1
00:00:00,120 --> 00:00:02,840
Speaker 1: Hey everyone, it's Robert and Joe here. Today we've got

2
00:00:02,840 --> 00:00:04,720
Speaker 1: something a little bit different to share with you. It

3
00:00:04,840 --> 00:00:08,000
Speaker 1: is a new season of the Smart Talks with IBM

4
00:00:08,119 --> 00:00:09,119
Speaker 1: podcast series.

5
00:00:09,600 --> 00:00:11,680
Speaker 2: Today we are witnessed to one of those rare moments

6
00:00:11,680 --> 00:00:14,360
Speaker 2: in history, the rise of an innovative technology with the

7
00:00:14,360 --> 00:00:18,680
Speaker 2: potential to radically transform business and society forever. The technology,

8
00:00:18,760 --> 00:00:22,200
Speaker 2: of course, is artificial intelligence, and it's the central focus

9
00:00:22,239 --> 00:00:24,800
Speaker 2: for this new season of Smart Talks with IBM.

10
00:00:25,320 --> 00:00:28,400
Speaker 1: Join hosts from your favorite Pushkin podcasts as they talk

11
00:00:28,480 --> 00:00:31,640
Speaker 1: with industry experts and leaders to explore how businesses can

12
00:00:31,680 --> 00:00:35,360
Speaker 1: integrate AI into their workflows and help drive real change

13
00:00:35,400 --> 00:00:38,200
Speaker 1: in this new era of AI. And of course, host

14
00:00:38,280 --> 00:00:40,440
Speaker 1: Malcolm Gladwell will be there to guide you through the

15
00:00:40,479 --> 00:00:42,640
Speaker 1: season and throw in his two cents as well.

16
00:00:43,080 --> 00:00:46,120
Speaker 2: Look out for new episodes of Smart Talks with IBM

17
00:00:46,400 --> 00:00:49,519
Speaker 2: every other week on the iHeartRadio app, Apple Podcasts, or

18
00:00:49,560 --> 00:00:53,360
Speaker 2: wherever you get your podcasts, and learn more at IBM

19
00:00:53,479 --> 00:00:55,480
Speaker 2: dot com slash smart Talks.

20
00:00:57,760 --> 00:01:03,920
Speaker 1: All right, Welcome everybody, you guys excited, here we go.

21
00:01:04,280 --> 00:01:07,800
Speaker 3: Hello, Hello, Welcome to Smart Talks with IBM, a podcast

22
00:01:07,800 --> 00:01:13,679
Speaker 3: from Pushkin Industries. iHeartRadio and IBM. I'm Malcolm Gladwell. This season,

23
00:01:13,680 --> 00:01:17,680
Speaker 3: we're continuing our conversations with new creators visionaries who are

24
00:01:17,720 --> 00:01:21,240
Speaker 3: creatively applying technology and business to drive change, but with

25
00:01:21,319 --> 00:01:25,640
Speaker 3: a focus on the transformative power of artificial intelligence and

26
00:01:25,680 --> 00:01:28,559
Speaker 3: what it means to leverage AI as a game changing

27
00:01:28,680 --> 00:01:33,319
Speaker 3: multiplier for your business. Today's episode is a bit different.

28
00:01:33,680 --> 00:01:37,560
Speaker 3: I was recently joined on stage by Dario Gill for

29
00:01:37,640 --> 00:01:40,000
Speaker 3: a conversation in front of a live audience at the

30
00:01:40,000 --> 00:01:44,120
Speaker 3: iHeartMedia headquarters in Manhattan. Dario is the senior vice president

31
00:01:44,319 --> 00:01:47,840
Speaker 3: and director of IBM Research, one of the world's largest

32
00:01:47,960 --> 00:01:52,240
Speaker 3: and most influential corporate research labs. We discussed the rise

33
00:01:52,280 --> 00:01:55,760
Speaker 3: of generative AI and what it means for business and society.

34
00:01:56,080 --> 00:01:59,560
Speaker 3: He also explained how organizations that leverage AI to create

35
00:01:59,640 --> 00:02:04,360
Speaker 3: value will dominate in the near future. Okay, let's get

36
00:02:04,360 --> 00:02:10,920
Speaker 3: on to the conversation. Hello everyone, welcome, and I'm here

37
00:02:11,000 --> 00:02:14,640
Speaker 3: with doctor Dario Gil and I wanted to say before

38
00:02:14,639 --> 00:02:16,280
Speaker 3: we get started. This is something I said backstage that

39
00:02:16,760 --> 00:02:23,000
Speaker 3: I feel very guilty today because you're the you know,

40
00:02:23,760 --> 00:02:25,800
Speaker 3: you know, arguably one of the most important figures in

41
00:02:26,160 --> 00:02:28,920
Speaker 3: AI research in the world, and we have taken you

42
00:02:28,960 --> 00:02:33,240
Speaker 3: away from your job for a morning. It's like, if

43
00:02:33,600 --> 00:02:37,440
Speaker 3: you know Oppenheimer's wife in nineteen forty four said let's

44
00:02:37,760 --> 00:02:41,600
Speaker 3: go and have a little getaway in the Bahamas. It's

45
00:02:41,600 --> 00:02:43,400
Speaker 3: that kind of thing. You know, what do you say

46
00:02:43,440 --> 00:02:46,519
Speaker 3: to your wife, I can't we have got to work

47
00:02:46,520 --> 00:02:48,800
Speaker 3: on this thing I can't tell you about. She's like

48
00:02:48,840 --> 00:02:51,200
Speaker 3: getting me out of Los Alamos. No, So I do

49
00:02:51,240 --> 00:02:56,160
Speaker 3: feel guilty. We've set back AI research by by about

50
00:02:56,160 --> 00:03:01,160
Speaker 3: four hours here. But I wanted to you've been up

51
00:03:01,160 --> 00:03:04,400
Speaker 3: with with ibo for twenty years, twenty years this summer.

52
00:03:04,800 --> 00:03:06,400
Speaker 3: So and how old were you when you Not to

53
00:03:06,440 --> 00:03:08,120
Speaker 3: give away your age, but you were how old when

54
00:03:08,160 --> 00:03:08,640
Speaker 3: you started?

55
00:03:08,919 --> 00:03:09,679
Speaker 4: I was twenty eight?

56
00:03:09,720 --> 00:03:11,480
Speaker 3: Okay, yeah, So I want to go back to your

57
00:03:11,480 --> 00:03:14,400
Speaker 3: twenty eight year old self. Now, if I asked you

58
00:03:14,440 --> 00:03:18,040
Speaker 3: about artificial intelligence, I asked twenty eight year old Dario,

59
00:03:18,919 --> 00:03:22,200
Speaker 3: what does the future hold for AI? How quickly will

60
00:03:22,639 --> 00:03:26,080
Speaker 3: this new technology transform our world? Et cetera, et cetera.

61
00:03:26,160 --> 00:03:28,160
Speaker 3: What would twenty eight year old Darigo said?

62
00:03:28,639 --> 00:03:30,880
Speaker 4: Well, I think the first thing is that even though

63
00:03:30,919 --> 00:03:33,280
Speaker 4: AI as a feel has been with us for a

64
00:03:33,280 --> 00:03:36,360
Speaker 4: long time, since the mid nineteen fifties, at that time,

65
00:03:36,800 --> 00:03:40,240
Speaker 4: AI was not a very polite word to say, meaning

66
00:03:40,320 --> 00:03:44,040
Speaker 4: within the scientific community, people didn't use sort of that term.

67
00:03:44,080 --> 00:03:45,960
Speaker 4: They would have said things like, you know, maybe I

68
00:03:46,080 --> 00:03:50,240
Speaker 4: do things relate to machine learning, right, or statistical techniques

69
00:03:50,280 --> 00:03:53,400
Speaker 4: in terms of classifiers and so on. But AI had

70
00:03:53,440 --> 00:03:56,520
Speaker 4: a mixed reputation, right, it had gone through different cycles

71
00:03:56,520 --> 00:04:00,800
Speaker 4: of hype, and it's also if moments of you know,

72
00:04:00,880 --> 00:04:05,760
Speaker 4: a lot of negativity towards it because of lack of success.

73
00:04:05,920 --> 00:04:07,600
Speaker 4: And so I think that would be the first thing

74
00:04:07,600 --> 00:04:10,440
Speaker 4: we probably say, like AI is like what is that? Like,

75
00:04:10,480 --> 00:04:14,120
Speaker 4: you know, respectable scientists are not working on AI the

76
00:04:14,200 --> 00:04:18,000
Speaker 4: finest side and that really changed over the last fifteen years. Only, right,

77
00:04:18,240 --> 00:04:20,839
Speaker 4: I would say, with the advent of deep learning over

78
00:04:20,880 --> 00:04:23,320
Speaker 4: the last decade, is when that re enter again the

79
00:04:23,400 --> 00:04:26,240
Speaker 4: lexicon of saying AI and that that was a legitimate

80
00:04:26,240 --> 00:04:28,920
Speaker 4: thing to work on. So I would say that that's

81
00:04:28,960 --> 00:04:30,440
Speaker 4: the first thing I think we would have noticed a

82
00:04:30,440 --> 00:04:31,480
Speaker 4: contrast twenty years ago.

83
00:04:31,760 --> 00:04:35,640
Speaker 3: Yeah, So what point in your twenty year tenure at

84
00:04:35,839 --> 00:04:39,599
Speaker 3: IBM would you say you kind of snapped into present

85
00:04:40,240 --> 00:04:41,520
Speaker 3: kind of wow mode.

86
00:04:42,880 --> 00:04:50,800
Speaker 4: I would say in a late two thousands, when IBM

87
00:04:51,600 --> 00:04:56,880
Speaker 4: was working on the Jeopardy project and just seeing the

88
00:04:56,920 --> 00:05:00,400
Speaker 4: demonstrations of what could be done in question answering.

89
00:05:00,800 --> 00:05:04,000
Speaker 3: It's literally Jeopardy. Is this crucial moment in the history

90
00:05:04,000 --> 00:05:04,560
Speaker 3: of YEA.

91
00:05:05,240 --> 00:05:07,640
Speaker 4: You know, there had been a long and wonderful history

92
00:05:08,320 --> 00:05:12,080
Speaker 4: inside IBM on AI. So for example, like you know,

93
00:05:12,160 --> 00:05:15,000
Speaker 4: in terms of like these grand challenges at the very

94
00:05:15,040 --> 00:05:18,000
Speaker 4: beginning of the field founding, which is this famous Dartmouth

95
00:05:18,000 --> 00:05:22,440
Speaker 4: conference that actually IBM sponsored h to create. There was

96
00:05:22,720 --> 00:05:27,000
Speaker 4: an IBM and there called Nathaniel Rochester, and there were

97
00:05:27,040 --> 00:05:30,040
Speaker 4: a few others who right after that they started thinking

98
00:05:30,040 --> 00:05:33,040
Speaker 4: about demonstrations of this field. And then for example, they

99
00:05:33,080 --> 00:05:36,479
Speaker 4: created the first you know game to play checkers and

100
00:05:36,560 --> 00:05:39,719
Speaker 4: to demonstrate that you could do machine learning on that.

101
00:05:40,400 --> 00:05:43,039
Speaker 4: Obviously we saw later in the nineties like chess that

102
00:05:43,120 --> 00:05:45,719
Speaker 4: was a very famous example of that, Deep Blue with

103
00:05:45,800 --> 00:05:49,039
Speaker 4: Deep Blue right and playing with Caspar and then but

104
00:05:49,120 --> 00:05:51,720
Speaker 4: I think the moment that was really those other ones

105
00:05:51,760 --> 00:05:54,240
Speaker 4: felt like, you know, kind of like brute force anticipating,

106
00:05:54,279 --> 00:05:56,679
Speaker 4: sort of like moves ahead. But this aspect of dealing

107
00:05:56,680 --> 00:06:00,960
Speaker 4: with language and question answering felt different, and I think

108
00:06:01,040 --> 00:06:03,599
Speaker 4: for for us internally and many others, was when a

109
00:06:03,600 --> 00:06:06,000
Speaker 4: moment of saying like, wow, you know, what are the

110
00:06:06,040 --> 00:06:09,440
Speaker 4: possibilities here? And then soon after that connected to the

111
00:06:09,480 --> 00:06:12,880
Speaker 4: sort of advancements in computing and with deep learning. The

112
00:06:12,960 --> 00:06:15,200
Speaker 4: last decade has just been an all out, you know,

113
00:06:15,279 --> 00:06:17,440
Speaker 4: sort of like front of advancements and that, and I

114
00:06:17,520 --> 00:06:19,560
Speaker 4: just continue to be more and more impressed. And the

115
00:06:19,640 --> 00:06:21,360
Speaker 4: last few years have been remarkable too.

116
00:06:21,560 --> 00:06:26,040
Speaker 3: Yeah. So I'll ask you three quick conceptual questions before

117
00:06:26,040 --> 00:06:27,880
Speaker 3: we dig into it, just so I sort of get

118
00:06:27,880 --> 00:06:32,120
Speaker 3: a we all get a feel for the shape of AI.

119
00:06:32,920 --> 00:06:36,960
Speaker 3: Question Number one is where are we in the evolution

120
00:06:37,080 --> 00:06:41,160
Speaker 3: of this? So you know the obvious question. We we're

121
00:06:41,160 --> 00:06:43,760
Speaker 3: all suddenly aware of it, we're talking about it. Can

122
00:06:43,800 --> 00:06:45,800
Speaker 3: you give us an analogy about where we are in

123
00:06:45,880 --> 00:06:50,680
Speaker 3: the kind of likely evolution of this is a technology.

