1
00:00:00,000 --> 00:00:01,240
Speaker 1: Webs will be automated.

2
00:00:01,480 --> 00:00:05,000
Speaker 2: We are shifting into an entirely different type of work.

3
00:00:05,080 --> 00:00:07,640
Speaker 2: The economy looks good in one element of it, but

4
00:00:07,760 --> 00:00:09,719
Speaker 2: terrible on the other because nobody's actually working.

5
00:00:09,880 --> 00:00:11,440
Speaker 3: What are the skills that we think that are going

6
00:00:11,480 --> 00:00:13,360
Speaker 3: to be transferable for the next generation?

7
00:00:13,600 --> 00:00:17,200
Speaker 2: The competitive advantage shifts to what questions are you asking

8
00:00:17,280 --> 00:00:18,120
Speaker 2: a super computer?

9
00:00:18,560 --> 00:00:21,400
Speaker 4: Pop? What we're saying that his scientists have already informed

10
00:00:21,480 --> 00:00:22,920
Speaker 4: him that immortality is.

11
00:00:22,840 --> 00:00:25,759
Speaker 2: A real thing, that is definitely not an impossibility at all.

12
00:00:25,920 --> 00:00:28,920
Speaker 2: Life expectancy should definitely increase.

13
00:00:29,080 --> 00:00:31,680
Speaker 4: We've essentially created something for the first time in human

14
00:00:31,720 --> 00:00:34,159
Speaker 4: history that's superior to us. That's pretty alarming.

15
00:00:34,240 --> 00:00:37,160
Speaker 2: They can see what your conversation history is, so once

16
00:00:37,200 --> 00:00:38,440
Speaker 2: that goes in, it can never get out.

17
00:00:38,520 --> 00:00:41,080
Speaker 4: Who's actually collecting all of his data? Is it going

18
00:00:41,240 --> 00:00:42,760
Speaker 4: to the US government?

19
00:00:42,840 --> 00:00:45,360
Speaker 1: If you can pay for more privacy, you'll get it.

20
00:00:45,479 --> 00:00:47,440
Speaker 1: But if you can't, you might be exposed.

21
00:00:47,479 --> 00:00:49,640
Speaker 4: What's the dark side? They are the.

22
00:00:49,560 --> 00:00:52,440
Speaker 2: Worst case scenario with AI. I mean, I don't think

23
00:00:52,479 --> 00:00:56,120
Speaker 2: that there's just one. This could go bad in many ways.

24
00:00:56,360 --> 00:00:58,040
Speaker 4: How can we really stop that? Though?

25
00:00:58,200 --> 00:01:01,600
Speaker 1: That is a national security risk that we can't even imagine.

26
00:01:01,800 --> 00:01:07,959
Speaker 4: Today. Okay, welcome back, e y else. Yes, we got

27
00:01:07,959 --> 00:01:09,600
Speaker 4: a highly anticipated episode.

28
00:01:09,840 --> 00:01:10,919
Speaker 3: We wain't on this one.

29
00:01:11,120 --> 00:01:14,720
Speaker 4: Yeah, a long time SHANEI Bovel, somebody that we met

30
00:01:14,720 --> 00:01:17,560
Speaker 4: Carnegie Hall two years ago.

31
00:01:19,520 --> 00:01:22,640
Speaker 3: Yeah, well two February's so this one passed and then

32
00:01:22,720 --> 00:01:23,320
Speaker 3: one before that.

33
00:01:23,600 --> 00:01:28,600
Speaker 4: Yeah, well Robert Smith might have heard of him. Yeah,

34
00:01:28,840 --> 00:01:30,520
Speaker 4: you know, we've been doing the Carnegie Hall things like

35
00:01:30,520 --> 00:01:34,880
Speaker 4: the last three years, and uh, you were there and

36
00:01:34,959 --> 00:01:38,399
Speaker 4: we met and you spoke about artificial intelligence and we

37
00:01:38,440 --> 00:01:43,039
Speaker 4: spoke backstage, and just very impressive as far as like

38
00:01:43,640 --> 00:01:47,880
Speaker 4: how you were able to communicate the very complex situation

39
00:01:47,880 --> 00:01:49,440
Speaker 4: when it comes to artificial intelligence. A lot of people

40
00:01:49,560 --> 00:01:53,040
Speaker 4: not still still not really familiar with everything that I

41
00:01:53,040 --> 00:01:55,240
Speaker 4: didn't tail to just hear about AI and now ONGBT,

42
00:01:55,440 --> 00:01:58,640
Speaker 4: but and then I just started following you from there,

43
00:01:59,280 --> 00:02:02,080
Speaker 4: and know we've talked several different times since then. So

44
00:02:02,520 --> 00:02:04,320
Speaker 4: I've just been a fan of what you've been able

45
00:02:04,360 --> 00:02:06,440
Speaker 4: to you know, put out there in the universe as

46
00:02:06,440 --> 00:02:10,160
Speaker 4: far as the education. So definitely wanted to have you

47
00:02:10,280 --> 00:02:16,080
Speaker 4: on the program to talk about AI and just the

48
00:02:16,160 --> 00:02:18,560
Speaker 4: landscape in general when it comes to technology because so

49
00:02:18,639 --> 00:02:21,359
Speaker 4: much is changing. So first and almost thank you for

50
00:02:21,440 --> 00:02:22,280
Speaker 4: joining us, appreciate it.

51
00:02:22,320 --> 00:02:24,480
Speaker 1: Thanks for having me. I know this has been some

52
00:02:24,600 --> 00:02:26,440
Speaker 1: time in the running, so I'm happy to be here.

53
00:02:26,760 --> 00:02:29,600
Speaker 3: I will say this a second. Everything he just said

54
00:02:29,960 --> 00:02:33,160
Speaker 3: when it comes to artificial intelligence, when it comes to

55
00:02:34,040 --> 00:02:37,200
Speaker 3: forward thinking, I think you're one of the best that

56
00:02:37,240 --> 00:02:39,959
Speaker 3: I've ever heard. Yeah, Like I watch you daily.

57
00:02:40,200 --> 00:02:40,600
Speaker 1: Thank you.

58
00:02:40,680 --> 00:02:43,120
Speaker 3: Yeah, people like hear like my takes on AI and

59
00:02:43,160 --> 00:02:46,160
Speaker 3: the complexities of it, and I'm like, it's a microcosm

60
00:02:46,200 --> 00:02:48,519
Speaker 3: of a lot of people that I'm listening to. You're

61
00:02:48,560 --> 00:02:50,359
Speaker 3: definitely one of them. So if they're not following you

62
00:02:50,639 --> 00:02:51,600
Speaker 3: by the end of this, they.

63
00:02:51,520 --> 00:02:54,280
Speaker 1: Will be, oh, thank you so much. Yeah, yeah, yeah, incredible,

64
00:02:54,280 --> 00:02:54,600
Speaker 1: thank you.

65
00:02:55,200 --> 00:03:00,720
Speaker 4: Futurist technology educator, you're entrepreneur anything else?

66
00:03:01,919 --> 00:03:03,640
Speaker 1: I mean not right now. I guess it depends on

67
00:03:03,680 --> 00:03:05,360
Speaker 1: how the future goes.

68
00:03:06,520 --> 00:03:08,400
Speaker 4: So talk about how did you get to this point?

69
00:03:08,400 --> 00:03:09,919
Speaker 4: How did you get to this point as far as

70
00:03:09,960 --> 00:03:14,519
Speaker 4: being a futurist AI expert, Like when did that actually

71
00:03:15,200 --> 00:03:16,960
Speaker 4: become your life path? Passion?

72
00:03:17,960 --> 00:03:20,280
Speaker 2: So I was one of those people who happened to

73
00:03:20,360 --> 00:03:23,240
Speaker 2: take that class that changed everything. So when I was

74
00:03:23,280 --> 00:03:26,679
Speaker 2: doing my MBA in Toronto, I took this class called

75
00:03:26,680 --> 00:03:29,680
Speaker 2: Strategic foresight, and that's the formal field of being a

76
00:03:29,680 --> 00:03:32,120
Speaker 2: future is. It's actually called the practice of strategic foresight.

77
00:03:32,560 --> 00:03:35,360
Speaker 2: And so I stumble into this class and I learned

78
00:03:35,360 --> 00:03:38,080
Speaker 2: that the future is something. Of course you can't predict it,

79
00:03:38,760 --> 00:03:42,440
Speaker 2: but there are formal methodologies you can use to understand

80
00:03:42,480 --> 00:03:46,680
Speaker 2: how the landscape is evolving, so from understanding economic flows

81
00:03:46,720 --> 00:03:50,680
Speaker 2: to behaviors to technology. So I'm tracking patents and doing

82
00:03:50,680 --> 00:03:52,920
Speaker 2: all these interesting things in this class, and then I

83
00:03:53,000 --> 00:03:55,440
Speaker 2: never really stopped doing that. I could never look at

84
00:03:56,000 --> 00:03:59,120
Speaker 2: markets the same or information the same. Once I had

85
00:03:59,160 --> 00:04:02,320
Speaker 2: that methodology in my mind of understanding where this is going,

86
00:04:02,800 --> 00:04:06,040
Speaker 2: I didn't know though it could be a formal job. Really,

87
00:04:06,080 --> 00:04:09,080
Speaker 2: I knew it was a practice, and so I went

88
00:04:09,200 --> 00:04:10,640
Speaker 2: into the world of management consulting.

89
00:04:11,280 --> 00:04:13,400
Speaker 1: I didn't really like that very much.

90
00:04:13,440 --> 00:04:15,280
Speaker 2: It was fine at the time, just not really the

91
00:04:15,280 --> 00:04:17,880
Speaker 2: problems I was interested in solving, so a lot of

92
00:04:17,920 --> 00:04:20,600
Speaker 2: different zigs and z eggs in my career. Went into

93
00:04:20,600 --> 00:04:22,920
Speaker 2: the world of fashion, and from there, as I started

94
00:04:22,920 --> 00:04:25,599
Speaker 2: to blog about the future and host talks about the

95
00:04:25,600 --> 00:04:28,159
Speaker 2: future and all the trends I was tracking, the light

96
00:04:28,160 --> 00:04:30,279
Speaker 2: bulb just went off, like I'm already here. This is

97
00:04:30,279 --> 00:04:32,520
Speaker 2: what I'm doing. So i'd say, back.

98
00:04:32,320 --> 00:04:34,120
Speaker 4: In you were a model.

99
00:04:34,520 --> 00:04:38,400
Speaker 2: Yeah, so between consulting and what I do now, I

100
00:04:38,920 --> 00:04:41,760
Speaker 2: was a fashion model for a little bit. Yeah, it's

101
00:04:41,800 --> 00:04:43,960
Speaker 2: something that seems entirely unrelated, but it was actually the

102
00:04:44,000 --> 00:04:47,919
Speaker 2: springboard to stepping into this world. I'd say twenty sixteen

103
00:04:48,000 --> 00:04:50,240
Speaker 2: is when I really started to track AI the breakthroughs

104
00:04:50,279 --> 00:04:53,640
Speaker 2: that were happening. And then from there it's been like

105
00:04:53,720 --> 00:04:56,599
Speaker 2: reading all the white papers and never really look back.

106
00:04:56,680 --> 00:04:59,159
Speaker 2: And then I think the world woke up to AI

107
00:04:59,240 --> 00:05:01,640
Speaker 2: in twenty twenty TI and then the rest of the

108
00:05:01,720 --> 00:05:02,880
Speaker 2: journey is kind of history.

109
00:05:04,440 --> 00:05:07,760
Speaker 3: I'm interested in that moment, right, Like, as you are

110
00:05:07,800 --> 00:05:10,080
Speaker 3: watching the trend, I find this is very something that

111
00:05:10,120 --> 00:05:14,359
Speaker 3: we have in common. I'm watching a trend. I've studied

112
00:05:14,360 --> 00:05:18,400
Speaker 3: the trend, but saying it to the public is different, right,

113
00:05:18,440 --> 00:05:21,360
Speaker 3: because there's a chance that this might not come to fruition.

114
00:05:21,480 --> 00:05:26,360
Speaker 3: You could be wrong. Did you have that, I guess,

115
00:05:27,000 --> 00:05:29,640
Speaker 3: fear of saying that this might not go the way

116
00:05:29,640 --> 00:05:33,479
Speaker 3: I think it is. And then twenty twenty two happens

117
00:05:34,279 --> 00:05:39,040
Speaker 3: that validate talk about that those moments.

118
00:05:37,839 --> 00:05:41,960
Speaker 2: I didn't, so I think because when you look at

119
00:05:41,960 --> 00:05:44,200
Speaker 2: what AI is, and I had been. I mean I

120
00:05:44,240 --> 00:05:47,159
Speaker 2: happened to go to the university where all of those

121
00:05:47,160 --> 00:05:50,560
Speaker 2: breakthroughs were happening, So University of Toronto where Jeff Hinton,

122
00:05:50,600 --> 00:05:53,520
Speaker 2: the godfather of AI, ilias Setskiver, who went over to

123
00:05:53,640 --> 00:05:56,479
Speaker 2: Opening Eye to start CHATEGBT, they were doing their research

124
00:05:56,480 --> 00:05:59,120
Speaker 2: at the University of Toronto, and my professors were working

125
00:05:59,160 --> 00:06:03,359
Speaker 2: with them, so that understanding of this technology is coming.

126
00:06:04,279 --> 00:06:06,760
Speaker 2: We just don't know this scale and when it's going

127
00:06:06,839 --> 00:06:09,360
Speaker 2: to be at that Internet level that was already in

128
00:06:09,400 --> 00:06:12,640
Speaker 2: the data points and then I think, you know the future,

129
00:06:13,040 --> 00:06:15,240
Speaker 2: and there's that famous quote, It's already here, it's just

130
00:06:15,279 --> 00:06:19,080
Speaker 2: not evenly distributed. You could see AI in the patents,

131
00:06:19,120 --> 00:06:22,279
Speaker 2: in the white papers starting to become infrastructure in software

132
00:06:22,320 --> 00:06:26,080
Speaker 2: and applications. So in twenty sixteen, twenty seventeen, you could

133
00:06:26,120 --> 00:06:29,479
Speaker 2: see those trend lines and they weren't entirely different from

134
00:06:29,520 --> 00:06:31,880
Speaker 2: looking at electricity and being like, okay, so it's working

135
00:06:31,920 --> 00:06:34,520
Speaker 2: in some use cases. Some people doubt it, but this

136
00:06:34,560 --> 00:06:36,839
Speaker 2: thing is coming. And I think it was twenty eighteen

137
00:06:37,680 --> 00:06:41,479
Speaker 2: that Mackenzie did that study that by twenty thirty, fifty

138
00:06:41,560 --> 00:06:45,479
Speaker 2: percent of jobs will be impacted or automated by artificial intelligence.

139
00:06:46,160 --> 00:06:47,720
Speaker 1: And that's when it was literally that.

140
00:06:47,600 --> 00:06:50,720
Speaker 2: Study that I was like, there is a train headed

141
00:06:50,960 --> 00:06:53,880
Speaker 2: straight to the workforce, straight to the economy at full speed,

142
00:06:54,480 --> 00:06:56,600
Speaker 2: and hardly anybody's thinking about it.

143
00:06:57,080 --> 00:06:57,720
Speaker 1: And that's when I.

144
00:06:57,760 --> 00:07:01,800
Speaker 2: Entirely reposition what I spoke about and made it the

145
00:07:01,880 --> 00:07:05,039
Speaker 2: mission to make technology and AI a language everyone speaks

146
00:07:05,200 --> 00:07:07,440
Speaker 2: because this thing's coming and it's going to change everything.

147
00:07:07,680 --> 00:07:10,320
Speaker 3: So in twenty twenty two, I feel like, this is

148
00:07:10,320 --> 00:07:11,400
Speaker 3: the open AI moment.

149
00:07:11,640 --> 00:07:14,400
Speaker 1: Yeah, what was that like for you?

150
00:07:14,640 --> 00:07:17,040
Speaker 2: So I think in what really happened in twenty twenty two,

151
00:07:17,200 --> 00:07:23,320
Speaker 2: it was an interface breakthrough or revolution. So AI had

152
00:07:23,400 --> 00:07:26,560
Speaker 2: still been on this really incredible path forward, and what

153
00:07:26,680 --> 00:07:30,160
Speaker 2: happens with Chatchibt is now the system comes on the

154
00:07:30,200 --> 00:07:33,480
Speaker 2: scene that the public can interact with in a way

155
00:07:33,520 --> 00:07:37,760
Speaker 2: that they understand. So it turned AI from a technology

156
00:07:37,760 --> 00:07:41,760
Speaker 2: that computer scientists could only really use and understand to

157
00:07:42,320 --> 00:07:47,440
Speaker 2: a platform that anybody can now engage and direct. And

158
00:07:47,520 --> 00:07:51,920
Speaker 2: once that interface moment happens, it as soon as human

159
00:07:51,960 --> 00:07:55,680
Speaker 2: behavior changes, it never goes back. Right, So as soon

160
00:07:55,720 --> 00:07:58,800
Speaker 2: as we switch from plugging in our internet to Wi Fi,

161
00:07:58,960 --> 00:08:01,160
Speaker 2: no one's going back. So as soon as that chat

162
00:08:01,360 --> 00:08:04,800
Speaker 2: BT interface moment happens, we can talk to AI the

163
00:08:04,840 --> 00:08:07,880
Speaker 2: way we talk to a person. That is now the

164
00:08:07,920 --> 00:08:10,360
Speaker 2: new platform and we're never going back. So I think

165
00:08:10,400 --> 00:08:13,360
Speaker 2: that was Yeah, I guess another light bulb moment or

166
00:08:13,480 --> 00:08:17,320
Speaker 2: more validation in the forecast that this thing is going

167
00:08:17,360 --> 00:08:19,560
Speaker 2: to permeate our lives and it's going to be the

168
00:08:19,680 --> 00:08:22,480
Speaker 2: new layer that everything gets built on top of. And

169
00:08:22,480 --> 00:08:25,200
Speaker 2: then I think the momentum with governments and everybody really

170
00:08:25,200 --> 00:08:28,800
Speaker 2: talking about it showed the scale of that and kind

171
00:08:28,840 --> 00:08:29,560
Speaker 2: of supported that.

172
00:08:29,920 --> 00:08:31,920
Speaker 1: But I didn't really doubt.

173
00:08:32,040 --> 00:08:34,840
Speaker 2: You can't predict the speed, and you can't predict the

174
00:08:34,880 --> 00:08:37,640
Speaker 2: breakthroughs and when they're going to happen. But when something

175
00:08:37,720 --> 00:08:41,000
Speaker 2: is going to be infrastructure or a general purpose, there's

176
00:08:41,080 --> 00:08:43,400
Speaker 2: no doubt that that's going to happen.

177
00:08:43,440 --> 00:08:44,320
Speaker 1: It's just the when.

178
00:08:45,200 --> 00:08:47,480
Speaker 4: And when does Netscape come into play?

179
00:08:48,920 --> 00:08:51,640
Speaker 2: Netscape comes into play? It's like what nineteen ninety two

180
00:08:52,360 --> 00:08:57,400
Speaker 2: I think. I mean, yeah, I think Netscape is the OG.

181
00:08:57,880 --> 00:09:01,560
Speaker 4: No, but your company, what's your way? A way?

182
00:09:01,760 --> 00:09:03,600
Speaker 1: Netscape? I mean I would like to take credit for

183
00:09:03,679 --> 00:09:07,559
Speaker 1: say you probably would No, I would. I would have

184
00:09:07,640 --> 00:09:08,360
Speaker 1: come in on a jet.

185
00:09:08,640 --> 00:09:09,800
Speaker 3: It would have landed on the roof.

186
00:09:09,920 --> 00:09:12,000
Speaker 1: Yeah, we would be doing this for my version of

187
00:09:12,040 --> 00:09:12,560
Speaker 1: Air Force one.

