1 00:00:02,759 --> 00:00:10,160 Speaker 1: Welcome, Welcome, Welcome to Smart Talks with IBM. 2 00:00:10,200 --> 00:00:13,840 Speaker 2: Hello, Hello, Welcome to Smart Talks with IBM, a podcast 3 00:00:14,000 --> 00:00:19,880 Speaker 2: from Pushkin Industries, iHeartRadio and IBM. I'm Malcolm Gladwell. This season, 4 00:00:20,120 --> 00:00:23,400 Speaker 2: we're diving back into the world of artificial intelligence, but 5 00:00:23,480 --> 00:00:29,120 Speaker 2: with a focus on the powerful concept of open its possibilities, implications, 6 00:00:29,160 --> 00:00:32,920 Speaker 2: and misconceptions. We'll look at openness from a variety of 7 00:00:32,960 --> 00:00:36,840 Speaker 2: angles and explore how the concept is already reshaping industries, 8 00:00:37,280 --> 00:00:40,879 Speaker 2: ways of doing business and our very notion of what's possible. 9 00:00:41,320 --> 00:00:43,839 Speaker 2: And for the first episode of this season, we're bringing 10 00:00:43,880 --> 00:00:48,040 Speaker 2: you a special conversation. I recently sat down with Rob Thomas. 11 00:00:48,520 --> 00:00:51,520 Speaker 2: Rob is the senior vice president of Software and chief 12 00:00:51,560 --> 00:00:55,120 Speaker 2: Commercial Officer of IBM. I spoke to him in front 13 00:00:55,160 --> 00:00:57,800 Speaker 2: of a live audience as part of New York Tech Week. 14 00:00:58,400 --> 00:01:02,040 Speaker 2: We discussed how business is can harness the immense productivity 15 00:01:02,080 --> 00:01:06,080 Speaker 2: benefits of AI while implementing it in a responsible and 16 00:01:06,160 --> 00:01:10,399 Speaker 2: ethical manner. We also broke down a fascinating concept that 17 00:01:10,520 --> 00:01:15,800 Speaker 2: Rob believes about AI, known as the productivity paradox. Okay, 18 00:01:16,040 --> 00:01:26,320 Speaker 2: let's get to the conversation. How are we doing good? 19 00:01:26,720 --> 00:01:26,920 Speaker 3: Rob? 20 00:01:26,959 --> 00:01:30,959 Speaker 2: This is our our second time. We did one of 21 00:01:31,000 --> 00:01:33,560 Speaker 2: these in the middle of the pandemic. But now it's 22 00:01:33,560 --> 00:01:35,200 Speaker 2: all such a blur now that us can figure out 23 00:01:35,200 --> 00:01:35,640 Speaker 2: when it was. 24 00:01:35,760 --> 00:01:38,120 Speaker 3: I know it's hard to those are like a blurry years. 25 00:01:38,160 --> 00:01:39,560 Speaker 3: You don't know what happened, right. 26 00:01:39,720 --> 00:01:42,640 Speaker 2: But well, it's good to see you, to meet you again. 27 00:01:44,000 --> 00:01:46,200 Speaker 2: I wanted to start by going back. You've been at 28 00:01:46,200 --> 00:01:47,440 Speaker 2: IBM twenty years. 29 00:01:48,080 --> 00:01:50,840 Speaker 3: Is that right? Twenty five in July, believe it or not. 30 00:01:51,040 --> 00:01:52,720 Speaker 2: So you were a kid when you joined. 31 00:01:52,800 --> 00:01:53,280 Speaker 3: I was four. 32 00:01:53,760 --> 00:01:59,560 Speaker 2: Yeah, So I want to contrast present day Rob and 33 00:02:00,280 --> 00:02:04,920 Speaker 2: twenty five years ago. Rob. When you arrive at IBM, 34 00:02:05,360 --> 00:02:07,280 Speaker 2: what do you think your job is going to be? 35 00:02:07,320 --> 00:02:08,120 Speaker 3: It, your career is going. 36 00:02:08,440 --> 00:02:10,400 Speaker 2: Where do you think the kind of problems you're going 37 00:02:10,480 --> 00:02:11,680 Speaker 2: to be addressing are? 38 00:02:13,520 --> 00:02:16,280 Speaker 1: Well, it's kind of surreal because I joined IBM Consulting 39 00:02:16,960 --> 00:02:21,359 Speaker 1: and I'm coming out of school and you quickly realize 40 00:02:21,400 --> 00:02:23,560 Speaker 1: what the job of a consultant is to tell other 41 00:02:23,600 --> 00:02:26,239 Speaker 1: companies what to do. And I was like, I literally 42 00:02:26,280 --> 00:02:29,880 Speaker 1: know nothing, and so you're immediately trying to figure out, 43 00:02:29,880 --> 00:02:31,480 Speaker 1: so how am I going to be relevant given that 44 00:02:31,520 --> 00:02:34,720 Speaker 1: I know absolutely nothing to advise other companies on what 45 00:02:34,720 --> 00:02:37,400 Speaker 1: they should be doing. And I remember it well, like 46 00:02:37,639 --> 00:02:40,680 Speaker 1: we were sitting in a room. When you're a consultant, 47 00:02:40,680 --> 00:02:42,480 Speaker 1: you're waiting for somebody else to find work for you. 48 00:02:43,560 --> 00:02:45,799 Speaker 1: A bunch of us sitting in a room, and somebody 49 00:02:45,880 --> 00:02:49,520 Speaker 1: walks in and says, we need somebody that knows Visio. 50 00:02:49,720 --> 00:02:52,480 Speaker 3: Does anybody know Visio? I'd never heard of Visio. 