124
00:06:50,720 --> 00:06:54,440
Speaker 4: So I think we're on a significant inflection point that

125
00:06:54,839 --> 00:06:58,920
Speaker 4: it feels the equivalent of the first browsers when they

126
00:06:58,960 --> 00:07:03,200
Speaker 4: appear and people imagine the possibilities of the Internet or

127
00:07:03,240 --> 00:07:07,120
Speaker 4: more imagined experience the internet. The Internet had been around,

128
00:07:07,320 --> 00:07:09,679
Speaker 4: right for quite a few decades. AI has been around

129
00:07:09,920 --> 00:07:12,560
Speaker 4: for many decades. I think the moment we find ourselves

130
00:07:12,680 --> 00:07:15,760
Speaker 4: is that people can touch it and they can Before

131
00:07:15,800 --> 00:07:17,920
Speaker 4: they were a systems that were like behind the scenes,

132
00:07:18,000 --> 00:07:22,120
Speaker 4: like your search results or translation systems, but they didn't

133
00:07:22,160 --> 00:07:24,040
Speaker 4: have the experience of like, this is what it feels

134
00:07:24,080 --> 00:07:27,040
Speaker 4: like to interact with this thing. So that's what I mean.

135
00:07:27,080 --> 00:07:29,200
Speaker 4: I think maybe that analogy of the browser is appropriate

136
00:07:29,240 --> 00:07:31,440
Speaker 4: because it's all of a sudden, it's like whoa, you know,

137
00:07:31,640 --> 00:07:35,240
Speaker 4: these network of machines and content can be distributed and

138
00:07:35,280 --> 00:07:37,920
Speaker 4: everybody can self publish, and there was a moment that

139
00:07:37,960 --> 00:07:40,080
Speaker 4: we all remember that, and I think that that is

140
00:07:40,200 --> 00:07:42,640
Speaker 4: what the world has experience over the last nine months

141
00:07:42,720 --> 00:07:46,520
Speaker 4: or so on. So but fundamentally, also what is important

142
00:07:46,600 --> 00:07:48,760
Speaker 4: is that this moment is where the ease of the

143
00:07:48,840 --> 00:07:53,440
Speaker 4: number of people that can build and use AI has skyrocketed.

144
00:07:54,000 --> 00:07:58,360
Speaker 4: So over the last decade, you know, technology firms that

145
00:07:58,480 --> 00:08:03,120
Speaker 4: had large research teams could build AI that worked really well, honestly,

146
00:08:03,560 --> 00:08:06,400
Speaker 4: but when you went down into say hey can everybody

147
00:08:06,520 --> 00:08:08,800
Speaker 4: use it? Can a data science team in a bank,

148
00:08:09,000 --> 00:08:11,640
Speaker 4: you know, go and develop these applications, it was like

149
00:08:11,800 --> 00:08:13,960
Speaker 4: more complicated. Some could do it, but it was more

150
00:08:14,040 --> 00:08:16,560
Speaker 4: the barrier of entry was high. Now is very different

151
00:08:16,920 --> 00:08:20,200
Speaker 4: because of foundation models and the implications that that has.

152
00:08:20,080 --> 00:08:23,720
Speaker 3: For at the moment where the technology is being democratized.

153
00:08:23,080 --> 00:08:28,560
Speaker 4: In demarketized, frankly, it works better for classes of problems

154
00:08:28,600 --> 00:08:31,200
Speaker 4: like programming and other things. Is really incredibly impressive what

155
00:08:31,240 --> 00:08:33,600
Speaker 4: it can do. So the accuracy and the performance of

156
00:08:33,640 --> 00:08:36,560
Speaker 4: it is much better, and the ease of use and

157
00:08:36,600 --> 00:08:38,760
Speaker 4: the number of use cases we can pursue it much bigger,

158
00:08:38,840 --> 00:08:40,560
Speaker 4: So that democratization is a big difference.

159
00:08:40,559 --> 00:08:42,440
Speaker 3: But when you say, when you make it an analogy

160
00:08:42,480 --> 00:08:46,760
Speaker 3: to the first browsers, if we do another one of

161
00:08:46,840 --> 00:08:49,520
Speaker 3: these time travel questions back at the beginning of the

162
00:08:49,520 --> 00:08:53,240
Speaker 3: first browsers, it's safe to say many of the potential

163
00:08:53,440 --> 00:08:56,800
Speaker 3: uses of the Internet and such we hadn't even begun.

164
00:08:56,880 --> 00:08:59,439
Speaker 3: We couldn't even anticipate, right, Right, So we're at the

165
00:08:59,480 --> 00:09:02,760
Speaker 3: point where the future direction is largely unpredictable.

166
00:09:03,000 --> 00:09:05,800
Speaker 4: Yeah, I think that that is right. Because it's such

167
00:09:05,800 --> 00:09:10,199
Speaker 4: a horizontal technology that the intersection of the horizontal capability,

168
00:09:10,280 --> 00:09:14,520
Speaker 4: which is about expanding our productivity and tasks that we

169
00:09:14,559 --> 00:09:17,360
Speaker 4: wouldn't be able to do efficiently without it, has to

170
00:09:17,440 --> 00:09:20,240
Speaker 4: marry now the use cases that reflect the diversity of

171
00:09:20,320 --> 00:09:23,360
Speaker 4: human experience, our institutional diversity. So as more and more

172
00:09:23,400 --> 00:09:26,000
Speaker 4: institutions said, you know, I'm focused on agriculture, you know,

173
00:09:26,160 --> 00:09:29,240
Speaker 4: to be able to improve seeds. You know, in these

174
00:09:29,320 --> 00:09:32,280
Speaker 4: kinds of environments, they'll find their own context in which

175
00:09:32,320 --> 00:09:34,120
Speaker 4: that matters that the creators of a I did not

176
00:09:34,160 --> 00:09:37,080
Speaker 4: anticipate at the beginning. So I think that that is

177
00:09:37,120 --> 00:09:39,520
Speaker 4: then the fruit of surprises will be like why I

178
00:09:39,559 --> 00:09:41,319
Speaker 4: wouldn't even think that it could be used for that.

179
00:09:41,559 --> 00:09:44,760
Speaker 4: And also clever people will create new business models as

180
00:09:44,880 --> 00:09:47,520
Speaker 4: associated with that, like it happened with the Internet of

181
00:09:47,559 --> 00:09:50,439
Speaker 4: course as well, and that will be its own source

182
00:09:50,480 --> 00:09:52,880
Speaker 4: of transformation and change in its own right. So I

183
00:09:52,920 --> 00:09:55,040
Speaker 4: think all of that is yet to unfold. Right, what

184
00:09:55,160 --> 00:09:58,040
Speaker 4: we're seeing is this catalyst moment of technology that works

185
00:09:58,040 --> 00:09:59,760
Speaker 4: well enough and it can be democratized.

186
00:10:00,080 --> 00:10:03,840
Speaker 3: Yeah, what next sort of conceptual question? You know, we

187
00:10:03,840 --> 00:10:10,320
Speaker 3: can loosely understand or categorize innovations in terms of their

188
00:10:10,360 --> 00:10:15,040
Speaker 3: impact on the kind of balance of power between haves

189
00:10:15,080 --> 00:10:20,000
Speaker 3: and have nots. Some innovations, you know, obviously favor those

190
00:10:20,040 --> 00:10:24,560
Speaker 3: who already have make the rich richer. Some some it's

191
00:10:24,559 --> 00:10:28,319
Speaker 3: arising to tie the lifts all boats, and some bias

192
00:10:28,320 --> 00:10:32,040
Speaker 3: in the other direction. They close the gap between is

193
00:10:32,080 --> 00:10:35,440
Speaker 3: it possible to say to predict which of those three

194
00:10:35,480 --> 00:10:37,200
Speaker 3: categories AI might fall into.

195
00:10:38,240 --> 00:10:41,880
Speaker 4: It's a great question, you know. A first observation I

196
00:10:41,920 --> 00:10:46,080
Speaker 4: would make on your first two categories is that it

197
00:10:46,160 --> 00:10:49,680
Speaker 4: will be both likely be true that the use of

198
00:10:49,720 --> 00:10:52,280
Speaker 4: AI will be highly democratized, meaning the number of people

199
00:10:52,280 --> 00:10:55,400
Speaker 4: that have access to its power to make improvements in

200
00:10:55,440 --> 00:10:58,239
Speaker 4: terms of efficiency and so on will be fairly universal,

201
00:10:59,040 --> 00:11:02,400
Speaker 4: and that the ones who are able to create AI

202
00:11:03,200 --> 00:11:06,400
Speaker 4: uh may be quite concentrated. So if you look at

203
00:11:06,400 --> 00:11:10,000
Speaker 4: it from the lens of who creates wealth and value

204
00:11:10,400 --> 00:11:14,400
Speaker 4: over sustained periods of time, particularly it's saying a context

205
00:11:14,440 --> 00:11:17,920
Speaker 4: like business, I think just being a user of AI

206
00:11:18,040 --> 00:11:22,280
Speaker 4: technology is an insufficient strategy and UH. And the reason

207
00:11:22,360 --> 00:11:24,200
Speaker 4: for that is like, yes, you will get the immediate

208
00:11:24,240 --> 00:11:27,200
Speaker 4: productivity boost of like just making API calls, and you

209
00:11:27,200 --> 00:11:29,840
Speaker 4: know that would be a new baseline for everybody, but

210
00:11:30,280 --> 00:11:34,040
Speaker 4: you're not accruing value in terms of representing your data

211
00:11:34,160 --> 00:11:36,640
Speaker 4: inside the AI in way that gives you a sustainable

212
00:11:36,640 --> 00:11:40,000
Speaker 4: competitive advantage. So I always try to tell people is

213
00:11:40,040 --> 00:11:42,080
Speaker 4: don't just be an AI user. We an you know,

214
00:11:42,120 --> 00:11:45,640
Speaker 4: AI value creator. And I think that that will have

215
00:11:45,720 --> 00:11:48,560
Speaker 4: a lot of consequences in terms of the haves and

216
00:11:48,640 --> 00:11:51,160
Speaker 4: have nots as an example, and that will apply both

217
00:11:51,160 --> 00:11:54,960
Speaker 4: to institutions and regions and countries, et cetera. So I

218
00:11:54,960 --> 00:11:57,480
Speaker 4: think it would be kind of a mistake, right to

219
00:11:57,760 --> 00:12:00,280
Speaker 4: just develop strategies that are just about.

220
00:12:00,160 --> 00:12:04,240
Speaker 3: Usage, but to to come back that question from them

221
00:12:04,440 --> 00:12:08,240
Speaker 3: to give you a specific suppose I'm a I'm an

222
00:12:08,280 --> 00:12:12,760
Speaker 3: industrial farmer in Iowa with ten million dollars of equipment

223
00:12:12,840 --> 00:12:16,400
Speaker 3: and move and I'm comparing it to a subsistence farmer

224
00:12:16,840 --> 00:12:20,160
Speaker 3: someone in the developing world who's got a cell phone.

225
00:12:20,920 --> 00:12:25,640
Speaker 3: Over the next five years, who's who's well being rises

226
00:12:25,720 --> 00:12:27,120
Speaker 3: by a greater amount.

227
00:12:28,160 --> 00:12:30,880
Speaker 4: Yeah, I think, I mean, it's a it's a good question,

228
00:12:30,960 --> 00:12:32,559
Speaker 4: but it might be hard to do a one to

229
00:12:32,679 --> 00:12:35,280
Speaker 4: one sort of like attribution to just one variable in

230
00:12:35,280 --> 00:12:39,520
Speaker 4: this case, which is AI. But again, provided that you

231
00:12:39,600 --> 00:12:42,440
Speaker 4: have access to a phone, right and some kind to

232
00:12:42,679 --> 00:12:45,120
Speaker 4: you know, be able to be connected. I do think

233
00:12:45,200 --> 00:12:48,200
Speaker 4: so for example, in that context we've developed we don't

234
00:12:48,200 --> 00:12:50,720
Speaker 4: work with NASA as an example, to build your spatial

235
00:12:50,760 --> 00:12:54,200
Speaker 4: models using some of these new techniques, and I think

236
00:12:54,240 --> 00:12:57,160
Speaker 4: for example, or ability to do flood prediction, I'll tell

237
00:12:57,160 --> 00:12:59,880
Speaker 4: you an advantage of why would be a democratization force

238
00:13:00,080 --> 00:13:03,680
Speaker 4: that context. Before to build a flowed model based on

239
00:13:03,760 --> 00:13:08,040
Speaker 4: satellite imagery was actually so onerous and so complicated and

240
00:13:08,080 --> 00:13:10,680
Speaker 4: difficult that you would just target to very specific regions

241
00:13:10,920 --> 00:13:14,080
Speaker 4: and then obviously countries prioritize their own right. But what

242
00:13:14,120 --> 00:13:16,960
Speaker 4: we've demonstrated is actually you can extend the technique to

243
00:13:17,000 --> 00:13:19,800
Speaker 4: have like global coverage around that. So in that context,

244
00:13:19,840 --> 00:13:21,960
Speaker 4: I would say it's a four stores and markeratization that

245
00:13:22,000 --> 00:13:25,439
Speaker 4: everybody sort of would have access if you have some connectivity,

246
00:13:25,640 --> 00:13:26,359
Speaker 4: as today.

247
00:13:26,080 --> 00:13:29,520
Speaker 3: Iowa farmer might have a flood model. The guy in

248
00:13:29,559 --> 00:13:31,800
Speaker 3: the developing world definitely didn't, and now he's a shot

249
00:13:31,800 --> 00:13:32,280
Speaker 3: of getting one.

250
00:13:32,320 --> 00:13:33,800
Speaker 4: Yeah, but now it has a shot of getting one.