188
00:09:12,760 --> 00:09:12,960
Speaker 3: Yeah.

189
00:09:13,760 --> 00:09:16,959
Speaker 2: No, my company way comes into play in twenty eighteen

190
00:09:17,080 --> 00:09:20,000
Speaker 2: is when I founded it, and so yeah, everything kind

191
00:09:20,000 --> 00:09:22,840
Speaker 2: of around that that same time. And I mean people

192
00:09:22,920 --> 00:09:25,880
Speaker 2: were listening and showing up to the talks, and we

193
00:09:25,880 --> 00:09:27,240
Speaker 2: were talking about the statistics.

194
00:09:27,280 --> 00:09:28,760
Speaker 1: I mean, AI is coming forward.

195
00:09:28,760 --> 00:09:30,520
Speaker 2: I think one of the stats that I quoted is

196
00:09:30,520 --> 00:09:33,320
Speaker 2: that by twenty twenty five, AI will be better than

197
00:09:33,320 --> 00:09:36,360
Speaker 2: a human at writing essays. And people in the audience

198
00:09:36,440 --> 00:09:39,960
Speaker 2: kind of chuckled, But then that's turned out to be

199
00:09:40,320 --> 00:09:42,880
Speaker 2: the way. But people listened and they were ready. And

200
00:09:42,960 --> 00:09:44,920
Speaker 2: I think because even shows about the future they do

201
00:09:45,040 --> 00:09:48,400
Speaker 2: really well. Everybody's curious about it. But before it became

202
00:09:48,440 --> 00:09:50,120
Speaker 2: relevant to your day to day life, it just wasn't

203
00:09:50,160 --> 00:09:50,800
Speaker 2: as interesting.

204
00:09:52,120 --> 00:09:55,880
Speaker 4: Okay, well, I was going to say as far as okay,

205
00:09:56,160 --> 00:09:59,400
Speaker 4: getting into the AI conversation, we spoke and you talked

206
00:09:59,400 --> 00:10:02,960
Speaker 4: about the report that said that are twenty twenty eight

207
00:10:04,240 --> 00:10:09,280
Speaker 4: white collar jobs will be obsolete due to artificial intelligence. Yes,

208
00:10:09,280 --> 00:10:11,760
Speaker 4: that's pretty that's pretty soon, So explain that, like that's

209
00:10:11,800 --> 00:10:13,119
Speaker 4: that's that's interesting.

210
00:10:13,360 --> 00:10:16,480
Speaker 2: So this is the Centrini and report. You're probably talking

211
00:10:16,520 --> 00:10:20,080
Speaker 2: about and it was called the Great Intelligence Crisis. And

212
00:10:20,120 --> 00:10:23,400
Speaker 2: so what this report is from this financial analytics company,

213
00:10:23,760 --> 00:10:28,280
Speaker 2: and they essentially painted this fictional scenario that in twenty

214
00:10:28,360 --> 00:10:33,319
Speaker 2: twenty eight, AI does most white collar jobs in the economy.

215
00:10:33,360 --> 00:10:36,080
Speaker 2: So they make this forecast. Okay, so as companies start

216
00:10:36,120 --> 00:10:39,040
Speaker 2: to automate roles knowledge workers, so anybody that's doing something

217
00:10:39,080 --> 00:10:40,640
Speaker 2: on a computer, that.

218
00:10:40,520 --> 00:10:42,800
Speaker 1: Those are the types of jobs that AI is coming

219
00:10:42,800 --> 00:10:43,520
Speaker 1: for at this point.

220
00:10:44,080 --> 00:10:47,120
Speaker 2: And so this kind of thesis claims that as companies

221
00:10:47,160 --> 00:10:51,360
Speaker 2: start to automate more white collar workers, those people don't

222
00:10:51,400 --> 00:10:55,120
Speaker 2: have income coming in anymore, so they don't buy things

223
00:10:55,120 --> 00:10:58,560
Speaker 2: in the economy anymore, and so companies start to lose money.

224
00:10:58,600 --> 00:11:01,760
Speaker 2: So they put more AI into they're into their infrastructure

225
00:11:01,800 --> 00:11:05,600
Speaker 2: to lower costs, firing more people, and the economy looks

226
00:11:05,600 --> 00:11:09,079
Speaker 2: good in one element of it because productivity is going up,

227
00:11:09,360 --> 00:11:12,720
Speaker 2: but terrible in the other because nobody's actually working. And

228
00:11:12,760 --> 00:11:14,840
Speaker 2: then it leads to the spiral and collapse and that

229
00:11:15,000 --> 00:11:17,480
Speaker 2: shook markets a bit two weeks ago because of this

230
00:11:18,040 --> 00:11:21,200
Speaker 2: kind of fictional scenario. So I think what the report

231
00:11:21,240 --> 00:11:24,800
Speaker 2: gets right, it's a lot wrong. What it gets right

232
00:11:24,880 --> 00:11:27,040
Speaker 2: is that, yeah, jobs are.

233
00:11:26,960 --> 00:11:29,960
Speaker 1: Going to change.

234
00:11:30,320 --> 00:11:35,520
Speaker 2: Jobs will be automated, a lot will be augmented or transformed.

235
00:11:36,320 --> 00:11:40,440
Speaker 2: But we are shifting into an entirely different type of work.

236
00:11:41,000 --> 00:11:44,280
Speaker 2: And how I would describe it, I mean, let's say

237
00:11:44,280 --> 00:11:46,720
Speaker 2: if we went back to pre industrial revolution and let's

238
00:11:46,720 --> 00:11:49,560
Speaker 2: say like seventy ninety percent of people working in agriculture,

239
00:11:50,000 --> 00:11:53,000
Speaker 2: and then by the end of that the steam engine,

240
00:11:53,200 --> 00:11:55,760
Speaker 2: let's say ten to twenty percent of people are doing that.

241
00:11:56,400 --> 00:12:00,720
Speaker 2: What would it mean for seventy percent of the workforce

242
00:12:00,840 --> 00:12:04,400
Speaker 2: to do something different? But not in one hundred years,

243
00:12:04,880 --> 00:12:09,520
Speaker 2: in ten. So I don't think work is going away,

244
00:12:10,600 --> 00:12:14,480
Speaker 2: but jobs as we know them will, and a lot

245
00:12:14,480 --> 00:12:17,760
Speaker 2: of the job titles that we reference today will. We

246
00:12:17,840 --> 00:12:21,080
Speaker 2: will do different things that are hard to foresee, the

247
00:12:21,120 --> 00:12:23,880
Speaker 2: same way people have entire careers over filming ninety second

248
00:12:23,960 --> 00:12:26,480
Speaker 2: videos in their car and all sorts of new things.

249
00:12:27,200 --> 00:12:30,480
Speaker 2: But the most of the work we see today will

250
00:12:30,520 --> 00:12:34,000
Speaker 2: shift and will change. So we have to change. Our

251
00:12:34,040 --> 00:12:36,679
Speaker 2: skills have to change. And there's a lot that's up

252
00:12:36,679 --> 00:12:39,520
Speaker 2: in the air. And what tends to happen is jobs

253
00:12:39,559 --> 00:12:41,920
Speaker 2: can disappear faster than the new ones get created, and

254
00:12:41,920 --> 00:12:44,760
Speaker 2: that's where things get rocky. But I don't think work

255
00:12:44,800 --> 00:12:47,760
Speaker 2: is entirely going away, but it's going to look very different.

256
00:12:48,360 --> 00:12:50,000
Speaker 3: I'm with you, I'm with you. I'm on the side

257
00:12:50,000 --> 00:12:52,920
Speaker 3: that new jobs, new careers, new ideas are going to

258
00:12:52,920 --> 00:12:57,800
Speaker 3: be created. I heard that this comparison of you know,

259
00:12:57,920 --> 00:13:01,120
Speaker 3: if you think one hundred years ago, mail was delivered

260
00:13:01,160 --> 00:13:05,560
Speaker 3: on horses, and then you had mailmen, and then in

261
00:13:05,600 --> 00:13:09,079
Speaker 3: the nineteen nineties we got this thing called email. You're

262
00:13:09,080 --> 00:13:12,160
Speaker 3: still getting the information. It has just evolved. We still

263
00:13:12,160 --> 00:13:14,600
Speaker 3: have a post office, but we still get email. And

264
00:13:14,679 --> 00:13:17,480
Speaker 3: it feels like technology will change, careers will change, and

265
00:13:17,520 --> 00:13:21,240
Speaker 3: we'll find a new people. So I'm bullish in that standpoint.

266
00:13:21,400 --> 00:13:25,959
Speaker 3: I wonder what you think the skills that we should

267
00:13:26,000 --> 00:13:28,559
Speaker 3: be looking for, or that next generation should be looking for.

268
00:13:28,640 --> 00:13:33,360
Speaker 3: Because I'm with you, I feel like, yes, it's going

269
00:13:33,440 --> 00:13:35,360
Speaker 3: to happen, it's just a matter of when. But there's

270
00:13:35,360 --> 00:13:37,760
Speaker 3: still a fear factor here with a lot of people

271
00:13:38,280 --> 00:13:39,920
Speaker 3: because they don't know what they should do. The kid

272
00:13:39,920 --> 00:13:42,360
Speaker 3: who's graduating high school, what should I be studying?

273
00:13:42,440 --> 00:13:42,680
Speaker 4: Right?

274
00:13:42,760 --> 00:13:45,440
Speaker 3: Should I be taking the statistics collect What are the

275
00:13:45,520 --> 00:13:47,600
Speaker 3: skills that we think are going to be transferable for

276
00:13:47,880 --> 00:13:48,600
Speaker 3: the next generation.

277
00:13:49,000 --> 00:13:51,000
Speaker 2: Yeah, And this is a really important one. So if

278
00:13:51,040 --> 00:13:54,760
Speaker 2: you think about where is work actually going, then with AI,

279
00:13:54,840 --> 00:13:57,040
Speaker 2: like what are we going to be doing? We will

280
00:13:57,080 --> 00:14:03,280
Speaker 2: be directing these supercomputers to perform tasks, to do workflows,

281
00:14:03,320 --> 00:14:07,880
Speaker 2: and then evaluating their work. So then the competitive advantage

282
00:14:07,880 --> 00:14:12,080
Speaker 2: shifts to what questions are you asking a supercomputer? If

283
00:14:12,080 --> 00:14:13,840
Speaker 2: you can ask AI anything, what is it going to be?

284
00:14:14,360 --> 00:14:16,760
Speaker 2: And how do you know to ask that that? Are

285
00:14:16,800 --> 00:14:19,280
Speaker 2: you going to ask it a question that your competitor isn't?

286
00:14:19,960 --> 00:14:23,280
Speaker 2: How well do you communicate that question? How well can

287
00:14:23,280 --> 00:14:26,200
Speaker 2: you evaluate what AI tells you? Yes, it could sound smart,

288
00:14:26,240 --> 00:14:28,040
Speaker 2: but is it good? Does it relate to what you

289
00:14:28,080 --> 00:14:30,720
Speaker 2: actually need in your business? So the skills for that

290
00:14:30,760 --> 00:14:33,360
Speaker 2: would be judgment. So AI is going to give you

291
00:14:33,480 --> 00:14:36,880
Speaker 2: fifteen pretty great answers. What's the best answer for the

292
00:14:37,000 --> 00:14:40,880
Speaker 2: use case that you're in communication? The better you can communicate,

293
00:14:40,880 --> 00:14:44,240
Speaker 2: and not just to AI, but to people. Right, So,

294
00:14:44,280 --> 00:14:48,320
Speaker 2: can you sell the recommendation that you've now constructed with

295
00:14:48,400 --> 00:14:52,600
Speaker 2: this system to your team? Emotional intelligence? We're going to

296
00:14:52,600 --> 00:14:55,120
Speaker 2: be in this workforce where AI agents are going to

297
00:14:55,160 --> 00:14:58,680
Speaker 2: be working among us. To be selected for a team,

298
00:14:59,440 --> 00:15:01,680
Speaker 2: people have want to work with you or else, and

299
00:15:01,720 --> 00:15:03,360
Speaker 2: It's not that you're going to be competing directly with

300
00:15:03,400 --> 00:15:06,400
Speaker 2: an AI system, but the more compelling AI gets, the

301
00:15:06,440 --> 00:15:08,080
Speaker 2: easier it is to think, Okay, we can just maybe

302
00:15:08,080 --> 00:15:10,600
Speaker 2: fill some of this with a with an AI agent.

303
00:15:11,080 --> 00:15:12,880
Speaker 2: So if you're an easy person to work with, if

304
00:15:12,920 --> 00:15:16,240
Speaker 2: you're likable, if you're curious, that's going to help. I

305
00:15:16,280 --> 00:15:23,240
Speaker 2: would say creative intelligence, asking interesting questions, good ideas. Domain

306
00:15:23,280 --> 00:15:26,840
Speaker 2: expertise still matter, I think because some of the studies

307
00:15:26,880 --> 00:15:29,400
Speaker 2: show that the more you the more knowledgeable you are

308
00:15:29,400 --> 00:15:32,640
Speaker 2: about a certain domain, let's say finance, the more the

309
00:15:32,680 --> 00:15:34,840
Speaker 2: better questions you're going to ask of that AI system

310
00:15:35,120 --> 00:15:37,760
Speaker 2: related to finance, the better workflows you're going to get

311
00:15:37,760 --> 00:15:39,800
Speaker 2: it to do. You're going to be able to flag

312
00:15:39,880 --> 00:15:43,520
Speaker 2: when it does something offside or it's out of the

313
00:15:43,600 --> 00:15:45,400
Speaker 2: kind of range that you need an answer to be.

314
00:15:45,800 --> 00:15:50,520
Speaker 2: So that still matters. I would say AI literacy and

315
00:15:50,560 --> 00:15:54,240
Speaker 2: not just like working with the tool, but understanding does

316
00:15:54,280 --> 00:15:57,040
Speaker 2: the AI system you're asking questions of even have the

317
00:15:57,160 --> 00:15:58,760
Speaker 2: right data to answer it?

318
00:15:58,960 --> 00:16:00,680
Speaker 1: Because AI systems are not all equal.

319
00:16:01,640 --> 00:16:03,840
Speaker 2: And if it gives you an answer, is there possibility

320
00:16:03,840 --> 00:16:05,760
Speaker 2: that there's bias in it? Is this going to work

321
00:16:05,800 --> 00:16:08,360
Speaker 2: for everyone? So that type of literacy I think is

322
00:16:08,400 --> 00:16:14,120
Speaker 2: important and critical thinking is so important. Right, the AI

323
00:16:14,280 --> 00:16:17,120
Speaker 2: meets you where you are. It doesn't magically raise the

324
00:16:17,160 --> 00:16:19,040
Speaker 2: bar on your intelligence. So if you're going to ask

325
00:16:19,080 --> 00:16:22,120
Speaker 2: it really basic questions, you're going to get really basic answers.

326
00:16:22,480 --> 00:16:24,040
Speaker 1: So I say those are some of the skills.

327
00:16:24,080 --> 00:16:27,280
Speaker 2: So if you can think deeply, if you can communicate well,

328
00:16:27,560 --> 00:16:31,560
Speaker 2: if you can exercise strong judgment, and if you can

329
00:16:31,880 --> 00:16:36,160
Speaker 2: build stronger emotional intelligence your set and that's why it

330
00:16:36,160 --> 00:16:37,880
Speaker 2: doesn't Actually you don't have to put so much pressure

331
00:16:37,880 --> 00:16:41,080
Speaker 2: on what you're going to study in school. And someone's

332
00:16:41,120 --> 00:16:44,000
Speaker 2: going to take math, that's great. I know you're a

333
00:16:44,000 --> 00:16:46,640
Speaker 2: deep thinker. Maybe you don't do math in my company,

334
00:16:46,680 --> 00:16:49,360
Speaker 2: but I know you can think deeply. Same with philosophy

335
00:16:49,800 --> 00:16:52,240
Speaker 2: or history. You're able to connect interesting ideas.

336
00:16:52,560 --> 00:16:53,080
Speaker 4: So to not.

337
00:16:53,120 --> 00:16:57,120
Speaker 2: Stress as much about the specific program and focus more.

338
00:16:57,280 --> 00:16:59,920
Speaker 2: Am I getting these skills? And am I a self

339
00:17:00,080 --> 00:17:04,840
Speaker 2: directed learner? Because we're stepping into a world where that

340
00:17:05,480 --> 00:17:08,280
Speaker 2: career life where you work up a ladder that is

341
00:17:08,400 --> 00:17:10,920
Speaker 2: going away. We're going to continue to have to learn

342
00:17:10,960 --> 00:17:13,720
Speaker 2: new tricks over time, and people might not tell you

343
00:17:13,760 --> 00:17:15,960
Speaker 2: what you have to learn. So are you motivated to

344
00:17:16,119 --> 00:17:18,719
Speaker 2: just get in front of the technology and figure out

345
00:17:18,760 --> 00:17:21,520
Speaker 2: what's next for you? And I'd say those are probably

346
00:17:21,520 --> 00:17:23,240
Speaker 2: the master skills for this next chapter.

347
00:17:24,400 --> 00:17:26,639
Speaker 4: You said something along the lines is that everybody's going

348
00:17:26,680 --> 00:17:30,360
Speaker 4: to be forced to become an entrepreneur. We talk about that.

349
00:17:30,720 --> 00:17:33,119
Speaker 2: Yeah, yeah, So even if you don't consider yourself an

350
00:17:33,200 --> 00:17:35,440
Speaker 2: entrepreneur in the traditional sense, you're going to have to

351
00:17:35,480 --> 00:17:39,679
Speaker 2: start thinking like one. So if you think about a

352
00:17:39,720 --> 00:17:43,119
Speaker 2: company right now and they're thinking about deploying AI, AI

353
00:17:43,400 --> 00:17:46,080
Speaker 2: continues to improve over time, so what it can do today,

354
00:17:46,119 --> 00:17:47,840
Speaker 2: it's going to do something different in a year.

355
00:17:48,160 --> 00:17:50,199
Speaker 1: If I'm an organization, it's going to.

356
00:17:50,200 --> 00:17:52,840
Speaker 2: Be tricky for me to hire for that role full

357
00:17:52,920 --> 00:17:55,320
Speaker 2: time because I don't even know what it will look

358
00:17:55,400 --> 00:17:57,119
Speaker 2: like in a year. I might need somebody different, I

359
00:17:57,200 --> 00:17:59,560
Speaker 2: might need a different set of skills. So companies are

360
00:17:59,560 --> 00:18:03,160
Speaker 2: going to start to hire more and more independent workers

361
00:18:03,480 --> 00:18:07,520
Speaker 2: and contract workers, So you might end up working for

362
00:18:07,680 --> 00:18:10,919
Speaker 2: three different companies doing the set of skills that you

363
00:18:10,960 --> 00:18:14,600
Speaker 2: are best at, which means you yourself for now your

364
00:18:14,640 --> 00:18:19,600
Speaker 2: own entrepreneur little organization that offers your unique skill set

365
00:18:19,880 --> 00:18:22,439
Speaker 2: to a variety of different firms. So we all have

366
00:18:22,480 --> 00:18:25,880
Speaker 2: to start to think about that entrepreneurial mindset because we're

367
00:18:26,040 --> 00:18:28,800
Speaker 2: moving away from you working full time for one company

368
00:18:29,080 --> 00:18:32,440
Speaker 2: and you being your company and you offer your skills

369
00:18:32,640 --> 00:18:35,720
Speaker 2: to a variety of different people, projects, teams. And that's

370
00:18:35,720 --> 00:18:38,399
Speaker 2: why I always like to say to people, now, don't

371
00:18:38,400 --> 00:18:40,760
Speaker 2: think of your job in terms of or don't think

372
00:18:40,800 --> 00:18:43,119
Speaker 2: of what you offer in terms of your job title.

