:04,080
Speaker 4: we don't have any plans to. But I have asked myself,

759
00:38:04,200 --> 00:38:06,839
Speaker 4: you know, even asked my team the question before. You know,

760
00:38:07,120 --> 00:38:09,239
Speaker 4: we don't need to now, but what if we were

761
00:38:09,280 --> 00:38:11,960
Speaker 4: backed against the wall? Yeah, I mean it does bring

762
00:38:12,040 --> 00:38:15,640
Speaker 4: up some very interesting questions. I mean, I'm all in

763
00:38:15,680 --> 00:38:20,239
Speaker 4: favor of regulation that kind of kind of modernizes how

764
00:38:20,280 --> 00:38:24,080
Speaker 4: we think about property and privacy. Whenever I bring up

765
00:38:24,120 --> 00:38:27,319
Speaker 4: this is what we do, people go, wow, there's application here,

766
00:38:27,360 --> 00:38:32,200
Speaker 4: there's application here. In my mind, I go, there could

767
00:38:32,239 --> 00:38:35,120
Speaker 4: be so much money made from the things that are necessary.

768
00:38:36,480 --> 00:38:38,719
Speaker 4: Why we would we need to turn to something that

769
00:38:38,800 --> 00:38:39,600
Speaker 4: is nefarious?

770
00:38:40,360 --> 00:38:41,120
Speaker 1: What do you mean by that?

771
00:38:41,600 --> 00:38:44,680
Speaker 4: So on the iOS side, we're able to tell you

772
00:38:44,760 --> 00:38:47,799
Speaker 4: whether you know the probability somebody likes you based on

773
00:38:47,840 --> 00:38:53,719
Speaker 4: the algorithm, and users pay for that. So the idea is,

774
00:38:54,480 --> 00:38:57,480
Speaker 4: with this information, it can be valuable to somebody, I

775
00:38:57,520 --> 00:38:59,719
Speaker 4: can sell it to them for what they're willing to pay,

776
00:39:01,040 --> 00:39:04,000
Speaker 4: as opposed to use it for to sell to somebody

777
00:39:04,000 --> 00:39:08,040
Speaker 4: else from marketing. I mean, it does bring up a

778
00:39:08,120 --> 00:39:12,480
Speaker 4: very good point of how do we monetize this. I

779
00:39:12,520 --> 00:39:17,680
Speaker 4: think it's important to be clear and to tell your

780
00:39:17,800 --> 00:39:22,160
Speaker 4: users what you're going to do. So we actually found

781
00:39:22,160 --> 00:39:24,200
Speaker 4: a solution about maybe two years ago.

782
00:39:25,920 --> 00:39:27,480
Speaker 2: We have a credit system in the app.

783
00:39:28,040 --> 00:39:30,520
Speaker 4: And what it does is when you give us your

784
00:39:30,600 --> 00:39:33,279
Speaker 4: text messages and when you label data for us, we

785
00:39:33,520 --> 00:39:37,000
Speaker 4: give you credits. And that's kind of like a placeholder

786
00:39:37,040 --> 00:39:39,920
Speaker 4: for what is the value of this data? I actually

787
00:39:39,920 --> 00:39:43,160
Speaker 4: don't know, right, so somebody uploads all of the conversations up.

788
00:39:43,880 --> 00:39:45,799
Speaker 4: You know, it contributes to the algorithms that might be

789
00:39:45,840 --> 00:39:48,560
Speaker 4: worth something someday. It contributes to the overall kind of

790
00:39:48,680 --> 00:39:51,280
Speaker 4: you know, AI and data ecosystem that might be worth

791
00:39:51,280 --> 00:39:54,040
Speaker 4: a lot someday. I don't know what that's worth, but

792
00:39:54,040 --> 00:39:56,319
Speaker 4: I'm going to give you a little placeholder that said, hey,

793
00:39:56,360 --> 00:39:59,560
Speaker 4: you gave me some information. We found that to be

794
00:39:59,840 --> 00:40:02,799
Speaker 4: a way that we can kind of return values to

795
00:40:02,840 --> 00:40:07,719
Speaker 4: the people that give us their data. So now in

796
00:40:07,760 --> 00:40:12,160
Speaker 4: the app, anytime you and goes, I think you seem

797
00:40:12,239 --> 00:40:13,720
Speaker 4: like you're a really empathetic person.

798
00:40:13,760 --> 00:40:14,560
Speaker 2: Am I right or wrong?

799
00:40:14,719 --> 00:40:18,640
Speaker 4: And you you know, agree disagree, You get a small credit.

800
00:40:19,320 --> 00:40:22,400
Speaker 4: And the idea is none of us know what this

801
00:40:22,640 --> 00:40:26,840
Speaker 4: data is worth, and maybe I for advertising reasons. Maybe

802
00:40:27,360 --> 00:40:29,840
Speaker 4: maybe this is a system that allows us to say, Okay,

803
00:40:29,840 --> 00:40:33,560
Speaker 4: well I choose to see an ad. You company are

804
00:40:33,600 --> 00:40:35,520
Speaker 4: making money for me seeing this ad.

805
00:40:36,480 --> 00:40:37,319
Speaker 2: We can give you a.

806
00:40:37,280 --> 00:40:40,840
Speaker 4: Small amount of those economics, and everybody is happy.