251
00:13:33,840 --> 00:13:35,640
Speaker 4: So there's aspects of it that so long as we

252
00:13:35,720 --> 00:13:38,680
Speaker 4: provide connectivity and access to it, that they can be

253
00:13:38,840 --> 00:13:41,840
Speaker 4: democratization forces. But I'll give you another example that that

254
00:13:41,840 --> 00:13:45,000
Speaker 4: can be quite concerning, which is language. Right, So there's

255
00:13:45,160 --> 00:13:49,480
Speaker 4: so much language in the you know, in English, and

256
00:13:49,960 --> 00:13:53,000
Speaker 4: there is sort of like this reinforcement loop that happens

257
00:13:53,000 --> 00:13:55,480
Speaker 4: that the more you concentrate, because it has obvious benefits

258
00:13:55,480 --> 00:13:59,120
Speaker 4: for global communication and standardization, the more you can enrich

259
00:13:59,280 --> 00:14:02,400
Speaker 4: like base aim models based on that capability. If you

260
00:14:02,520 --> 00:14:06,720
Speaker 4: have very resource cars languages, you tend to develop less

261
00:14:06,760 --> 00:14:10,320
Speaker 4: powerful AI with those languages and so on. So one

262
00:14:10,400 --> 00:14:14,600
Speaker 4: has to actually worry and focus on the ability to

263
00:14:14,760 --> 00:14:18,120
Speaker 4: actually represent you know that in that case is language

264
00:14:18,120 --> 00:14:20,920
Speaker 4: as a piece of culture. Also in the AI sets

265
00:14:20,960 --> 00:14:24,000
Speaker 4: that everybody can benefit from it too. So there's a

266
00:14:24,120 --> 00:14:27,640
Speaker 4: lot of considerations in terms of equity about the data

267
00:14:27,680 --> 00:14:30,840
Speaker 4: and the data sets that we accrue and what problems

268
00:14:30,880 --> 00:14:33,040
Speaker 4: are we trying to solve. I mean, you mentioned agriculture

269
00:14:33,120 --> 00:14:35,600
Speaker 4: or healthcare and so on. If we only solve problems

270
00:14:35,640 --> 00:14:38,320
Speaker 4: that are related to marketing as an example, that would

271
00:14:38,320 --> 00:14:40,880
Speaker 4: be a less rich world in terms of opportunity that

272
00:14:40,880 --> 00:14:43,760
Speaker 4: if we incorporate many many other broad set of problems.

273
00:14:44,800 --> 00:14:46,680
Speaker 3: Who do you think what do you think are the

274
00:14:46,720 --> 00:14:52,080
Speaker 3: biggest impediments to the adoption of AI as you would

275
00:14:52,200 --> 00:14:54,600
Speaker 3: like as you think AIR to be adopted. I mean,

276
00:14:54,640 --> 00:14:57,119
Speaker 3: if you look, what are the sticking points that you would.

277
00:14:57,960 --> 00:15:00,480
Speaker 4: Look Indiana, I'm going to give a non time technological

278
00:15:00,520 --> 00:15:03,080
Speaker 4: answer as a first one has to do with workflow, right,

279
00:15:03,320 --> 00:15:07,920
Speaker 4: So even if the technology is very capable, the organizational

280
00:15:08,040 --> 00:15:11,280
Speaker 4: change inside a company to incorporate into the natural workflow

281
00:15:11,320 --> 00:15:14,840
Speaker 4: of people and how we work is it's a lesson

282
00:15:14,880 --> 00:15:17,400
Speaker 4: we have learned over the last decade is hugely important.

283
00:15:18,040 --> 00:15:21,720
Speaker 4: So there's a lot of design considerations. There's a lot

284
00:15:21,760 --> 00:15:24,560
Speaker 4: of how do people want to work right? How do

285
00:15:24,640 --> 00:15:27,000
Speaker 4: they work today? And what is the natural entry points

286
00:15:27,000 --> 00:15:29,600
Speaker 4: for AI? So that's like number one, and then the

287
00:15:29,680 --> 00:15:33,600
Speaker 4: second one is, you know, for the broad value creation

288
00:15:33,720 --> 00:15:37,720
Speaker 4: aspect of it is the understanding inside the companies of

289
00:15:37,960 --> 00:15:42,240
Speaker 4: how you have to curate and create data to combine

290
00:15:42,240 --> 00:15:45,120
Speaker 4: it with external data says that you can have powerful

291
00:15:45,160 --> 00:15:48,800
Speaker 4: AI models that actually fit your need. And that aspect

292
00:15:48,880 --> 00:15:51,920
Speaker 4: of what it takes to actually create and curate the

293
00:15:52,040 --> 00:15:56,920
Speaker 4: data for this modern AI, it's still working progress, right.

294
00:15:57,240 --> 00:15:59,680
Speaker 4: I think part of the problem that happens very often

295
00:15:59,720 --> 00:16:03,040
Speaker 4: when I talk to institutions is that they say yea yeah, yah, yeah,

296
00:16:03,040 --> 00:16:05,600
Speaker 4: I'm doing it. I've been doing it for a long time.

297
00:16:06,360 --> 00:16:08,760
Speaker 4: And the reality is that that answer can sometimes be

298
00:16:08,760 --> 00:16:10,840
Speaker 4: a little of our cop out, right, is like I

299
00:16:10,880 --> 00:16:13,480
Speaker 4: know you were doing machine learning, you were doing some

300
00:16:13,560 --> 00:16:16,440
Speaker 4: of these things, but actually the leader's version of AI,

301
00:16:16,560 --> 00:16:19,280
Speaker 4: what what's happening with foundation models? Not only is it

302
00:16:19,400 --> 00:16:22,800
Speaker 4: very new. It's very hard to do, and honestly, if

303
00:16:22,800 --> 00:16:25,560
Speaker 4: you haven't been you know, assembling very large teams and

304
00:16:25,560 --> 00:16:28,080
Speaker 4: spending hundreds of millions of dollars of compute, and so

305
00:16:28,240 --> 00:16:31,000
Speaker 4: you're probably not doing it right. You're doing something else

306
00:16:31,040 --> 00:16:33,680
Speaker 4: that is in the broad category. And I think the

307
00:16:33,800 --> 00:16:36,600
Speaker 4: lessons about what it means to make this transition to

308
00:16:36,640 --> 00:16:39,479
Speaker 4: this new wave is still in early phases of understanding.

309
00:16:39,680 --> 00:16:41,280
Speaker 3: So what would you say? I want to give you

310
00:16:41,280 --> 00:16:44,800
Speaker 3: a couple of examples of people with kind of real

311
00:16:44,840 --> 00:16:47,880
Speaker 3: world in real world positions of responsibility. Imagine I'm sitting

312
00:16:47,960 --> 00:16:50,840
Speaker 3: right here, So imagine that I am the president of

313
00:16:50,840 --> 00:16:53,160
Speaker 3: a small liberal arts college and I come to you

314
00:16:53,200 --> 00:16:55,200
Speaker 3: and I say, Dario, I keep hearing about a AI

315
00:16:56,000 --> 00:16:58,800
Speaker 3: my college has you know I don't make it. You know,

316
00:16:58,880 --> 00:17:01,840
Speaker 3: I'm I'm not I'm making this much money. If that

317
00:17:01,960 --> 00:17:06,399
Speaker 3: every year enroments declining, I feel like this maybe is

318
00:17:06,440 --> 00:17:09,760
Speaker 3: an opportunity. What is the opportunity for me? What would

319
00:17:09,800 --> 00:17:10,600
Speaker 3: you say?

320
00:17:11,720 --> 00:17:14,040
Speaker 4: So, it's probably in a couple of segments around that.

321
00:17:14,280 --> 00:17:17,720
Speaker 4: Right one has to do is well, what is the

322
00:17:17,720 --> 00:17:21,919
Speaker 4: implications of this technology inside the institution itself instead of

323
00:17:22,000 --> 00:17:25,639
Speaker 4: the college, And how we operate and can we improve

324
00:17:25,680 --> 00:17:28,480
Speaker 4: for example, efficiency, like if you have in very low

325
00:17:28,880 --> 00:17:31,919
Speaker 4: levels of sort of margin to be able to reinvest

326
00:17:32,520 --> 00:17:37,119
Speaker 4: is you know you run it, you run you know infrastructure,

327
00:17:37,200 --> 00:17:39,240
Speaker 4: you run many things inside the college. What are the

328
00:17:39,280 --> 00:17:43,560
Speaker 4: opportunities to increase the productivity or automate and drive savings

329
00:17:43,560 --> 00:17:46,200
Speaker 4: such that you can reinvest that money into the mission

330
00:17:46,200 --> 00:17:46,679
Speaker 4: of education?

331
00:17:46,840 --> 00:17:49,320
Speaker 3: Right as an example, So number one is operational efficiency.

332
00:17:49,320 --> 00:17:52,760
Speaker 4: Operational efficiency is a big one. I think the second

333
00:17:52,800 --> 00:17:55,240
Speaker 4: one is within the context of the college, there's implications

334
00:17:55,240 --> 00:17:58,000
Speaker 4: for the educational mission on its own, right, How will

335
00:17:58,359 --> 00:18:01,000
Speaker 4: you know how does a correct need to evolve or not?

336
00:18:01,440 --> 00:18:04,160
Speaker 4: What are acceptable use policies or of someone these ai

337
00:18:04,280 --> 00:18:06,320
Speaker 4: I think we've all read a lot about like what

338
00:18:06,359 --> 00:18:08,760
Speaker 4: can happen in terms of exams and so on and

339
00:18:08,800 --> 00:18:11,200
Speaker 4: cheating and not cheating, or what are the actually positive

340
00:18:11,240 --> 00:18:13,600
Speaker 4: elements of it in terms of how curriculum should be

341
00:18:13,640 --> 00:18:17,280
Speaker 4: developed and professions sustain around that. And then there's another

342
00:18:17,280 --> 00:18:19,960
Speaker 4: third dimension, which is the outdoor oriented element of it,

343
00:18:20,000 --> 00:18:23,879
Speaker 4: which is like prospect students right, so, which is frankly speaking,

344
00:18:23,920 --> 00:18:26,040
Speaker 4: a big use case that is happening right now, which

345
00:18:26,080 --> 00:18:28,560
Speaker 4: in the broader industry is called customer care or client

346
00:18:28,560 --> 00:18:30,959
Speaker 4: care or citizen care. So in this question will be education,

347
00:18:31,080 --> 00:18:33,680
Speaker 4: like you know, hey, are you reaching the right students

348
00:18:34,359 --> 00:18:37,320
Speaker 4: around that that may apply to the college. How can

349
00:18:37,320 --> 00:18:39,760
Speaker 4: you create them? For example, an environment to interact with

350
00:18:39,760 --> 00:18:41,800
Speaker 4: the college and answering questions that could be a chat

351
00:18:41,840 --> 00:18:44,879
Speaker 4: bought or something like that to learn about it. And personalization.

352
00:18:45,280 --> 00:18:47,720
Speaker 4: So I would say there's like at least three lenses

353
00:18:47,800 --> 00:18:49,399
Speaker 4: with which I would give advice, right.

354
00:18:49,320 --> 00:18:54,960
Speaker 3: The positive seglee because it's really interesting. So I really

355
00:18:54,960 --> 00:18:58,800
Speaker 3: can't as sign an essay anymore? Can I?

356
00:18:58,840 --> 00:19:00,119
Speaker 4: Can I sign an essay?

357
00:19:00,280 --> 00:19:03,120
Speaker 3: Can I say? Rend me a research paper? And come

358
00:19:03,119 --> 00:19:03,760
Speaker 3: back to being three?

359
00:19:03,760 --> 00:19:03,840
Speaker 4: We?

360
00:19:03,920 --> 00:19:04,920
Speaker 3: Can I do that anymore?

361
00:19:05,000 --> 00:19:05,600
Speaker 4: I think you can?

362
00:19:05,720 --> 00:19:06,520
Speaker 3: How do I do that?

363
00:19:06,640 --> 00:19:10,280
Speaker 4: And then you can that Look, there's there's two questions

364
00:19:10,280 --> 00:19:14,440
Speaker 4: around that. I think that if one goes and explains

365
00:19:14,440 --> 00:19:16,240
Speaker 4: in the context like what is it? Why are we here?

366
00:19:16,280 --> 00:19:18,280
Speaker 4: Why are in this class? What is the purpose of this?

367
00:19:19,160 --> 00:19:22,679
Speaker 4: And and one starts with assuming like an element of

368
00:19:22,720 --> 00:19:24,840
Speaker 4: like decency and people are people are there like to

369
00:19:24,920 --> 00:19:27,280
Speaker 4: learn and so on, and you just give it this disclaimer. Look,

370
00:19:27,520 --> 00:19:30,040
Speaker 4: I know that one option you have is like just

371
00:19:30,119 --> 00:19:32,280
Speaker 4: you know, put the essay question and click, go and

372
00:19:32,320 --> 00:19:34,280
Speaker 4: like and give an answer. You know, but that is

373
00:19:34,359 --> 00:19:36,440
Speaker 4: not why we're here, and that is not the intent

374
00:19:36,480 --> 00:19:38,040
Speaker 4: of what we're trying to do. So first I would

375
00:19:38,080 --> 00:19:42,400
Speaker 4: start with the sort of like the norms of intent

376
00:19:42,560 --> 00:19:45,560
Speaker 4: and decency and appeal to those as step number one.

377
00:19:46,200 --> 00:19:48,359
Speaker 4: Then we all know that there will be a distribution

378
00:19:48,400 --> 00:19:50,439
Speaker 4: of use cases of people like that will come in

379
00:19:50,480 --> 00:19:52,320
Speaker 4: one year and come out of the other and do that.