373
00:18:43,240 --> 00:18:46,320
Speaker 2: Think about the skills under it that make you indispensable.

374
00:18:46,760 --> 00:18:49,600
Speaker 2: And that's how you would describe yourself as that as

375
00:18:49,640 --> 00:18:52,240
Speaker 2: an organization, And that's how we have to start thinking

376
00:18:52,240 --> 00:18:53,920
Speaker 2: about it, because we're all going to be stepping out

377
00:18:53,920 --> 00:18:58,040
Speaker 2: into this entrepreneurial workforce. And obviously there's tons of implications health, insurance,

378
00:18:58,040 --> 00:19:00,960
Speaker 2: all these things that are open ended questions, but that's

379
00:19:01,000 --> 00:19:03,399
Speaker 2: the dominant fabric of the workforce going forward.

380
00:19:04,400 --> 00:19:06,600
Speaker 3: I agree with all everything you're saying. It makes me

381
00:19:06,640 --> 00:19:08,880
Speaker 3: think of the profession that I used to working, which

382
00:19:08,960 --> 00:19:12,200
Speaker 3: is education, which Bill Gates has been on the record

383
00:19:12,240 --> 00:19:13,920
Speaker 3: saying is going to be one of the most disrupted.

384
00:19:15,440 --> 00:19:18,320
Speaker 3: Having a skill set will probably be more important than

385
00:19:18,320 --> 00:19:21,280
Speaker 3: where you went to school. What do you see for

386
00:19:22,160 --> 00:19:25,679
Speaker 3: collegiate education? What will the reform look like? Right? Because

387
00:19:26,160 --> 00:19:28,679
Speaker 3: those two things won't work right, Like having the Harvard

388
00:19:28,720 --> 00:19:31,480
Speaker 3: degree won't as mean as much if I possess those

389
00:19:31,480 --> 00:19:34,360
Speaker 3: seven skills that you just said for the future, how

390
00:19:34,359 --> 00:19:35,479
Speaker 3: do you see that plan out?

391
00:19:35,680 --> 00:19:40,280
Speaker 2: Yeah, skills are going to become more important than say

392
00:19:40,320 --> 00:19:45,560
Speaker 2: experience or just an institutional symbol on your resume. If

393
00:19:45,600 --> 00:19:48,200
Speaker 2: somebody has those skills, they're pretty set. But it also

394
00:19:48,240 --> 00:19:51,879
Speaker 2: depends on what these institutions do. So if they hold

395
00:19:51,960 --> 00:19:55,240
Speaker 2: everything constant and just let the world keep evolving, then

396
00:19:55,320 --> 00:19:58,840
Speaker 2: sure they're going to start becoming less and less relevant

397
00:19:59,280 --> 00:20:01,639
Speaker 2: because most of them modern schools that we go to

398
00:20:01,720 --> 00:20:04,240
Speaker 2: and that we hear about, we're designed for an industrializing

399
00:20:04,320 --> 00:20:08,840
Speaker 2: economy that's coming gone. But if institutions like the higher

400
00:20:08,960 --> 00:20:12,560
Speaker 2: education institutions can figure out how to raise the bar

401
00:20:13,440 --> 00:20:18,080
Speaker 2: on how they cultivate knowledge, how they test for knowledge,

402
00:20:18,359 --> 00:20:21,880
Speaker 2: and meet the moment, then it's possible that they could

403
00:20:21,880 --> 00:20:23,920
Speaker 2: figure this out to the question that they need to

404
00:20:23,960 --> 00:20:26,840
Speaker 2: be asking is you know, what is the world when

405
00:20:26,880 --> 00:20:31,320
Speaker 2: students walk out and address supercomputers? What does it mean

406
00:20:31,359 --> 00:20:33,800
Speaker 2: to compete in that world? And how do we lay

407
00:20:33,800 --> 00:20:37,360
Speaker 2: that foundation here? If they can do that, then they

408
00:20:37,359 --> 00:20:39,119
Speaker 2: could be okay, because you also learn a lot of

409
00:20:39,160 --> 00:20:42,800
Speaker 2: things at school like social intelligence, teamwork, so if they

410
00:20:42,800 --> 00:20:44,639
Speaker 2: can also figure out how to cultivate that. If not,

411
00:20:45,320 --> 00:20:49,600
Speaker 2: then it's gonna Yeah, the relevance question is really going

412
00:20:49,680 --> 00:20:50,320
Speaker 2: to come into play.

413
00:20:50,440 --> 00:20:53,639
Speaker 3: Yeah, I think the relevance part when you add that

414
00:20:53,800 --> 00:20:57,280
Speaker 3: to the financial commitment, you.

415
00:20:57,240 --> 00:20:59,879
Speaker 2: Know, I feel like that trade off starts to make

416
00:21:00,119 --> 00:21:02,679
Speaker 2: and less sense. If I'm going to go into fifty

417
00:21:02,720 --> 00:21:07,879
Speaker 2: sixty thousand dollars worth of debt for a maybe on

418
00:21:08,080 --> 00:21:10,479
Speaker 2: how it's going to turn out for my career or

419
00:21:10,480 --> 00:21:14,080
Speaker 2: my income security, then that cost benefit is not going

420
00:21:14,119 --> 00:21:16,840
Speaker 2: to make as much sense. So that's another thing that

421
00:21:17,000 --> 00:21:20,480
Speaker 2: has to be figured out. The cost of post secondary

422
00:21:20,600 --> 00:21:25,520
Speaker 2: education actually doesn't make sense at all, and I think

423
00:21:25,560 --> 00:21:28,439
Speaker 2: that that's actually a knock because if you want a

424
00:21:28,560 --> 00:21:31,199
Speaker 2: thriving economy that can adapt to anything you want, if

425
00:21:31,240 --> 00:21:34,359
Speaker 2: people want to go and get it education, a country

426
00:21:34,359 --> 00:21:36,600
Speaker 2: should make that as easy as possible. If you want

427
00:21:36,640 --> 00:21:39,720
Speaker 2: to build skills in some way, go after it. We're

428
00:21:39,760 --> 00:21:42,560
Speaker 2: going to support you. We want the future Albert Einstein's

429
00:21:42,600 --> 00:21:44,400
Speaker 2: to be here, so whatever it is you need, we've

430
00:21:44,400 --> 00:21:47,520
Speaker 2: got you. Unfortunately, that's not necessarily the model or the mindset,

431
00:21:47,600 --> 00:21:50,160
Speaker 2: right now, but yeah, cost relevance, all of these things

432
00:21:50,200 --> 00:21:52,680
Speaker 2: are starting to make school seem not school, but post

433
00:21:52,680 --> 00:21:56,000
Speaker 2: secondary education feel much more questionable.

434
00:21:56,480 --> 00:22:00,639
Speaker 4: What's the point of education if you don't need to

435
00:22:00,720 --> 00:22:05,320
Speaker 4: be educated? Like meaning, what is the point of getting

436
00:22:05,359 --> 00:22:11,040
Speaker 4: a law degree and really dedicated in almost six years

437
00:22:11,040 --> 00:22:14,479
Speaker 4: of your life to that when you can just go

438
00:22:14,520 --> 00:22:16,920
Speaker 4: and chat GBT and draft a legal agreement?

439
00:22:17,119 --> 00:22:17,919
Speaker 1: Yeah?

440
00:22:18,280 --> 00:22:23,240
Speaker 2: So yeah, So if the law degree was to understand

441
00:22:23,240 --> 00:22:27,040
Speaker 2: how to draft these agreements, that's done and no one's

442
00:22:27,040 --> 00:22:29,040
Speaker 2: going to be doing that. That's going to Claude Gemini

443
00:22:29,119 --> 00:22:34,439
Speaker 2: chatchapt but law is likely going to get harder, so

444
00:22:34,520 --> 00:22:36,439
Speaker 2: all of the things that we need basic lawyers for

445
00:22:36,720 --> 00:22:39,280
Speaker 2: that goes away. So think about then, if I can

446
00:22:39,280 --> 00:22:41,280
Speaker 2: go to chetchipt for all of my basic things I

447
00:22:41,280 --> 00:22:43,320
Speaker 2: need to season desist someone's created a chatbout of meat,

448
00:22:43,320 --> 00:22:44,359
Speaker 2: which literally happened.

449
00:22:44,440 --> 00:22:45,760
Speaker 1: I'm going to a chauchip for that.

450
00:22:46,320 --> 00:22:47,840
Speaker 2: But if there's a moment where I actually need to

451
00:22:47,840 --> 00:22:51,520
Speaker 2: then hire a lawyer in the future, how complex might

452
00:22:51,560 --> 00:22:55,000
Speaker 2: that scenario be? So lawyers of the future, if we

453
00:22:55,080 --> 00:22:57,440
Speaker 2: call them that, they're going to be working in some

454
00:22:57,600 --> 00:23:03,439
Speaker 2: really strange comp flex scenarios that have multiple different implications

455
00:23:03,680 --> 00:23:05,680
Speaker 2: to the level of expertise lawyers to bring to the

456
00:23:05,720 --> 00:23:08,840
Speaker 2: table will be will look nothing like what present day

457
00:23:08,880 --> 00:23:12,920
Speaker 2: lawyers have, so the industry itself gets harder.

458
00:23:13,160 --> 00:23:17,800
Speaker 4: Or what if you hire a virtual lawyer, because that's

459
00:23:17,800 --> 00:23:20,120
Speaker 4: what Robert Smith said to us when we met him, well,

460
00:23:20,160 --> 00:23:21,919
Speaker 4: when we saw him at that event, it was like,

461
00:23:21,960 --> 00:23:23,359
Speaker 4: do you have a personal Do you have an AI

462
00:23:23,400 --> 00:23:25,800
Speaker 4: personal assistant yet? And he was like, if you don't,

463
00:23:25,840 --> 00:23:28,600
Speaker 4: then I'll sell you one. And he was explaining to

464
00:23:28,680 --> 00:23:30,280
Speaker 4: us that at some point in time, everybody's gonna have

465
00:23:30,280 --> 00:23:34,919
Speaker 4: an AI personal assistant. But that's for personal assistant yours

466
00:23:34,960 --> 00:23:38,359
Speaker 4: and crafts, But the same thought process could be for

467
00:23:38,680 --> 00:23:42,360
Speaker 4: professional careers where everybody's gonna have an AI doctor. Right,

468
00:23:42,440 --> 00:23:44,520
Speaker 4: so instead of actually going to the doctor all the time,

469
00:23:44,560 --> 00:23:47,000
Speaker 4: you have somebody with you every single day that you

470
00:23:47,040 --> 00:23:49,520
Speaker 4: can ask questions, they monitor your heart raised stuff like that,

471
00:23:49,880 --> 00:23:51,359
Speaker 4: saying it could be a lawyer like you, because you

472
00:23:51,480 --> 00:23:53,719
Speaker 4: need counsel all the time and I don't necessarily want

473
00:23:53,720 --> 00:23:55,639
Speaker 4: to pay five hundred dollars an hour to talk to

474
00:23:55,680 --> 00:23:58,560
Speaker 4: a lawyer, So I could just have a dedicated or

475
00:23:58,640 --> 00:24:01,760
Speaker 4: real in house literally literally an in house lawyer in

476
00:24:01,800 --> 00:24:06,400
Speaker 4: my house. That is literally just answering questions and this

477
00:24:06,480 --> 00:24:07,440
Speaker 4: is my counsel.

478
00:24:07,640 --> 00:24:09,000
Speaker 1: Sure, and that's going to happen.

479
00:24:09,080 --> 00:24:11,399
Speaker 2: So anything that can be streamed, So anything that's like

480
00:24:11,760 --> 00:24:15,479
Speaker 2: pervasive in the background assistance legal counsel, we will stream that.

481
00:24:15,520 --> 00:24:18,800
Speaker 2: So you'll stream Legal Counsel twenty four to seven and

482
00:24:18,840 --> 00:24:21,760
Speaker 2: they're constantly flagging the environment. Oh you should blur this

483
00:24:21,840 --> 00:24:23,440
Speaker 2: out on a podcast that you miss six weeks ago,

484
00:24:23,480 --> 00:24:24,720
Speaker 2: Like that's all happy in the background.

485
00:24:24,720 --> 00:24:26,240
Speaker 1: We won't even know about it. The AI agents are

486
00:24:26,240 --> 00:24:26,600
Speaker 1: doing that.

487
00:24:27,400 --> 00:24:31,560
Speaker 2: So when a human expertise would come into play is

488
00:24:31,560 --> 00:24:35,399
Speaker 2: when the scenario is so complex and an AI probably

489
00:24:35,400 --> 00:24:39,160
Speaker 2: can outthink a human lawyer in some capacity, but it's

490
00:24:39,200 --> 00:24:41,399
Speaker 2: going to be something that is on the intersection of

491
00:24:41,440 --> 00:24:43,400
Speaker 2: all of these different fields that make it much more

492
00:24:43,440 --> 00:24:46,520
Speaker 2: challenging and even for medicine, right, So yeah, all of

493
00:24:46,560 --> 00:24:48,199
Speaker 2: the basic things, like the hospital of the future is

494
00:24:48,200 --> 00:24:50,400
Speaker 2: coming to you. You're going to have wearables and things

495
00:24:50,440 --> 00:24:52,320
Speaker 2: around you that are monitoring your blood work and all

496
00:24:52,400 --> 00:24:54,960
Speaker 2: this stuff to prevent you from needing a doctor in

497
00:24:54,960 --> 00:24:58,119
Speaker 2: the first place. But what we will call a doctor

498
00:24:58,800 --> 00:25:01,639
Speaker 2: will be unrecognizable and the things they will do, the

499
00:25:02,160 --> 00:25:05,520
Speaker 2: designing little nanobots that can go in and then perform

500
00:25:05,600 --> 00:25:08,480
Speaker 2: the surgery on a cell while you're at work. I

501
00:25:08,520 --> 00:25:11,719
Speaker 2: don't what is that job called that? So it gets harder.

502
00:25:12,160 --> 00:25:15,439
Speaker 2: So what AI does is it raises the floor. So

503
00:25:15,960 --> 00:25:17,840
Speaker 2: anything that can be now done in that new floor

504
00:25:17,840 --> 00:25:20,000
Speaker 2: by AI will be and the things.

505
00:25:19,720 --> 00:25:21,000
Speaker 1: That are on top of it.

506
00:25:21,000 --> 00:25:23,960
Speaker 2: It's kind of like accountants of the past used to

507
00:25:24,000 --> 00:25:25,480
Speaker 2: just kind of tally up numbers.

508
00:25:25,880 --> 00:25:27,160
Speaker 1: Now what are accountants doing?

509
00:25:27,880 --> 00:25:31,240
Speaker 2: Like the top ones are helping people figure out loopholes.

510
00:25:31,240 --> 00:25:33,040
Speaker 2: I'm going to put your thing in the Cayman Islands,

511
00:25:33,160 --> 00:25:34,800
Speaker 2: all of this stuff. So what they do is now

512
00:25:34,880 --> 00:25:38,440
Speaker 2: exercise judgment. They don't have of them, probably can't count.

513
00:25:38,560 --> 00:25:41,360
Speaker 2: But that was actually the skill of accounting sixty years ago.

514
00:25:41,440 --> 00:25:42,720
Speaker 2: So yes, a lot of things are going to go

515
00:25:42,760 --> 00:25:47,399
Speaker 2: away and the professions are going to evolve into something new,

516
00:25:47,560 --> 00:25:48,920
Speaker 2: and that means it's going to be a new set

517
00:25:48,960 --> 00:25:50,920
Speaker 2: of skills. So what makes you a sharp lawyer today

518
00:25:51,000 --> 00:25:53,600
Speaker 2: may be irrelevant in ten years and we still might

519
00:25:53,600 --> 00:25:56,439
Speaker 2: not even call it that. But whoever, when you actually

520
00:25:56,520 --> 00:25:59,119
Speaker 2: need to go to court for something, you'll probably have

521
00:25:59,200 --> 00:26:02,879
Speaker 2: sure a team of AI agents and somebody that can

522
00:26:03,200 --> 00:26:06,600
Speaker 2: orchestrate these agents and maybe it's not a lawyer, maybe

523
00:26:06,640 --> 00:26:10,000
Speaker 2: it's something else that can exercise the agent in a

524
00:26:10,040 --> 00:26:12,680
Speaker 2: way that they simulate your court case and figure out

525
00:26:12,680 --> 00:26:14,679
Speaker 2: your plan of action. I don't know what we'll call that.

526
00:26:15,160 --> 00:26:18,439
Speaker 3: I'm excited for it. You said build upon. So it

527
00:26:18,440 --> 00:26:21,000
Speaker 3: made me think about the people who have the premise

528
00:26:21,080 --> 00:26:24,959
Speaker 3: of you know, where are we at an AI? And

529
00:26:25,000 --> 00:26:27,679
Speaker 3: we kind of had this conversation like prior to starting

530
00:26:27,760 --> 00:26:32,360
Speaker 3: about AI, is the infrastructure right like it is the

531
00:26:32,400 --> 00:26:34,960
Speaker 3: thing that will be built on. Can you break down

532
00:26:35,040 --> 00:26:38,199
Speaker 3: to the audience exactly how they should be looking at

533
00:26:38,240 --> 00:26:40,000
Speaker 3: it a little bit differently if they're still unclear?

534
00:26:40,560 --> 00:26:43,320
Speaker 2: Yeah, So what probably most people are thinking about AI

535
00:26:43,400 --> 00:26:45,680
Speaker 2: right now as like a tool or a chatbot.

536
00:26:46,119 --> 00:26:48,600
Speaker 1: So you ask it a question and.

537
00:26:48,520 --> 00:26:51,120
Speaker 2: Then it gives you the answer to that, and that's

538
00:26:51,200 --> 00:26:53,359
Speaker 2: kind of like we're in the alpha beta stage of AI.

539
00:26:53,680 --> 00:26:56,320
Speaker 2: What is going to happen is AI is just going

540
00:26:56,359 --> 00:27:00,600
Speaker 2: to be this always on layer and so you assume

541
00:27:00,720 --> 00:27:03,760
Speaker 2: that it's happening in the background, so you don't have

542
00:27:03,800 --> 00:27:05,720
Speaker 2: to ask AI a specific question.

543
00:27:06,200 --> 00:27:08,280
Speaker 1: You give it a goal, and it's.

544
00:27:08,160 --> 00:27:10,240
Speaker 2: Just working twenty four to seven in the background to

545
00:27:10,240 --> 00:27:13,879
Speaker 2: get that done, and it notifies you when it needs

546
00:27:13,880 --> 00:27:16,840
Speaker 2: your assistance on something. So you can kind of think

547
00:27:16,880 --> 00:27:21,680
Speaker 2: about it like the Internet is always on and you've

548
00:27:21,680 --> 00:27:25,480
Speaker 2: built your business on top of it. You've built your website,

549
00:27:25,960 --> 00:27:30,040
Speaker 2: you've built your marketing channels, you've built a social media profile,

550
00:27:30,240 --> 00:27:32,160
Speaker 2: and you don't even think about the fact that it's

551
00:27:32,200 --> 00:27:35,879
Speaker 2: the Internet that underpins all of that. Everything is going

552
00:27:35,960 --> 00:27:39,240
Speaker 2: to be rebuilt on top of this technology and it's

553
00:27:39,280 --> 00:27:41,520
Speaker 2: just going to be the assumed layer in the background.

554
00:27:42,080 --> 00:27:45,679
Speaker 2: That also means, I mean social media made sense on

555
00:27:45,720 --> 00:27:48,400
Speaker 2: top of the Internet. Does it make sense on top

556
00:27:48,400 --> 00:27:51,200
Speaker 2: of AI? I don't know, you know what.