807
00:40:41,760 --> 00:40:44,520
Speaker 1: Right like this the idea that you kind of take

808
00:40:44,560 --> 00:40:48,919
Speaker 1: issues with the way that that companies take data right now,

809
00:40:48,920 --> 00:40:50,680
Speaker 1: It's like they take your data and we don't know

810
00:40:50,680 --> 00:40:53,040
Speaker 1: what's happening to it, and then we're advertising. It's under

811
00:40:53,040 --> 00:40:54,880
Speaker 1: the guys that it's free, right.

812
00:40:57,680 --> 00:41:02,359
Speaker 3: What Okay, does it say that you're secretly in love

813
00:41:02,400 --> 00:41:02,600
Speaker 3: with me?

814
00:41:02,920 --> 00:41:06,319
Speaker 1: No, it's just my predicted age is forty four? Are

815
00:41:06,320 --> 00:41:07,520
Speaker 1: you serious?

816
00:41:08,040 --> 00:41:09,479
Speaker 3: That's definitely not on a maturity level.

817
00:41:09,560 --> 00:41:12,759
Speaker 1: Okay, so my predicted age is forty four. I guess

818
00:41:12,800 --> 00:41:15,560
Speaker 1: for our listeners, I'm thirty four and it says your

819
00:41:15,560 --> 00:41:16,879
Speaker 1: predicted age is thirty six.

820
00:41:17,880 --> 00:41:19,320
Speaker 3: Whoa, I am thirty six?

821
00:41:19,600 --> 00:41:20,520
Speaker 2: How would it know that?

822
00:41:23,160 --> 00:41:26,759
Speaker 1: I want to get into my personal experience with May

823
00:41:26,920 --> 00:41:31,200
Speaker 1: because I think it's I thought it was super interesting

824
00:41:31,440 --> 00:41:34,480
Speaker 1: and it was such a personal experience because what we did,

825
00:41:34,560 --> 00:41:36,239
Speaker 1: I have an iPhone, So like what we did, just

826
00:41:36,239 --> 00:41:39,360
Speaker 1: for full disclosure, is we sent you guys the messages

827
00:41:39,440 --> 00:41:42,200
Speaker 1: to analyze, and I sent you a year's worth of

828
00:41:42,239 --> 00:41:45,120
Speaker 1: messages with Derek, who's a good friend but also my

829
00:41:45,160 --> 00:41:47,440
Speaker 1: co founder, and we thought it'd be interesting to do

830
00:41:47,480 --> 00:41:50,319
Speaker 1: it because we've been working together, we're also friends, to

831
00:41:50,360 --> 00:41:55,439
Speaker 1: see what your algorithm picked up about us. But like, man,

832
00:41:55,560 --> 00:41:58,480
Speaker 1: it took an hour and a half for my messages

833
00:41:58,520 --> 00:42:01,000
Speaker 1: to just download because we've down loaded all my messages,

834
00:42:01,120 --> 00:42:05,880
Speaker 1: which was horrifying, simply horrifying. When you talk about like novels,

835
00:42:05,880 --> 00:42:08,040
Speaker 1: like it was like the Great American Novel minus, like

836
00:42:08,080 --> 00:42:10,240
Speaker 1: any thing classy or interesting.

837
00:42:10,400 --> 00:42:11,719
Speaker 2: Did you read through some of it?

838
00:42:12,719 --> 00:42:15,680
Speaker 1: You read through some of it? Yeah, there were really.

839
00:42:15,680 --> 00:42:18,480
Speaker 1: I mean it's it's fascinating to see these like snippets

840
00:42:18,480 --> 00:42:21,040
Speaker 1: of conversations, especially with you, like years ago, I.

841
00:42:21,000 --> 00:42:22,839
Speaker 2: Mean you were like was that me? Could I have

842
00:42:23,640 --> 00:42:24,160
Speaker 2: something like that?

843
00:42:24,239 --> 00:42:26,400
Speaker 1: I mean, like unfortunately, I'm like, yes, that was me.

844
00:42:27,600 --> 00:42:30,160
Speaker 1: You know, So I'm curious to see what your algorithm

845
00:42:30,360 --> 00:42:32,880
Speaker 1: showed because I think, like, what you're doing is fascinating

846
00:42:32,880 --> 00:42:35,200
Speaker 1: because it's so beyond just to someone like me. It's like,

847
00:42:35,520 --> 00:42:38,040
Speaker 1: how do you better working relationships, friendships, all those kind

848
00:42:38,040 --> 00:42:39,400
Speaker 1: of stuff? So should we try it?

849
00:42:39,760 --> 00:42:41,040
Speaker 2: Yeah? Yeah, we should dive in.

850
00:42:41,200 --> 00:42:43,480
Speaker 1: I mean, and have you done this interview like knowing

851
00:42:43,680 --> 00:42:45,240
Speaker 1: my personality traits already?

852
00:42:45,760 --> 00:42:49,120
Speaker 2: No? No, no, I'm not. I'm not that creepy, organized

853
00:42:49,239 --> 00:42:49,799
Speaker 2: or any Yeah.

854
00:42:49,920 --> 00:42:51,400
Speaker 1: I didn't mean to say creepy organized.

855
00:42:52,800 --> 00:42:54,480
Speaker 2: No I haven't. I haven't looked at any of these.

856
00:42:54,600 --> 00:42:57,640
Speaker 4: Okay, So I did the same thing and downloaded your

857
00:42:57,640 --> 00:42:59,960
Speaker 4: conversations into a new phone of mind.