51 00:02:52,520 --> 00:02:55,000 Speaker 1: I don't know if anybody in the room has. So 52 00:02:55,080 --> 00:02:58,720 Speaker 1: everybody's like sitting around looking at their shoes. So finally 53 00:02:58,760 --> 00:03:01,600 Speaker 1: I was like, I know it. So I raised my hand. 54 00:03:01,600 --> 00:03:03,960 Speaker 1: They're like, great, we got a project for you next week. 55 00:03:04,840 --> 00:03:06,880 Speaker 1: So I was like, all right, I have like three 56 00:03:06,960 --> 00:03:10,880 Speaker 1: days to figure out what visio is, and I hope 57 00:03:10,880 --> 00:03:12,480 Speaker 1: I can actually figure out how to use it now. 58 00:03:12,560 --> 00:03:14,320 Speaker 3: Luckily, it wasn't like. 59 00:03:14,480 --> 00:03:17,040 Speaker 1: A programming language. I mean, it's pretty much a drag 60 00:03:17,120 --> 00:03:21,360 Speaker 1: and drop capability. And so I literally left the office, 61 00:03:21,720 --> 00:03:24,800 Speaker 1: went to a bookstore, bought the first three books on 62 00:03:24,880 --> 00:03:27,360 Speaker 1: Visio I could find, spent the whole week in reading 63 00:03:27,360 --> 00:03:30,800 Speaker 1: the books, and showed up and got to work on 64 00:03:30,840 --> 00:03:31,399 Speaker 1: the project. 65 00:03:31,520 --> 00:03:33,840 Speaker 3: And so it was a bit of a risky move, 66 00:03:33,960 --> 00:03:36,800 Speaker 3: but I think that's kind of you. 67 00:03:38,400 --> 00:03:40,560 Speaker 1: This well, but if you don't take risk you'll never 68 00:03:40,920 --> 00:03:44,880 Speaker 1: you'll never achieve, and so does some extent. Everybody's making 69 00:03:44,920 --> 00:03:47,640 Speaker 1: everything up all the time. It's like, can you learn 70 00:03:47,680 --> 00:03:51,360 Speaker 1: faster than somebody else? Is what the difference is in 71 00:03:51,440 --> 00:03:55,000 Speaker 1: almost every part of life. And so it was not planned, 72 00:03:55,040 --> 00:03:57,400 Speaker 1: but it was an accident, but it kind of forced 73 00:03:57,440 --> 00:03:59,240 Speaker 1: me to figure out that you're gonna have to figure 74 00:03:59,280 --> 00:03:59,680 Speaker 1: things out. 75 00:04:00,080 --> 00:04:02,520 Speaker 2: You know, we're here to talk about AI. And I'm 76 00:04:02,560 --> 00:04:08,920 Speaker 2: curious about the evolution of your understanding or IBM's understanding 77 00:04:08,960 --> 00:04:11,400 Speaker 2: of my AI. At what point in the last twenty 78 00:04:11,400 --> 00:04:13,960 Speaker 2: five years do you begin to think, oh, this is 79 00:04:14,000 --> 00:04:15,720 Speaker 2: really going to be at the core of what we 80 00:04:16,080 --> 00:04:19,760 Speaker 2: think about and work on at this company. 81 00:04:20,000 --> 00:04:24,159 Speaker 1: The computer scientist John McCarthy, he was he's the person 82 00:04:24,200 --> 00:04:27,839 Speaker 1: that's credited with coining the phrase artificial intelligence. It's like 83 00:04:27,880 --> 00:04:33,039 Speaker 1: in the fifties, and he made an interesting comedy said 84 00:04:33,240 --> 00:04:35,560 Speaker 1: he said, once it works, it's no longer called AI, 85 00:04:37,800 --> 00:04:41,000 Speaker 1: and that then became it's called like the AI effect, 86 00:04:41,080 --> 00:04:44,839 Speaker 1: which is it seems very difficult, very mysterious, but once 87 00:04:44,880 --> 00:04:49,040 Speaker 1: it becomes commonplace, it's just no longer what it is. 88 00:04:49,400 --> 00:04:51,839 Speaker 1: And so if you put that frame on it, I 89 00:04:51,839 --> 00:04:54,160 Speaker 1: think we've always been doing AI at some level, and 90 00:04:54,320 --> 00:04:56,320 Speaker 1: I even think back to when I joined IBM in 91 00:04:56,560 --> 00:04:57,200 Speaker 1: ninety nine. 92 00:04:57,839 --> 00:04:59,120 Speaker 3: At that point there. 93 00:04:59,080 --> 00:05:04,200 Speaker 1: Was work on rules based engines, analytics. 94 00:05:04,360 --> 00:05:05,440 Speaker 3: All of this was happening. 95 00:05:05,960 --> 00:05:10,200 Speaker 1: So it all depends on how you really define that term. 96 00:05:10,480 --> 00:05:15,880 Speaker 1: You could argue that elements of statistics, probability, it's not 97 00:05:15,920 --> 00:05:18,920 Speaker 1: exactly AI, but it certainly feeds into it. And so 98 00:05:19,240 --> 00:05:22,480 Speaker 1: I feel like we've been working on this topic of 99 00:05:22,520 --> 00:05:28,400 Speaker 1: how do we deliver better insights better automation since IBM 100 00:05:28,480 --> 00:05:31,040 Speaker 1: was formed. If you read about what Thomas Watson Junior did, 101 00:05:31,040 --> 00:05:35,719 Speaker 1: that was all about automating tasks that AI well, probably 102 00:05:35,880 --> 00:05:39,680 Speaker 1: certainly not by today's definition, but it's in the same 103 00:05:39,760 --> 00:05:40,360 Speaker 1: zip code. 104 00:05:40,480 --> 00:05:42,640 Speaker 2: So from your perspective, it feels a lot more like 105 00:05:42,680 --> 00:05:44,080 Speaker 2: an evolution than a revolution. 106 00:05:44,240 --> 00:05:47,400 Speaker 1: Is that a fair statement, yes, which I think most 107 00:05:47,480 --> 00:05:51,680 Speaker 1: great things in technology tend to happen that way. Many 108 00:05:51,720 --> 00:05:54,720 Speaker 1: of the revolutions, if you will, tend to fizzle out. 109 00:05:55,760 --> 00:05:57,760 Speaker 2: But even given that is there, I guess what I'm 110 00:05:57,800 --> 00:06:00,560 Speaker 2: asking is, I'm curious about whether there was a a 111 00:06:00,680 --> 00:06:04,480 Speaker 2: moment in that evolution when you had to readjust your 112 00:06:04,480 --> 00:06:08,479 Speaker 2: expectations about what AI was going to be capable of. 113 00:06:08,600 --> 00:06:12,000 Speaker 2: I mean, was there, you know, was there a particular 114 00:06:12,360 --> 00:06:15,800 Speaker 2: innovation or a particular problem that was solved that made 115 00:06:15,800 --> 00:06:18,960 Speaker 2: you think, oh, this is different than what I thought. 