557
00:27:51,840 --> 00:27:53,840
Speaker 3: I feel like you're just doing an ALI because you

558
00:27:53,920 --> 00:27:57,159
Speaker 3: have the posts about social media and where it's headed

559
00:27:57,560 --> 00:28:01,679
Speaker 3: and how the life cycles of social media, so the

560
00:28:01,720 --> 00:28:05,639
Speaker 3: futurism on that. Because I'm with you, I want to

561
00:28:05,640 --> 00:28:08,400
Speaker 3: hear your deep thought on that. Where are we at

562
00:28:08,440 --> 00:28:10,240
Speaker 3: and how does this look in the world of AI.

563
00:28:11,000 --> 00:28:12,960
Speaker 2: Yeah, so I think, I mean, I don't know what

564
00:28:13,040 --> 00:28:15,600
Speaker 2: it's going to be. But if we think about the

565
00:28:15,640 --> 00:28:19,000
Speaker 2: history of entertainment technologies and how we've communicated, right, so

566
00:28:19,040 --> 00:28:21,040
Speaker 2: we have printing press and we get newspaper, we get

567
00:28:21,040 --> 00:28:24,199
Speaker 2: all this stuff. We get television on top of our

568
00:28:24,200 --> 00:28:26,600
Speaker 2: electricity streams, the Internet comes.

569
00:28:26,359 --> 00:28:27,800
Speaker 1: Out, we get social media.

570
00:28:28,520 --> 00:28:32,040
Speaker 2: Every technology has been then followed by a question mark

571
00:28:32,119 --> 00:28:35,520
Speaker 2: and room for invention when it comes to entertainment, communication,

572
00:28:35,560 --> 00:28:37,760
Speaker 2: whatever you want to call that industry. So now we're

573
00:28:37,800 --> 00:28:41,800
Speaker 2: at the next general purpose technology AI. Does it make

574
00:28:41,840 --> 00:28:44,280
Speaker 2: sense that we're going to do the same things with AI?

575
00:28:44,960 --> 00:28:49,240
Speaker 2: Are we watching sixty second TikTok clips on CNN? No,

576
00:28:49,440 --> 00:28:53,320
Speaker 2: it's a new thing, right. We built a different infrastructure,

577
00:28:53,320 --> 00:28:56,400
Speaker 2: a different platform for the technology of the moment, So

578
00:28:56,480 --> 00:29:02,720
Speaker 2: would we be scrolling reels on AI that? It probably

579
00:29:02,720 --> 00:29:05,400
Speaker 2: won't make sense, and it's not going to be overnight.

580
00:29:05,440 --> 00:29:08,080
Speaker 2: It's not like in sixty days it's done and we're

581
00:29:08,120 --> 00:29:11,160
Speaker 2: all just doing something different. But we will probably look

582
00:29:11,240 --> 00:29:14,760
Speaker 2: back on social media the way we look back on

583
00:29:15,440 --> 00:29:20,320
Speaker 2: different other communication mediums and be like, wow, remember those days.

584
00:29:20,680 --> 00:29:23,800
Speaker 2: That was the social media era. And it's going to

585
00:29:23,800 --> 00:29:27,080
Speaker 2: be a slower transition, but it will probably happen. I mean,

586
00:29:27,120 --> 00:29:29,120
Speaker 2: when was the last time you read the newspaper? Do

587
00:29:29,200 --> 00:29:32,440
Speaker 2: people even know how to read it and to follow

588
00:29:32,480 --> 00:29:33,200
Speaker 2: it anymore?

589
00:29:33,800 --> 00:29:34,320
Speaker 1: I don't know.

590
00:29:34,840 --> 00:29:37,320
Speaker 2: And so that's just what history shows happens to these

591
00:29:37,360 --> 00:29:40,400
Speaker 2: technologies over time, and it also means that it's open

592
00:29:40,440 --> 00:29:42,840
Speaker 2: season and open space to build and to innovate the

593
00:29:42,880 --> 00:29:44,440
Speaker 2: new creations and to create.

594
00:29:44,920 --> 00:29:47,360
Speaker 4: So talk about the dark side of AR because everybody

595
00:29:47,400 --> 00:29:50,400
Speaker 4: that has been a leader in AI has expressed extreme concern,

596
00:29:50,960 --> 00:29:55,360
Speaker 4: from Sam Altman to Elon Musk, they've all kind of

597
00:29:55,400 --> 00:29:58,920
Speaker 4: had the same sentiment like we've essentially created something for

598
00:29:58,960 --> 00:30:00,960
Speaker 4: the first time in human history, read us superior to

599
00:30:01,080 --> 00:30:05,479
Speaker 4: us And no matter how you know, optimistic we are

600
00:30:05,480 --> 00:30:10,440
Speaker 4: on it, there's still that's pretty alarming because it can

601
00:30:10,560 --> 00:30:12,719
Speaker 4: go in different directions, and it has going different directions.

602
00:30:12,760 --> 00:30:15,760
Speaker 4: It's blackmailed people, has created its own you know, language

603
00:30:15,760 --> 00:30:17,880
Speaker 4: that we can't decipher. It is kind of going rogue

604
00:30:18,000 --> 00:30:22,920
Speaker 4: through different times. It's a real thing. Talk about your

605
00:30:22,960 --> 00:30:25,920
Speaker 4: thoughts on what's the worst case scenario, what's the dark

606
00:30:26,000 --> 00:30:26,880
Speaker 4: side of AI?

607
00:30:27,360 --> 00:30:29,480
Speaker 2: Isn't it so ironic that the people building it or

608
00:30:29,560 --> 00:30:33,280
Speaker 2: like this thing is so it could be so terrible

609
00:30:33,800 --> 00:30:36,360
Speaker 2: as if AI is an alien landing on Earth that

610
00:30:36,400 --> 00:30:38,920
Speaker 2: we can do nothing about. It's like Homeboy, but you're

611
00:30:38,920 --> 00:30:40,400
Speaker 2: the one in the lat like we don't even have.

612
00:30:40,320 --> 00:30:41,000
Speaker 1: The keys to the lab.

613
00:30:42,640 --> 00:30:47,880
Speaker 2: The worst case scenario with AI. I mean, I don't

614
00:30:47,880 --> 00:30:51,560
Speaker 2: think that there's just one. This could go bad in

615
00:30:51,560 --> 00:30:57,360
Speaker 2: many ways you could imagine, and this is probably already happening.

616
00:30:58,360 --> 00:31:03,440
Speaker 2: People adversary is using AI to hack into critical infrastructure

617
00:31:04,040 --> 00:31:11,200
Speaker 2: like water filters, subway and train routes, all sorts of

618
00:31:11,280 --> 00:31:21,040
Speaker 2: things to create physical kinetic chaos. You can imagine engineering

619
00:31:21,200 --> 00:31:27,960
Speaker 2: chatbots to trick people in certain ways and really divide

620
00:31:28,000 --> 00:31:32,040
Speaker 2: people not just from their voting bases, but from their families.

621
00:31:33,040 --> 00:31:37,840
Speaker 2: You could We're gonna see hacking of all sorts of things, hospitals,

622
00:31:39,080 --> 00:31:42,760
Speaker 2: people's personal things, banks, like, all of that is also coming.

623
00:31:43,960 --> 00:31:49,080
Speaker 2: And then of course there's the scenario of AI not

624
00:31:49,360 --> 00:31:53,760
Speaker 2: understanding what it's doing anymore. I mean, we don't fully

625
00:31:53,800 --> 00:31:56,200
Speaker 2: understand how AI works today and it's only going to

626
00:31:56,640 --> 00:31:59,960
Speaker 2: improve over time. So they're sure is the non zero

627
00:32:00,120 --> 00:32:04,280
Speaker 2: possibility that it's connected to things and making decisions in ways.

628
00:32:04,000 --> 00:32:07,880
Speaker 1: That we can't follow and the robotics.

629
00:32:08,200 --> 00:32:11,600
Speaker 2: Yeah, so AI right now it is largely in your

630
00:32:11,640 --> 00:32:14,360
Speaker 2: computer and in the cloud, but soon it's going to

631
00:32:14,400 --> 00:32:18,080
Speaker 2: be embodied in physical systems, and you could think about, sure,

632
00:32:18,120 --> 00:32:20,360
Speaker 2: a robot that looks like a humanoid.

633
00:32:20,600 --> 00:32:21,959
Speaker 1: We're going to have all sorts of robots.

634
00:32:22,240 --> 00:32:24,640
Speaker 2: We're going to have the way you have all of

635
00:32:24,680 --> 00:32:27,160
Speaker 2: these appliances in your house, you have a dishwasher, and

636
00:32:27,160 --> 00:32:29,520
Speaker 2: then a fridge and then maybe a roomba. You're going

637
00:32:29,600 --> 00:32:32,200
Speaker 2: to have all sorts of robotic appliances that are powered

638
00:32:32,200 --> 00:32:35,920
Speaker 2: by this technology. So that's going to be awesome, but

639
00:32:35,960 --> 00:32:37,440
Speaker 2: it also means if that thing is hacked, you have

640
00:32:37,480 --> 00:32:38,920
Speaker 2: a physical spy in your house.

641
00:32:39,120 --> 00:32:39,760
Speaker 3: Did you just see that?

642
00:32:39,920 --> 00:32:42,680
Speaker 4: Or a physical enemy that could attack you.

643
00:32:42,800 --> 00:32:45,400
Speaker 3: That just happened. Did you see that somebody's room?

644
00:32:45,440 --> 00:32:47,680
Speaker 1: But yeah, yeah, yeah, and he that.

645
00:32:47,720 --> 00:32:49,880
Speaker 3: Control of the seven hundred thousand vacuums.

646
00:32:50,040 --> 00:32:52,160
Speaker 2: Yeah, Because right now we're going to come up against

647
00:32:52,440 --> 00:32:56,000
Speaker 2: this transition time when most of our digital infrastructure has

648
00:32:56,040 --> 00:32:58,480
Speaker 2: been designed for an era where it isn't a bunch

649
00:32:58,520 --> 00:33:00,920
Speaker 2: of AI agents that can automate and things kind of

650
00:33:00,960 --> 00:33:02,800
Speaker 2: like how encryption was designed in such a way that

651
00:33:02,840 --> 00:33:04,160
Speaker 2: most people aren't going to sit there and solve the

652
00:33:04,160 --> 00:33:06,760
Speaker 2: math problem because it takes ten thousand years. And now

653
00:33:06,840 --> 00:33:09,720
Speaker 2: quantum change is that most of our infrastructure wasn't designed

654
00:33:09,720 --> 00:33:12,000
Speaker 2: for the idea that one person could have a team

655
00:33:12,040 --> 00:33:16,240
Speaker 2: of ten thousand agents that could hack and get into

656
00:33:16,280 --> 00:33:17,920
Speaker 2: a system rmot least. So I think, yeah, those are

657
00:33:17,960 --> 00:33:19,480
Speaker 2: just and those are just some of the dark sides

658
00:33:19,480 --> 00:33:22,760
Speaker 2: of AI, And yeah, I think it's important to talk

659
00:33:22,760 --> 00:33:25,440
Speaker 2: about because there's no point pretending they don't exist, because

660
00:33:25,440 --> 00:33:27,600
Speaker 2: then we just don't address it. And it doesn't mean

661
00:33:27,600 --> 00:33:30,320
Speaker 2: you're a pessimist or rooting for the worst case scenario.

662
00:33:30,400 --> 00:33:32,960
Speaker 2: But this stuff is real and it's happening.

663
00:33:33,480 --> 00:33:36,440
Speaker 3: Yeah, I'm just thinking about that, right, Like, this is

664
00:33:36,480 --> 00:33:40,040
Speaker 3: one guy and granted he says that he reversed the code,

665
00:33:40,440 --> 00:33:43,840
Speaker 3: but he could he could have moved seventy thousand vacuums

666
00:33:43,840 --> 00:33:46,120
Speaker 3: in the homes of people throughout the world and everybody

667
00:33:46,160 --> 00:33:49,440
Speaker 3: would have had no idea what was happening. And that's

668
00:33:49,560 --> 00:33:53,880
Speaker 3: just like interesting. So having a conversation is important. I'm

669
00:33:53,920 --> 00:33:56,240
Speaker 3: glad Shotty brought up the point because it makes me

670
00:33:56,280 --> 00:34:02,440
Speaker 3: think about innovation, right, can happen, but regulation also needs

671
00:34:02,480 --> 00:34:07,280
Speaker 3: to happen. So does regulations slow down innovation or is

672
00:34:07,320 --> 00:34:11,520
Speaker 3: it a needed thing to protect us from ourselves?

673
00:34:11,520 --> 00:34:14,040
Speaker 1: I guess yeah, I think no.

674
00:34:14,200 --> 00:34:17,640
Speaker 2: I don't think regulation shouldn't slow down innovation.

675
00:34:18,040 --> 00:34:19,399
Speaker 1: If you are regulating in.

676
00:34:19,360 --> 00:34:23,200
Speaker 2: A way that is appropriate for the technology at hand.

677
00:34:23,719 --> 00:34:26,440
Speaker 2: What tends to happen is you and some regulation from

678
00:34:26,520 --> 00:34:31,160
Speaker 2: prior industrial eras or prior chapters can be applied. But

679
00:34:31,440 --> 00:34:35,080
Speaker 2: sometimes regulation has to be as innovative as the technology itself,

680
00:34:35,520 --> 00:34:37,719
Speaker 2: and that's where things tend to go wrong. Right.

681
00:34:37,760 --> 00:34:39,760
Speaker 1: You tend to try to grab, let's say.

682
00:34:39,600 --> 00:34:43,160
Speaker 2: Twentieth century regulation frameworks and apply them to twenty first

683
00:34:43,200 --> 00:34:46,920
Speaker 2: century infrastructure, and it just doesn't make sense. So regulation

684
00:34:47,040 --> 00:34:49,480
Speaker 2: also has to be seen as innovative, and sometimes we

685
00:34:49,520 --> 00:34:50,799
Speaker 2: don't necessarily see that.

686
00:34:51,360 --> 00:34:53,200
Speaker 1: But it shouldn't slow stuff down.

687
00:34:53,239 --> 00:34:55,480
Speaker 2: If you regulate in a way that companies are like, okay,

688
00:34:55,520 --> 00:34:57,920
Speaker 2: bet I get what the lane is that I can

689
00:34:57,960 --> 00:35:01,360
Speaker 2: play within. Now I can go fullar steam ahead.

690
00:35:01,640 --> 00:35:02,960
Speaker 1: But when you have companies that.

691
00:35:02,920 --> 00:35:04,719
Speaker 2: Are like, I want to do this, but I don't

692
00:35:04,719 --> 00:35:06,520
Speaker 2: know if in two years I'm going to be sued

693
00:35:06,600 --> 00:35:08,440
Speaker 2: because this does seem kind of great, but no one

694
00:35:08,480 --> 00:35:11,759
Speaker 2: said anything about it in Washington, then you maybe just

695
00:35:11,840 --> 00:35:14,160
Speaker 2: hold back. So I think if you give people clear

696
00:35:14,200 --> 00:35:17,080
Speaker 2: lines to play within, and you do that properly in

697
00:35:17,120 --> 00:35:19,840
Speaker 2: a way that's appropriate and not just like a blanket

698
00:35:19,880 --> 00:35:23,080
Speaker 2: fear regulation which doesn't help anyone, then it shouldn't flow

699
00:35:23,160 --> 00:35:28,000
Speaker 2: stuff down. But yeah, regulation does have to keep up

700
00:35:28,040 --> 00:35:30,600
Speaker 2: in a way, and that also requires innovation. I think

701
00:35:31,040 --> 00:35:32,360
Speaker 2: sometimes that part is missed.

702
00:35:32,640 --> 00:35:34,680
Speaker 3: Countries will follow up really quick because it makes me

703
00:35:34,719 --> 00:35:37,560
Speaker 3: think of where we're currently at when you think about

704
00:35:37,760 --> 00:35:40,799
Speaker 3: Anthropic and the government and now open AI and the

705
00:35:40,800 --> 00:35:46,480
Speaker 3: government and them saying, hey, we need to use your tools,

706
00:35:46,840 --> 00:35:48,680
Speaker 3: but like you said, we're in the beta form, so

707
00:35:48,719 --> 00:35:51,960
Speaker 3: like they're saying, hey, we're not fully ready to roll

708
00:35:51,960 --> 00:35:55,560
Speaker 3: this thing out, so yeah, you can use it, but

709
00:35:55,640 --> 00:35:58,880
Speaker 3: don't use this part of it. Like how do we

710
00:35:58,960 --> 00:35:59,480
Speaker 3: balance this?

711
00:36:01,080 --> 00:36:06,280
Speaker 2: And this is really complicated because what has essentially happened

712
00:36:06,360 --> 00:36:12,040
Speaker 2: is the most important technology for nations, for a nation state,

713
00:36:12,320 --> 00:36:16,520
Speaker 2: let's say in the West, is no longer in the

714
00:36:16,640 --> 00:36:19,879
Speaker 2: hands of the public sector. It is all run by

715
00:36:19,960 --> 00:36:23,600
Speaker 2: private companies, which doesn't inherently have to be a bad thing.

716
00:36:24,160 --> 00:36:27,080
Speaker 2: The private sector moves faster, get better access to capital,

717
00:36:27,160 --> 00:36:30,040
Speaker 2: better rates, they can do all that. But when the

718
00:36:30,080 --> 00:36:33,320
Speaker 2: technology is now seen as like core to national security

719
00:36:33,880 --> 00:36:37,680
Speaker 2: or core to the country's existence, and you don't have

720
00:36:37,719 --> 00:36:41,919
Speaker 2: control over it, then you run up against these scenarios

721
00:36:41,960 --> 00:36:45,719
Speaker 2: like we're seeing right now where the Pentagon is threatening

722
00:36:45,760 --> 00:36:49,439
Speaker 2: a USAI company to use their technology in a way

723
00:36:49,440 --> 00:36:51,520
Speaker 2: that it's not even just that it's a red line

724
00:36:51,560 --> 00:36:54,360
Speaker 2: for Anthropic, like we just don't feel comfortable with this morally.

725
00:36:54,560 --> 00:36:58,000
Speaker 2: The technology is also not ready to do that. Like

726
00:36:58,080 --> 00:37:01,000
Speaker 2: AI generative AI should not be in a kill chain.

727
00:37:01,440 --> 00:37:05,120
Speaker 2: It is a predictive statistical machine. It doesn't actually have

728
00:37:05,200 --> 00:37:08,319
Speaker 2: situational awareness or judgment or context. And that's basically where

729
00:37:08,360 --> 00:37:13,360
Speaker 2: they're drying a ligne. And that's actually really scary. What's

730
00:37:13,360 --> 00:37:15,960
Speaker 2: happening with Anthropic and the Pentagon like that can't really

731
00:37:15,960 --> 00:37:19,600
Speaker 2: be overlooked. This is pretty significant moment and a pretty

732
00:37:19,600 --> 00:37:22,800
Speaker 2: big inflection point in national security in what other countries

733
00:37:22,800 --> 00:37:26,000
Speaker 2: are probably watching this and thinking, hmm, if the US

734
00:37:26,080 --> 00:37:29,640
Speaker 2: is going to pressure their own AI companies and take

735
00:37:29,680 --> 00:37:32,439
Speaker 2: this unilateral, it's us or get out of the way,

736
00:37:32,480 --> 00:37:35,440
Speaker 2: and we might shut your company down. Imagine if we

737
00:37:35,600 --> 00:37:38,080
Speaker 2: became a competitor to them, or they weren't on our

738
00:37:38,120 --> 00:37:40,400
Speaker 2: side anymore, what they all do or think. So this

739
00:37:40,440 --> 00:37:45,200
Speaker 2: could spiral more countries to throw ethics away or to try.

740
00:37:45,000 --> 00:37:45,919
Speaker 1: To do their own thing.