858
00:43:00,520 --> 00:43:03,120
Speaker 1: And is it going to show me or is this

859
00:43:03,120 --> 00:43:04,719
Speaker 1: also going to show me and Derek? Because if it's

860
00:43:04,719 --> 00:43:07,880
Speaker 1: going to show Derek, he's in the room, so we

861
00:43:08,000 --> 00:43:11,520
Speaker 1: might as well just bring him over. Right, he's glaring

862
00:43:11,560 --> 00:43:13,480
Speaker 1: at me, so that means we should absolutely bring him

863
00:43:13,520 --> 00:43:19,400
Speaker 1: over so grumpy. I wonder if his personality profiles is that.

864
00:43:19,840 --> 00:43:22,200
Speaker 2: Oh, I guess we will see.

865
00:43:23,840 --> 00:43:25,840
Speaker 4: So it is able to kind of give insights about

866
00:43:25,960 --> 00:43:29,239
Speaker 4: both Derek and you. So I guess not that you're

867
00:43:29,480 --> 00:43:30,960
Speaker 4: sitting down, Derek, you.

868
00:43:30,880 --> 00:43:32,200
Speaker 2: Can be the first rate.

869
00:43:32,280 --> 00:43:34,320
Speaker 1: So for our listeners, he's pressing a button.

870
00:43:34,600 --> 00:43:39,120
Speaker 4: Yes, I'm pressing a button. And it says you seem

871
00:43:39,239 --> 00:43:40,960
Speaker 4: passionate about their work.

872
00:43:41,600 --> 00:43:44,280
Speaker 1: I do, Derek does. Oh that's good.

873
00:43:44,760 --> 00:43:45,200
Speaker 3: That's good.

874
00:43:45,400 --> 00:43:47,480
Speaker 1: Yeah. I mean that's helpful given that we just launched

875
00:43:47,600 --> 00:43:50,560
Speaker 1: a company. We launched our mediacompany dot dot dot. So

876
00:43:50,880 --> 00:43:52,640
Speaker 1: I'm glad you're passionate about it. Thank god.

877
00:43:53,160 --> 00:43:55,239
Speaker 2: You see that. It says agree and disagree and this

878
00:43:55,280 --> 00:43:57,359
Speaker 2: is kind of our way of getting better. Right.

879
00:43:57,480 --> 00:44:00,640
Speaker 1: Oh, okay, So I'm going to go with agree on that.

880
00:44:02,000 --> 00:44:04,880
Speaker 4: So it says that you seem like you're more philosophical

881
00:44:04,960 --> 00:44:05,680
Speaker 4: than concrete.

882
00:44:07,440 --> 00:44:08,239
Speaker 3: Let me ponder that.

883
00:44:10,640 --> 00:44:11,279
Speaker 2: Agree on that?

884
00:44:12,120 --> 00:44:14,359
Speaker 1: And so what is it doing now? It's taking us

885
00:44:14,400 --> 00:44:16,080
Speaker 1: through an exercise for Derek.

886
00:44:17,000 --> 00:44:19,839
Speaker 4: It's it's trying to understand things about Derek, so that

887
00:44:19,920 --> 00:44:23,319
Speaker 4: if it figures out that there was this big personality

888
00:44:23,360 --> 00:44:26,320
Speaker 4: difference between the two of you and that might cause

889
00:44:26,520 --> 00:44:30,880
Speaker 4: you to kind of misunderstand, it'll try to give advice

890
00:44:30,880 --> 00:44:32,600
Speaker 4: on how to bridge that communication gap.

891
00:44:32,680 --> 00:44:34,760
Speaker 1: Okay, sounds useful.

892
00:44:36,320 --> 00:44:36,560
Speaker 2: Well.

893
00:44:37,080 --> 00:44:41,000
Speaker 1: From Laurie, why are you looking awkward?

894
00:44:41,560 --> 00:44:42,680
Speaker 2: I don't know how old you are.

895
00:44:43,760 --> 00:44:45,560
Speaker 1: It got it wrong by the way I looked at

896
00:44:45,560 --> 00:44:48,040
Speaker 1: the data. It got it wrong, and that's good. I'm

897
00:44:48,040 --> 00:44:50,120
Speaker 1: glad you felt a little awkward looking at me thinking

898
00:44:50,120 --> 00:44:52,600
Speaker 1: that that's not how old I was, because that's that's

899
00:44:52,680 --> 00:44:54,640
Speaker 1: ten years older than I actually am ass.

900
00:44:55,239 --> 00:44:58,000
Speaker 4: Yeah, maybe it's Maybe it's because it's a professional relationship.

901
00:44:58,120 --> 00:44:59,800
Speaker 1: Do you think it says I have an old soul?

902
00:45:00,120 --> 00:45:03,960
Speaker 1: What we were discussing earlier in the office. Potentially, so

903
00:45:04,239 --> 00:45:07,840
Speaker 1: for our listeners, your algorithm thinks that I'm forty four

904
00:45:08,120 --> 00:45:10,640
Speaker 1: and I'm actually thirty four, So why do you think

905
00:45:10,680 --> 00:45:11,520
Speaker 1: it picked that up?

906
00:45:12,360 --> 00:45:13,480
Speaker 2: So usually you.

907
00:45:13,480 --> 00:45:15,400
Speaker 4: Would have a lot more information to go off of.

908
00:45:15,760 --> 00:45:18,440
Speaker 4: This was just one conversation, so you can imagine if

909
00:45:18,480 --> 00:45:21,160
Speaker 4: it was done fifty times. Maybe the way you text

910
00:45:21,800 --> 00:45:24,360
Speaker 4: Derek might be different from the way you text other people.