116 00:06:22,080 --> 00:06:25,760 Speaker 1: I would say the moments that caught our attention certainly 117 00:06:27,080 --> 00:06:31,119 Speaker 1: casper Off winning the chess tournament Nobody or Deep Blue 118 00:06:31,120 --> 00:06:34,039 Speaker 1: beating casper Off. I should say, nobody really thought that 119 00:06:34,120 --> 00:06:39,719 Speaker 1: was possible before that, and then it was Watson winning Jeopardy. 120 00:06:40,120 --> 00:06:42,800 Speaker 1: These were moments that said, maybe there's more here than 121 00:06:42,800 --> 00:06:47,000 Speaker 1: we even thought was possible. And so I do think 122 00:06:47,040 --> 00:06:53,400 Speaker 1: there's points in time where we realized maybe way more could. 123 00:06:53,279 --> 00:06:55,359 Speaker 3: Be done than we had even imagined. 124 00:06:56,040 --> 00:07:00,280 Speaker 1: But I do think it's consistent progress every month and 125 00:07:00,360 --> 00:07:04,000 Speaker 1: every year versus some seminal moment. 126 00:07:04,320 --> 00:07:04,520 Speaker 3: Now. 127 00:07:04,600 --> 00:07:08,200 Speaker 1: Certainly, large language models as of recent have caught everybody's 128 00:07:08,200 --> 00:07:12,760 Speaker 1: attention because it has a direct consumer application. But I 129 00:07:12,760 --> 00:07:16,920 Speaker 1: would almost think of that as what Netscape was for 130 00:07:16,960 --> 00:07:20,000 Speaker 1: the for the web browser. Yeah, it brought the Internet 131 00:07:20,040 --> 00:07:24,360 Speaker 1: to everybody, but that didn't become the Internet per se. 132 00:07:25,160 --> 00:07:25,400 Speaker 3: Yeah. 133 00:07:25,800 --> 00:07:28,400 Speaker 2: I have a cousin who worked for IBM for forty 134 00:07:28,440 --> 00:07:31,239 Speaker 2: one years. I saw him this weekend. He's in Toronto. 135 00:07:31,840 --> 00:07:34,720 Speaker 2: By the way, I said, do you work for Rob Thomas. 136 00:07:34,840 --> 00:07:35,640 Speaker 3: He went like this. 137 00:07:35,800 --> 00:07:43,120 Speaker 2: He goes, he said, I'm five layers down. But so 138 00:07:43,160 --> 00:07:45,080 Speaker 2: I always whenever I see my cousin, I ask him, 139 00:07:45,360 --> 00:07:46,880 Speaker 2: can you tell me again what you do? Because it's 140 00:07:46,920 --> 00:07:49,400 Speaker 2: always changing, right, I guess this is a function of 141 00:07:49,440 --> 00:07:52,920 Speaker 2: working at IBM. So eventually he just gives up and says, 142 00:07:53,000 --> 00:07:54,920 Speaker 2: you know, we're just solving problems. So what we're doing, 143 00:07:55,240 --> 00:07:57,800 Speaker 2: which I sort of loved as a kind of frame, 144 00:07:58,320 --> 00:08:01,160 Speaker 2: And I was curious, What's what's the coolest problem you 145 00:08:01,240 --> 00:08:05,760 Speaker 2: ever worked on? Not biggest, not most important, but the coolest, 146 00:08:05,760 --> 00:08:08,280 Speaker 2: the one that's like that sort of makes you smile 147 00:08:08,320 --> 00:08:09,520 Speaker 2: when you think back on it. 148 00:08:09,600 --> 00:08:14,360 Speaker 1: Probably when I was in microelectronics, because it was a 149 00:08:14,360 --> 00:08:17,440 Speaker 1: world I had no exposure to. I hadn't studied computer science, 150 00:08:18,520 --> 00:08:24,640 Speaker 1: and we were building a lot of high performance semiconductor technology, 151 00:08:24,720 --> 00:08:27,520 Speaker 1: so just chips that do a really great job of 152 00:08:28,040 --> 00:08:32,040 Speaker 1: processing something or other. And we figured out that there 153 00:08:32,120 --> 00:08:35,920 Speaker 1: was a market in consumer gaming that was starting to happen, 154 00:08:36,679 --> 00:08:39,400 Speaker 1: and we got to the point where we became the 155 00:08:39,480 --> 00:08:47,160 Speaker 1: chip inside the Nintendo. We the Microsoft Xbox Sony PlayStation, 156 00:08:47,360 --> 00:08:50,360 Speaker 1: so we basically had the entire gaming market running on 157 00:08:50,480 --> 00:08:52,360 Speaker 1: IBM chips and. 158 00:08:52,520 --> 00:08:57,359 Speaker 2: To use every parent basically is pointing at you and saying. 159 00:08:57,440 --> 00:09:02,200 Speaker 1: You're the Probably well, they would have found it from anybody. 160 00:09:02,520 --> 00:09:06,760 Speaker 1: But it was the first time I could explain my 161 00:09:06,920 --> 00:09:08,880 Speaker 1: job to my kids, who were quite young at that time, 162 00:09:09,040 --> 00:09:11,680 Speaker 1: like what I did, Like it was more tangible for 163 00:09:11,720 --> 00:09:14,160 Speaker 1: them than saying we solve problems or douce you know, 164 00:09:14,200 --> 00:09:18,520 Speaker 1: build solutions like it became very tangible for them, and 165 00:09:18,880 --> 00:09:21,599 Speaker 1: I think that's, you know, a rewarding part of the 166 00:09:21,679 --> 00:09:24,480 Speaker 1: job is when you can help your family actually understand 167 00:09:24,520 --> 00:09:26,040 Speaker 1: what you do. Most people can't do that. It's probably 168 00:09:26,080 --> 00:09:27,840 Speaker 1: easier for you. They can, they can see the books, 169 00:09:29,280 --> 00:09:32,040 Speaker 1: but for for some of us in the business the 170 00:09:32,120 --> 00:09:34,320 Speaker 1: business world, it's not always as obvious. So that was 171 00:09:34,360 --> 00:09:37,160 Speaker 1: like one example where the dots really connected. 