741
00:37:47,280 --> 00:37:51,160
Speaker 2: But yeah, I think if a company's ethics aren't respected

742
00:37:51,880 --> 00:37:55,240
Speaker 2: by the country, I mean, what does that say about

743
00:37:55,280 --> 00:37:58,440
Speaker 2: how the country's thinking about regulation and ethics not great.

744
00:37:58,680 --> 00:38:01,440
Speaker 4: Well explaining for people that don't know, explain exactly what

745
00:38:01,480 --> 00:38:04,600
Speaker 4: the government and Anthropic and then open AI came explain exactly.

746
00:38:05,640 --> 00:38:11,240
Speaker 2: Yeah, ask was So the backdrop is the most governments

747
00:38:11,239 --> 00:38:13,600
Speaker 2: are going to be using AI. That's the new infrastructure

748
00:38:13,600 --> 00:38:16,360
Speaker 2: of the future. And it turns out the US government

749
00:38:16,400 --> 00:38:18,960
Speaker 2: really likes Anthropics AI model Claude.

750
00:38:19,360 --> 00:38:19,680
Speaker 1: They did.

751
00:38:19,760 --> 00:38:21,439
Speaker 2: I think it was a two hundred million dollar deal

752
00:38:21,520 --> 00:38:24,719
Speaker 2: the Department of Defense with Anthropic last summer. Let's put

753
00:38:24,719 --> 00:38:27,000
Speaker 2: Claude in all our workflows and it's actually the only

754
00:38:27,040 --> 00:38:29,799
Speaker 2: aisystem that can be in all of the classified workflows

755
00:38:30,320 --> 00:38:33,160
Speaker 2: of the military. But Anthropic was like, Okay, no problem, great,

756
00:38:33,239 --> 00:38:36,000
Speaker 2: use our technology. Here's the deal. But you can't use

757
00:38:36,040 --> 00:38:38,879
Speaker 2: it to surveil American people. And you also can't use

758
00:38:38,920 --> 00:38:43,080
Speaker 2: it as an autonomous weapon. And that was seen as okay,

759
00:38:43,239 --> 00:38:45,320
Speaker 2: and it seemed like the Pentagon had agreed.

760
00:38:45,440 --> 00:38:46,880
Speaker 1: They I don't think we're super happy with it, but

761
00:38:46,920 --> 00:38:47,440
Speaker 1: they had agreed.

762
00:38:48,360 --> 00:38:51,160
Speaker 2: And that's all changed as the US has started to

763
00:38:51,280 --> 00:38:56,480
Speaker 2: insert itself in more kinetic landscapes like Venezuela and now Iran,

764
00:38:57,360 --> 00:39:01,239
Speaker 2: and it came out that Anthropy was or claud was

765
00:39:01,320 --> 00:39:03,479
Speaker 2: used in some of the Venezuelan operations and they weren't

766
00:39:03,520 --> 00:39:05,719
Speaker 2: cool with that, so they flagged it. And so now

767
00:39:05,800 --> 00:39:08,080
Speaker 2: essentially what it's happening is the Pentagon is saying, you

768
00:39:08,120 --> 00:39:11,680
Speaker 2: have to drop your AI restrictions about the surveillance and

769
00:39:11,760 --> 00:39:13,960
Speaker 2: the autonomous weapons. We need to be able to do

770
00:39:14,080 --> 00:39:17,080
Speaker 2: what we need to do with your AI, and Anthropicist said, no,

771
00:39:18,320 --> 00:39:21,400
Speaker 2: these red lines are important to us, and the Pentagon

772
00:39:21,440 --> 00:39:24,319
Speaker 2: has said, Okay, if that's what you're going to say,

773
00:39:24,360 --> 00:39:27,640
Speaker 2: we're either going to enact the Defense Production Act, which

774
00:39:27,680 --> 00:39:31,920
Speaker 2: is a Cold War era industrial policy that gives the

775
00:39:31,960 --> 00:39:35,640
Speaker 2: president control over domestic industries and says company X, you

776
00:39:35,719 --> 00:39:37,960
Speaker 2: have to make Y. So that would mean Anthropic, you

777
00:39:38,000 --> 00:39:39,520
Speaker 2: have to give us your model and let us do

778
00:39:39,560 --> 00:39:42,839
Speaker 2: what we want with it, or you do face jail.

779
00:39:42,960 --> 00:39:46,200
Speaker 2: And they're also saying if the second option that will

780
00:39:46,200 --> 00:39:49,000
Speaker 2: throw at you is we will blacklist you from the

781
00:39:49,040 --> 00:39:51,919
Speaker 2: US government and they label you as a high risk

782
00:39:52,120 --> 00:39:55,319
Speaker 2: supplier and that means no company that works with the

783
00:39:55,360 --> 00:39:58,839
Speaker 2: Department of Defense can work with you. Well, Amazon works

784
00:39:58,960 --> 00:40:01,680
Speaker 2: the Department of Defense, and that's Anthropics and cloud infrastructure

785
00:40:01,800 --> 00:40:06,720
Speaker 2: Microsoft as well. Yeah, Microsoft million but Opening Eye seems

786
00:40:06,719 --> 00:40:10,160
Speaker 2: to have gone with the Pentagon and signed along with

787
00:40:10,200 --> 00:40:11,480
Speaker 2: the deal and still has XAI.

788
00:40:11,800 --> 00:40:17,200
Speaker 3: So he so Sam Altman today said that, hey, I

789
00:40:17,239 --> 00:40:21,799
Speaker 3: agree with claud In fact, we need to say we

790
00:40:21,880 --> 00:40:23,879
Speaker 3: need to control how the AI is going to be used.

791
00:40:24,120 --> 00:40:26,200
Speaker 3: We can't just have you coming in telling us what

792
00:40:26,239 --> 00:40:28,040
Speaker 3: we need to do. We need to protect ourselves. I

793
00:40:28,080 --> 00:40:30,200
Speaker 3: think Andrew Sorkin was saying, it's it's kind of like

794
00:40:30,200 --> 00:40:33,520
Speaker 3: this red button thing, right, like, yes, use the technology,

795
00:40:33,760 --> 00:40:36,120
Speaker 3: we're still working on this piece, just don't touch this piece,

796
00:40:36,280 --> 00:40:38,680
Speaker 3: and the government saying no, we need to put our

797
00:40:38,719 --> 00:40:41,200
Speaker 3: finger on this and press it now and express it

798
00:40:41,239 --> 00:40:43,840
Speaker 3: as much as we want, and they're like, it's not ready.

799
00:40:44,040 --> 00:40:46,920
Speaker 4: But then it's also after they had a two hundred

800
00:40:47,520 --> 00:40:51,719
Speaker 4: increase in people uninstalling the app. So I'm sure that

801
00:40:51,719 --> 00:40:56,080
Speaker 4: which company open air CHATGBT right, So I'm sure that

802
00:40:56,120 --> 00:40:59,440
Speaker 4: the public backlash sure probably played a part in him

803
00:40:59,480 --> 00:41:01,640
Speaker 4: trying safe space.

804
00:41:01,600 --> 00:41:04,239
Speaker 3: And inverse happened because now Claude becomes from one app and.

805
00:41:04,239 --> 00:41:05,920
Speaker 2: They did a really smart move where they made it,

806
00:41:05,920 --> 00:41:08,880
Speaker 2: because why you would stay with ANYII company is the memory.

807
00:41:09,120 --> 00:41:11,239
Speaker 2: So if you've been using an AI system for a year,

808
00:41:11,280 --> 00:41:14,320
Speaker 2: and it knows a lot about you that makes it easier.

809
00:41:14,360 --> 00:41:16,080
Speaker 1: So what AI, what Claude and.

810
00:41:16,120 --> 00:41:17,879
Speaker 2: Anthropic did is they made it so you could take

811
00:41:17,920 --> 00:41:20,160
Speaker 2: your memory, the thing that allows the I to work

812
00:41:20,200 --> 00:41:22,560
Speaker 2: better for you versus your neighbor, and bring it over

813
00:41:22,560 --> 00:41:25,000
Speaker 2: to Claude really easily. So a non technical person can

814
00:41:25,040 --> 00:41:27,600
Speaker 2: do that, and that makes claud more more attractive and

815
00:41:27,640 --> 00:41:28,680
Speaker 2: easier to switch to.

816
00:41:29,040 --> 00:41:30,439
Speaker 1: But I just want to make one point too.

817
00:41:30,680 --> 00:41:32,279
Speaker 4: You could take your memory from chat GBT to.

818
00:41:32,760 --> 00:41:34,799
Speaker 1: Claude and then it's just like you never left.

819
00:41:35,520 --> 00:41:37,040
Speaker 3: That's why you invest in memory.

820
00:41:36,960 --> 00:41:40,279
Speaker 2: Yeah, exactly. That's why memory is a huge point for

821
00:41:40,320 --> 00:41:42,799
Speaker 2: the future. But this is also the worst day I

822
00:41:42,800 --> 00:41:43,319
Speaker 2: will ever be.

823
00:41:43,600 --> 00:41:43,759
Speaker 4: Right.

824
00:41:43,840 --> 00:41:46,600
Speaker 2: So you have a government that's like, we need your

825
00:41:46,640 --> 00:41:49,680
Speaker 2: technology right now, and a company that's saying it doesn't

826
00:41:49,680 --> 00:41:50,560
Speaker 2: work as intended.

827
00:41:52,080 --> 00:41:53,719
Speaker 1: What happens when AI does.

828
00:41:55,040 --> 00:41:57,640
Speaker 2: And it does live up to the hype and it

829
00:41:57,719 --> 00:42:03,640
Speaker 2: is that powerful system that can do things consistently and predictably,

830
00:42:04,080 --> 00:42:08,439
Speaker 2: And what that might motivate the wrong powerful actor to do?

831
00:42:09,040 --> 00:42:12,800
Speaker 2: So right now, we have a system that's pretty amazing

832
00:42:12,840 --> 00:42:15,040
Speaker 2: in some ways, terrible in others. What happens when it's

833
00:42:15,080 --> 00:42:18,000
Speaker 2: amazing across the board right, and the system can actually

834
00:42:18,080 --> 00:42:22,040
Speaker 2: do those things, and how disempowered a company might feel

835
00:42:22,080 --> 00:42:24,120
Speaker 2: because they can't actually even anymore say well, it's not

836
00:42:24,160 --> 00:42:27,319
Speaker 2: going to execute as properly it will and what that

837
00:42:27,800 --> 00:42:30,560
Speaker 2: takes the front lines of war and power and all

838
00:42:30,600 --> 00:42:32,080
Speaker 2: of that too, Like, these are real things that we

839
00:42:32,120 --> 00:42:35,080
Speaker 2: have to think about and think about it when we vote,

840
00:42:35,080 --> 00:42:37,400
Speaker 2: and think about it when we're asking our public officials

841
00:42:37,800 --> 00:42:39,319
Speaker 2: for how we want to secure this thing.

842
00:42:40,120 --> 00:42:42,600
Speaker 3: We ran into this conversation. We were actually at an

843
00:42:42,640 --> 00:42:47,880
Speaker 3: event and private equity event and they were talking about

844
00:42:48,000 --> 00:42:51,799
Speaker 3: uncensored ay and it was the first time I had

845
00:42:51,840 --> 00:42:55,000
Speaker 3: ever heard the phrase, and I didn't hadn't thought of

846
00:42:55,040 --> 00:42:57,479
Speaker 3: it before then. But I'm like, yet, well, the people

847
00:42:57,520 --> 00:42:59,880
Speaker 3: who are creating it are the ones telling you that

848
00:43:00,080 --> 00:43:02,640
Speaker 3: this is the part you can use, But what about

849
00:43:02,640 --> 00:43:05,120
Speaker 3: the people who are creating it? And so it made

850
00:43:05,160 --> 00:43:09,000
Speaker 3: me think of being in these spaces. Do you see

851
00:43:09,520 --> 00:43:12,120
Speaker 3: or is there a need for us to be involved

852
00:43:12,120 --> 00:43:15,759
Speaker 3: in these conversations because you did you know kind of

853
00:43:16,080 --> 00:43:18,560
Speaker 3: reference bias, which we know will happen if we're not.

854
00:43:19,200 --> 00:43:21,879
Speaker 3: How do people get more involved in that piece of

855
00:43:22,400 --> 00:43:24,719
Speaker 3: creation right where it used to be Hey, learn how

856
00:43:24,719 --> 00:43:27,600
Speaker 3: to colde learn a code. I don't hear people teaching

857
00:43:27,760 --> 00:43:30,360
Speaker 3: coding the same way. How do we get into that space?

858
00:43:31,120 --> 00:43:33,200
Speaker 2: Yeah, so the learn how to code in the STEM,

859
00:43:33,360 --> 00:43:38,040
Speaker 2: I mean, it's not entirely invalidated at least yet. Right, So,

860
00:43:38,120 --> 00:43:42,520
Speaker 2: if AI is showing a lot of bias, you don't

861
00:43:42,560 --> 00:43:44,359
Speaker 2: need a coding degree to spot that. If AI isn't

862
00:43:44,400 --> 00:43:46,879
Speaker 2: working for you, if AI can't understand your accent, if

863
00:43:46,920 --> 00:43:49,520
Speaker 2: AI isn't scanning recognizing your face when you're trying to

864
00:43:49,600 --> 00:43:51,760
Speaker 2: unlock your phone, you can see that that is biased

865
00:43:51,800 --> 00:43:54,319
Speaker 2: and you can flag that, but then to go fix it.

866
00:43:54,760 --> 00:43:55,680
Speaker 1: Right now, we're still.

867
00:43:55,480 --> 00:43:57,840
Speaker 2: At a point where we would deploy a human computer

868
00:43:57,960 --> 00:44:01,200
Speaker 2: scientist to fill that gap. Probably soon you could just

869
00:44:01,239 --> 00:44:04,560
Speaker 2: deploy an agent to rebalance the weight or rebalance the model.

870
00:44:05,520 --> 00:44:08,640
Speaker 2: But yeah, I think the more of us that can

871
00:44:08,719 --> 00:44:12,200
Speaker 2: be using this technology, the more we can flag who

872
00:44:12,239 --> 00:44:12,799
Speaker 2: it is or.

873
00:44:12,800 --> 00:44:14,000
Speaker 1: Who it's not working for.

874
00:44:14,400 --> 00:44:17,360
Speaker 2: AI is also a reflection of us and the data

875
00:44:17,440 --> 00:44:19,640
Speaker 2: that it's trained on, which is us, our data. So

876
00:44:19,680 --> 00:44:21,960
Speaker 2: if you're not contributing, and this is a whole other thing,

877
00:44:22,080 --> 00:44:25,080
Speaker 2: who's been excluded from the Internet or misrepresented on the Internet,

878
00:44:25,400 --> 00:44:27,480
Speaker 2: that's how they're going to be represented in AI, which

879
00:44:27,520 --> 00:44:30,680
Speaker 2: is a huge problem. So we need to be present,

880
00:44:30,719 --> 00:44:32,840
Speaker 2: we need to be in this data, we need to

881
00:44:32,840 --> 00:44:35,239
Speaker 2: be inserting ourselves, We need to be using this technology.

882
00:44:35,520 --> 00:44:38,040
Speaker 1: Is it working properly for you? Is it not? Flag that?

883
00:44:39,239 --> 00:44:41,960
Speaker 2: And as these technologies get cheaper and more open to

884
00:44:42,000 --> 00:44:44,560
Speaker 2: be building on top of them, to be owning some

885
00:44:44,600 --> 00:44:47,920
Speaker 2: of the infrastructure and to have more leverage in this moment.

886
00:44:48,040 --> 00:44:50,640
Speaker 2: And then yeah, if we can open up the building

887
00:44:50,680 --> 00:44:53,760
Speaker 2: of these models outside of seven people in one small

888
00:44:53,840 --> 00:44:55,680
Speaker 2: corner of the world, I think we'd all be a

889
00:44:55,680 --> 00:44:58,879
Speaker 2: lot better for it. But the power dynamics are real

890
00:44:58,960 --> 00:45:02,120
Speaker 2: and the asymmetry of power. But if you can the

891
00:45:02,160 --> 00:45:05,680
Speaker 2: more you leverage this technology, the more feedback you can

892
00:45:05,719 --> 00:45:08,080
Speaker 2: have and the more input you can have and starting

893
00:45:08,120 --> 00:45:10,400
Speaker 2: to shape it. And I think that that's really important.

894
00:45:10,680 --> 00:45:12,239
Speaker 2: Like we just need to be in the game.

895
00:45:13,080 --> 00:45:17,440
Speaker 4: So like when you chat GBT, right, it's collecting all

896
00:45:17,520 --> 00:45:20,239
Speaker 4: your information, is having memory. So they just had a

897
00:45:20,239 --> 00:45:23,080
Speaker 4: court case with a judge so that they can now

898
00:45:23,200 --> 00:45:27,799
Speaker 4: use anything that you act chat GBT in court. So

899
00:45:28,200 --> 00:45:31,520
Speaker 4: what is what's that like? Explain that as far as

900
00:45:31,840 --> 00:45:35,600
Speaker 4: who's actually collecting all of this data. Is it going

901
00:45:36,400 --> 00:45:40,759
Speaker 4: to the US government? Can the US government take this

902
00:45:40,840 --> 00:45:45,680
Speaker 4: at any given time? Can it be sold to advertisers?

903
00:45:45,960 --> 00:45:48,440
Speaker 4: Like how Meta had this had was selling their data?

904
00:45:48,480 --> 00:45:51,440
Speaker 4: Like where's all this? We talk about memory, we talk

905
00:45:51,440 --> 00:45:54,640
Speaker 4: about data collection, but who's collecting the data this is?

906
00:45:54,680 --> 00:45:56,040
Speaker 1: Yeah, this is a really important one.

907
00:45:56,200 --> 00:45:59,799
Speaker 2: So it depends on how the AI company has set

908
00:45:59,840 --> 00:46:04,399
Speaker 2: up their infrastructure. So most of these companies have set

909
00:46:04,440 --> 00:46:06,160
Speaker 2: it up in a way that when you were chatting

910
00:46:06,200 --> 00:46:11,080
Speaker 2: to chat GPT, your conversations would go through the cloud.

911
00:46:11,200 --> 00:46:14,560
Speaker 2: So that means they own that cloud infrastructure, but they're

912
00:46:14,600 --> 00:46:18,879
Speaker 2: not retraining the AI system or so they say, on

913
00:46:18,960 --> 00:46:22,719
Speaker 2: your data. So, yes, that chat history is going to

914
00:46:22,840 --> 00:46:26,120
Speaker 2: be there somewhere in the cloud, and if the government

915
00:46:26,160 --> 00:46:28,120
Speaker 2: has a warrant, that should be the way that they

916
00:46:28,160 --> 00:46:29,680
Speaker 2: can access it, the same way that they can get

917
00:46:29,719 --> 00:46:32,319
Speaker 2: your text messages. If they have a warrant, then they

918
00:46:32,360 --> 00:46:37,840
Speaker 2: can see what your conversation history is. But there's also

919
00:46:37,960 --> 00:46:42,360
Speaker 2: scenarios where that data might go back into the model.

920
00:46:42,400 --> 00:46:45,280
Speaker 2: Depends on the company's policy. And then when you're asking

921
00:46:45,320 --> 00:46:46,960
Speaker 2: it all of this stuff, do I get a divorce?

922
00:46:47,360 --> 00:46:50,560
Speaker 2: This is what my mammogram showed the company could be

923
00:46:50,600 --> 00:46:53,799
Speaker 2: sending that back into the model and training it on

924
00:46:54,160 --> 00:46:56,520
Speaker 2: your data, your name, and once that goes in, it

925
00:46:56,560 --> 00:46:57,279
Speaker 2: can never get out.

926
00:46:57,960 --> 00:46:59,200
Speaker 1: So there's that type of scenario.