911
00:45:24,760 --> 00:45:27,239
Speaker 1: Maybe because we've been more professional over the last year.

912
00:45:27,280 --> 00:45:29,400
Speaker 1: I bet if we'd given it our twelve years, it

913
00:45:29,440 --> 00:45:31,399
Speaker 1: certainly would not have thought I was forty four.

914
00:45:31,920 --> 00:45:33,960
Speaker 3: But then so I also looked at the data and

915
00:45:34,000 --> 00:45:37,720
Speaker 3: it predicted that I'm thirty six, and that is correct.

916
00:45:38,520 --> 00:45:41,480
Speaker 1: Oh wow, all right, it's like your AI favors Derek,

917
00:45:41,560 --> 00:45:43,200
Speaker 1: which is no big deal. Let's keep going.

918
00:45:43,600 --> 00:45:46,719
Speaker 4: So the personality profiles, I actually think I included that

919
00:45:46,880 --> 00:45:50,239
Speaker 4: when I I sent over the information, what did you

920
00:45:50,239 --> 00:45:50,839
Speaker 4: think about that?

921
00:45:51,200 --> 00:45:52,759
Speaker 1: It was good? It was pretty spot on.

922
00:45:53,320 --> 00:45:59,200
Speaker 4: It says your top personality traits are empathetic, emotionally aware, dutyful,

923
00:45:59,480 --> 00:46:01,000
Speaker 4: altruously sick, and philosophical.

924
00:46:01,280 --> 00:46:04,280
Speaker 3: I think that's totally spot on for her. I've always

925
00:46:04,360 --> 00:46:07,000
Speaker 3: said that Laurie is one of the most empathetic people

926
00:46:07,000 --> 00:46:09,840
Speaker 3: I know, So the fact that empathy was the first

927
00:46:10,719 --> 00:46:12,879
Speaker 3: characteristic that it listed, I thought was spot on.

928
00:46:14,080 --> 00:46:15,280
Speaker 2: What did you think about yours?

929
00:46:15,440 --> 00:46:21,200
Speaker 4: It says that your top five are philosophical, empathetic, energetic, ambitious,

930
00:46:21,239 --> 00:46:21,920
Speaker 4: and fearless.

931
00:46:22,200 --> 00:46:24,520
Speaker 3: Yeah, I think so. Maybe I'd like some more of

932
00:46:24,520 --> 00:46:25,120
Speaker 3: that energy.

933
00:46:27,040 --> 00:46:29,759
Speaker 1: I like those for him? Yeah, I like those.

934
00:46:30,120 --> 00:46:33,120
Speaker 4: So it'll also look at your personnelity, and this is

935
00:46:33,200 --> 00:46:35,680
Speaker 4: just pure math when it just shows you how similar

936
00:46:35,719 --> 00:46:36,520
Speaker 4: you guys are.

937
00:46:37,480 --> 00:46:39,239
Speaker 2: A ninety similarity.

938
00:46:39,400 --> 00:46:41,560
Speaker 1: I mean, that's kind of great, right, How does I

939
00:46:41,600 --> 00:46:43,239
Speaker 1: compare to other people's similarities?

940
00:46:43,640 --> 00:46:46,120
Speaker 2: That's pretty good, pretty well.

941
00:46:46,200 --> 00:46:49,160
Speaker 4: I mean, here's one thing that we tend to try

942
00:46:49,200 --> 00:46:52,240
Speaker 4: to empathize and match the way the other person speaks.

943
00:46:53,360 --> 00:46:56,480
Speaker 4: So just looking at this slice, it kind of shows

944
00:46:56,560 --> 00:46:58,560
Speaker 4: you know whether whether or not you are these things.

945
00:46:58,600 --> 00:47:00,560
Speaker 4: This is how you want to come across to the

946
00:47:00,600 --> 00:47:04,920
Speaker 4: other person. And so yeah, I mean that's a good score.

947
00:47:05,160 --> 00:47:06,319
Speaker 1: It's great. Keep going.

948
00:47:06,640 --> 00:47:09,719
Speaker 4: Oh, let's see, let's look at the probability that this

949
00:47:09,800 --> 00:47:16,720
Speaker 4: is a romantic relationship. So we're gonna ask, does Derek

950
00:47:16,760 --> 00:47:21,000
Speaker 4: have a crush on me? So it is a thirty

951
00:47:21,040 --> 00:47:24,800
Speaker 4: six percent chance. It's a small chance, but it doesn't

952
00:47:24,800 --> 00:47:25,359
Speaker 4: seem likely.

953
00:47:25,440 --> 00:47:29,120
Speaker 1: I think his husband would agree with that truthfully.

954
00:47:30,000 --> 00:47:32,880
Speaker 3: Yes, that's very true. But I mean I would say

955
00:47:33,080 --> 00:47:34,840
Speaker 3: I have a crush on you in other ways.

956
00:47:35,000 --> 00:47:38,160
Speaker 1: Yeah, thank you.

957
00:47:38,200 --> 00:47:40,560
Speaker 3: No, I think that's amazing though, I mean we have

958
00:47:40,920 --> 00:47:43,919
Speaker 3: after looking at thousands of text messages over a year,

959
00:47:44,040 --> 00:47:47,879
Speaker 3: it's able to determine that basically we're friends. Yeah, there

960
00:47:47,920 --> 00:47:48,680
Speaker 3: is any more there?