380
00:19:52,880 --> 00:19:55,359
Speaker 4: And so for a subset of that, you know, I

381
00:19:55,400 --> 00:19:57,320
Speaker 4: think the technology is going to have all in such

382
00:19:57,359 --> 00:19:59,720
Speaker 4: a way that we will have more and more of

383
00:19:59,760 --> 00:20:02,719
Speaker 4: the to discern, right, you know, when that has been

384
00:20:02,760 --> 00:20:06,880
Speaker 4: AI generated right and uncreated, it won't be perfect, right,

385
00:20:06,960 --> 00:20:09,480
Speaker 4: But there's some elements that you can imagine in putting

386
00:20:09,480 --> 00:20:11,520
Speaker 4: the essay and you say, hey, this is likely to

387
00:20:11,520 --> 00:20:14,680
Speaker 4: be generated right around that. And for example, one way

388
00:20:14,680 --> 00:20:16,080
Speaker 4: you can do that, just to give you an intuition,

389
00:20:16,160 --> 00:20:18,480
Speaker 4: you could just have an essay that you write with

390
00:20:18,600 --> 00:20:21,920
Speaker 4: pencil and paper. At the beginning, you get a baseline

391
00:20:21,960 --> 00:20:24,600
Speaker 4: of what you're writing is like, and then later when

392
00:20:24,640 --> 00:20:28,040
Speaker 4: you you know generate it, there will be obvious differences

393
00:20:28,160 --> 00:20:30,560
Speaker 4: around what kind of writing has been generating on the

394
00:20:30,560 --> 00:20:30,880
Speaker 4: other way.

395
00:20:30,880 --> 00:20:34,480
Speaker 3: But you've turned it's everything you're describing makes sense put

396
00:20:34,520 --> 00:20:38,399
Speaker 3: it greatly in this respect, at least, it seems to

397
00:20:38,440 --> 00:20:41,400
Speaker 3: greatly complicate the life of the teacher, whereas the other

398
00:20:41,440 --> 00:20:46,199
Speaker 3: two use cases seem to kind of clarify and simplify

399
00:20:46,280 --> 00:20:51,119
Speaker 3: the role. Right suddenly, you know, reaching student perspective students,

400
00:20:51,160 --> 00:20:53,119
Speaker 3: sounds like I can do that much more kind of

401
00:20:53,160 --> 00:20:55,840
Speaker 3: efficient in a lite. Yeah, I can bring you administration costs,

402
00:20:55,880 --> 00:20:57,840
Speaker 3: but the teaching thing is tricky.

403
00:20:58,960 --> 00:21:02,720
Speaker 4: Well, until we developed the new norms, right, I mean again,

404
00:21:02,880 --> 00:21:05,160
Speaker 4: I mean, I know it's not abuse analogy, but calculators

405
00:21:05,200 --> 00:21:08,760
Speaker 4: we deal. We've done with that too, right, And it says, well, calculator,

406
00:21:08,800 --> 00:21:10,199
Speaker 4: what is the purpose of math? How are we going

407
00:21:10,280 --> 00:21:10,600
Speaker 4: to do this?

408
00:21:10,680 --> 00:21:13,920
Speaker 3: And so can I tell you my dad's calculator story?

409
00:21:14,040 --> 00:21:14,639
Speaker 4: Yes? Please.

410
00:21:14,880 --> 00:21:18,879
Speaker 3: My father was a mathematician, taught mathematics at University of Waterloo, Agada,

411
00:21:19,560 --> 00:21:22,879
Speaker 3: and in the seventies when people started to get pocket calculators,

412
00:21:23,359 --> 00:21:25,920
Speaker 3: his students demanded that they'd be able to use them,

413
00:21:26,119 --> 00:21:27,879
Speaker 3: and he said no, and they took him to the

414
00:21:27,880 --> 00:21:33,720
Speaker 3: administration and he lost. So he then changed completely throughout

415
00:21:33,720 --> 00:21:36,800
Speaker 3: all of his old exams and introduced new exams where

416
00:21:37,160 --> 00:21:41,400
Speaker 3: there was no calculation. It was all like deep think,

417
00:21:41,520 --> 00:21:43,840
Speaker 3: you know, figure out the problem on a conceptual level

418
00:21:43,840 --> 00:21:47,120
Speaker 3: and describe it to me. And they were all students

419
00:21:47,160 --> 00:21:49,440
Speaker 3: deeply unhappy that he'd made their lives for computation.

420
00:21:49,800 --> 00:21:51,320
Speaker 4: But it's to other.

421
00:21:51,280 --> 00:21:54,760
Speaker 3: Point, to your point, I mean, he probably the result

422
00:21:54,840 --> 00:21:58,919
Speaker 3: was probably a better education. He just removed the element

423
00:21:58,960 --> 00:22:01,840
Speaker 3: that they could gain with their pocket calculators. I suppose

424
00:22:01,880 --> 00:22:02,600
Speaker 3: it's a version of.

425
00:22:02,720 --> 00:22:04,359
Speaker 4: I think it's a version of that, And so I

426
00:22:04,359 --> 00:22:06,480
Speaker 4: think they will develop the equivalent of what your father did.

427
00:22:06,480 --> 00:22:07,960
Speaker 4: And I think people say, you know what if like

428
00:22:08,040 --> 00:22:10,360
Speaker 4: these kinds of things, everybody's doing it generically and none

429
00:22:10,400 --> 00:22:12,600
Speaker 4: of us have any meaning because all you're doing is

430
00:22:12,640 --> 00:22:14,800
Speaker 4: pressing buttons, and like the intent of this was something

431
00:22:14,800 --> 00:22:16,439
Speaker 4: which was to teach you how to write or to

432
00:22:16,520 --> 00:22:18,760
Speaker 4: think or something. There may be a variant of how

433
00:22:18,800 --> 00:22:20,920
Speaker 4: we do all of this. I mean, obviously some version

434
00:22:20,960 --> 00:22:22,960
Speaker 4: of that that has happened is like Okay, we're all

435
00:22:22,960 --> 00:22:24,600
Speaker 4: going to sit down and doing with pencil on paper

436
00:22:24,640 --> 00:22:27,359
Speaker 4: and computers in their classroom. But there'll be other variants

437
00:22:27,400 --> 00:22:29,879
Speaker 4: of creativity that people will put forth to say, you

438
00:22:29,880 --> 00:22:31,919
Speaker 4: know what, you know, that's a way to solve that

439
00:22:31,960 --> 00:22:32,480
Speaker 4: problem too.

440
00:22:32,560 --> 00:22:35,600
Speaker 3: But this is interesting because to stay on this analogy,

441
00:22:36,240 --> 00:22:41,639
Speaker 3: we're really talking about a profound rethinking. Just using college

442
00:22:41,640 --> 00:22:45,400
Speaker 3: as an example, a real profound rethinking of the way,

443
00:22:46,000 --> 00:22:49,760
Speaker 3: there's no part of this college it's unaffected by aia B.

444
00:22:50,760 --> 00:22:53,880
Speaker 3: In one case, I've made everyone's job easier. In one case,

445
00:22:53,920 --> 00:22:57,000
Speaker 3: I've made I'm asking us to really rethink from the

446
00:22:57,000 --> 00:23:01,360
Speaker 3: ground up what teaching means. In another case, I've automated

447
00:23:01,359 --> 00:23:03,040
Speaker 3: systems that I didn't think of it. I mean, it's like,

448
00:23:03,359 --> 00:23:06,280
Speaker 3: that's right, that's it's not that's a lot to ask

449
00:23:06,359 --> 00:23:10,000
Speaker 3: someone who got a PhD in medieval language literature, you know,

450
00:23:10,119 --> 00:23:10,920
Speaker 3: forty years ago.

451
00:23:11,520 --> 00:23:13,720
Speaker 4: Yeah. But you know, I'll tell you a positive sort

452
00:23:13,760 --> 00:23:16,399
Speaker 4: of development that I'm seeing the sciences around this, which

453
00:23:16,480 --> 00:23:19,840
Speaker 4: is you're seeing as you see more and more examples

454
00:23:20,160 --> 00:23:24,400
Speaker 4: of applying AI technology within the context of like historians

455
00:23:24,400 --> 00:23:27,879
Speaker 4: to as an example, right, you have archival and you know,

456
00:23:27,920 --> 00:23:29,560
Speaker 4: and you have all these books and being able to

457
00:23:29,640 --> 00:23:32,719
Speaker 4: actually help you as an assistant right around that, but

458
00:23:32,760 --> 00:23:35,960
Speaker 4: not only with text now, but with diagrams, right, And

459
00:23:36,320 --> 00:23:40,080
Speaker 4: I've seen it in anthropology too, Rite and archaeology with

460
00:23:40,200 --> 00:23:43,439
Speaker 4: examples of engravings and translations and things that can happen.

461
00:23:43,800 --> 00:23:47,360
Speaker 4: So so as you see in diverse fields people applying

462
00:23:47,400 --> 00:23:49,920
Speaker 4: these techniques to advance and how to do physics or

463
00:23:49,920 --> 00:23:53,239
Speaker 4: how to do chemistry, they inspire each other, right, and

464
00:23:53,240 --> 00:23:55,080
Speaker 4: they said, you know, how does it apply actually to

465
00:23:55,160 --> 00:23:58,760
Speaker 4: my area? So once as that happens, it becomes less

466
00:23:58,760 --> 00:24:00,320
Speaker 4: of a chore of like my god, you know, how

467
00:24:00,359 --> 00:24:02,439
Speaker 4: do I have to deal with this? But actually it's

468
00:24:02,440 --> 00:24:06,159
Speaker 4: triggered by curiosity, is triggered by you know, they'll be like,

469
00:24:06,280 --> 00:24:07,960
Speaker 4: you know, faculty that will be like you know what,

470
00:24:08,040 --> 00:24:10,240
Speaker 4: you know, let me explore what this means for my area,

471
00:24:10,640 --> 00:24:13,000
Speaker 4: and they will adapt it to the local context, to

472
00:24:13,040 --> 00:24:16,600
Speaker 4: the local you know, language, and the professional itself. So

473
00:24:16,840 --> 00:24:19,679
Speaker 4: I see that as a positive vector that is not

474
00:24:19,720 --> 00:24:21,639
Speaker 4: all going to feel like homework, you know, it's not

475
00:24:21,800 --> 00:24:24,000
Speaker 4: going to feel like, oh my god, this is so overwhelming,

476
00:24:24,280 --> 00:24:26,639
Speaker 4: but rather to be very practical to see what works,

477
00:24:26,760 --> 00:24:28,920
Speaker 4: What have I seen others to do that is inspiring,

478
00:24:29,080 --> 00:24:31,120
Speaker 4: and what am I inspired to do? You know what?

479
00:24:31,119 --> 00:24:33,080
Speaker 4: What is how is this going to help my career?

480
00:24:33,400 --> 00:24:35,119
Speaker 4: I think that that's going to be an interesting question

481
00:24:35,280 --> 00:24:37,399
Speaker 4: for for you know, those faculty members for the.

482
00:24:37,400 --> 00:24:40,639
Speaker 3: Students, the professional Sorry, I'm gonna stick with this example

483
00:24:40,640 --> 00:24:43,399
Speaker 3: along because it's really interesting. I'm curious, following up on

484
00:24:43,440 --> 00:24:46,240
Speaker 3: what you just said, that one of the most persistent

485
00:24:46,440 --> 00:24:50,679
Speaker 3: critiques of academic but also of many of many corporate

486
00:24:50,720 --> 00:24:56,240
Speaker 3: institutions in these years, has been siloing. Right, Different parts

487
00:24:56,280 --> 00:24:59,520
Speaker 3: of the of the organization are going off on their own,

488
00:24:59,560 --> 00:25:03,520
Speaker 3: and that's to each other. Is a potential is a

489
00:25:03,680 --> 00:25:08,880
Speaker 3: real potential benefit to AI the kind of breaking down

490
00:25:09,440 --> 00:25:12,479
Speaker 3: a simple tool for breaking down those kinds of barriers.

491
00:25:12,560 --> 00:25:14,439
Speaker 3: Is that a very Is that an elegant way of

492
00:25:14,440 --> 00:25:16,359
Speaker 3: sort of saying what.

493
00:25:15,760 --> 00:25:17,679
Speaker 4: I really think? And I was actually just having a

494
00:25:17,680 --> 00:25:20,720
Speaker 4: conversation with Provos stuff and very much on this topic,

495
00:25:20,840 --> 00:25:24,600
Speaker 4: very recently, exactly on that which is all these this,

496
00:25:24,800 --> 00:25:27,600
Speaker 4: you know, this appetite right to collaborate across disciplines. There's

497
00:25:27,600 --> 00:25:31,840
Speaker 4: a lot of attempts stores a goal, right, creating interdisciplinary centers,

498
00:25:32,400 --> 00:25:36,520
Speaker 4: creating dual degree programs or dual appointment programs. But actually

499
00:25:37,080 --> 00:25:41,200
Speaker 4: in a lot of progress in academia happens by methodology too.

500
00:25:41,440 --> 00:25:44,040
Speaker 4: Write like a new you know, when when some methodology

501
00:25:44,040 --> 00:25:46,960
Speaker 4: gets adopted, I mean the most famous example of that

502
00:25:47,040 --> 00:25:49,399
Speaker 4: is a scientific method as an example of that. But

503
00:25:49,480 --> 00:25:52,320
Speaker 4: when you have a methodology that gets adopted, it also

504
00:25:52,359 --> 00:25:56,639
Speaker 4: provides a way to speak to your colleagues across different disciplines,

505
00:25:56,880 --> 00:25:59,399
Speaker 4: and I think what's happened in AI is linked to

506
00:25:59,440 --> 00:26:02,560
Speaker 4: that that within the context of the scientific method as

507
00:26:02,560 --> 00:26:08,040
Speaker 4: an example, the methodology about we about what we do discovery,

508
00:26:08,400 --> 00:26:11,119
Speaker 4: the role of data, the role of these neural networks,

509
00:26:11,119 --> 00:26:13,760
Speaker 4: of how we actually find proximity to concepts to one

510
00:26:13,760 --> 00:26:19,320
Speaker 4: another is actually fundamentally different than how we've traditionally applied it.