172 00:09:38,120 --> 00:09:41,240 Speaker 2: There were a couple there's a couple of stuck about 173 00:09:41,280 --> 00:09:43,760 Speaker 2: a little bit of this in the context of of AI. 174 00:09:43,840 --> 00:09:47,040 Speaker 2: I love because I love the frame of problem solving 175 00:09:47,080 --> 00:09:49,400 Speaker 2: as a way of understanding what the function of the 176 00:09:49,440 --> 00:09:52,480 Speaker 2: technology is. So I know that you guys did something, 177 00:09:52,920 --> 00:09:56,360 Speaker 2: did some work with I never know how to pronounce 178 00:09:56,400 --> 00:10:00,400 Speaker 2: it is it Sevilla Sevilla with the football club Severe 179 00:10:00,679 --> 00:10:03,840 Speaker 2: in Spain. Tell me about Tell me a little bit 180 00:10:03,840 --> 00:10:06,640 Speaker 2: about that. What problem were they trying to solve and 181 00:10:06,679 --> 00:10:07,600 Speaker 2: why did they call you? 182 00:10:07,679 --> 00:10:15,679 Speaker 1: In Every sports franchise is trying to get an advantage, right, 183 00:10:15,720 --> 00:10:21,719 Speaker 1: Let's just be that clear. Everybody's how can I use data, analytics, insights, 184 00:10:21,800 --> 00:10:24,920 Speaker 1: anything that will make us one percent better on the 185 00:10:24,960 --> 00:10:30,720 Speaker 1: field at some point in the future. And Seville reached 186 00:10:30,720 --> 00:10:33,520 Speaker 1: out to us because they had seen some of the 187 00:10:33,600 --> 00:10:35,800 Speaker 1: We've done some work with the Toronto Raptors in the 188 00:10:35,800 --> 00:10:40,520 Speaker 1: past and others, and their thought was maybe there's something 189 00:10:40,520 --> 00:10:44,000 Speaker 1: we could do. They'd heard all about generative AI, they 190 00:10:44,040 --> 00:10:47,319 Speaker 1: heard about large language models. And the problem, back to 191 00:10:47,360 --> 00:10:51,240 Speaker 1: your point on solving problems, was we want to do 192 00:10:51,280 --> 00:10:56,120 Speaker 1: a way better job of assessing talent, because really the 193 00:10:56,800 --> 00:10:59,719 Speaker 1: lifeblood of a sports franchise is can you continue to 194 00:10:59,720 --> 00:11:03,800 Speaker 1: cult a talent, Can you find talent that others don't find? 195 00:11:04,440 --> 00:11:06,920 Speaker 1: Can you see something in somebody that they don't see 196 00:11:06,920 --> 00:11:08,200 Speaker 1: in themselves or maybe no other. 197 00:11:08,120 --> 00:11:09,240 Speaker 3: Team season them. 198 00:11:09,600 --> 00:11:13,520 Speaker 1: And we ended up building somebody with them called Scout Advisor, 199 00:11:14,400 --> 00:11:18,800 Speaker 1: which is built on Watson X, which basically just ingests 200 00:11:19,960 --> 00:11:23,559 Speaker 1: tons and tons of data, and we like to think 201 00:11:23,559 --> 00:11:25,760 Speaker 1: of it as finding you know, the needle in the 202 00:11:25,800 --> 00:11:30,240 Speaker 1: haystack of you know, here's three players that aren't being considered. 203 00:11:30,559 --> 00:11:34,880 Speaker 1: They're not on the top teams today, and I think 204 00:11:35,000 --> 00:11:37,320 Speaker 1: working with them together we found some pretty good insights 205 00:11:37,320 --> 00:11:37,959 Speaker 1: that's helped them out. 206 00:11:38,160 --> 00:11:40,720 Speaker 2: How What was intriguing to me was we're not just 207 00:11:40,760 --> 00:11:45,840 Speaker 2: talking about quantitative data. We're also talking about qualitative data. 208 00:11:46,480 --> 00:11:48,839 Speaker 2: But that's the puzzle part of the thing that fastens me. 209 00:11:49,080 --> 00:11:52,720 Speaker 2: How does one incorporate qualitative analysis into that sort of 210 00:11:53,320 --> 00:11:56,400 Speaker 2: so you just feeding in scouting reports and things like that. 211 00:11:58,280 --> 00:11:59,760 Speaker 1: I got to realize, think about how much I can 212 00:11:59,800 --> 00:12:05,080 Speaker 1: act actually disclosed it. But if you think about so, 213 00:12:05,280 --> 00:12:11,080 Speaker 1: quantitative is relatively easy. Every team collects that, you know, 214 00:12:11,640 --> 00:12:14,680 Speaker 1: what's their forty yard dash? They use that term, certainly 215 00:12:14,720 --> 00:12:19,920 Speaker 1: not in Spain. That's all quantitative. Qualitative is what's happening 216 00:12:19,960 --> 00:12:24,119 Speaker 1: off the field. It could be diet, it could be habits, 217 00:12:24,280 --> 00:12:28,000 Speaker 1: it could be behavior. You can imagine a range of 218 00:12:28,040 --> 00:12:32,800 Speaker 1: things that would all feed into an athlete's performance and 219 00:12:32,880 --> 00:12:34,400 Speaker 1: so relationships. 220 00:12:35,600 --> 00:12:37,120 Speaker 3: There's many different aspects, and. 221 00:12:37,080 --> 00:12:40,439 Speaker 1: So it's trying to figure out the right blend of 222 00:12:40,559 --> 00:12:43,680 Speaker 1: quantitative and qualitative that gives you a unique insight. 