961
00:47:48,880 --> 00:47:51,920
Speaker 1: Yeah, right, that's because it did. It categorized what we

962
00:47:51,960 --> 00:47:54,400
Speaker 1: are because technically we're like business partners if you were

963
00:47:54,440 --> 00:47:56,279
Speaker 1: to look at us on paper, but like, actually we're

964
00:47:56,320 --> 00:47:59,320
Speaker 1: really good friends, and that's how it categorized us from.

965
00:47:59,120 --> 00:48:00,439
Speaker 2: That, right.

966
00:48:00,680 --> 00:48:02,560
Speaker 1: Yeah, although I mean it did say I was ten

967
00:48:02,640 --> 00:48:03,799
Speaker 1: years older than I am, So.

968
00:48:04,280 --> 00:48:05,920
Speaker 2: All right, we're going to have to go back in

969
00:48:05,960 --> 00:48:06,600
Speaker 2: tweak that way.

970
00:48:06,719 --> 00:48:09,680
Speaker 1: Yeah, what about these other there were some graphs so.

971
00:48:09,960 --> 00:48:12,439
Speaker 4: What's on the app. There's kind of a lot more

972
00:48:12,719 --> 00:48:16,040
Speaker 4: behind the scenes. So those graphs were things that, you know,

973
00:48:16,160 --> 00:48:19,239
Speaker 4: are all things that contribute to the analysis that we

974
00:48:19,280 --> 00:48:23,080
Speaker 4: give to users, but it might also be information overload

975
00:48:23,080 --> 00:48:25,000
Speaker 4: for people who aren't looking for it.

976
00:48:25,440 --> 00:48:27,319
Speaker 1: But like, is it because I want to know, Like

977
00:48:27,400 --> 00:48:29,319
Speaker 1: could it tell us, like how I could communicate better

978
00:48:29,400 --> 00:48:31,880
Speaker 1: with him, or like if there were issues that we

979
00:48:31,960 --> 00:48:34,440
Speaker 1: had recently, Like it's okay, you know, like.

980
00:48:34,400 --> 00:48:36,080
Speaker 3: I'm please help, please help us.

981
00:48:36,280 --> 00:48:39,560
Speaker 1: Yeah, we're okay, Like we've been through some stuff together.

982
00:48:39,800 --> 00:48:42,440
Speaker 3: Well, what's so interesting to me about this technology is

983
00:48:42,440 --> 00:48:45,520
Speaker 3: that Laurie is someone who I communicate with probably more

984
00:48:45,560 --> 00:48:48,879
Speaker 3: than anybody, especially at this current phase of building out

985
00:48:48,920 --> 00:48:53,399
Speaker 3: dot dot dot, But with anybody you can communicate better, right,

986
00:48:53,440 --> 00:48:55,759
Speaker 3: And so if this technology can help us learn how

987
00:48:55,800 --> 00:48:58,080
Speaker 3: to communicate better to each other, and we're already good

988
00:48:58,120 --> 00:49:00,759
Speaker 3: friends and we're already high performing business partner, right, well,

989
00:49:00,800 --> 00:49:03,440
Speaker 3: then that just becomes that much more valuable that this

990
00:49:03,520 --> 00:49:06,400
Speaker 3: can give us insights about the way we communicate to

991
00:49:06,440 --> 00:49:08,800
Speaker 3: each other, which is predominantly over text message.

992
00:49:09,000 --> 00:49:09,120
Speaker 2: Right.

993
00:49:09,320 --> 00:49:12,000
Speaker 3: Laurie sends probably ten text messages per my one text

994
00:49:12,000 --> 00:49:13,280
Speaker 3: message that's just her stay.

995
00:49:13,480 --> 00:49:16,200
Speaker 1: The data shows I send three point six text messages.

996
00:49:16,280 --> 00:49:17,640
Speaker 1: Is that what it showed, Yeah.

997
00:49:17,480 --> 00:49:20,239
Speaker 2: Per day, whereas he was you know, half of.

998
00:49:20,200 --> 00:49:24,120
Speaker 3: That, which in this case is not not a indicator

999
00:49:24,120 --> 00:49:26,160
Speaker 3: of the fact that I'm not interested in her, it's

1000
00:49:26,200 --> 00:49:28,360
Speaker 3: just simply my style of communication.

1001
00:49:28,800 --> 00:49:29,080
Speaker 1: Right.

1002
00:49:30,000 --> 00:49:31,880
Speaker 4: Yeah, there's actually a tool in the hair that shows

1003
00:49:31,880 --> 00:49:36,000
Speaker 4: the relationship balance, and that's actually something that was carried

1004
00:49:36,040 --> 00:49:38,719
Speaker 4: over from crush. So what crush did was it not

1005
00:49:38,760 --> 00:49:40,960
Speaker 4: only showed you the balance of the relationship, but it

1006
00:49:40,960 --> 00:49:44,120
Speaker 4: actually showed you how that changed over time. And I

1007
00:49:44,160 --> 00:49:47,120
Speaker 4: think it's in a point you make that Okay, you

1008
00:49:47,200 --> 00:49:50,120
Speaker 4: might send a lot more messages than Lorie does.

1009
00:49:50,080 --> 00:49:52,200
Speaker 1: Right, and so like I send more messages.