511
00:26:19,720 --> 00:26:23,040
Speaker 4: So as we see across more professions, people applying this

512
00:26:23,160 --> 00:26:26,520
Speaker 4: methodology is also going to give some element of common

513
00:26:26,600 --> 00:26:29,960
Speaker 4: language to each other. Right And in fact, you know,

514
00:26:30,200 --> 00:26:33,480
Speaker 4: in this very high dimensional representation of information that is

515
00:26:33,520 --> 00:26:37,520
Speaker 4: pressent to neural networks, we may find amazing adjacencies or

516
00:26:37,560 --> 00:26:41,280
Speaker 4: connections of themes and topics in ways that the individual

517
00:26:41,359 --> 00:26:45,080
Speaker 4: practitioners cannot describe, but yet will be latent in these

518
00:26:45,160 --> 00:26:48,280
Speaker 4: large cal neural networks. We are going to suffer a

519
00:26:48,320 --> 00:26:50,880
Speaker 4: little bit from causality, from the problem of like, hey,

520
00:26:50,880 --> 00:26:53,720
Speaker 4: what's the root cause of that? Because I think one

521
00:26:53,760 --> 00:26:57,959
Speaker 4: of the unsatisfying aspects that this methodology will provide is

522
00:26:58,040 --> 00:26:59,960
Speaker 4: they may give you answers for which they don't give

523
00:26:59,960 --> 00:27:04,120
Speaker 4: you good reasons for where the answers came from and uh,

524
00:27:04,160 --> 00:27:06,679
Speaker 4: and then there will be the traditional process of discovery

525
00:27:06,760 --> 00:27:09,320
Speaker 4: of saying, if that is the answer, what are the reasons?

526
00:27:09,760 --> 00:27:11,600
Speaker 4: So we're going to have to do this sort of

527
00:27:11,720 --> 00:27:15,360
Speaker 4: hybrid way of understanding the world. But I do think

528
00:27:15,400 --> 00:27:18,080
Speaker 4: that common layer of AI is a powerful new thing.

529
00:27:18,400 --> 00:27:21,520
Speaker 3: Yeah. Well, a couple of random questions. I couldn't mind

530
00:27:21,520 --> 00:27:24,439
Speaker 3: as you talk. In the In the Writer's strike that

531
00:27:24,560 --> 00:27:27,000
Speaker 3: just ended in Hollywood, one of the sticking points was

532
00:27:27,560 --> 00:27:31,639
Speaker 3: how the studios and writers would treat AI generated content. Right,

533
00:27:31,800 --> 00:27:35,480
Speaker 3: good writers get credit if their material with somehow the

534
00:27:35,560 --> 00:27:40,119
Speaker 3: source for a but more broadly, did the writers need

535
00:27:40,160 --> 00:27:42,520
Speaker 3: protections against the use of I could go on. You

536
00:27:42,520 --> 00:27:45,160
Speaker 3: know what, probably we're familiar with all of this. Had

537
00:27:45,200 --> 00:27:47,440
Speaker 3: you been I don't know whether you were, but had

538
00:27:47,720 --> 00:27:52,080
Speaker 3: either side called you in for advice during that the writers,

539
00:27:52,080 --> 00:27:54,720
Speaker 3: had the writers called you and said, Daria, what should

540
00:27:54,760 --> 00:27:57,520
Speaker 3: we do about AI? And how should we that should

541
00:27:57,520 --> 00:27:59,960
Speaker 3: be ref how should that be reflected in our content

542
00:28:00,000 --> 00:28:01,760
Speaker 3: track negotiations? What would you've told.

543
00:28:01,600 --> 00:28:06,000
Speaker 4: Them the way I think about that is that I died.

544
00:28:06,119 --> 00:28:08,360
Speaker 4: I would divide it into two pieces. First, is what's

545
00:28:08,400 --> 00:28:13,159
Speaker 4: technically possible, right, and anticipate scenarios like you know, what

546
00:28:13,240 --> 00:28:16,120
Speaker 4: can you do with voice cloning for example? You know, now,

547
00:28:16,160 --> 00:28:20,879
Speaker 4: for example it is possible there being dubbing right legist

548
00:28:20,880 --> 00:28:22,800
Speaker 4: take that topic right around the world. There was all

549
00:28:22,840 --> 00:28:26,560
Speaker 4: these folks that would dub people in other languages. Well,

550
00:28:26,600 --> 00:28:29,520
Speaker 4: now you can do these incredible rendering in some I mean,

551
00:28:29,520 --> 00:28:31,600
Speaker 4: I know if you've seen them, where you know you

552
00:28:31,720 --> 00:28:34,520
Speaker 4: match the lips is your original voice, but speaking any

553
00:28:34,600 --> 00:28:37,000
Speaker 4: language that you want. As an example, so busy that

554
00:28:37,080 --> 00:28:38,880
Speaker 4: has a set of implications around that. I mean, just

555
00:28:38,880 --> 00:28:40,520
Speaker 4: to give an example, So I would say, create a

556
00:28:40,560 --> 00:28:45,080
Speaker 4: taxonomy that describes technical capabilities that we know of today

557
00:28:45,920 --> 00:28:49,960
Speaker 4: and applications to the industry, and two examples of like, hey,

558
00:28:50,000 --> 00:28:51,520
Speaker 4: you know I could film you for five minutes and

559
00:28:51,520 --> 00:28:53,480
Speaker 4: I could generate two hours of content of you and

560
00:28:53,520 --> 00:28:55,440
Speaker 4: I don't have to you know, then if you get

561
00:28:55,480 --> 00:28:57,440
Speaker 4: paid by the hour, obviously I'm not paying you for

562
00:28:57,480 --> 00:29:00,400
Speaker 4: that other thing. So I would say technological capability and

563
00:29:00,440 --> 00:29:03,920
Speaker 4: then map with their expertise consequences of how it changes

564
00:29:03,960 --> 00:29:06,320
Speaker 4: the way they work or the way they interact or

565
00:29:06,360 --> 00:29:08,360
Speaker 4: the way they negotiate and so on. So that would

566
00:29:08,360 --> 00:29:10,640
Speaker 4: be one element of it, and then the other one

567
00:29:10,760 --> 00:29:13,320
Speaker 4: is like a non technology related matter, which is an

568
00:29:13,320 --> 00:29:16,080
Speaker 4: element of almost of distributed justice, is like who deserves

569
00:29:16,120 --> 00:29:18,400
Speaker 4: what right and who has the power to get what?

570
00:29:19,280 --> 00:29:23,000
Speaker 4: And then that's a completely different discussion. That is to say, well,

571
00:29:23,040 --> 00:29:25,840
Speaker 4: if this is the scenario of what's possible, you know,

572
00:29:25,920 --> 00:29:28,920
Speaker 4: what do we want and what are we able to get?

573
00:29:29,080 --> 00:29:31,200
Speaker 4: And I think that that's a different discussion, which is

574
00:29:31,400 --> 00:29:32,320
Speaker 4: which we all as.

575
00:29:32,200 --> 00:29:33,320
Speaker 3: Life, which when do you do?

576
00:29:33,400 --> 00:29:38,320
Speaker 4: First? I think it's very helpful to have an understanding

577
00:29:38,600 --> 00:29:42,520
Speaker 4: of what's possible and how it changes a landscape as

578
00:29:42,600 --> 00:29:46,480
Speaker 4: part of a broader discussion, right, and a broad negotiation

579
00:29:47,080 --> 00:29:50,800
Speaker 4: because you also have to see the opportunities because there

580
00:29:50,800 --> 00:29:53,480
Speaker 4: will be a lot of ground to say, actually, you know,

581
00:29:53,880 --> 00:29:56,360
Speaker 4: if we can do it in this way and we

582
00:29:56,400 --> 00:29:58,760
Speaker 4: can all be that much more efficient in getting this

583
00:29:58,920 --> 00:30:02,640
Speaker 4: piece work done on this but we have a reasonable

584
00:30:02,720 --> 00:30:06,520
Speaker 4: agreement about how we both sides benefit from it, right,

585
00:30:07,280 --> 00:30:09,600
Speaker 4: then that's a win win for everybody. Yeah, Right, So

586
00:30:09,680 --> 00:30:12,640
Speaker 4: that's I think that would be a golden triangle, right.

587
00:30:12,560 --> 00:30:15,120
Speaker 3: Here's my reading, and I would like you to correct

588
00:30:15,120 --> 00:30:16,880
Speaker 3: me if I'm wrong, and I'm likely to be wrong.

589
00:30:18,000 --> 00:30:20,120
Speaker 3: When I looked at that strike, I said, if they're

590
00:30:20,120 --> 00:30:24,160
Speaker 3: worried about AI, the writers are worried about AI. That

591
00:30:24,280 --> 00:30:26,840
Speaker 3: seems silly. It should be the studios who are worried

592
00:30:26,840 --> 00:30:29,280
Speaker 3: about the economic impact of AI. Does it in the

593
00:30:29,320 --> 00:30:32,360
Speaker 3: long run AI put the studios out of business long

594
00:30:32,400 --> 00:30:34,320
Speaker 3: before it puts the writers out of business. I only

595
00:30:34,360 --> 00:30:38,360
Speaker 3: need the studio because the costs of production are as

596
00:30:38,440 --> 00:30:41,000
Speaker 3: high as the sky, and the cost of production are overwhelming.

597
00:30:41,040 --> 00:30:44,280
Speaker 3: And whereas if I don't, if I have a tool

598
00:30:44,320 --> 00:30:49,680
Speaker 3: which brings introduces massive technological efficiencies to the production of movies,

599
00:30:50,120 --> 00:30:52,600
Speaker 3: then I don't. Why don't need a studio? Why would

600
00:30:52,600 --> 00:30:53,440
Speaker 3: they the scared ones?

601
00:30:53,600 --> 00:30:55,640
Speaker 4: Or maybe maybe you need like a different kind of

602
00:30:55,680 --> 00:30:57,560
Speaker 4: studio or a different kind of different kind of study.

603
00:30:57,640 --> 00:31:01,320
Speaker 3: But I mean the in the but in in this strike,

604
00:31:01,400 --> 00:31:04,680
Speaker 3: the fright the frightened ones with the writers and the

605
00:31:05,200 --> 00:31:08,200
Speaker 3: you know, with the studios. Wasn't that backwards?

606
00:31:09,600 --> 00:31:11,960
Speaker 4: I haven't thought about it. Uh, it can be about

607
00:31:11,960 --> 00:31:13,800
Speaker 4: the implications of it. It goes back to we're talking

608
00:31:13,840 --> 00:31:16,960
Speaker 4: before the implications because are so horizontal. It is right

609
00:31:17,000 --> 00:31:18,520
Speaker 4: to think about it, like what does it do to

610
00:31:18,560 --> 00:31:19,680
Speaker 4: the studios as well? Right?

611
00:31:19,800 --> 00:31:21,000
Speaker 3: Yeah?

612
00:31:21,040 --> 00:31:23,680
Speaker 4: But then you know, the reason why that happens is

613
00:31:23,720 --> 00:31:28,160
Speaker 4: that it's the order of either negotiations or who first

614
00:31:28,160 --> 00:31:31,840
Speaker 4: got concerned about it and did something about it, right,

615
00:31:31,880 --> 00:31:34,680
Speaker 4: which is in the context of the strike. You know,

616
00:31:34,800 --> 00:31:37,160
Speaker 4: I don't know what the equivalent conversations are going inside

617
00:31:37,200 --> 00:31:39,120
Speaker 4: the studio and whether they have a workroom saying what

618
00:31:39,160 --> 00:31:41,160
Speaker 4: this is going to mean to us? Right, but it

619
00:31:41,200 --> 00:31:43,720
Speaker 4: doesn't get exercise through a strike, but maybe through a

620
00:31:43,840 --> 00:31:46,160
Speaker 4: task force inside you know, the companies about what are

621
00:31:46,160 --> 00:31:46,920
Speaker 4: they going to do? Right?

622
00:31:47,120 --> 00:31:49,360
Speaker 3: Well, And to go back to your thing you said,

623
00:31:49,400 --> 00:31:50,680
Speaker 3: the first thing you do is you make a list

624
00:31:50,680 --> 00:31:55,400
Speaker 3: of what technological capabilities are. But don't technological capabilities change

625
00:31:55,400 --> 00:31:58,960
Speaker 3: every I mean they do. You're raising ahead so fast,

626
00:31:59,160 --> 00:32:01,959
Speaker 3: so you can't. Can you you have a contract? I'm

627
00:32:01,960 --> 00:32:04,080
Speaker 3: sorry for getting into little weeds here, but this is interesting.

628
00:32:04,520 --> 00:32:07,360
Speaker 3: Can you you can't have a five year contract if

629
00:32:07,360 --> 00:32:11,080
Speaker 3: the contract is based on an assessment of technological capabilities

630
00:32:11,120 --> 00:32:13,840
Speaker 3: in twenty twenty three, because by the time get to

631
00:32:13,880 --> 00:32:18,680
Speaker 3: twenty eight, twenty three eight, it's totally.

632
00:32:18,320 --> 00:32:21,120
Speaker 4: Different, right, yeah, But like you know, I mean where

633
00:32:21,120 --> 00:32:24,720
Speaker 4: I was going is like there are some abstractions around

634
00:32:24,720 --> 00:32:27,560
Speaker 4: that is like, you know, one can we do with

635
00:32:27,760 --> 00:32:28,360
Speaker 4: my image?

636
00:32:28,440 --> 00:32:28,640
Speaker 3: Right?

637
00:32:28,680 --> 00:32:31,240
Speaker 4: Like if I generally get the category that my image

638
00:32:31,240 --> 00:32:34,080
Speaker 4: can be reproduced, generated contents and so on, it's like,

639
00:32:34,200 --> 00:32:36,760
Speaker 4: let's talk about the abstract notion about who has rights

640
00:32:36,800 --> 00:32:39,600
Speaker 4: to that or do we both get to benefit from that?