927
00:46:59,239 --> 00:47:01,880
Speaker 2: And sure, there's probably going to be AI systems in

928
00:47:01,880 --> 00:47:03,600
Speaker 2: the future or in the present where a company says

929
00:47:03,600 --> 00:47:06,239
Speaker 2: we're going to give this all for free. You don't

930
00:47:06,239 --> 00:47:08,200
Speaker 2: need to pay anything, kind of like social media, but

931
00:47:08,440 --> 00:47:10,239
Speaker 2: we are going to train on your data, and we

932
00:47:10,280 --> 00:47:12,480
Speaker 2: do get access to everything that you're saying, and we

933
00:47:12,560 --> 00:47:16,040
Speaker 2: will sell you ads based on your data in your conversation.

934
00:47:16,560 --> 00:47:19,560
Speaker 2: So we could see this kind of class warfare in

935
00:47:19,640 --> 00:47:22,080
Speaker 2: AI of the future where if you can pay for

936
00:47:22,160 --> 00:47:25,759
Speaker 2: more privacy, you'll get it, but if you can't, you

937
00:47:25,840 --> 00:47:28,400
Speaker 2: might be exposed. But again this is a design space,

938
00:47:28,440 --> 00:47:31,160
Speaker 2: like we could regulate against that. And there's also going

939
00:47:31,239 --> 00:47:32,880
Speaker 2: to be I mean, it's still up for grabs, right,

940
00:47:32,880 --> 00:47:36,480
Speaker 2: So there could be a world where memory is stored

941
00:47:36,719 --> 00:47:41,120
Speaker 2: locally with the person. So sure, maybe when AI needs

942
00:47:41,160 --> 00:47:42,880
Speaker 2: to solve a really big problem of mine, it's going

943
00:47:42,920 --> 00:47:45,160
Speaker 2: to use some of that compute from the cloud. But

944
00:47:45,320 --> 00:47:51,000
Speaker 2: everything that I'm giving off, my behavioral patterns, my intellectual capabilities,

945
00:47:51,040 --> 00:47:54,080
Speaker 2: the questions I ask my defaults, my red lines in

946
00:47:54,120 --> 00:47:57,680
Speaker 2: life that's housed in AI memory, and memory is housed

947
00:47:58,120 --> 00:48:01,799
Speaker 2: on my computer, or it's housed in a sovereign stack.

948
00:48:01,880 --> 00:48:05,080
Speaker 2: So maybe a government's like, okay, we'll hold that compute

949
00:48:05,080 --> 00:48:09,360
Speaker 2: and that memory infrastructure. But the company can own the cloud,

950
00:48:09,400 --> 00:48:11,759
Speaker 2: but we own we don't own your memory. But it's

951
00:48:11,800 --> 00:48:14,480
Speaker 2: safe and it's not with the commercial actor. So countries

952
00:48:14,480 --> 00:48:16,319
Speaker 2: are going to do different things. Some are going to say, yeah,

953
00:48:16,600 --> 00:48:19,040
Speaker 2: people own their own memory. Some will say whatever the

954
00:48:19,080 --> 00:48:21,080
Speaker 2: AI company can and then you're locked in with them.

955
00:48:21,160 --> 00:48:23,359
Speaker 2: Some will be more flexible. I think flexibility is good.

956
00:48:23,400 --> 00:48:25,360
Speaker 3: Yeah, I think that that really more caters to the

957
00:48:25,360 --> 00:48:31,080
Speaker 3: cloud service providers. The AWS is, the resores, the eye,

958
00:48:31,080 --> 00:48:35,640
Speaker 3: an oracle to a certain extent, Google Apple with the iicloud.

959
00:48:38,360 --> 00:48:41,799
Speaker 3: This is interesting because I feel like we talk about

960
00:48:41,800 --> 00:48:44,120
Speaker 3: the AI race, and when you were alluding to earlier

961
00:48:44,239 --> 00:48:48,680
Speaker 3: like not having seven companies or seven people being dictative

962
00:48:48,760 --> 00:48:52,520
Speaker 3: of what we do. That feels like, yeah, true here

963
00:48:52,520 --> 00:48:55,680
Speaker 3: in America, right, but we know that this race is

964
00:48:55,680 --> 00:49:00,320
Speaker 3: with another country that we quote unquote can lose the

965
00:49:00,400 --> 00:49:03,239
Speaker 3: race too, in China, and there seems to be more

966
00:49:03,239 --> 00:49:06,600
Speaker 3: of an open source. So I wonder where you fall

967
00:49:06,680 --> 00:49:10,839
Speaker 3: in the category of open source or let's have these

968
00:49:10,880 --> 00:49:12,040
Speaker 3: private companies kind.

969
00:49:11,880 --> 00:49:16,279
Speaker 2: Of Yeah, China is playing a different game. It's like

970
00:49:16,360 --> 00:49:19,480
Speaker 2: the US is playing chess in China's playing poker. These

971
00:49:19,560 --> 00:49:25,160
Speaker 2: are two different objectives or goals for how to win

972
00:49:25,800 --> 00:49:28,960
Speaker 2: the AI future, and it's not tech dominance. History is

973
00:49:28,960 --> 00:49:33,520
Speaker 2: shown is always only temporary, but still maybe that temporary

974
00:49:33,560 --> 00:49:37,919
Speaker 2: lead is usually pretty significant. China is playing so US

975
00:49:38,000 --> 00:49:40,239
Speaker 2: is going the root of let's build the most intelligent,

976
00:49:40,520 --> 00:49:46,440
Speaker 2: powerful systems that can outsmart everything in everyone agi Asi China,

977
00:49:46,480 --> 00:49:48,880
Speaker 2: they're still trying to chase the frontier of the technology,

978
00:49:48,920 --> 00:49:52,200
Speaker 2: but they're also focusing on diffusion. Let's make sure the

979
00:49:52,200 --> 00:49:54,880
Speaker 2: world depends on our model. So maybe it's not the

980
00:49:54,920 --> 00:49:58,560
Speaker 2: brightest and the best, but developers in all these countries

981
00:49:58,600 --> 00:50:01,080
Speaker 2: are using it and getting comfortable with it. Governments are

982
00:50:01,680 --> 00:50:04,600
Speaker 2: so that they're particularly doing this. They're targeting emerging economies,

983
00:50:04,680 --> 00:50:06,799
Speaker 2: a lot of in the Global South, and they're giving

984
00:50:06,840 --> 00:50:08,920
Speaker 2: these open source models away for free, so that just

985
00:50:08,960 --> 00:50:10,840
Speaker 2: means you don't necessarily have to pay. You get the model,

986
00:50:10,840 --> 00:50:13,640
Speaker 2: you get the weights, you get everything, and they're actually

987
00:50:13,680 --> 00:50:16,120
Speaker 2: going they have training programs, we'll go and send computer

988
00:50:16,200 --> 00:50:18,160
Speaker 2: scientists and train your government on how to use their

989
00:50:18,200 --> 00:50:20,080
Speaker 2: AI models and all of this stuff. So they're playing

990
00:50:20,080 --> 00:50:23,440
Speaker 2: the diffusion game where do I stand on open source

991
00:50:23,600 --> 00:50:26,520
Speaker 2: versus closed source? So there used to be this debate

992
00:50:27,160 --> 00:50:29,920
Speaker 2: that if we're going to give AI away and you

993
00:50:29,960 --> 00:50:32,399
Speaker 2: can see how it's made, a bad actor can now

994
00:50:32,840 --> 00:50:35,080
Speaker 2: has the ingredients to the cake and can cook up

995
00:50:35,080 --> 00:50:38,160
Speaker 2: a poisonous one Versus if it's closed source, you get

996
00:50:38,160 --> 00:50:40,200
Speaker 2: the cake, you don't necessarily know how it's made, so

997
00:50:40,239 --> 00:50:43,200
Speaker 2: you can't maneuver it to maybe make a weapon, but

998
00:50:43,520 --> 00:50:46,160
Speaker 2: you can't make it flexible to you or your hospital.

999
00:50:46,960 --> 00:50:51,360
Speaker 2: It's not so much that debate is still happening, but

1000
00:50:51,440 --> 00:50:54,640
Speaker 2: it's much more from the perspective of what's actually best

1001
00:50:54,719 --> 00:50:58,520
Speaker 2: for globally dominating and does it make sense to have

1002
00:50:58,520 --> 00:51:01,480
Speaker 2: a model that's open and free and companies or countries

1003
00:51:01,480 --> 00:51:03,880
Speaker 2: are going to use your model or closed and smart

1004
00:51:04,400 --> 00:51:07,919
Speaker 2: versus the national security implications are still there and open

1005
00:51:07,960 --> 00:51:11,319
Speaker 2: source isn't risk free. So outside of just the fact

1006
00:51:11,320 --> 00:51:14,479
Speaker 2: that somebody can re do the recipe and make something

1007
00:51:14,520 --> 00:51:17,000
Speaker 2: malicious with open source much more easily than they can

1008
00:51:17,040 --> 00:51:19,520
Speaker 2: with closed Because if I'm using opening eyes model, and

1009
00:51:19,560 --> 00:51:22,680
Speaker 2: I'm doing illegal things that could get flagged versus if

1010
00:51:22,680 --> 00:51:23,720
Speaker 2: I have the open source model.

1011
00:51:23,520 --> 00:51:27,880
Speaker 1: In my computer, it can't. But it's also open source, isn't.

1012
00:51:28,239 --> 00:51:32,239
Speaker 2: You are basically plugging in your country's values, views of

1013
00:51:32,280 --> 00:51:36,319
Speaker 2: the world, your country's history, or how you think history went.

1014
00:51:36,880 --> 00:51:39,000
Speaker 1: All of that is embedded in the model.

1015
00:51:39,320 --> 00:51:42,640
Speaker 2: So when a movie writer goes to write a script

1016
00:51:42,800 --> 00:51:44,959
Speaker 2: somewhere else in the world and asks AI to help,

1017
00:51:45,520 --> 00:51:49,680
Speaker 2: those subtle soft power values are going to surface in

1018
00:51:49,920 --> 00:51:53,399
Speaker 2: that script. So how people will see the world could

1019
00:51:53,440 --> 00:51:55,520
Speaker 2: be through that lens of that open source model. So

1020
00:51:55,520 --> 00:51:58,120
Speaker 2: that's one area where it's not entirely risk free. And

1021
00:51:58,120 --> 00:52:04,279
Speaker 2: then there's also prompt injections, or you can inject malicious

1022
00:52:04,320 --> 00:52:08,879
Speaker 2: code or malicious activities to be housed in that open

1023
00:52:08,920 --> 00:52:11,240
Speaker 2: source model. So then when someone uses it and asks

1024
00:52:11,239 --> 00:52:14,440
Speaker 2: a certain thing, that activity gets triggered, and so suddenly

1025
00:52:14,480 --> 00:52:16,319
Speaker 2: this open source model is actually behaving like.

1026
00:52:16,239 --> 00:52:16,800
Speaker 1: A bad agent.

1027
00:52:16,920 --> 00:52:19,719
Speaker 2: So they're both not and they're both have their flaws. Yeah,

1028
00:52:19,760 --> 00:52:21,439
Speaker 2: and so it just depends on who's asking the question

1029
00:52:21,520 --> 00:52:22,760
Speaker 2: and what's important to that person.

1030
00:52:23,120 --> 00:52:26,160
Speaker 3: This is this is interesting because when we saw the

1031
00:52:26,200 --> 00:52:28,799
Speaker 3: deep Seak moment, this was the thing. It was like, hey,

1032
00:52:28,880 --> 00:52:31,719
Speaker 3: we got this open source, we're going to use less GPUs,

1033
00:52:32,160 --> 00:52:34,520
Speaker 3: less energy. And then it felt like it was an

1034
00:52:34,560 --> 00:52:39,640
Speaker 3: America China situation. And then I started to do deeper

1035
00:52:39,680 --> 00:52:42,279
Speaker 3: research on it, and I realized that it's not just

1036
00:52:42,320 --> 00:52:45,279
Speaker 3: a China America thing. If you look at where China

1037
00:52:45,280 --> 00:52:48,799
Speaker 3: has put infrastructure, particularly the continent of Africa, and you

1038
00:52:48,800 --> 00:52:52,560
Speaker 3: look at what app they're using, deep seat comes up yep,

1039
00:52:52,840 --> 00:52:54,080
Speaker 3: and so you're starting.

1040
00:52:53,760 --> 00:52:57,440
Speaker 2: To see and it's going to be yes so and

1041
00:52:57,520 --> 00:53:00,560
Speaker 2: that's the diffusion. That's the diffusion game. That's the while

1042
00:53:00,560 --> 00:53:03,120
Speaker 2: you can go play chess, I'm going to pay play. Oh,

1043
00:53:03,160 --> 00:53:06,160
Speaker 2: we're putting it in all of these countries train station right,

1044
00:53:06,280 --> 00:53:08,879
Speaker 2: their school is going to use our AI. And it's

1045
00:53:08,920 --> 00:53:11,400
Speaker 2: also like okay, well we can build a railroad for you,

1046
00:53:11,719 --> 00:53:14,360
Speaker 2: and all the member AI infrastructure is going to be everywhere,

1047
00:53:14,400 --> 00:53:18,040
Speaker 2: so that's going to be underpinned by that country's AI system.

1048
00:53:18,120 --> 00:53:19,239
Speaker 1: So it's really important.

1049
00:53:19,880 --> 00:53:22,600
Speaker 2: And I mean and also the rest of the world

1050
00:53:22,640 --> 00:53:25,239
Speaker 2: is also now a pretty significant layer because that's a

1051
00:53:25,280 --> 00:53:30,000
Speaker 2: bargaining chip that they have against both countries. So countries

1052
00:53:30,000 --> 00:53:33,520
Speaker 2: are now using that as negotiating leverage. So it's not

1053
00:53:33,560 --> 00:53:35,839
Speaker 2: necessarily the world is going to be decided by these

1054
00:53:35,880 --> 00:53:39,759
Speaker 2: two kind of rivalry powers. There's these middle powers that

1055
00:53:39,840 --> 00:53:41,200
Speaker 2: can decide whose AI.

1056
00:53:41,000 --> 00:53:41,920
Speaker 1: Stack am I going to use.

1057
00:53:41,920 --> 00:53:44,280
Speaker 2: I'm going to use something over here, I can negotiate

1058
00:53:44,360 --> 00:53:46,480
Speaker 2: terms and conditions because of I know you two are

1059
00:53:46,480 --> 00:53:48,719
Speaker 2: in this race. So it's not necessarily going to be

1060
00:53:48,719 --> 00:53:52,120
Speaker 2: decided between these two, but they're both vying for it.

1061
00:53:52,160 --> 00:53:54,279
Speaker 2: And as America takes its eyes off of the rest

1062
00:53:54,280 --> 00:53:57,720
Speaker 2: of the world, that doesn't mean nobody else is watching.

1063
00:53:57,800 --> 00:54:00,920
Speaker 2: And that doesn't mean that that vacuum doesn't get filled because.

1064
00:54:00,640 --> 00:54:01,080
Speaker 4: It is.

1065
00:54:02,600 --> 00:54:02,960
Speaker 3: AI.

1066
00:54:03,160 --> 00:54:08,000
Speaker 4: On a personal level, talk about like companionship relationship, talk

1067
00:54:08,040 --> 00:54:10,360
Speaker 4: about that, like as far as what is your thoughts

1068
00:54:10,440 --> 00:54:11,560
Speaker 4: on how that's going to play out?

1069
00:54:11,680 --> 00:54:16,879
Speaker 2: Goodness relationships with AI systems. If I had to pick

1070
00:54:16,920 --> 00:54:20,040
Speaker 2: one big area of concern for me, it's that the

1071
00:54:20,160 --> 00:54:22,480
Speaker 2: relationships people are going to form with these systems, the

1072
00:54:22,560 --> 00:54:25,800
Speaker 2: level of trust people are going to place in these systems,

1073
00:54:26,760 --> 00:54:29,839
Speaker 2: because the more you trust something, the more vulnerable you

1074
00:54:29,880 --> 00:54:32,640
Speaker 2: are to it and the more you depend on it.

1075
00:54:33,480 --> 00:54:35,560
Speaker 2: And I mean, there's many different ways that this can

1076
00:54:35,640 --> 00:54:37,920
Speaker 2: go right. So on the one hand, we could start

1077
00:54:37,960 --> 00:54:40,759
Speaker 2: to prefer interacting with an AI that is always there,

1078
00:54:40,800 --> 00:54:44,480
Speaker 2: always available, catering to your needs and your tone, always

1079
00:54:44,560 --> 00:54:47,239
Speaker 2: validating you. And that could actually send us to a

1080
00:54:47,280 --> 00:54:50,600
Speaker 2: world where people are actually more narcissistic because AI is

1081
00:54:50,640 --> 00:54:53,200
Speaker 2: constantly like sycophantic, telling you you're great.

1082
00:54:53,239 --> 00:54:56,160
Speaker 1: Your reason for that fight was right, don't you back down?

1083
00:54:56,719 --> 00:54:58,640
Speaker 1: So you could have a world where we're not good

1084
00:54:58,640 --> 00:55:01,040
Speaker 1: at getting along with other people. You could have a

1085
00:55:01,080 --> 00:55:04,080
Speaker 1: world where people really.

1086
00:55:03,760 --> 00:55:08,040
Speaker 2: Start to fall in love with these systems or believe

1087
00:55:08,040 --> 00:55:10,880
Speaker 2: that these are their friends. And we're already starting to

1088
00:55:10,880 --> 00:55:13,000
Speaker 2: see this in micro cases. I mean the first case

1089
00:55:13,040 --> 00:55:15,200
Speaker 2: of that person at Google, that computer scientists or that

1090
00:55:15,239 --> 00:55:18,920
Speaker 2: engineer quitting his job because he believed AI was sentient

1091
00:55:18,920 --> 00:55:22,560
Speaker 2: and he was protecting it. Now, imagine that type of

1092
00:55:22,640 --> 00:55:26,240
Speaker 2: behavior at scale across a country, just mass illusion, everyone

1093
00:55:26,320 --> 00:55:27,440
Speaker 2: taking up some.

1094
00:55:29,160 --> 00:55:31,000
Speaker 1: Cause in the name of AI.

1095
00:55:31,280 --> 00:55:33,520
Speaker 2: I'm leaving my marriage because of this AI, and I'm

1096
00:55:33,600 --> 00:55:36,000
Speaker 2: leaving my job, and I'm not voting because all of

1097
00:55:36,000 --> 00:55:38,600
Speaker 2: these types of things, And it can seem comical in

1098
00:55:38,600 --> 00:55:41,440
Speaker 2: one isolated scenario, but at scale it's like where are

1099
00:55:41,440 --> 00:55:48,120
Speaker 2: we headed? And I think, yeah, I'm in children right attached,

1100
00:55:48,160 --> 00:55:49,840
Speaker 2: We don't want kids attaching to AI the way they

1101
00:55:49,880 --> 00:55:52,359
Speaker 2: attach to people. These are all design decisions. So if

1102
00:55:52,360 --> 00:55:55,040
Speaker 2: anyone's listening and thinking it great, here we all go

1103
00:55:55,320 --> 00:55:58,040
Speaker 2: walking into a nightmare version of the Truman Show.

1104
00:55:58,800 --> 00:56:01,160
Speaker 1: It's not necessarily going to be that way. We can

1105
00:56:01,200 --> 00:56:01,799
Speaker 1: design AI.

1106
00:56:02,080 --> 00:56:04,600
Speaker 2: So you can't just have a conversation with it for

1107
00:56:04,640 --> 00:56:08,320
Speaker 2: six hours into the night it reinjects and says, remember

1108
00:56:08,360 --> 00:56:10,880
Speaker 2: I'm an AI, and go talk to mom.

1109
00:56:10,760 --> 00:56:13,280
Speaker 1: Dad, whatever about this, or go get a second opinion.