223 00:12:44,320 --> 00:12:46,680 Speaker 2: How transparent is that kind of system? I mean, is 224 00:12:46,679 --> 00:12:51,040 Speaker 2: it telling you it's saying pick this guy not this guy, 225 00:12:51,080 --> 00:12:52,880 Speaker 2: But is it telling you why it prefers this guy 226 00:12:52,920 --> 00:12:53,319 Speaker 2: to this guy? 227 00:12:53,480 --> 00:12:53,599 Speaker 3: Is that? 228 00:12:54,840 --> 00:12:57,000 Speaker 1: I think for anything in the realm of AI, you 229 00:12:57,080 --> 00:13:00,640 Speaker 1: have to answer the why question, otherwise you fall into 230 00:13:00,640 --> 00:13:04,800 Speaker 1: the trap of the you know, the proverbial black box, 231 00:13:05,080 --> 00:13:07,840 Speaker 1: and then wait, I made this decision, I'd never understood 232 00:13:07,840 --> 00:13:09,280 Speaker 1: why it didn't work out. 233 00:13:09,520 --> 00:13:11,880 Speaker 3: So you always have to answer why without a doubt? 234 00:13:12,840 --> 00:13:14,160 Speaker 2: And how is why? Answered? 235 00:13:16,679 --> 00:13:20,679 Speaker 1: Sources of data, the reasoning that went into it, and 236 00:13:20,800 --> 00:13:24,040 Speaker 1: so it's basically just tracing back the chain of how 237 00:13:24,080 --> 00:13:26,960 Speaker 1: you got to the answer. And in the case of 238 00:13:27,160 --> 00:13:29,640 Speaker 1: what we do in Watson X is we have IBM models. 239 00:13:30,080 --> 00:13:32,719 Speaker 1: We also use some other open source models, So it 240 00:13:32,720 --> 00:13:35,560 Speaker 1: would be which model was used, what was the data 241 00:13:35,600 --> 00:13:37,800 Speaker 1: set that was fed into that model, How is it 242 00:13:37,840 --> 00:13:38,600 Speaker 1: making decisions? 243 00:13:38,600 --> 00:13:41,840 Speaker 3: How is it performing? Is it robust? 244 00:13:42,040 --> 00:13:44,280 Speaker 1: Meaning is it reliable in terms of if you feed 245 00:13:44,320 --> 00:13:46,080 Speaker 1: it two of the same data set, do you get 246 00:13:46,120 --> 00:13:49,040 Speaker 1: the same answer. These are all the you know, the 247 00:13:49,080 --> 00:13:51,040 Speaker 1: technical aspects of understanding the why. 248 00:13:52,120 --> 00:13:56,240 Speaker 2: How quickly do you expect all professional sports franchises to 249 00:13:56,320 --> 00:13:58,440 Speaker 2: adopt some kind of are they already there? If I 250 00:13:58,480 --> 00:14:02,120 Speaker 2: went out and pulled the general managers of the one 251 00:14:02,200 --> 00:14:05,080 Speaker 2: hundred most valuable sports franchises in the world, how many 252 00:14:05,120 --> 00:14:07,720 Speaker 2: of them would be using some kind of AI system 253 00:14:07,760 --> 00:14:09,000 Speaker 2: to assist in their efforts. 254 00:14:10,880 --> 00:14:14,600 Speaker 1: One hundred and twenty percent would, meaning that everybody's doing it, 255 00:14:14,640 --> 00:14:16,480 Speaker 1: and some think they're doing way more than they probably 256 00:14:16,520 --> 00:14:20,120 Speaker 1: actually are. So everybody's doing it. I think what's weird 257 00:14:20,160 --> 00:14:25,680 Speaker 1: about sports is everybody's so convinced that what they're doing 258 00:14:25,800 --> 00:14:30,120 Speaker 1: is unique that they generally speaking, don't want to work 259 00:14:30,160 --> 00:14:32,520 Speaker 1: with a third party to do it because they're afraid 260 00:14:32,680 --> 00:14:35,320 Speaker 1: that that would expose them. But in reality, I think 261 00:14:35,360 --> 00:14:38,200 Speaker 1: most are doing eighty to ninety percent of the same things. 262 00:14:39,840 --> 00:14:42,640 Speaker 3: So but without a doubt, everybody's doing it. Yeah. 263 00:14:43,240 --> 00:14:47,200 Speaker 2: Yeah. The other I say that I loved was there 264 00:14:47,240 --> 00:14:51,120 Speaker 2: was one but a shipping line tricon on the Mississippi River. 265 00:14:52,120 --> 00:14:53,920 Speaker 2: Tell me a little bit about that project. What problem 266 00:14:54,000 --> 00:14:54,840 Speaker 2: were they trying to solve? 267 00:14:56,920 --> 00:15:00,280 Speaker 1: Think about the problem that I would say every body 268 00:15:00,360 --> 00:15:04,080 Speaker 1: noticed if you go back to twenty twenty was things 269 00:15:04,120 --> 00:15:06,760 Speaker 1: are getting hold held up in ports. It was actually 270 00:15:06,760 --> 00:15:09,040 Speaker 1: an article in the paper this morning kind of tracing 271 00:15:09,040 --> 00:15:12,520 Speaker 1: the history of what happened twenty twenty twenty one and 272 00:15:12,760 --> 00:15:15,520 Speaker 1: why ships were basically sitting at seas for months at 273 00:15:15,520 --> 00:15:19,160 Speaker 1: a time. And at that stage we just we had 274 00:15:19,200 --> 00:15:24,640 Speaker 1: a massive throughput issue. But moving even beyond the pandemic, 275 00:15:24,720 --> 00:15:28,880 Speaker 1: you can see it now with ships getting through like 276 00:15:28,960 --> 00:15:32,160 Speaker 1: Panama Canal, there's like a narrow window where you can 277 00:15:32,200 --> 00:15:35,760 Speaker 1: get through, and if you don't have your paperwork done, 278 00:15:36,440 --> 00:15:38,600 Speaker 1: you don't have the right approvals, you're not going through 279 00:15:38,640 --> 