641
00:32:39,800 --> 00:32:42,600
Speaker 4: If you get that straight, Yes, the nature of how

642
00:32:42,640 --> 00:32:46,000
Speaker 4: the image gets alter created as something will change underneath,

643
00:32:46,160 --> 00:32:48,760
Speaker 4: but the concept will stay the same. And so I

644
00:32:48,760 --> 00:32:50,960
Speaker 4: think it's what's important is to get the categories right.

645
00:32:51,280 --> 00:32:56,480
Speaker 3: Yeah. Yeah, If you just think about the biggest technological

646
00:32:58,000 --> 00:33:02,480
Speaker 3: revolutions of the post war era last seventy five years,

647
00:33:03,400 --> 00:33:05,280
Speaker 3: you can all come up with a list. Actually, it's

648
00:33:05,280 --> 00:33:06,640
Speaker 3: really fun to come up with the list. I was

649
00:33:06,720 --> 00:33:10,360
Speaker 3: thinking about this when we were you know, containerized shipping

650
00:33:10,440 --> 00:33:17,280
Speaker 3: is my favorite, the green revolution, the internet is Where

651
00:33:17,360 --> 00:33:18,400
Speaker 3: is the I in that list?

652
00:33:21,080 --> 00:33:23,880
Speaker 4: So I would put it first in that context that

653
00:33:23,920 --> 00:33:28,000
Speaker 4: you put forth over since World War Two, undoubtedly, like

654
00:33:28,200 --> 00:33:32,320
Speaker 4: computing as a category is one of those trajectories that

655
00:33:32,520 --> 00:33:36,400
Speaker 4: has reshaped right or world. And I think we think computing,

656
00:33:36,960 --> 00:33:41,040
Speaker 4: I would say the role that semiconductors have had has

657
00:33:41,120 --> 00:33:45,160
Speaker 4: been incrowdly defining. I would say AI is the second

658
00:33:45,400 --> 00:33:49,719
Speaker 4: example of that as a core architecture that is going

659
00:33:49,760 --> 00:33:52,680
Speaker 4: to have an equivalent level of impact. And then the

660
00:33:52,720 --> 00:33:54,520
Speaker 4: third leg I would put to that equation will be

661
00:33:54,600 --> 00:33:57,560
Speaker 4: quantum and quantum information. And that's sort of like I

662
00:33:57,640 --> 00:33:59,720
Speaker 4: like to summarize that the future of computing it's bits,

663
00:33:59,760 --> 00:34:03,360
Speaker 4: neural and cubits, and is that idea of high precision computation,

664
00:34:04,080 --> 00:34:06,720
Speaker 4: the world of neural networks and artificial intelligence, and the

665
00:34:06,760 --> 00:34:10,239
Speaker 4: world of quantum and the combination of those things is

666
00:34:10,239 --> 00:34:12,240
Speaker 4: going to be the defining force over the next hundred

667
00:34:12,320 --> 00:34:14,960
Speaker 4: years in that category of computing. But it makes a

668
00:34:15,040 --> 00:34:15,719
Speaker 4: list for sure.

669
00:34:15,960 --> 00:34:18,360
Speaker 3: If it's that high up on the list. This is

670
00:34:18,360 --> 00:34:22,200
Speaker 3: a total hypothetical. Would you if you were starting over,

671
00:34:22,680 --> 00:34:26,160
Speaker 3: if you're starting IBM right now, would you say, oh,

672
00:34:26,440 --> 00:34:29,920
Speaker 3: our AI operations actually should be way bigger, Like how

673
00:34:29,920 --> 00:34:31,960
Speaker 3: many how many thousands of people working for you?

674
00:34:32,840 --> 00:34:36,160
Speaker 4: So within the research division it's about like three thousand,

675
00:34:36,200 --> 00:34:37,120
Speaker 4: five hundred scientists.

676
00:34:37,120 --> 00:34:39,279
Speaker 3: So in a perfect world, would you if it's that big,

677
00:34:39,400 --> 00:34:42,400
Speaker 3: isn't that too small as a group?

678
00:34:42,800 --> 00:34:44,920
Speaker 4: Yeah, Well, that's like in the RICAR division. I mean

679
00:34:45,000 --> 00:34:48,879
Speaker 4: IBM overall, But I mean.

680
00:34:48,840 --> 00:34:51,680
Speaker 3: Like, so starting from first, so you have a you

681
00:34:51,880 --> 00:34:56,400
Speaker 3: we've got a technology that you're ranking with Compute and

682
00:34:56,719 --> 00:34:58,759
Speaker 3: you know, up there with as in terms of a

683
00:34:59,080 --> 00:35:04,120
Speaker 3: world changer. Are we So what I'm basically asking is

684
00:35:04,160 --> 00:35:07,560
Speaker 3: are we underinvested in this huge you know?

685
00:35:07,680 --> 00:35:10,360
Speaker 4: But so so yeah, it's a good question. So like

686
00:35:10,400 --> 00:35:12,239
Speaker 4: what I would say is that I think we should

687
00:35:12,320 --> 00:35:15,880
Speaker 4: segment how many people do you need on the creation

688
00:35:16,680 --> 00:35:19,400
Speaker 4: of the technology itself, and what is the right size

689
00:35:19,440 --> 00:35:22,239
Speaker 4: of research and engineers and compute to do that? And

690
00:35:22,360 --> 00:35:24,600
Speaker 4: how many people do you need in the sort of

691
00:35:24,719 --> 00:35:29,760
Speaker 4: application of the technology to create better products, to deliver

692
00:35:29,920 --> 00:35:33,319
Speaker 4: services and consulting and then ultimately to diffuse it through

693
00:35:33,440 --> 00:35:35,920
Speaker 4: you know, sort of all feheres of society. And the

694
00:35:36,040 --> 00:35:38,520
Speaker 4: numbers are very different, and that is not different than

695
00:35:38,560 --> 00:35:41,040
Speaker 4: anywhere else. I mean, I mean, if you give examples

696
00:35:41,040 --> 00:35:43,279
Speaker 4: of since you were talking about in context of World

697
00:35:43,280 --> 00:35:45,600
Speaker 4: War two, how many people does it take to create,

698
00:35:45,920 --> 00:35:48,640
Speaker 4: you know, an atomic weapon as an example, it's a

699
00:35:48,719 --> 00:35:51,120
Speaker 4: large number. I mean, it wasn't just Los animals. There's

700
00:35:51,120 --> 00:35:53,200
Speaker 4: a lot of people in Okay, it's a large number,

701
00:35:53,280 --> 00:35:56,839
Speaker 4: but it wasn't a million people, right, So you could

702
00:35:56,880 --> 00:36:01,520
Speaker 4: have highly concentrated teams of people that, with enough resources,

703
00:36:01,520 --> 00:36:05,680
Speaker 4: can do extraordinary scientific and technological achievements. And that's always,

704
00:36:05,719 --> 00:36:07,680
Speaker 4: by definition, is going to be a fraction of like

705
00:36:07,760 --> 00:36:10,400
Speaker 4: one percent compared to the total volume that is going

706
00:36:10,440 --> 00:36:11,719
Speaker 4: to require to then deal with it.

707
00:36:11,960 --> 00:36:15,040
Speaker 3: Yeah, but the application side is infinite almost.

708
00:36:14,840 --> 00:36:17,520
Speaker 4: That's exactly. So that is where like in the end

709
00:36:17,560 --> 00:36:21,440
Speaker 4: the bottleneck really is. So with you know, thousands of

710
00:36:21,600 --> 00:36:25,279
Speaker 4: scientists and engineers, you can create world class AI, right,

711
00:36:25,640 --> 00:36:27,799
Speaker 4: And so no, you don't need ten thousand to be

712
00:36:27,840 --> 00:36:30,240
Speaker 4: able to create the large language model and the generatic

713
00:36:30,280 --> 00:36:33,360
Speaker 4: model and some but you need thousands, and you need

714
00:36:33,520 --> 00:36:35,680
Speaker 4: you know, very significant amount of computer and data. You

715
00:36:35,760 --> 00:36:39,880
Speaker 4: need that The rest is Okay, I build software, I

716
00:36:39,880 --> 00:36:43,280
Speaker 4: build databases, or I build a software product that allows

717
00:36:43,280 --> 00:36:46,080
Speaker 4: you to do inventory management, or I build you a

718
00:36:46,120 --> 00:36:51,960
Speaker 4: photo editor and so on. Now that product incorporating the AI, modifying,

719
00:36:52,080 --> 00:36:54,920
Speaker 4: expanding it and so on. Well, now you're talking about

720
00:36:55,000 --> 00:36:57,920
Speaker 4: the entire software industries. So now you're talking about millions

721
00:36:57,920 --> 00:37:00,600
Speaker 4: of people right who are nested, you know, who are

722
00:37:00,640 --> 00:37:03,799
Speaker 4: required to bring AI into their product. Then you go

723
00:37:03,840 --> 00:37:07,280
Speaker 4: on a step beyond the technology creators in terms of software,

724
00:37:07,280 --> 00:37:09,799
Speaker 4: and you say, well, okay, now what the skills to

725
00:37:09,880 --> 00:37:13,960
Speaker 4: help organizations go undeployed in the department of you know,

726
00:37:14,040 --> 00:37:16,399
Speaker 4: the interior, right, And then I said, okay, well, now

727
00:37:16,400 --> 00:37:19,839
Speaker 4: you need like consultants and experts and people to work

728
00:37:19,880 --> 00:37:22,239
Speaker 4: they are to integer into the workflow. So now you're

729
00:37:22,280 --> 00:37:25,040
Speaker 4: talking into the many tens of millions of people around that.

730
00:37:25,320 --> 00:37:27,719
Speaker 4: So I see it as these concentric circles of it,

731
00:37:28,160 --> 00:37:31,120
Speaker 4: but to some degree in many of these core technology areas,

732
00:37:31,160 --> 00:37:32,680
Speaker 4: just saying like well, I need a team of like

733
00:37:32,680 --> 00:37:35,040
Speaker 4: one hundred thousand people to create like AI or a

734
00:37:35,360 --> 00:37:37,879
Speaker 4: or a new transistor or a new quantum computer. It's

735
00:37:37,920 --> 00:37:40,120
Speaker 4: actually a diminished in return right in the end, like

736
00:37:40,200 --> 00:37:42,560
Speaker 4: too many people connecting with each other's very difficult.

737
00:37:42,560 --> 00:37:45,399
Speaker 3: But on the application side of just to go back

738
00:37:45,440 --> 00:37:51,200
Speaker 3: to our example of that college, just the task of

739
00:37:51,280 --> 00:37:55,799
Speaker 3: sitting down with a faculty and working with them to

740
00:37:55,880 --> 00:37:59,520
Speaker 3: reimagine what they do with these new set of tools

741
00:37:59,600 --> 00:38:02,200
Speaker 3: in mind, and with the understanding that the students coming

742
00:38:02,200 --> 00:38:03,759
Speaker 3: in are probably going to know more about it than

743
00:38:03,800 --> 00:38:06,840
Speaker 3: they do that a lot, I mean, that's a that

744
00:38:07,000 --> 00:38:09,759
Speaker 3: is a curricullion people problem.

745
00:38:09,960 --> 00:38:11,960
Speaker 4: It's a people problem. Yeah, that's why I started in

746
00:38:12,080 --> 00:38:13,880
Speaker 4: terms of the barriers of adoption of that. I mean

747
00:38:13,920 --> 00:38:17,160
Speaker 4: the context of IBM an example, that's why we have

748
00:38:17,640 --> 00:38:21,440
Speaker 4: a consulting organization, IVAN Consulting that complements ib AND technology,

749
00:38:21,960 --> 00:38:24,759
Speaker 4: and the IVAN Consulting Organization has over one hundred and

750
00:38:24,800 --> 00:38:28,480
Speaker 4: fifty thousand employees because of this question, right, because you

751
00:38:28,560 --> 00:38:30,960
Speaker 4: have to sit down and you say, Okay, what problem

752
00:38:30,960 --> 00:38:33,520
Speaker 4: are you trying to solve, what is the methodology we're

753
00:38:33,520 --> 00:38:35,480
Speaker 4: going to do, and here's the technology options that we

754
00:38:35,600 --> 00:38:37,680
Speaker 4: have to be able to bring into the table. In

755
00:38:37,760 --> 00:38:42,960
Speaker 4: the end, the adoption across or society will be limited

756
00:38:43,200 --> 00:38:46,560
Speaker 4: by this part. The technology is going to make it easier,

757
00:38:46,640 --> 00:38:51,479
Speaker 4: more cost effective to implement those solutions, but you first

758
00:38:51,560 --> 00:38:53,480
Speaker 4: have to think about what you want to do, how

759
00:38:53,520 --> 00:38:55,040
Speaker 4: you're going to do it, and how are you going

760
00:38:55,080 --> 00:38:56,719
Speaker 4: to bring it into a life of this in this

761
00:38:56,800 --> 00:39:00,480
Speaker 4: context faculty member or you know, the administrator and so

762
00:39:00,600 --> 00:39:01,360
Speaker 4: on in these colleges.

763
00:39:01,600 --> 00:39:04,360
Speaker 3: With that Hollywood that that notion I thought, which was

764
00:39:04,400 --> 00:39:09,680
Speaker 3: absolutely I thought really interesting that in a Hollywood strike

765
00:39:09,760 --> 00:39:13,240
Speaker 3: you have to have this conversation about a distributive justice,

766
00:39:13,440 --> 00:39:16,920
Speaker 3: conversation about how do we that's it's a really hard conversation,

767
00:39:17,360 --> 00:39:19,719
Speaker 3: right to have. And uh so this brings me to

768
00:39:19,719 --> 00:39:21,399
Speaker 3: my nee, which is that you we were talking about

769
00:39:21,440 --> 00:39:25,480
Speaker 3: stage you have. You have two daughters, one in college,

770
00:39:25,480 --> 00:39:28,359
Speaker 3: one about to go to college. That's right, so they're

771
00:39:28,360 --> 00:39:32,799
Speaker 3: both science minded. So tell me about the conversations you

772
00:39:32,800 --> 00:39:35,680
Speaker 3: you have with your daughter. You have a unique conversation

773
00:39:35,760 --> 00:39:38,839
Speaker 3: with your daughters because your conversation, your advice to them

774
00:39:38,960 --> 00:39:42,360
Speaker 3: is is influenced by what you do for a living.