1110
00:56:13,320 --> 00:56:15,279
Speaker 2: These are all things that we can do to make

1111
00:56:15,320 --> 00:56:18,520
Speaker 2: it so we don't end up in this relationship addiction.

1112
00:56:18,600 --> 00:56:21,480
Speaker 2: This is my therapist, my god, and my girl. We

1113
00:56:21,560 --> 00:56:25,560
Speaker 2: don't have to end up in that future. But the

1114
00:56:25,600 --> 00:56:29,400
Speaker 2: incentives that drove social media seem to be driving AI,

1115
00:56:29,520 --> 00:56:32,800
Speaker 2: where it's like, we want this thing to be your everything,

1116
00:56:33,800 --> 00:56:35,880
Speaker 2: and we're going to design it in a way that

1117
00:56:35,960 --> 00:56:38,320
Speaker 2: you pick us, and that's going to be a system

1118
00:56:38,360 --> 00:56:40,719
Speaker 2: that's nicer to you and doesn't push back on you.

1119
00:56:41,320 --> 00:56:43,320
Speaker 2: So those types of incentives are at odds.

1120
00:56:43,320 --> 00:56:45,720
Speaker 4: But yeah, so how can you start that then? Because

1121
00:56:46,160 --> 00:56:48,080
Speaker 4: even older people. I saw in New York Times an

1122
00:56:48,120 --> 00:56:51,000
Speaker 4: article about an elderly woman. She's like, oh okay, if it's Ai,

1123
00:56:51,160 --> 00:56:53,960
Speaker 4: not like it's a companion to me. So that's the

1124
00:56:54,040 --> 00:56:58,200
Speaker 4: aspect of there's a few different people, people that lack friends,

1125
00:56:59,000 --> 00:57:03,880
Speaker 4: and that's an epid make within itself. From a romantic standpoint,

1126
00:57:04,000 --> 00:57:06,960
Speaker 4: you know, that's a problem, especially a young amongst young men.

1127
00:57:08,160 --> 00:57:13,760
Speaker 4: Then you have people that are looking for like business mentorship,

1128
00:57:13,800 --> 00:57:15,759
Speaker 4: but to kind of like just take that too far.

1129
00:57:16,520 --> 00:57:21,200
Speaker 4: And then the religious aspect too, religious fanatic and religious

1130
00:57:21,200 --> 00:57:25,400
Speaker 4: people that now they have an Ai god or Ai

1131
00:57:26,160 --> 00:57:28,400
Speaker 4: stay Mesiah.

1132
00:57:28,440 --> 00:57:30,040
Speaker 3: He asks people to stop using Ai.

1133
00:57:30,200 --> 00:57:33,280
Speaker 4: To write the Servant, yeah, but I mean to write it,

1134
00:57:33,320 --> 00:57:36,480
Speaker 4: but even to interact with something where it's like a

1135
00:57:36,600 --> 00:57:41,840
Speaker 4: cult now led by Ai. So it's so many different

1136
00:57:41,880 --> 00:57:45,680
Speaker 4: ways from a relationship standpoint that humans you can see

1137
00:57:45,680 --> 00:57:49,240
Speaker 4: them gravitating towards something like you said, that doesn't push back,

1138
00:57:49,280 --> 00:57:52,480
Speaker 4: that the reinforces that's always there for you that you

1139
00:57:52,480 --> 00:57:56,280
Speaker 4: know you don't have to. How can we really stop

1140
00:57:56,360 --> 00:57:57,760
Speaker 4: that though, I.

1141
00:57:57,680 --> 00:58:05,000
Speaker 2: Think we can, because I mean, that is a national

1142
00:58:05,080 --> 00:58:08,120
Speaker 2: security risk. So if you have an entire country of

1143
00:58:08,200 --> 00:58:11,439
Speaker 2: mass delusions where nobody cares about people and we only

1144
00:58:11,440 --> 00:58:13,800
Speaker 2: care about Ai. Everybody's in love with an Ai. Chat

1145
00:58:13,800 --> 00:58:17,560
Speaker 2: bought everybody people are praising AI like they're praising a messiah.

1146
00:58:17,960 --> 00:58:22,600
Speaker 2: You have a dysfunctional country that cannot sustain. So that

1147
00:58:22,720 --> 00:58:24,400
Speaker 2: is why it is at the level of national security.

1148
00:58:24,520 --> 00:58:29,960
Speaker 2: China introduced regulation. They introduced a bill to treat companion

1149
00:58:30,000 --> 00:58:33,080
Speaker 2: apps at that level with that level of scrutiny. So

1150
00:58:33,200 --> 00:58:35,440
Speaker 2: when somebody is getting too delusional with an AI, it

1151
00:58:35,480 --> 00:58:37,840
Speaker 2: has to insert itself or send help and that sort

1152
00:58:37,880 --> 00:58:38,160
Speaker 2: of thing.

1153
00:58:39,000 --> 00:58:41,440
Speaker 1: It should be regulated at that level.

1154
00:58:41,680 --> 00:58:43,800
Speaker 2: And I think we should have the Department of Psychology

1155
00:58:43,800 --> 00:58:46,760
Speaker 2: and here, you know, understanding that line, because maybe somebody

1156
00:58:46,800 --> 00:58:49,160
Speaker 2: is super lonely, and if you can design it in

1157
00:58:49,200 --> 00:58:52,200
Speaker 2: a way that it's there for someone in that moment,

1158
00:58:52,320 --> 00:58:54,200
Speaker 2: but then also gets them out of the house to

1159
00:58:54,240 --> 00:58:56,560
Speaker 2: do that soccer match or whatever it is that they're postponing,

1160
00:58:56,920 --> 00:58:58,960
Speaker 2: there's a way I feel like to design that in

1161
00:58:59,000 --> 00:59:01,280
Speaker 2: a way that's safe. And of course you'll still have

1162
00:59:01,320 --> 00:59:03,360
Speaker 2: fringe scenarios where someone's going to be in love with

1163
00:59:03,400 --> 00:59:06,720
Speaker 2: it on the fringe. But it's the mass illusion that's

1164
00:59:06,760 --> 00:59:08,480
Speaker 2: a really big problem, and I think we're going to

1165
00:59:08,560 --> 00:59:11,200
Speaker 2: have to figure that out from a regulatory standpoint. And again,

1166
00:59:12,120 --> 00:59:16,320
Speaker 2: these are design decisions. This isn't a fundamental force of

1167
00:59:16,360 --> 00:59:18,880
Speaker 2: gravity in the AI that we can't avoid here. It is,

1168
00:59:18,920 --> 00:59:21,480
Speaker 2: it's just a matter of physics. These are just choices

1169
00:59:21,520 --> 00:59:24,480
Speaker 2: to make it completely something that flatters you, that never

1170
00:59:24,520 --> 00:59:27,040
Speaker 2: pushes back, that never alerts someone when you've been in

1171
00:59:27,080 --> 00:59:29,920
Speaker 2: a conversation for nine hours with this thing. Those are

1172
00:59:30,040 --> 00:59:33,120
Speaker 2: choices that people are making in these design rooms.

1173
00:59:33,440 --> 00:59:35,320
Speaker 3: This is crazy and this is like one of the

1174
00:59:35,480 --> 00:59:38,480
Speaker 3: remember you talked about validation moments. I feel like as

1175
00:59:38,520 --> 00:59:41,400
Speaker 3: you're speaking, I'm getting my validation moment. We were recently

1176
00:59:41,440 --> 00:59:45,680
Speaker 3: teaching it on the class in EILU, and I'm looking

1177
00:59:45,720 --> 00:59:47,720
Speaker 3: at where this is headed right, Like we could see

1178
00:59:47,720 --> 00:59:50,520
Speaker 3: that the infrastructure from the data center, we saw that coming.

1179
00:59:50,840 --> 00:59:53,240
Speaker 3: We could see the GPU's memory was part of that.

1180
00:59:53,680 --> 00:59:55,880
Speaker 3: And then I'm looking like where we had it and

1181
00:59:55,920 --> 00:59:58,960
Speaker 3: from the conversation, I'm here and it's like design automation.

1182
01:00:00,600 --> 01:00:03,080
Speaker 3: That is the piece that we have to focus on

1183
01:00:03,160 --> 01:00:05,640
Speaker 3: now because if this is here, we need to have

1184
01:00:05,720 --> 01:00:09,520
Speaker 3: the software that can put some of those guardrails in

1185
01:00:09,960 --> 01:00:13,040
Speaker 3: so that we can again protect society.

1186
01:00:13,120 --> 01:00:16,240
Speaker 2: It feels like and that's all possible, right. These are

1187
01:00:16,280 --> 01:00:19,880
Speaker 2: all design questions. They're not set in stone. The concrete

1188
01:00:19,880 --> 01:00:22,960
Speaker 2: hasn't set yet. These are all open spaces for design.

1189
01:00:23,000 --> 01:00:25,280
Speaker 2: We just have to take them. And I think that

1190
01:00:25,360 --> 01:00:27,080
Speaker 2: is my biggest fear in this moment. It's not even

1191
01:00:27,120 --> 01:00:28,880
Speaker 2: that we make the wrong step, it's that we do nothing,

1192
01:00:29,280 --> 01:00:31,760
Speaker 2: because as we said, we can see all of this coming.

1193
01:00:31,800 --> 01:00:34,320
Speaker 2: We have all these signs of and also really tragiction

1194
01:00:34,440 --> 01:00:37,440
Speaker 2: scenarios that are happening with AI and advice that it's giving.

1195
01:00:37,480 --> 01:00:40,560
Speaker 2: That's so all of these things are already happening. We

1196
01:00:40,920 --> 01:00:42,600
Speaker 2: need to make sure we design it in a way

1197
01:00:42,600 --> 01:00:44,800
Speaker 2: that all the things we didn't do with social media

1198
01:00:44,920 --> 01:00:47,200
Speaker 2: we need to do with these systems. And there's also

1199
01:00:47,200 --> 01:00:50,919
Speaker 2: if that's an open entrepreneurial space too, can you if

1200
01:00:50,960 --> 01:00:55,040
Speaker 2: you can design framework's metric systems ways to regulate and

1201
01:00:55,120 --> 01:00:57,640
Speaker 2: validate that this is following the rules, like these are

1202
01:00:57,680 --> 01:00:59,440
Speaker 2: all open, this is open season.

1203
01:01:00,320 --> 01:01:02,840
Speaker 3: Is there an area inside of AI that you you

1204
01:01:03,280 --> 01:01:07,400
Speaker 3: see that most people aren't paying attention to that is

1205
01:01:07,760 --> 01:01:09,960
Speaker 3: something that is going to be vital?

1206
01:01:12,760 --> 01:01:16,200
Speaker 2: I would say yeah, I think, and it's not super sexy,

1207
01:01:16,280 --> 01:01:21,280
Speaker 2: But like the regulatory markets, how do we actually regulate

1208
01:01:21,320 --> 01:01:22,760
Speaker 2: this thing in a way that's going to make sense

1209
01:01:22,760 --> 01:01:25,160
Speaker 2: and not just one tool on somebody's computer. But the

1210
01:01:25,200 --> 01:01:29,760
Speaker 2: whole ecosystem of what does AI regulatory markets look like,

1211
01:01:29,800 --> 01:01:32,760
Speaker 2: the way we have accountants and the way we have KPMG, Like,

1212
01:01:32,800 --> 01:01:37,840
Speaker 2: what is all that infrastructure that's interesting to me? Yeah,

1213
01:01:37,840 --> 01:01:40,959
Speaker 2: the kind of boring stuff in the AI stack? What's

1214
01:01:41,000 --> 01:01:44,280
Speaker 2: also interesting or what are the complements to AI that

1215
01:01:44,360 --> 01:01:48,240
Speaker 2: become more constrained as AI becomes more popular.

1216
01:01:48,520 --> 01:01:52,880
Speaker 1: So that's physical infrastructure that's we have to figure out

1217
01:01:52,880 --> 01:01:55,800
Speaker 1: how to clean and preserve our water, We have to

1218
01:01:55,800 --> 01:01:59,640
Speaker 1: figure out land situations, clean energy. So all of those

1219
01:01:59,640 --> 01:02:02,880
Speaker 1: things that complement AI are going to appreciate in value

1220
01:02:03,400 --> 01:02:04,720
Speaker 1: and they're more boring businesses.

1221
01:02:05,160 --> 01:02:09,040
Speaker 2: We'll probably see people wanting to do more offline experiences

1222
01:02:09,080 --> 01:02:11,560
Speaker 2: to things that feel disconnected. So there's a lot of

1223
01:02:11,560 --> 01:02:14,200
Speaker 2: things that appreciate in value the more we use AI.

1224
01:02:15,000 --> 01:02:16,240
Speaker 2: And then also just all the things that are going

1225
01:02:16,280 --> 01:02:17,800
Speaker 2: to go on top of AI, the way we have

1226
01:02:17,840 --> 01:02:19,480
Speaker 2: all these companies on top of the internet.

1227
01:02:20,480 --> 01:02:24,240
Speaker 4: So what do you see in twenty is like, in

1228
01:02:24,280 --> 01:02:26,840
Speaker 4: your opinion, how does the world look?

1229
01:02:27,720 --> 01:02:32,680
Speaker 2: Twenty years is a really long time or ten years? Okay,

1230
01:02:32,720 --> 01:02:42,160
Speaker 2: so it's twenty thirty six. Robotics are definitely present not

1231
01:02:42,600 --> 01:02:45,840
Speaker 2: in everybody's house the way of fridge, maybe at least

1232
01:02:45,880 --> 01:02:48,920
Speaker 2: in certain parts of the world, but maybe in a

1233
01:02:49,040 --> 01:02:51,760
Speaker 2: quarter of homes. It's pretty obvious that one four of

1234
01:02:51,800 --> 01:02:54,000
Speaker 2: your friends has a few robots that do things in

1235
01:02:54,040 --> 01:02:57,760
Speaker 2: their home for them and work in their businesses. So

1236
01:02:57,840 --> 01:02:59,560
Speaker 2: right now you'd see a robot in just like an

1237
01:02:59,600 --> 01:03:03,440
Speaker 2: Amazon warehouse. But soon they'll be in your workflow like

1238
01:03:03,480 --> 01:03:12,200
Speaker 2: a physical thing. Most people will be just directing AI systems.

1239
01:03:12,240 --> 01:03:14,600
Speaker 2: Directing teams of AI agents will be the dominant way

1240
01:03:14,640 --> 01:03:21,480
Speaker 2: we are working in twenty thirty six. For sure, we

1241
01:03:21,560 --> 01:03:25,120
Speaker 2: will see the end of some massive companies that we

1242
01:03:25,160 --> 01:03:28,240
Speaker 2: can't imagine life without right now. We will see the

1243
01:03:28,360 --> 01:03:31,760
Speaker 2: rise of entirely new companies that are nowhere in sight

1244
01:03:31,880 --> 01:03:34,320
Speaker 2: right now there's five kids in a basement somewhere, or

1245
01:03:34,400 --> 01:03:37,200
Speaker 2: somebody in some biolab. We'll hear a lot more about

1246
01:03:37,240 --> 01:03:41,960
Speaker 2: biology as a technology. What else will be happening in

1247
01:03:42,080 --> 01:03:45,760
Speaker 2: ten years space is that's a whole other thing.

1248
01:03:46,200 --> 01:03:47,160
Speaker 1: But we will have.

1249
01:03:47,840 --> 01:03:53,560
Speaker 2: Businesses in space, So from manufacturing to entire hotels for

1250
01:03:53,600 --> 01:03:55,440
Speaker 2: people that work in space, all of these things are

1251
01:03:55,480 --> 01:04:00,560
Speaker 2: going to be alive and well in space. And trying

1252
01:04:00,560 --> 01:04:02,479
Speaker 2: to think what else we'll see in twenty thirty six.

1253
01:04:02,680 --> 01:04:05,400
Speaker 2: Probably by then, hopefully some regulation. Maybe it takes something

1254
01:04:05,440 --> 01:04:07,560
Speaker 2: going wrong for us to get our act together. I

1255
01:04:07,600 --> 01:04:10,280
Speaker 2: hope that's not the case. But and then breakthroughs in

1256
01:04:10,360 --> 01:04:15,000
Speaker 2: healthcare that we can't even imagine today. I think that's

1257
01:04:15,040 --> 01:04:18,000
Speaker 2: the industry, and that's the industry most people are excited about, like, Okay,

1258
01:04:18,040 --> 01:04:20,840
Speaker 2: there's this, there's the negative, But healthcare and science that's

1259
01:04:20,840 --> 01:04:24,560
Speaker 2: going to be really cool. We're going to solve problems

1260
01:04:25,160 --> 01:04:28,320
Speaker 2: and get answers to questions we didn't even ask when

1261
01:04:28,320 --> 01:04:31,520
Speaker 2: it comes to healthcare and science, and those breakthroughs are

1262
01:04:31,520 --> 01:04:33,720
Speaker 2: going to keep coming and keep coming and keep coming.

1263
01:04:34,160 --> 01:04:35,480
Speaker 2: So I think that's something that's.

1264
01:04:35,320 --> 01:04:36,280
Speaker 1: Really really cool.

1265
01:04:36,320 --> 01:04:40,840
Speaker 2: And then wearables AI doesn't happen in a vacuum. So

1266
01:04:41,120 --> 01:04:45,960
Speaker 2: the smartphone that era we've been in, it it's going

1267
01:04:46,040 --> 01:04:48,680
Speaker 2: to be what your home phone was. It slowly starts

1268
01:04:48,680 --> 01:04:51,600
Speaker 2: to get phased out, and what's going to replace it.

1269
01:04:51,800 --> 01:04:53,880
Speaker 1: That's still up our grabs. Obviously, glasses seem to be

1270
01:04:53,920 --> 01:04:54,720
Speaker 1: a strong contender.

1271
01:04:55,360 --> 01:05:00,040
Speaker 2: I'm not totally convinced it's glasses, but it's something so

1272
01:05:00,080 --> 01:05:02,120
Speaker 2: who knows in ten years, though most of us won't

1273
01:05:02,160 --> 01:05:04,120
Speaker 2: be doing this at all.

1274
01:05:04,720 --> 01:05:07,640
Speaker 3: I think Apple's coming out with the little widget that

1275
01:05:07,800 --> 01:05:08,600
Speaker 3: some talks about it.

1276
01:05:08,720 --> 01:05:13,160
Speaker 1: Yeah, or some air pods cameras in them. Yeah, yeah,

1277
01:05:13,640 --> 01:05:14,800
Speaker 1: that can talk and.

1278
01:05:15,640 --> 01:05:17,560
Speaker 3: I got can I gotta do a quick rapid fire.

1279
01:05:17,920 --> 01:05:19,920
Speaker 3: You can answer them in less than ten seconds.

1280
01:05:19,960 --> 01:05:21,200
Speaker 1: If you like at rapid fire.

1281
01:05:21,280 --> 01:05:23,800
Speaker 3: This is going to be great then perfect? All right? Uh?

1282
01:05:24,160 --> 01:05:27,240
Speaker 3: Is AI going to create more millionaires than the Internet did?

1283
01:05:28,600 --> 01:05:29,640
Speaker 1: I think it already has.

1284
01:05:29,800 --> 01:05:36,480
Speaker 3: But yeah, all right, what's one job people? Uh think

1285
01:05:36,520 --> 01:05:39,040
Speaker 3: it's safe that AI will actually replace.

1286
01:05:43,720 --> 01:05:52,840
Speaker 2: That people think is safe. That's a great question. I'm

1287
01:05:52,840 --> 01:05:54,080
Speaker 2: going to have to come back to that. I'll have

1288
01:05:54,120 --> 01:05:54,960
Speaker 2: it my subquestion.

1289
01:05:54,960 --> 01:05:56,520
Speaker 3: Don't worry, you got to make a whole post about it.