1010
00:49:52,480 --> 00:49:55,560
Speaker 4: Sorry, yes, And people have said before like, okay, well

1011
00:49:55,560 --> 00:49:58,680
Speaker 4: that's why crush doesn't actually really work, because I'm just

1012
00:49:58,880 --> 00:50:01,640
Speaker 4: terrible at this. But if you actually look at your

1013
00:50:01,800 --> 00:50:05,160
Speaker 4: relationship balance over time, right, Like I've looked in enough

1014
00:50:05,200 --> 00:50:07,040
Speaker 4: of these graphs, and it's almost like you can kind

1015
00:50:07,040 --> 00:50:10,719
Speaker 4: of see like the ebb and flow of relationships. It's

1016
00:50:10,760 --> 00:50:13,799
Speaker 4: almost like a dance in a way, right, Like we've

1017
00:50:13,800 --> 00:50:18,120
Speaker 4: seen relationships where you actually memory first employee a week

1018
00:50:18,160 --> 00:50:20,319
Speaker 4: into the job. I was like you wanted to see

1019
00:50:20,320 --> 00:50:22,560
Speaker 4: this in action that I can actually just you know,

1020
00:50:22,640 --> 00:50:25,560
Speaker 4: look at your relationship with your boyfriend. And we went

1021
00:50:25,600 --> 00:50:27,200
Speaker 4: through and I was like, something happened on that day,

1022
00:50:27,200 --> 00:50:28,839
Speaker 4: on that day, on that day, and on that day.

1023
00:50:29,400 --> 00:50:32,440
Speaker 4: She was like, oh, yeah, shit, yeah, we had a

1024
00:50:32,480 --> 00:50:36,040
Speaker 4: fight on those days. And then I told her like, yeah,

1025
00:50:36,160 --> 00:50:39,440
Speaker 4: the relationship balance has never been so in your favor,

1026
00:50:39,480 --> 00:50:43,319
Speaker 4: like he's never tried so hard, and she said, yeah,

1027
00:50:43,360 --> 00:50:45,680
Speaker 4: we actually had a fight on Saturday and he's taking

1028
00:50:45,680 --> 00:50:47,080
Speaker 4: me out to dinner on Thursday.

1029
00:50:47,440 --> 00:50:49,799
Speaker 1: That's interesting though, like that you could actually look at

1030
00:50:49,800 --> 00:50:52,600
Speaker 1: the patterns of like over a long time and see

1031
00:50:52,719 --> 00:50:55,279
Speaker 1: kind of inherently what I mean, I just wonder, like

1032
00:50:56,040 --> 00:50:57,560
Speaker 1: is that a lot of power that you want or

1033
00:50:57,600 --> 00:51:00,000
Speaker 1: don't want? And like could it be a self fulfilling profit.

1034
00:51:00,280 --> 00:51:01,920
Speaker 1: I don't know, I'm interesting.

1035
00:51:01,960 --> 00:51:04,239
Speaker 3: My feeling is this type of data is already in

1036
00:51:04,280 --> 00:51:07,399
Speaker 3: the hands of people like advertisers, and so why can't

1037
00:51:07,440 --> 00:51:10,120
Speaker 3: we as consumers have this data in a way that

1038
00:51:10,440 --> 00:51:15,000
Speaker 3: is actionable and beneficial to ourselves and actionable? I mean,

1039
00:51:15,000 --> 00:51:16,800
Speaker 3: we all have people in our lives who we communicate

1040
00:51:16,840 --> 00:51:21,239
Speaker 3: with in a way that is either frustrating or causes anxiety,

1041
00:51:21,480 --> 00:51:23,720
Speaker 3: and if we had technology that could help us manage

1042
00:51:23,760 --> 00:51:26,560
Speaker 3: those relationships in that communication better, then that would be

1043
00:51:26,560 --> 00:51:27,440
Speaker 3: extremely powerful.

1044
00:51:28,239 --> 00:51:30,239
Speaker 1: But to give you the flip side of it, I

1045
00:51:30,280 --> 00:51:32,040
Speaker 1: read something that a Wired writer wrote and he was

1046
00:51:32,080 --> 00:51:34,319
Speaker 1: talking about it's almost like tear out cards, right, Like

1047
00:51:34,880 --> 00:51:37,319
Speaker 1: it is big data analytics, and I believe in it,

1048
00:51:37,560 --> 00:51:40,080
Speaker 1: and I've been obsessed with this idea for a while.

1049
00:51:40,120 --> 00:51:43,000
Speaker 1: And like, I was interviewing a guy years ago who

1050
00:51:43,040 --> 00:51:46,400
Speaker 1: does predictive data analytics to determine if something bad happens,

1051
00:51:46,800 --> 00:51:48,839
Speaker 1: And in the middle of the interview, he's like, I

1052
00:51:48,920 --> 00:51:51,560
Speaker 1: like he could predict if like a suicide bombing or

1053
00:51:51,600 --> 00:51:53,839
Speaker 1: something really terrible was going to happen in the area,

1054
00:51:53,840 --> 00:51:55,600
Speaker 1: and he had like a certain level of accuracy, and

1055
00:51:55,640 --> 00:51:57,759
Speaker 1: I think you would probably you would probably be like, Oh,

1056
00:51:57,760 --> 00:51:59,480
Speaker 1: that makes a lot of sense given what you do.

1057
00:52:00,239 --> 00:52:01,440
Speaker 1: And in the middle of the interview, he's like, I

1058
00:52:01,520 --> 00:52:04,839
Speaker 1: analyzed all your data on Facebook, Instagram, Twitter, everything you've

1059
00:52:04,880 --> 00:52:07,200
Speaker 1: said over the last seven or eight years, and he's like,

1060
00:52:07,280 --> 00:52:09,399
Speaker 1: I predict you're unhappy in your relationship and you're growing

1061
00:52:09,480 --> 00:52:12,239
Speaker 1: unhappy at your job. And I was like, WHOA, that's

1062
00:52:12,400 --> 00:52:15,480
Speaker 1: fascinating When I left my job and that relationship. I

1063
00:52:15,520 --> 00:52:16,840
Speaker 1: called him up. I was like, had you do it?