775
00:39:42,640 --> 00:39:43,480
Speaker 4: Yes, it's true.

776
00:39:43,640 --> 00:39:48,000
Speaker 3: So did you warn your daughters away from certain fields?

777
00:39:48,040 --> 00:39:53,600
Speaker 3: Did you say whatever you do, don't be you know, no.

778
00:39:52,560 --> 00:39:55,120
Speaker 4: No, no, that's not my style. I mean for me, no,

779
00:39:55,400 --> 00:39:57,520
Speaker 4: I try not to be like you know, preachy about that.

780
00:39:58,480 --> 00:40:01,000
Speaker 4: So for me, it was just about by example of

781
00:40:01,120 --> 00:40:04,160
Speaker 4: things I love, right and things I care about, and

782
00:40:04,200 --> 00:40:06,360
Speaker 4: then you know, bringing them to the lab and seeing

783
00:40:06,440 --> 00:40:09,680
Speaker 4: things and then the natural conversations of things working on

784
00:40:10,000 --> 00:40:13,040
Speaker 4: or interesting people I meet. So to the extent that

785
00:40:13,080 --> 00:40:15,400
Speaker 4: they have chosen that and obviously this has an influence

786
00:40:15,440 --> 00:40:19,040
Speaker 4: on them. It has been through seeing it, you know,

787
00:40:19,080 --> 00:40:21,040
Speaker 4: perhaps through my eyes, right, and what do you see

788
00:40:21,080 --> 00:40:22,960
Speaker 4: me do? And that I like my profession, right.

789
00:40:22,760 --> 00:40:25,680
Speaker 3: But one of your daughters. You said, is thinking that

790
00:40:25,760 --> 00:40:29,000
Speaker 3: she wants to be a doctor. But being a doctor

791
00:40:29,120 --> 00:40:31,600
Speaker 3: in a post AI world it's surely a very different

792
00:40:31,600 --> 00:40:34,759
Speaker 3: proposition than being a doctor in a PREAI world. Do

793
00:40:34,800 --> 00:40:38,800
Speaker 3: you think have you tried to prepare her for that difference?

794
00:40:39,000 --> 00:40:41,040
Speaker 3: Have you explained to her what you think will happen

795
00:40:41,080 --> 00:40:42,399
Speaker 3: to this profession she might enter.

796
00:40:42,920 --> 00:40:45,840
Speaker 4: Yeah, I mean not in like, you know, incredible amount

797
00:40:45,880 --> 00:40:49,640
Speaker 4: of detail, but yes, at the level of understanding what

798
00:40:49,760 --> 00:40:53,239
Speaker 4: is changing, like this lens of the information, lens with

799
00:40:53,280 --> 00:40:55,279
Speaker 4: which you can look at the world and what is

800
00:40:55,480 --> 00:40:58,839
Speaker 4: possible and what it can do, Like what is our

801
00:40:58,960 --> 00:41:01,279
Speaker 4: role and what is all of the technology and how

802
00:41:01,320 --> 00:41:04,399
Speaker 4: that shapes At that level of abstraction, for sure, but

803
00:41:04,440 --> 00:41:06,840
Speaker 4: not at the level of like, don't be a radiologist,

804
00:41:06,960 --> 00:41:08,719
Speaker 4: you know, because this is what we want for you.

805
00:41:08,760 --> 00:41:10,600
Speaker 3: I was gonna say, if you're not't happy with your

806
00:41:10,640 --> 00:41:13,160
Speaker 3: current job, you could do a podcast called Parenting Tips

807
00:41:13,160 --> 00:41:17,279
Speaker 3: with Dario, which is just an AI person gives you

808
00:41:17,360 --> 00:41:19,880
Speaker 3: advice on what your kids should do based on exactly this,

809
00:41:20,040 --> 00:41:22,920
Speaker 3: Like should I be a readiologist? Dario tell me, like

810
00:41:23,160 --> 00:41:26,799
Speaker 3: it seems to be a really important question. Yeah, let

811
00:41:26,840 --> 00:41:28,560
Speaker 3: me ask this question in a more I'm joking, but

812
00:41:28,640 --> 00:41:32,880
Speaker 3: in a more serious way. Surely it would if I

813
00:41:32,880 --> 00:41:34,520
Speaker 3: don't mean to use your daughter as an example. But

814
00:41:34,560 --> 00:41:36,759
Speaker 3: let's imagine we're giving advice to somebody who wants to

815
00:41:36,880 --> 00:41:41,719
Speaker 3: enter medicine. A really useful conversation to have is what

816
00:41:41,800 --> 00:41:46,279
Speaker 3: are the skills that are will be most prized in

817
00:41:46,320 --> 00:41:49,760
Speaker 3: that profession fifteen years from now, and are they different

818
00:41:49,800 --> 00:41:52,000
Speaker 3: from the skills that are prized now. How would you

819
00:41:52,040 --> 00:41:52,960
Speaker 3: answer that question?

820
00:41:53,840 --> 00:41:58,080
Speaker 4: Yeah, I think for example, this goes back to how

821
00:41:58,160 --> 00:42:01,080
Speaker 4: is this scientific method in this context, like the practice

822
00:42:01,120 --> 00:42:03,759
Speaker 4: of medicine going to change? I think we will see

823
00:42:03,760 --> 00:42:06,279
Speaker 4: more changes on how we practice a scientific method and

824
00:42:06,320 --> 00:42:10,160
Speaker 4: so on as a consequence of what is happening with

825
00:42:10,360 --> 00:42:13,440
Speaker 4: the world of computing and information, how we represent information,

826
00:42:13,560 --> 00:42:16,840
Speaker 4: how we represent knowledge, how we extract meaning from knowledge

827
00:42:17,000 --> 00:42:20,040
Speaker 4: as a method than we have seen in the last

828
00:42:20,080 --> 00:42:23,480
Speaker 4: two hundred years. So therefore, what I would like strongly

829
00:42:23,560 --> 00:42:25,920
Speaker 4: encourage is not about like, hey, use this tool for

830
00:42:25,960 --> 00:42:29,040
Speaker 4: doing this or doing that, but in the curriculum itself,

831
00:42:29,080 --> 00:42:32,840
Speaker 4: in understanding how we do problems solving in the age

832
00:42:32,920 --> 00:42:35,920
Speaker 4: of like data and data representation and so on, that

833
00:42:36,000 --> 00:42:39,319
Speaker 4: needs to be embedded in the curriculum of everybody you

834
00:42:39,360 --> 00:42:41,840
Speaker 4: know that is I would say, actually quite horizontally, but

835
00:42:41,920 --> 00:42:44,759
Speaker 4: certainly in the context of medicine and scientists and so on,

836
00:42:44,880 --> 00:42:48,760
Speaker 4: for sure, And to the extent that that gets ingrained,

837
00:42:48,920 --> 00:42:51,000
Speaker 4: that will give us a lens that no matter what

838
00:42:51,840 --> 00:42:55,040
Speaker 4: specialty they go within medicine, they will say, actually, the

839
00:42:55,120 --> 00:42:57,600
Speaker 4: way I want to be able to tackle improving the

840
00:42:57,680 --> 00:43:00,160
Speaker 4: quality of care, the way to do that is in

841
00:43:00,200 --> 00:43:02,920
Speaker 4: addition to all the elements that we have practiced in

842
00:43:02,960 --> 00:43:05,839
Speaker 4: the field of medicine. Is this new lens? And are

843
00:43:05,840 --> 00:43:08,080
Speaker 4: we representing the data the right way? Do we have

844
00:43:08,160 --> 00:43:10,880
Speaker 4: the right tools to be able to represent that knowledge?

845
00:43:11,000 --> 00:43:14,080
Speaker 4: Am I incorporating that in my own so with my

846
00:43:14,120 --> 00:43:16,719
Speaker 4: own knowledge in a way that gives me better outcomes? Right?

847
00:43:16,880 --> 00:43:20,200
Speaker 4: Do I have the rigor of benchmarking too? And quality

848
00:43:20,480 --> 00:43:23,000
Speaker 4: of the results? So that is what needs to be incorporated.

849
00:43:23,120 --> 00:43:28,839
Speaker 3: How in a perfect world, if I asked you your

850
00:43:28,880 --> 00:43:33,360
Speaker 3: team to rewrite curriculum for American medical schools, how dramatic

851
00:43:33,520 --> 00:43:36,760
Speaker 3: a revision is that? Are we tinkering with ten percent

852
00:43:36,760 --> 00:43:38,960
Speaker 3: of the curriculum or we tinkering with fifty percent of it?

853
00:43:41,200 --> 00:43:45,400
Speaker 4: I think they would be a subset of classes. That

854
00:43:45,480 --> 00:43:47,920
Speaker 4: is about the method the methodology, what has changed, like

855
00:43:48,400 --> 00:43:52,120
Speaker 4: have these lens of it to understand and then within

856
00:43:52,320 --> 00:43:56,880
Speaker 4: each class that methodology will represent something that is embedded

857
00:43:56,960 --> 00:44:01,880
Speaker 4: in it. Right, Well, it will be substantive but not

858
00:44:02,080 --> 00:44:05,480
Speaker 4: but doesn't mean replacing the specialization and the context and

859
00:44:05,520 --> 00:44:08,440
Speaker 4: the knowledge of each domain. But I do think everybody

860
00:44:08,520 --> 00:44:11,920
Speaker 4: should have sort of a basic knowledge of the horizontal, right,

861
00:44:12,000 --> 00:44:14,560
Speaker 4: what is it, how does it work? What tools you have,

862
00:44:14,960 --> 00:44:16,840
Speaker 4: what is the technology, and like you know what are

863
00:44:16,880 --> 00:44:19,600
Speaker 4: the dos and don'ts around that? And then every area

864
00:44:19,680 --> 00:44:21,640
Speaker 4: you say, and you know that thing that you learn,

865
00:44:21,760 --> 00:44:24,759
Speaker 4: this is how it applies to anatomy, and this is

866
00:44:24,760 --> 00:44:27,000
Speaker 4: how you know how it applies to you know, radiology

867
00:44:27,000 --> 00:44:29,080
Speaker 4: if you studying that, or or this is how you

868
00:44:29,120 --> 00:44:31,360
Speaker 4: apply you know, in the context of discovery right of

869
00:44:31,440 --> 00:44:33,360
Speaker 4: self structure and this is how we can use it,

870
00:44:33,480 --> 00:44:36,480
Speaker 4: or protein folding and this is how it does so

871
00:44:36,520 --> 00:44:39,840
Speaker 4: that way you'll see a connecting tissue through throughout the

872
00:44:39,880 --> 00:44:40,319
Speaker 4: whole thing.

873
00:44:40,520 --> 00:44:43,440
Speaker 3: Yeah, I mean I would add to that because I

874
00:44:43,480 --> 00:44:47,840
Speaker 3: was sticking it the sch that it's also this incredible

875
00:44:47,840 --> 00:44:50,960
Speaker 3: opportunity to do what doctors are supposed to do but

876
00:44:51,040 --> 00:44:53,600
Speaker 3: don't have time to do now, which is they're so

877
00:44:53,719 --> 00:44:58,799
Speaker 3: consumed with figuring out what's wrong with you that they

878
00:44:58,800 --> 00:45:01,560
Speaker 3: have little, little time to talk about the implications of

879
00:45:01,600 --> 00:45:04,600
Speaker 3: the diagnosisness and what we really want to if we

880
00:45:04,640 --> 00:45:08,080
Speaker 3: can freedom of some of the burden of what is

881
00:45:08,080 --> 00:45:10,279
Speaker 3: actually quite a prosaic question of what's wrong with you

882
00:45:10,719 --> 00:45:14,040
Speaker 3: and leave the hard human thing of let make should

883
00:45:14,040 --> 00:45:17,759
Speaker 3: you be scared or hopeful? Should you you know? What

884
00:45:17,800 --> 00:45:18,560
Speaker 3: do you need to do?

885
00:45:18,640 --> 00:45:18,879
Speaker 4: Or what?

886
00:45:19,160 --> 00:45:20,640
Speaker 3: Let me put this in the context of all the

887
00:45:20,640 --> 00:45:23,640
Speaker 3: patients I've seen, that conversation, which is the most important one,

888
00:45:23,680 --> 00:45:26,279
Speaker 3: is the one that seems to me so like, if

889
00:45:26,280 --> 00:45:29,600
Speaker 3: I had to, I would add, if we're reimagining the

890
00:45:29,760 --> 00:45:33,920
Speaker 3: curriculum of med school, I'd like, with whatever this, by

891
00:45:33,920 --> 00:45:35,719
Speaker 3: the way, very little time. Maybe we have to add

892
00:45:35,719 --> 00:45:38,560
Speaker 3: two more years to med school, but like.

893
00:45:38,440 --> 00:45:39,200
Speaker 4: A whole.

894
00:45:40,960 --> 00:45:44,959
Speaker 3: But the whole thing about bringing back the human side

895
00:45:45,040 --> 00:45:49,080
Speaker 3: of you know, now, if I can give you ten

896
00:45:49,080 --> 00:45:52,160
Speaker 3: more minutes, how do you use that ten more minutes in.

897
00:45:52,120 --> 00:45:55,520
Speaker 4: That in that reconceptualization that you just did is what

898
00:45:55,560 --> 00:45:57,440
Speaker 4: we should be doing around that, Because I think the

899
00:45:57,480 --> 00:46:00,759
Speaker 4: debate as to like, well, I'm I need doctors or

900
00:46:00,800 --> 00:46:03,359
Speaker 4: not it's actually a not very useful debate. But rather

901
00:46:03,440 --> 00:46:06,080
Speaker 4: this other question is how is your time being spent?