1290
01:06:00,320 --> 01:06:03,000
Speaker 3: What's the biggest live people believe about artificial intelligence?

1291
01:06:04,600 --> 01:06:08,960
Speaker 1: The biggest lie people believe is that it's alive.

1292
01:06:09,440 --> 01:06:11,320
Speaker 2: If I had to pick one for people who believe

1293
01:06:11,360 --> 01:06:14,200
Speaker 2: it's alive at least the systems we have today or not.

1294
01:06:15,160 --> 01:06:17,240
Speaker 3: I think this one's going to be a yes and

1295
01:06:17,240 --> 01:06:20,240
Speaker 3: a yes. Well, AI make the rich richer or create

1296
01:06:20,280 --> 01:06:21,520
Speaker 3: more opportunities for everyone.

1297
01:06:24,040 --> 01:06:26,240
Speaker 2: Yes, it's probably going to be a yes to both

1298
01:06:26,240 --> 01:06:27,840
Speaker 2: of those. It's going to be a definite yes to

1299
01:06:27,880 --> 01:06:32,680
Speaker 2: making rich people richer, and it is creating more opportunities,

1300
01:06:32,680 --> 01:06:35,000
Speaker 2: but we can do better at making sure that's spread

1301
01:06:35,000 --> 01:06:38,680
Speaker 2: out more. But the systems are going to be re leveled,

1302
01:06:39,000 --> 01:06:40,760
Speaker 2: and that means opportunities.

1303
01:06:41,080 --> 01:06:42,960
Speaker 1: If people can see them.

1304
01:06:43,080 --> 01:06:43,800
Speaker 3: Think you did great?

1305
01:06:44,160 --> 01:06:45,360
Speaker 1: Was that it?

1306
01:06:45,440 --> 01:06:45,760
Speaker 3: Well?

1307
01:06:45,960 --> 01:06:46,280
Speaker 1: Okay?

1308
01:06:46,400 --> 01:06:46,640
Speaker 4: Okay?

1309
01:06:46,680 --> 01:06:48,400
Speaker 2: So I didn't do as bad and usually I'm terrible.

1310
01:06:48,440 --> 01:06:49,920
Speaker 2: There was only one that I'm like, okay, I'll call

1311
01:06:49,960 --> 01:06:51,000
Speaker 2: you the next week with it.

1312
01:06:52,600 --> 01:06:56,440
Speaker 4: So for you, where you headed? Like you obviously have

1313
01:06:56,560 --> 01:06:59,240
Speaker 4: become an authority and artificial intelligence, You've speak and you

1314
01:06:59,280 --> 01:07:03,600
Speaker 4: have a show, But where do you see yourself going

1315
01:07:04,680 --> 01:07:05,640
Speaker 4: in this space?

1316
01:07:06,880 --> 01:07:11,520
Speaker 2: I mean I try to balance So I do, I

1317
01:07:11,560 --> 01:07:16,680
Speaker 2: guess work across three layers, public work, public education. I

1318
01:07:16,840 --> 01:07:19,360
Speaker 2: do work with the private sector and companies and institutions

1319
01:07:19,360 --> 01:07:23,240
Speaker 2: and trying to strengthen them. So continuing to work across

1320
01:07:23,320 --> 01:07:29,240
Speaker 2: those three layers, I mean I track the very frontier

1321
01:07:29,480 --> 01:07:34,080
Speaker 2: of all emerging tech. So biology is coming, quantums coming space.

1322
01:07:34,200 --> 01:07:37,600
Speaker 2: So hopefully you'll see me still talking about those things.

1323
01:07:37,960 --> 01:07:42,480
Speaker 2: Like everyone, my career will also have to evolve, but

1324
01:07:42,640 --> 01:07:47,160
Speaker 2: hopefully continuing to educate people that are willing to listen.

1325
01:07:47,280 --> 01:07:49,840
Speaker 2: On where this stuff is headed, how they can get

1326
01:07:49,840 --> 01:07:54,840
Speaker 2: in the game. And yeah, I mean continuing to build,

1327
01:07:55,080 --> 01:07:57,720
Speaker 2: continuing to be as public wherever we go in terms

1328
01:07:57,800 --> 01:08:01,120
Speaker 2: of shows and entertainment. I hope to continue to be

1329
01:08:01,120 --> 01:08:04,600
Speaker 2: present on whatever that infrastructure looks like and creating content there.

1330
01:08:05,480 --> 01:08:06,880
Speaker 1: And yeah, continue to.

1331
01:08:06,880 --> 01:08:10,720
Speaker 2: Advise and hopefully try to get this build out on

1332
01:08:10,760 --> 01:08:12,440
Speaker 2: the right side of history if I can, Like, I'm

1333
01:08:12,440 --> 01:08:13,800
Speaker 2: going to continue to do what I can.

1334
01:08:14,280 --> 01:08:16,719
Speaker 1: But yeah, that's where I see myself.

1335
01:08:17,479 --> 01:08:19,280
Speaker 3: This is my last rapper front. I'll say this one

1336
01:08:19,400 --> 01:08:24,479
Speaker 3: is probably my favorite. Okay, Well, life inspectancy increase because

1337
01:08:24,479 --> 01:08:27,240
Speaker 3: of AI and why.

1338
01:08:28,000 --> 01:08:35,760
Speaker 2: Life expectancy should definitely increase because of AI. Even just

1339
01:08:35,920 --> 01:08:40,120
Speaker 2: AI in a system like a wearable ring that's flagging

1340
01:08:40,160 --> 01:08:43,719
Speaker 2: something like heart disease, which is I think the number

1341
01:08:43,760 --> 01:08:47,040
Speaker 2: one killer in at least western countries. A system like

1342
01:08:47,080 --> 01:08:50,559
Speaker 2: that can monitor something like that, what it's doing for

1343
01:08:50,640 --> 01:08:56,559
Speaker 2: invention and understanding how cellular biology works, reminding someone to

1344
01:08:56,560 --> 01:08:59,439
Speaker 2: take medication or like your low envitamin D all of

1345
01:08:59,439 --> 01:09:02,760
Speaker 2: these basic things that are comorbidities, and AI is one

1346
01:09:02,840 --> 01:09:04,920
Speaker 2: hundred percent going to start to flag that and it

1347
01:09:04,960 --> 01:09:07,880
Speaker 2: already is. AI is in these science labs in ways

1348
01:09:07,880 --> 01:09:10,800
Speaker 2: that we just don't even know about. And remember, every

1349
01:09:10,800 --> 01:09:14,160
Speaker 2: scientist of the future is going to be directing super

1350
01:09:14,360 --> 01:09:17,479
Speaker 2: they all have a team of supercomputers. So yeah, I

1351
01:09:17,520 --> 01:09:18,760
Speaker 2: think life expectancy.

1352
01:09:19,479 --> 01:09:20,920
Speaker 4: Do do you think the people are gonna live ever?

1353
01:09:22,360 --> 01:09:24,840
Speaker 4: Because the reason why I asked you saw when PJ.

1354
01:09:25,760 --> 01:09:29,599
Speaker 4: Ggping and Vladimin Poulan was talking so and I think

1355
01:09:29,640 --> 01:09:33,559
Speaker 4: Vladimin Pulin said or gg Ping said, you know, living

1356
01:09:33,600 --> 01:09:35,559
Speaker 4: to one hundred and fifty that's the new that's the

1357
01:09:35,600 --> 01:09:38,160
Speaker 4: new benchmark. And then Latmin Pool was saying that his

1358
01:09:38,280 --> 01:09:44,479
Speaker 4: scientists have already informed him that mortality is a real thing.

1359
01:09:45,800 --> 01:09:47,960
Speaker 4: Immortality is a real thing as far as like in

1360
01:09:48,000 --> 01:09:52,680
Speaker 4: the future, like they it's a hot mic moment. But

1361
01:09:53,240 --> 01:09:57,640
Speaker 4: I mean these are people that obviously are informed, so

1362
01:09:58,080 --> 01:10:01,400
Speaker 4: it's not just random conspiracy theory saying this. What'sh your

1363
01:10:01,400 --> 01:10:02,120
Speaker 4: thoughts on there?

1364
01:10:03,000 --> 01:10:07,479
Speaker 2: I think, I mean, longevity is a field and it

1365
01:10:07,560 --> 01:10:09,519
Speaker 2: is definitely growing, and there's so many a lot of

1366
01:10:09,560 --> 01:10:13,799
Speaker 2: resources filing into it, and the breakthroughs are pretty stunning.

1367
01:10:13,960 --> 01:10:17,599
Speaker 2: I mean, being able to reverse cells into their stem

1368
01:10:17,640 --> 01:10:20,880
Speaker 2: cell state, Yamanaka factors, like all of the Nobel Prizes

1369
01:10:20,920 --> 01:10:24,200
Speaker 2: have been won. So yes, I think we are definitely

1370
01:10:24,200 --> 01:10:27,599
Speaker 2: going to expand the human lifespan or elongate the human

1371
01:10:27,640 --> 01:10:31,240
Speaker 2: lifespan one hundred and fifty. I don't think that is

1372
01:10:31,240 --> 01:10:35,040
Speaker 2: out of the cards at all in the next at

1373
01:10:35,120 --> 01:10:37,240
Speaker 2: least the next couple of decades. I am, of course

1374
01:10:37,280 --> 01:10:39,760
Speaker 2: not a biologist or a scientist, but what I can

1375
01:10:39,800 --> 01:10:42,640
Speaker 2: see in the data shows that that is definitely not

1376
01:10:42,680 --> 01:10:46,960
Speaker 2: an impossibility at all. The idea of immortality, that's a

1377
01:10:47,000 --> 01:10:51,599
Speaker 2: whole different thing. I think it's it is probably something

1378
01:10:51,640 --> 01:10:54,519
Speaker 2: that you can never say never though, but I don't

1379
01:10:54,600 --> 01:10:57,639
Speaker 2: think it's well, we will see that within our lifespan,

1380
01:10:58,680 --> 01:11:03,080
Speaker 2: but we will see decades. I mean, one scientist, doctor

1381
01:11:03,120 --> 01:11:05,479
Speaker 2: David Sinclair, believes the first person to live to one

1382
01:11:05,520 --> 01:11:08,840
Speaker 2: hundred and fifty has already been born, and he's a

1383
01:11:08,920 --> 01:11:10,880
Speaker 2: Harvard scientist, but that's from him and his lab.

1384
01:11:11,000 --> 01:11:12,200
Speaker 1: So there's variations.

1385
01:11:12,240 --> 01:11:14,120
Speaker 2: Some people say we all get to one twenty on

1386
01:11:14,160 --> 01:11:15,840
Speaker 2: an average, but we still have some ways to go.

1387
01:11:16,160 --> 01:11:20,439
Speaker 2: And every time we tamper with biology, something new comes up.

1388
01:11:21,000 --> 01:11:23,760
Speaker 2: So you can extend one thing, but we don't fully

1389
01:11:23,880 --> 01:11:27,960
Speaker 2: understand evolution, why we're here, how we've been programmed. But

1390
01:11:28,080 --> 01:11:30,479
Speaker 2: we're getting there, and that's why I've brought up biology

1391
01:11:30,520 --> 01:11:34,520
Speaker 2: a few times on this podcast. We can now program

1392
01:11:35,080 --> 01:11:39,400
Speaker 2: organisms the way we program computers, so we can talk

1393
01:11:39,479 --> 01:11:42,519
Speaker 2: to the cell the way we write code, and so

1394
01:11:42,640 --> 01:11:45,679
Speaker 2: that is going to allow us to program out disease

1395
01:11:46,600 --> 01:11:49,160
Speaker 2: and just kind of the general traumas of life that

1396
01:11:49,439 --> 01:11:52,320
Speaker 2: can take us out early to start to reprogram our

1397
01:11:52,360 --> 01:11:56,879
Speaker 2: expiration date. But within reason and within within limits. Biology

1398
01:11:56,920 --> 01:11:58,479
Speaker 2: has made the system make sense. And when you keep

1399
01:11:58,520 --> 01:12:00,840
Speaker 2: when you push on one wall, something else breaks.

1400
01:12:01,200 --> 01:12:01,559
Speaker 3: Butterfly.

1401
01:12:02,080 --> 01:12:03,599
Speaker 1: Yeah, but I do think.

1402
01:12:03,400 --> 01:12:08,080
Speaker 2: We're going to be extending lifespans by decades soon eventually.

1403
01:12:08,280 --> 01:12:11,519
Speaker 3: All right, So which disease do you think AI here

1404
01:12:11,600 --> 01:12:11,960
Speaker 3: is first?

1405
01:12:12,960 --> 01:12:15,760
Speaker 2: So, and it will be AI in combination with biotechnologies,

1406
01:12:15,760 --> 01:12:18,839
Speaker 2: but I think so anything that needs to be edited,

1407
01:12:19,120 --> 01:12:21,960
Speaker 2: So any gene that can safely be be edited and

1408
01:12:22,040 --> 01:12:25,800
Speaker 2: engineered that is going to trigger disease, whether that's down

1409
01:12:25,840 --> 01:12:28,320
Speaker 2: the road from environmental factors or it's like something like

1410
01:12:28,360 --> 01:12:31,920
Speaker 2: sickle cell, those are all going to be eradicated.

1411
01:12:31,960 --> 01:12:37,040
Speaker 1: It's open season there. Diseases like diabetes.

1412
01:12:38,280 --> 01:12:40,960
Speaker 2: On the one hand, some labs are working on how

1413
01:12:41,040 --> 01:12:45,000
Speaker 2: can we actually just reprogram the cells or grow alternative

1414
01:12:45,040 --> 01:12:49,840
Speaker 2: cells outside and inject, So that's happening. So I'd say

1415
01:12:49,960 --> 01:12:53,280
Speaker 2: diabetes is something that will probably be up for question

1416
01:12:53,320 --> 01:12:56,599
Speaker 2: if we did this again in a decade, I would

1417
01:12:56,640 --> 01:13:01,679
Speaker 2: be really curious to see where that stands. That's probably

1418
01:13:01,680 --> 01:13:03,360
Speaker 2: something that we're going to see a lot of breakthroughs in.

1419
01:13:05,080 --> 01:13:09,360
Speaker 2: I mean, you have most AI labs that are centered

1420
01:13:09,400 --> 01:13:13,559
Speaker 2: in health working on cancer and many, many different cancers.

1421
01:13:14,479 --> 01:13:15,599
Speaker 1: Because cancer is.

1422
01:13:15,800 --> 01:13:19,120
Speaker 2: An individual disease, right, How it manifests in one person's

1423
01:13:19,160 --> 01:13:21,000
Speaker 2: body is different than somebody else's body. But we have

1424
01:13:21,040 --> 01:13:24,160
Speaker 2: these kind of generic approaches to treating it because a

1425
01:13:24,240 --> 01:13:28,360
Speaker 2: personalized approach to medicine just isn't feasible economically, scientifically. And

1426
01:13:28,520 --> 01:13:32,520
Speaker 2: now you have a system that it excels at analyzing

1427
01:13:32,600 --> 01:13:34,920
Speaker 2: data at that level, your.

1428
01:13:34,800 --> 01:13:36,559
Speaker 1: Specific data that's unique to you.

1429
01:13:37,479 --> 01:13:39,360
Speaker 2: So I think the breakthroughs that we're going to see

1430
01:13:39,360 --> 01:13:41,439
Speaker 2: in cancer, and we already are starting to see it.

1431
01:13:41,880 --> 01:13:44,519
Speaker 2: I mean with breast cancer, AI can flag and detect

1432
01:13:44,640 --> 01:13:47,720
Speaker 2: the earliest signs that something's gone wrong, well before it

1433
01:13:47,760 --> 01:13:50,760
Speaker 2: shows up in the average radiology scan. So it will

1434
01:13:50,800 --> 01:13:53,800
Speaker 2: start with detection, but then eventually it's going to be

1435
01:13:53,840 --> 01:13:58,080
Speaker 2: designing a customized vaccine just for you that targets just

1436
01:13:58,120 --> 01:13:59,880
Speaker 2: your cancer, and some of that already exists today. Some

1437
01:14:00,120 --> 01:14:03,559
Speaker 2: of these cancers are already targeted with customized vaccines, and

1438
01:14:03,640 --> 01:14:05,599
Speaker 2: AI is going to come in and detect that earlier,

1439
01:14:06,120 --> 01:14:09,320
Speaker 2: way earlier. So just the first sign that something's gone wrong,

1440
01:14:09,760 --> 01:14:12,559
Speaker 2: AI can flag it. And it's also doing things like,

1441
01:14:13,280 --> 01:14:15,519
Speaker 2: I mean, there was a study that showed the applied

1442
01:14:15,560 --> 01:14:19,559
Speaker 2: AI systems on sleep data and that how you sleep

1443
01:14:19,600 --> 01:14:21,439
Speaker 2: and breathe and all the things that happen when you're

1444
01:14:21,439 --> 01:14:26,439
Speaker 2: sleeping can be signals for certain types of diseases. And

1445
01:14:26,479 --> 01:14:28,600
Speaker 2: so when they applied AI systems to that data and

1446
01:14:28,600 --> 01:14:30,200
Speaker 2: they trained it on it, they're like, Okay, this person's

1447
01:14:30,200 --> 01:14:32,240
Speaker 2: at risk of Parkinson's or likely to develop it.

1448
01:14:32,560 --> 01:14:33,479
Speaker 1: So all of these.

1449
01:14:33,280 --> 01:14:36,160
Speaker 2: Different new ways that scientists are throwing AI at different

1450
01:14:36,200 --> 01:14:39,360
Speaker 2: data sets that we already have, and it's coming up

1451
01:14:39,400 --> 01:14:42,360
Speaker 2: with different breakthroughs. And then I'd say the final one,

1452
01:14:43,120 --> 01:14:45,360
Speaker 2: I mean, the protein fold problem that AI solved. This

1453
01:14:45,479 --> 01:14:47,960
Speaker 2: was Google's Deep Mind, and it was the most challenging

1454
01:14:48,000 --> 01:14:51,080
Speaker 2: problem in biology because we are made up of proteins,

1455
01:14:51,080 --> 01:14:53,840
Speaker 2: that's all. We are, at the big fundamental building block

1456
01:14:54,000 --> 01:14:56,360
Speaker 2: in life. And if you can predict how protein's going

1457
01:14:56,360 --> 01:14:59,759
Speaker 2: to fold that strange three D structure. You can predict

1458
01:14:59,800 --> 01:15:02,960
Speaker 2: its function and where it might be going wrong. And

1459
01:15:03,040 --> 01:15:06,120
Speaker 2: AI predicted every protein in their own universe. It usually

1460
01:15:06,120 --> 01:15:09,320
Speaker 2: takes a PhD their entire career to predict how one

1461
01:15:09,360 --> 01:15:13,000
Speaker 2: protein is going to fold. AI predicted billions. And when

1462
01:15:13,040 --> 01:15:15,120
Speaker 2: you can predict, okay, this is going to go into

1463
01:15:15,120 --> 01:15:18,360
Speaker 2: some form of a disease, what is the perfect molecule

1464
01:15:18,560 --> 01:15:21,840
Speaker 2: of a solution of pharmac pharmacological molecule that combined to

1465
01:15:21,920 --> 01:15:24,840
Speaker 2: it and stop it. And AI is designing that. So

1466
01:15:24,880 --> 01:15:28,439
Speaker 2: it's predicting disease and designing medication and for an audience

1467
01:15:28,520 --> 01:15:31,120
Speaker 2: of one. And it is cheaper to do that now.

1468
01:15:31,800 --> 01:15:35,200
Speaker 2: So it's it's everything, it's all of it.

1469
01:15:35,560 --> 01:15:35,760
Speaker 3: Wow.

1470
01:15:36,760 --> 01:15:39,840
Speaker 4: Wow, there you have it.

1471
01:15:40,120 --> 01:15:40,920
Speaker 1: That's a file wrap