1064
00:52:16,840 --> 00:52:19,520
Speaker 1: And he talked a lot about you know, he talked

1065
00:52:19,560 --> 00:52:21,000
Speaker 1: a lot about how a lot of the stuff you're

1066
00:52:21,040 --> 00:52:23,600
Speaker 1: talking about negative and positive words, the time you tweet,

1067
00:52:23,600 --> 00:52:26,600
Speaker 1: the time you post, all this information that's out there.

1068
00:52:27,239 --> 00:52:29,120
Speaker 1: And I've talked with my tech context about it, and I

1069
00:52:29,160 --> 00:52:31,120
Speaker 1: think there's a lot there. But then there's also a

1070
00:52:31,120 --> 00:52:34,000
Speaker 1: lot of skepticism too, being like it's also like a

1071
00:52:34,080 --> 00:52:36,320
Speaker 1: terror card, right, like where you give us some things like,

1072
00:52:36,360 --> 00:52:39,200
Speaker 1: oh yeah, that totally makes sense. And a Wired writer

1073
00:52:39,400 --> 00:52:42,120
Speaker 1: wrote and so you have to respond to this, but

1074
00:52:42,160 --> 00:52:45,279
Speaker 1: he said whether a text analyzer reveals anything real or not,

1075
00:52:45,480 --> 00:52:48,520
Speaker 1: using one seems to offer a false sense of predictability

1076
00:52:48,560 --> 00:52:52,680
Speaker 1: and a semblance of control over otherwise messy human relationships.

1077
00:52:53,080 --> 00:52:56,120
Speaker 1: Does the emoji mean it's true love? Did the double

1078
00:52:56,160 --> 00:52:58,799
Speaker 1: text ruin the mood? Am I doing this right? The

1079
00:52:58,880 --> 00:53:01,320
Speaker 1: answer is displeasingly, never live in the app. The guidance

1080
00:53:01,320 --> 00:53:04,120
Speaker 1: there is about as useful as a deck of tarot cards.

1081
00:53:04,200 --> 00:53:07,080
Speaker 1: So what do you think, Well.

1082
00:53:07,200 --> 00:53:11,680
Speaker 4: You know the tarot card and the horoscope industry is huge.

1083
00:53:12,000 --> 00:53:14,600
Speaker 4: Then you know whether you believe it or not right,

1084
00:53:14,920 --> 00:53:19,319
Speaker 4: people are actually just looking for another data point. It's

1085
00:53:19,400 --> 00:53:21,719
Speaker 4: up to them whether they want to believe it or not.

1086
00:53:22,239 --> 00:53:24,319
Speaker 1: So, and then one other thing I'll go into is

1087
00:53:24,320 --> 00:53:27,480
Speaker 1: I think there is a The one thing after Cambridge

1088
00:53:27,480 --> 00:53:32,279
Speaker 1: Analytica and the whole tobaco with privacy and Facebook that

1089
00:53:32,440 --> 00:53:35,480
Speaker 1: really I think wasn't covered enough was this fine line

1090
00:53:35,520 --> 00:53:38,319
Speaker 1: between micro targeting and manipulation. This idea that we know

1091
00:53:38,480 --> 00:53:41,439
Speaker 1: so much about people and we can tell so much

1092
00:53:41,480 --> 00:53:43,880
Speaker 1: that just like you're tought, we've spent this whole time

1093
00:53:43,960 --> 00:53:46,239
Speaker 1: talking about we can tell so much about people. Now

1094
00:53:46,840 --> 00:53:48,760
Speaker 1: there is kind of a fine line with this AI

1095
00:53:49,000 --> 00:53:52,960
Speaker 1: being able to micro target or to manipulate, right, Like

1096
00:53:53,080 --> 00:53:55,400
Speaker 1: do you worry about kind of the future and some

1097
00:53:55,440 --> 00:53:56,720
Speaker 1: of those ethical lines.

1098
00:53:58,120 --> 00:54:02,320
Speaker 4: I think every piece of technology is controversial and worrying.

1099
00:54:03,320 --> 00:54:05,719
Speaker 4: I mean, I think the more powerful technology is, the

1100
00:54:05,719 --> 00:54:09,480
Speaker 4: more controversial it is, and it always came down to

1101
00:54:09,520 --> 00:54:14,920
Speaker 4: how people used it. I think it's inevitability, and I

1102
00:54:14,960 --> 00:54:18,160
Speaker 4: think we just we should just embrace the things that

1103
00:54:18,200 --> 00:54:21,759
Speaker 4: are and try to understand it, talk about it, yeah,

1104
00:54:21,840 --> 00:54:24,000
Speaker 4: and see how we can kind of live in the

1105
00:54:24,040 --> 00:54:26,880
Speaker 4: world with this tech, because it's not going away.

1106
00:54:27,640 --> 00:54:29,719
Speaker 1: How will we use it. I'm sitting here, Derek, how

1107
00:54:29,719 --> 00:54:31,719
Speaker 1: are we using this? This is like virgin, This is

1108
00:54:31,719 --> 00:54:34,040
Speaker 1: like what one point oh right, like two point oh

1109
00:54:34,120 --> 00:54:36,560
Speaker 1: one point oh somewhere? You know, what does this look

1110
00:54:36,640 --> 00:54:39,359
Speaker 1: like in ten fifteen years? How are we using this?