902
00:46:06,239 --> 00:46:09,120
Speaker 4: What problems are you getting stuck? I mean I generalize

903
00:46:09,160 --> 00:46:11,719
Speaker 4: this by like the obvious observation that if you look

904
00:46:11,719 --> 00:46:14,480
Speaker 4: around in your professions, in our daily lives, we have

905
00:46:14,520 --> 00:46:16,719
Speaker 4: not run out of problems to solve. So as an

906
00:46:16,760 --> 00:46:18,880
Speaker 4: example of that is, hey, if I'm spending all my

907
00:46:18,920 --> 00:46:20,840
Speaker 4: time trying to do diagnosis, and I could do that

908
00:46:20,880 --> 00:46:23,600
Speaker 4: ten times faster, and it allows me actually to go,

909
00:46:24,680 --> 00:46:26,560
Speaker 4: you know, and take care of the patients and all

910
00:46:26,600 --> 00:46:28,439
Speaker 4: the next steps of what we have to do about it.

911
00:46:28,560 --> 00:46:30,880
Speaker 4: That's probably a trade off that a lot of doctors

912
00:46:30,920 --> 00:46:33,480
Speaker 4: would take, right. And then you say, well, you know,

913
00:46:33,520 --> 00:46:35,279
Speaker 4: to what degree that does it allow me to do that?

914
00:46:35,320 --> 00:46:37,319
Speaker 4: And I can do these other things and these other

915
00:46:37,360 --> 00:46:40,880
Speaker 4: things are critically important for my profession around that. So

916
00:46:41,200 --> 00:46:44,480
Speaker 4: when you actually become less abstract and like we get

917
00:46:44,840 --> 00:46:47,440
Speaker 4: past the futile conversation of like, oh, there's no more

918
00:46:47,520 --> 00:46:49,000
Speaker 4: jobs and I'm going to take it all of it,

919
00:46:49,040 --> 00:46:52,120
Speaker 4: which is kind of nonsense, is you go back to say,

920
00:46:52,160 --> 00:46:56,239
Speaker 4: in practice in your context, right, for you, what does

921
00:46:56,280 --> 00:46:58,560
Speaker 4: it mean? How do you work? What can you do

922
00:46:58,560 --> 00:47:01,560
Speaker 4: differently around that. Actually, that's a much richer conversation, and

923
00:47:01,640 --> 00:47:03,840
Speaker 4: very often we would find ourselves that there's a portion

924
00:47:03,920 --> 00:47:05,799
Speaker 4: of the work we do that we say I would

925
00:47:05,880 --> 00:47:07,919
Speaker 4: rather do less of that. This is this other part

926
00:47:08,160 --> 00:47:10,920
Speaker 4: I like a lot, And if it is possible that

927
00:47:10,920 --> 00:47:13,440
Speaker 4: technology could help us make that trade off, I'll take

928
00:47:13,440 --> 00:47:17,799
Speaker 4: it in a heartbeat. Now, poorly implemented technology can also

929
00:47:17,840 --> 00:47:19,920
Speaker 4: create another problem. You say, Hey, this was supposed to

930
00:47:19,960 --> 00:47:23,319
Speaker 4: solve me things, but the way it's being implemented is

931
00:47:23,360 --> 00:47:26,280
Speaker 4: not helping me, right, it's making my life more more miserable,

932
00:47:26,400 --> 00:47:29,000
Speaker 4: or so on, or I've lost connection in how I

933
00:47:29,120 --> 00:47:32,480
Speaker 4: used to work, et cetera. So that is why design

934
00:47:33,160 --> 00:47:36,279
Speaker 4: is so important. That is why I also workflow is

935
00:47:36,320 --> 00:47:39,080
Speaker 4: so important in being able to solve these problems. But

936
00:47:39,920 --> 00:47:43,440
Speaker 4: it begins by, you know, going from the intergalactic to

937
00:47:43,520 --> 00:47:45,920
Speaker 4: the reality of it, of that faculty member in the

938
00:47:45,920 --> 00:47:48,880
Speaker 4: Liberal Arts college or you know, or a you know,

939
00:47:49,000 --> 00:47:51,680
Speaker 4: a practitioner in medicine in a hospital and what it

940
00:47:51,800 --> 00:47:52,799
Speaker 4: means for them. Right.

941
00:47:53,480 --> 00:47:57,680
Speaker 3: Yeah. What struck me AREIO throughout our conversation is how

942
00:47:57,800 --> 00:48:02,960
Speaker 3: much of this revolution and is non technical? Is to say,

943
00:48:03,280 --> 00:48:05,200
Speaker 3: you guys are doing the technical thing here, but the

944
00:48:05,239 --> 00:48:08,200
Speaker 3: real the revolution is going to require a whole range

945
00:48:08,200 --> 00:48:11,920
Speaker 3: of people doing things that have nothing to do with software,

946
00:48:12,160 --> 00:48:15,640
Speaker 3: that have to do with working out new new human arrangements.

947
00:48:16,080 --> 00:48:18,719
Speaker 3: Talking about that, I mean, does keep coming back to

948
00:48:18,760 --> 00:48:21,480
Speaker 3: the Hollywood strike thing that you have to have a

949
00:48:21,520 --> 00:48:28,240
Speaker 3: conversation about our values. Is creators of of of of movies?

950
00:48:28,280 --> 00:48:31,000
Speaker 3: How are we going to divide up the exactly credit

951
00:48:31,040 --> 00:48:35,439
Speaker 3: and the like. That's a that's a conversation about philosophy,

952
00:48:35,520 --> 00:48:37,399
Speaker 3: and you know it is.

953
00:48:37,360 --> 00:48:40,680
Speaker 4: And is it's in the grand tradition of why you know,

954
00:48:42,160 --> 00:48:46,440
Speaker 4: a liberal education is so important in the broadest possible sense, Right,

955
00:48:46,840 --> 00:48:50,640
Speaker 4: there's no common conception of the good, right, that is

956
00:48:50,640 --> 00:48:54,880
Speaker 4: always a contested dialogue that happens within our society. And

957
00:48:55,000 --> 00:48:57,279
Speaker 4: technology is going to fit in that context too, right.

958
00:48:57,400 --> 00:48:59,560
Speaker 4: So that's why I personally, as a philosophy I'm not

959
00:48:59,600 --> 00:49:03,360
Speaker 4: a technological determinists, right, And I don't like when colleagues

960
00:49:03,360 --> 00:49:06,040
Speaker 4: in my profession right starts saying like, well, this is

961
00:49:06,080 --> 00:49:08,640
Speaker 4: the way the technology is going to be, and by consequence,

962
00:49:08,880 --> 00:49:10,840
Speaker 4: this is how society is going to be. I'm like,

963
00:49:10,920 --> 00:49:14,000
Speaker 4: that's a highly contested goal. And if you want to

964
00:49:14,120 --> 00:49:16,640
Speaker 4: enter into realm of politics or the real other ones,

965
00:49:16,680 --> 00:49:19,279
Speaker 4: go and stand up on a stool and discuss with it.

966
00:49:19,280 --> 00:49:21,440
Speaker 4: That's what society wants. You will find that it's a

967
00:49:21,520 --> 00:49:25,680
Speaker 4: huge diversity of opinions and perspective and that's what makes

968
00:49:25,719 --> 00:49:28,319
Speaker 4: you know, you know, in a democracy, the richness of

969
00:49:28,320 --> 00:49:30,520
Speaker 4: our society. And in the end, that is going to

970
00:49:30,560 --> 00:49:33,520
Speaker 4: be the centerpiece of the conversation what do we want?

971
00:49:34,400 --> 00:49:37,000
Speaker 4: You know, who gets what? And so on? And that

972
00:49:37,320 --> 00:49:39,560
Speaker 4: is actually I don't think it's anything negative. That's acid

973
00:49:39,640 --> 00:49:42,560
Speaker 4: should be because in the end is anchor of who

974
00:49:42,560 --> 00:49:46,080
Speaker 4: we want as humans, you know, you know, as friends, family, citizens,

975
00:49:46,280 --> 00:49:49,080
Speaker 4: and we have many overlapping sets of responsibilities, right and

976
00:49:49,120 --> 00:49:52,239
Speaker 4: as a technology creator, my only responsibility is not just

977
00:49:52,239 --> 00:49:55,080
Speaker 4: as a scientist and a technology creator. I'm also a

978
00:49:55,080 --> 00:49:57,120
Speaker 4: member of family. I'm a citizen, and I'm many other

979
00:49:57,160 --> 00:49:59,239
Speaker 4: things that I care about. And I think that that

980
00:49:59,440 --> 00:50:04,879
Speaker 4: sometimes the debate of the technological determinists they start now

981
00:50:04,960 --> 00:50:10,080
Speaker 4: budding into what is the realm of you know, justice

982
00:50:10,320 --> 00:50:13,920
Speaker 4: and you know, in society and philosophy and democracy, And

983
00:50:13,960 --> 00:50:16,480
Speaker 4: that's where they get the most uncomfortable because it's like

984
00:50:16,480 --> 00:50:19,560
Speaker 4: I'm just telling you, like you know what's possible, and

985
00:50:19,640 --> 00:50:23,640
Speaker 4: when there's pushback, it's like, yeah, but now we're talking

986
00:50:23,640 --> 00:50:27,480
Speaker 4: about how we live and how we work and how

987
00:50:27,560 --> 00:50:31,160
Speaker 4: much I get paid or not paid. So that technology

988
00:50:31,200 --> 00:50:34,480
Speaker 4: is important. Technology shapes that conversation, but we're going to

989
00:50:34,560 --> 00:50:38,160
Speaker 4: have the conversation with a different language, as it should be,

990
00:50:38,680 --> 00:50:41,160
Speaker 4: and technologies need to get accustomed to if they want

991
00:50:41,160 --> 00:50:44,080
Speaker 4: to participate in that world with the broad consequences. Hey,

992
00:50:44,280 --> 00:50:46,920
Speaker 4: get a custom to deal with the complexity of that

993
00:50:47,000 --> 00:50:51,399
Speaker 4: world of politics, society, institutions, unions, all that stuff. And

994
00:50:51,440 --> 00:50:53,440
Speaker 4: you know, you can be like whiny about it. It's

995
00:50:53,440 --> 00:50:56,000
Speaker 4: like they're not adopting my technology. That's what it takes

996
00:50:56,040 --> 00:50:57,360
Speaker 4: to bring technology into the world.

997
00:50:58,280 --> 00:51:04,840
Speaker 3: Yeah, well said, thank you Dario for this wonderful conversation.

998
00:51:05,040 --> 00:51:08,720
Speaker 3: Thank you to all of you for coming and listening,

999
00:51:08,880 --> 00:51:10,360
Speaker 3: and thank you.

1000
00:51:10,640 --> 00:51:10,920
Speaker 4: Thank you.

1001
00:51:14,440 --> 00:51:17,560
Speaker 3: Dario gild transformed how I think about the future of AI.

1002
00:51:18,280 --> 00:51:20,560
Speaker 3: He explained to me how huge of a leap it

1003
00:51:20,760 --> 00:51:24,080
Speaker 3: was when we went from chess playing models to language

1004
00:51:24,160 --> 00:51:27,400
Speaker 3: learning models, and he talked about how we still have

1005
00:51:27,480 --> 00:51:30,120
Speaker 3: a lot of room to grow. That's why it's important

1006
00:51:30,440 --> 00:51:33,759
Speaker 3: that we get things right. The future of AI is

1007
00:51:33,840 --> 00:51:38,120
Speaker 3: impossible to predict, but the technology has so much potential

1008
00:51:38,239 --> 00:51:41,960
Speaker 3: in every industry. Zooming into an academic or medical setting

1009
00:51:42,200 --> 00:51:45,280
Speaker 3: showed just how close we are to the widespread adoption

1010
00:51:45,640 --> 00:51:49,279
Speaker 3: of AI. Even Hollywood is being forced to figure this out.

1011
00:51:49,920 --> 00:51:52,640
Speaker 3: Institutions of all sorts will have to be at the

1012
00:51:52,719 --> 00:51:56,360
Speaker 3: forefront of integration in order to unlock the full power

1013
00:51:56,400 --> 00:52:01,360
Speaker 3: of AI thoughtfully and responsibly. Humans have the power and

1014
00:52:01,400 --> 00:52:05,360
Speaker 3: the responsibility to shape the tech for our world. I

1015
00:52:05,600 --> 00:52:08,480
Speaker 3: for one, I'm excited to see how things play out.

1016
00:52:09,840 --> 00:52:13,960
Speaker 3: Smart Talks with IBM is produced by Matt Romano, Joey Fishground,

1017
00:52:14,200 --> 00:52:18,600
Speaker 3: David jaw and Jacob Goldstein. We're edited by Lydia Jane Kott.

1018
00:52:18,920 --> 00:52:23,400
Speaker 3: Our engineers are Jason Gambrel, Sarah Bruguier, and Ben Holliday.

1019
00:52:24,040 --> 00:52:29,360
Speaker 3: Theme song by Gramoscope. Special thanks to Andy Kelly, Kathy Callahan,

1020
00:52:29,640 --> 00:52:32,480
Speaker 3: and the eight Bar and IBM teams, as well as

1021
00:52:32,480 --> 00:52:36,239
Speaker 3: the Pushkin marketing team. Smart Talks with IBM is a

1022
00:52:36,239 --> 00:52:40,840
Speaker 3: production of Pushkin Industries and Ruby Studio at iHeartMedia. To

1023
00:52:40,880 --> 00:52:45,799
Speaker 3: find more Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts,

1024
00:52:46,080 --> 00:52:51,319
Speaker 3: or wherever you listen to podcasts. I'm Malcolm Gladwell. This

1025
00:52:51,440 --> 00:52:58,319
Speaker 3: is a paid advertisement from IBM.