00:15:40,080 Speaker 1: and it may cost you a day or two and 280 00:15:40,080 --> 00:15:43,000 Speaker 1: that's a lot of money. In the shipping industry and 281 00:15:43,080 --> 00:15:46,800 Speaker 1: the Tricon example, it's really just about when you're pulling 282 00:15:46,840 --> 00:15:51,520 Speaker 1: into a port, if you have the right paperwork done, 283 00:15:52,200 --> 00:15:56,040 Speaker 1: you can get goods off the ship very quickly. They 284 00:15:56,120 --> 00:16:00,600 Speaker 1: ship a lot of food, which by definition, since it's 285 00:16:00,600 --> 00:16:04,080 Speaker 1: not packaged food, it's fresh food, there is an expiration 286 00:16:04,160 --> 00:16:08,240 Speaker 1: period and so if it takes them an extra two hours, 287 00:16:09,320 --> 00:16:12,800 Speaker 1: certainly multiple hours or a day, they have a massive 288 00:16:12,840 --> 00:16:15,200 Speaker 1: problem because then you're going to deal with spoilage and 289 00:16:15,240 --> 00:16:17,840 Speaker 1: so it's going to set you back. And what we've 290 00:16:17,840 --> 00:16:21,280 Speaker 1: worked with them on is using an assistant that we've 291 00:16:21,280 --> 00:16:25,680 Speaker 1: built in Watson X called orchestrate, which basically is just 292 00:16:26,360 --> 00:16:31,960 Speaker 1: AI doing digital labor, so we can replicate nearly any 293 00:16:32,080 --> 00:16:35,920 Speaker 1: repetitive task and do that with software. 294 00:16:35,720 --> 00:16:36,560 Speaker 3: Instead of humans. 295 00:16:37,480 --> 00:16:40,960 Speaker 1: So, as you may imagine, shipping industry still has a 296 00:16:40,960 --> 00:16:43,920 Speaker 1: lot of paperwork that goes on, and so being able 297 00:16:44,000 --> 00:16:47,000 Speaker 1: to take forms that normally would be multiple hours of 298 00:16:47,080 --> 00:16:49,120 Speaker 1: filling it out, Oh this isn't right, send it back. 299 00:16:49,640 --> 00:16:53,280 Speaker 1: We've basically built that as a digital skill inside of 300 00:16:53,600 --> 00:16:58,040 Speaker 1: watsonex orchestrate, and so now it's done in minutes. 301 00:16:59,040 --> 00:17:01,720 Speaker 2: They did Did they realize that they could have that 302 00:17:01,800 --> 00:17:04,080 Speaker 2: kind of efficiency by teaming up with you or is 303 00:17:04,119 --> 00:17:08,119 Speaker 2: that something you came to them and said, guys, we 304 00:17:08,160 --> 00:17:09,480 Speaker 2: can do this way better than you think. 305 00:17:09,640 --> 00:17:10,080 Speaker 3: What's the. 306 00:17:11,920 --> 00:17:15,439 Speaker 1: I'd say it's always, it's always both sides coming together 307 00:17:15,600 --> 00:17:18,520 Speaker 1: at a moment that for some reason makes sense because 308 00:17:19,720 --> 00:17:21,520 Speaker 1: you could say, why didn't this happen like five years ago, 309 00:17:21,600 --> 00:17:25,880 Speaker 1: like seems so obvious. Well, technology wasn't quite ready then, 310 00:17:26,400 --> 00:17:28,560 Speaker 1: I would say, But they knew they had a need 311 00:17:29,040 --> 00:17:32,600 Speaker 1: because I forget what the precise number is, but you know, 312 00:17:32,840 --> 00:17:36,639 Speaker 1: reduction of spoilage has massive impact on their bottom line, 313 00:17:38,640 --> 00:17:41,320 Speaker 1: and so they knew they had a need, we. 314 00:17:41,280 --> 00:17:44,480 Speaker 3: Thought we could solve it, and the two together. 315 00:17:44,920 --> 00:17:47,879 Speaker 2: Who did you guys go to them thought? Or did 316 00:17:47,880 --> 00:17:48,520 Speaker 2: they come to you? 317 00:17:48,800 --> 00:17:52,159 Speaker 1: I recall that this one was an inbound meaning they 318 00:17:52,160 --> 00:17:55,199 Speaker 1: had reached out to IBM and that we'd like to 319 00:17:55,200 --> 00:17:57,159 Speaker 1: solve this problem. I think it went into one of 320 00:17:57,200 --> 00:17:59,800 Speaker 1: our digital centers, if I if I recall so literary, 321 00:17:59,800 --> 00:18:01,200 Speaker 1: I call yeah. 322 00:18:01,240 --> 00:18:05,480 Speaker 2: But the other the reverse is more interesting to me 323 00:18:05,840 --> 00:18:08,000 Speaker 2: because there seems to be a very very large universe 324 00:18:08,040 --> 00:18:10,800 Speaker 2: of people who have problems that could be solved this 325 00:18:10,840 --> 00:18:12,240 Speaker 2: way and they don't realize it. 326 00:18:13,119 --> 00:18:13,800 Speaker 3: What's your. 327 00:18:15,359 --> 00:18:18,320 Speaker 2: Is there a shining example of this of someone you 328 00:18:18,440 --> 00:18:20,760 Speaker 2: just can't you just think could benefit so much and 329 00:18:20,960 --> 00:18:22,120 Speaker 2: isn't benefiting right now? 330 00:18:24,880 --> 00:18:26,320 Speaker 3: Maybe I'll answer it slightly differently. 331 00:18:26,480 --> 00:18:31,280 Speaker 1: I'm I'm surprised by how many people can benefit that 332 00:18:31,359 --> 00:18:33,080 Speaker 1: you wouldn't even logically think of. 333 00:18:33,520 --> 00:18:34,920 Speaker 3: First, let me give you an example. 334 00:18:35,960 --> 00:18:43,000 Speaker 1: There's a franchiser of hair salons, sport Clips is the name. 335 00:18:44,200 --> 00:18:46,359 Speaker 1: My sons used to go there for haircuts because they 336 00:18:46,359 --> 00:18:48,719 Speaker 1: have like TVs and you can watch sports, so they 337 00:18:48,760 --> 00:18:50,959 Speaker 1: loved that they got entertained while they would get their haircut. 