1111
00:54:40,239 --> 00:54:46,600
Speaker 4: I think this technology has the ability to simplify our lives.

1112
00:54:47,400 --> 00:54:50,160
Speaker 4: I think the choice will always still be ours on

1113
00:54:50,239 --> 00:54:52,719
Speaker 4: whether we take its advice or not. Yeah, I mean,

1114
00:54:52,880 --> 00:54:54,960
Speaker 4: I don't know where this is going to go. I

1115
00:54:55,040 --> 00:54:59,399
Speaker 4: just think if the people who have this information kind

1116
00:54:59,440 --> 00:55:01,760
Speaker 4: of do it, I guess, do it for good reason.

1117
00:55:02,480 --> 00:55:05,160
Speaker 4: That's all we could really ask for. I can see

1118
00:55:05,160 --> 00:55:08,160
Speaker 4: that didn't seem like a very gratifying response.

1119
00:55:08,360 --> 00:55:11,640
Speaker 1: No, it's not that it's not gratifying. I think it's optimistic.

1120
00:55:12,560 --> 00:55:24,080
Speaker 2: I think there's a double lets sword with every technology.

1121
00:55:24,239 --> 00:55:27,400
Speaker 1: The microphone is off and then s says something that

1122
00:55:27,640 --> 00:55:31,239
Speaker 1: I thought was fascinating, so I asked him permission to

1123
00:55:31,280 --> 00:55:34,799
Speaker 1: include it in the episode, since technically the interview was over.

1124
00:55:35,239 --> 00:55:38,560
Speaker 1: He said yes. S says May has over thirty thousand

1125
00:55:38,640 --> 00:55:42,280
Speaker 1: comments from users, and he said of those comments, only

1126
00:55:42,320 --> 00:55:46,120
Speaker 1: one percent are about privacy. Many of his users who've

1127
00:55:46,160 --> 00:55:49,239
Speaker 1: experienced May don't mind giving over their data. In fact,

1128
00:55:49,800 --> 00:55:52,799
Speaker 1: one of the most common requests, he says, is to

1129
00:55:52,840 --> 00:55:56,799
Speaker 1: give more, although you have to remember May's users have

1130
00:55:56,880 --> 00:55:59,640
Speaker 1: shown a willingness to share more than the average person.

1131
00:56:00,360 --> 00:56:03,080
Speaker 1: As says, users asked to just turn on their microphone

1132
00:56:03,160 --> 00:56:06,399
Speaker 1: and record all day get a much larger data set

1133
00:56:06,440 --> 00:56:10,920
Speaker 1: to analyze. So could more data points about our conversations

1134
00:56:11,040 --> 00:56:14,400
Speaker 1: give us more feedback that could lead to stronger relationships.

1135
00:56:14,960 --> 00:56:17,840
Speaker 1: And when given the choice, if we saw the power

1136
00:56:17,880 --> 00:56:20,399
Speaker 1: of May to help us learn more about each other

1137
00:56:20,520 --> 00:56:24,720
Speaker 1: to communicate better, is the trade off wortha We often

1138
00:56:24,760 --> 00:56:28,560
Speaker 1: hear these very strong arguments for privacy, but as s says,

1139
00:56:28,960 --> 00:56:32,160
Speaker 1: the feedback he gets is I'll give more if you

1140
00:56:32,280 --> 00:56:35,880
Speaker 1: give me more information that's valuable to me, although of

1141
00:56:35,920 --> 00:56:39,520
Speaker 1: course there's a lot of gray area in between. I'll

1142
00:56:39,560 --> 00:56:43,160
Speaker 1: leave you with that. I'm Lori Siegel and this is

1143
00:56:43,160 --> 00:56:47,719
Speaker 1: First Contact. For more about the guests you hear on

1144
00:56:47,760 --> 00:56:50,399
Speaker 1: First Contact, sign up for our newsletter. Go to First

1145
00:56:50,400 --> 00:56:54,000
Speaker 1: contactpodcast dot com to subscribe. Follow me I'm at Lori

1146
00:56:54,120 --> 00:56:56,919
Speaker 1: Siegel on Twitter and Instagram and the show is at

1147
00:56:56,920 --> 00:56:59,880
Speaker 1: First Contact Podcast. If you like the show, I want

1148
00:56:59,920 --> 00:57:01,600
Speaker 1: to hear from you, leave us a review, on the

1149
00:57:01,640 --> 00:57:05,080
Speaker 1: Apple Podcast app or wherever you listen and don't forget

1150
00:57:05,120 --> 00:57:08,240
Speaker 1: to subscribe see you don't miss an episode. First Contact

1151
00:57:08,280 --> 00:57:11,200
Speaker 1: is a production of dot dot dot Media. Executive produced

1152
00:57:11,200 --> 00:57:14,400
Speaker 1: by Lori Siegel and Derek Dodge. Original theme music by

1153
00:57:14,480 --> 00:57:22,800
Speaker 1: Xander Singh. Visit us at First contactpodcast dot com. First

1154
00:57:22,840 --> 00:57:25,080
Speaker 1: Contact with Lori Siegel is a production of Dot dot

1155
00:57:25,120 --> 00:57:26,800
Speaker 1: Dot Media and iHeartRadio