338 00:18:51,960 --> 00:18:53,879 Speaker 1: I think the last place that you would think is 339 00:18:53,960 --> 00:19:00,320 Speaker 1: using AI today would be a franchiser of hair salons. Yeah, 340 00:18:59,760 --> 00:19:04,280 Speaker 1: but just follow it through. The biggest part of how 341 00:19:04,320 --> 00:19:06,440 Speaker 1: they run their business is can I get people to 342 00:19:06,480 --> 00:19:10,600 Speaker 1: cut hair? And this is the high turnover industry because 343 00:19:10,600 --> 00:19:12,080 Speaker 1: there's a lot of different places you can work if 344 00:19:12,119 --> 00:19:14,560 Speaker 1: you want to cut hair. People actually get injured cutting 345 00:19:14,560 --> 00:19:16,280 Speaker 1: hair because you're on your feet all day, that type 346 00:19:16,280 --> 00:19:21,480 Speaker 1: of thing. And they're using same technology orchestrate as part 347 00:19:21,560 --> 00:19:25,360 Speaker 1: of their recruiting process. How can they automate a lot 348 00:19:25,400 --> 00:19:31,240 Speaker 1: of people submitting resumes, who they speak to, how they qualify. 349 00:19:30,800 --> 00:19:31,760 Speaker 3: Them for the position. 350 00:19:32,520 --> 00:19:35,080 Speaker 1: And so the reason I give that example is the 351 00:19:35,520 --> 00:19:40,960 Speaker 1: opportunity for AI, which is unlike other technologies, is truly unlimited. 352 00:19:42,560 --> 00:19:46,159 Speaker 1: It will touch every single business. It's not the realm 353 00:19:46,200 --> 00:19:49,000 Speaker 1: of the fortune five hundred or the fortune one thousand. 354 00:19:49,800 --> 00:19:53,240 Speaker 1: This is the fortune any size. And I think that 355 00:19:53,280 --> 00:19:56,119 Speaker 1: may be one thing that people underestimate about AI. 356 00:19:56,640 --> 00:19:59,359 Speaker 2: Yeah, what about I mean I was thinking about education 357 00:19:59,680 --> 00:20:02,480 Speaker 2: as as a kind of I mean, education is a 358 00:20:02,560 --> 00:20:08,240 Speaker 2: perennial whipping boy for you guys that are living in 359 00:20:08,240 --> 00:20:11,760 Speaker 2: the nineteenth century, right. I'm just curious about if a 360 00:20:13,400 --> 00:20:16,040 Speaker 2: superintendent of a public school system or the president of 361 00:20:16,080 --> 00:20:19,320 Speaker 2: the university sat down and had lunch with you and said, 362 00:20:21,000 --> 00:20:23,480 Speaker 2: do the university first. My cost are out of control, 363 00:20:24,160 --> 00:20:29,960 Speaker 2: my enrollment is down, my students hate me, and my 364 00:20:30,040 --> 00:20:31,200 Speaker 2: board is revolting. 365 00:20:31,359 --> 00:20:31,639 Speaker 3: Help. 366 00:20:33,560 --> 00:20:37,640 Speaker 2: How would you think about helping someone in that situation. 367 00:20:39,240 --> 00:20:41,720 Speaker 3: I spend some time with universities. I like to go 368 00:20:41,800 --> 00:20:42,520 Speaker 3: back and there's. 369 00:20:42,359 --> 00:20:46,840 Speaker 1: Alma maters where I went to school, and so I 370 00:20:46,880 --> 00:20:50,000 Speaker 1: do that every year. The challenge I have hall of 371 00:20:50,040 --> 00:20:53,200 Speaker 1: Seming University is there has to be a will. Yeah, 372 00:20:53,680 --> 00:20:55,840 Speaker 1: and I'm not sure the incentives are quite right today 373 00:20:57,040 --> 00:21:00,679 Speaker 1: because bringing in new technology, say we want to go 374 00:21:00,720 --> 00:21:05,080 Speaker 1: after we can help you figure out student recruiting or 375 00:21:05,119 --> 00:21:10,119 Speaker 1: how you automate more of your education, everybody suddenly feels 376 00:21:10,119 --> 00:21:11,120 Speaker 1: threatened that university. 377 00:21:11,680 --> 00:21:12,760 Speaker 3: Hold on, that's my job. 378 00:21:13,320 --> 00:21:15,680 Speaker 1: I'm the one that decides that, or I'm the one 379 00:21:15,720 --> 00:21:18,760 Speaker 1: that wants to dictate the course. So there has to 380 00:21:18,760 --> 00:21:22,639 Speaker 1: be a will. So I think it's very possible, and 381 00:21:23,520 --> 00:21:25,960 Speaker 1: I do think over the next decade you will see 382 00:21:26,000 --> 00:21:28,360 Speaker 1: some universities that jump all over this and they will 383 00:21:28,400 --> 00:21:30,760 Speaker 1: move ahead, and you see others that do not. 384 00:21:31,640 --> 00:21:33,560 Speaker 3: Because it's very possible. 385 00:21:35,160 --> 00:21:38,119 Speaker 2: Where how does when you say there has to be 386 00:21:38,160 --> 00:21:41,040 Speaker 2: a will? Is that the kind is that a kind 387 00:21:41,080 --> 00:21:43,280 Speaker 2: of thing that that people that IBM to think about, 388 00:21:44,000 --> 00:21:47,560 Speaker 2: Like when in this conversation you hypothetical conversation you might 389 00:21:47,600 --> 00:21:51,320 Speaker 2: have with the university president, would you give advice on 390 00:21:51,920 --> 00:21:55,320 Speaker 2: where the will comes from? 391 00:21:55,400 --> 00:21:57,720 Speaker 1: I don't do that as much in a university context. 392 00:21:57,720 --> 00:22:01,520 Speaker 1: I do that every day in a business context, because 393 00:22:02,280 --> 00:22:04,320 Speaker 1: if you can find the right person in a business 394 00:22:04,320 --> 00:22:08,440 Speaker 1: that wants to focus on growth or the bottom line 395 00:22:08,960 --> 00:22:11,560 Speaker 1: or how do you create more productivity. Yes, it's going 396 00:22:11,560 --> 00:22:15,679 Speaker 1: to create a lot of organizational resistance potentially, but you 397 00:22:15,720 --> 00:22:17,840 Speaker 1: can find somebody that will figure out how to push 398 00:22:17,920 --> 00:22:23,000 Speaker 1: that through. I think for universities, I think that's also possible. 399 00:22:23,280 --> 00:22:26,160 Speaker 1: I'm not sure there's there's there's a return on investment for. 400 00:22:26,160 --> 00:22:26,679 Speaker 3: Us to do that. 401 00:22:27,000 --> 00:22:34,000 Speaker 2: Yeah, yeah, yeah, God, let's let's find some terms. AI 402 00:22:34,160 --> 00:22:37,360 Speaker 2: years I told you'd like to use What does that mean? 403 00:22:39,119 --> 00:22:41,840 Speaker 1: We just started using this term literally in the last 404 00:22:41,880 --> 00:22:47,640 Speaker 1: three months, and it was it was what we observed internally, 405 00:22:48,640 --> 00:22:51,760 Speaker 1: which is most technology you build, you say, all right, 406 00:22:51,800 --> 00:22:54,480 Speaker 1: what's going to happen in year one, year two, year three, 407 00:22:54,800 --> 00:22:59,080 Speaker 1: and it's you know, largely by by a calendar. AI 408 00:22:59,240 --> 00:23:01,199 Speaker 1: years are the idea that what used to be a 409 00:23:01,280 --> 00:23:05,440 Speaker 1: year is now like a week. And that is how 410 00:23:05,480 --> 00:23:06,960 Speaker 1: fast the technology is moving. 411 00:23:07,680 --> 00:23:09,720 Speaker 3: And do you give you an example. We had one 412 00:23:09,760 --> 00:23:10,760 Speaker 3: client we're working with. 413 00:23:11,640 --> 00:23:15,080 Speaker 1: They're using one of our granite models, and the results 414 00:23:15,119 --> 00:23:17,760 Speaker 1: they were getting were not very good. Accuracy was not there, 415 00:23:17,840 --> 00:23:20,720 Speaker 1: their performance was not there. So I was like scratching 416 00:23:20,760 --> 00:23:23,679 Speaker 1: my head. I was like, what is going on? They 417 00:23:23,760 --> 00:23:27,320 Speaker 1: were financial services, the bank, So I'm scratching my head, 418 00:23:27,359 --> 00:23:29,239 Speaker 1: like what is going on? Everybody else is getting this 419 00:23:29,359 --> 00:23:33,119 Speaker 1: and like these results are horrible. And I said to 420 00:23:33,119 --> 00:23:35,560 Speaker 1: the team, which version of the model are you using? 421 00:23:36,680 --> 00:23:40,240 Speaker 1: This was in February, Like we're using the one from October. 422 00:23:41,440 --> 00:23:43,399 Speaker 1: I was like, all right, now we know precisely the 423 00:23:43,440 --> 00:23:47,480 Speaker 1: problem because the model from October is effectively useless now 424 00:23:47,520 --> 00:23:48,639 Speaker 1: since we're here in February. 425 00:23:49,240 --> 00:23:53,320 Speaker 2: Serious, actually useless, completely useless. 426 00:23:53,520 --> 00:23:56,480 Speaker 1: Yeah, that is how fast this is changing. And so 427 00:23:56,960 --> 00:24:01,160 Speaker 1: the minute, same use case, same day, you give them 428 00:24:01,160 --> 00:24:05,720 Speaker 1: the model from late January instead of October, the results 429 00:24:05,760 --> 00:24:06,480 Speaker 1: are off the charts. 430 00:24:07,040 --> 00:24:07,520 Speaker 3: Yeah. 431 00:24:07,720 --> 00:24:10,640 Speaker 2: Wait, so what exactly happened between October and January? 432 00:24:10,840 --> 00:24:11,920 Speaker 3: The model got way better? 433 00:24:12,480 --> 00:24:14,199 Speaker 2: Could dig into that, Like, what do you mean by 434 00:24:14,240 --> 00:24:14,639 Speaker 2: the way. 435 00:24:14,520 --> 00:24:15,720 Speaker 3: We are constant. 436 00:24:15,760 --> 00:24:20,360 Speaker 1: We have built large compute infrastructure where we're doing model training. 437 00:24:21,000 --> 00:24:23,439 Speaker 1: And to be clear, model training is the realm of 438 00:24:23,560 --> 00:24:27,760 Speaker 1: probably in the world my guess is five to ten companies. 439 00:24:28,840 --> 00:24:29,200 Speaker 3: And so. 440 00:24:30,720 --> 00:24:33,880 Speaker 1: You build a model, you're constantly training it, you're doing 441 00:24:33,920 --> 00:24:37,440 Speaker 1: fine tuning, you're doing more training, you're adding data every day, 442 00:24:37,480 --> 00:24:41,720 Speaker 1: every hour it gets better. And so how does it 443 00:24:41,760 --> 00:24:44,280 Speaker 1: do that. You're feeding it more data, you're feeding it 444 00:24:44,359 --> 00:24:49,320 Speaker 1: more live examples. We're using things like synthetic data at 445 00:24:49,320 --> 00:24:51,520 Speaker 1: this point, which is we're basically creating data to do 446 00:24:51,560 --> 00:24:54,800 Speaker 1: the training as well. All of this feeds into how 447 00:24:54,880 --> 00:24:59,359 Speaker 1: useful the model is. And so using the October model, 448 00:24:59,400 --> 00:25:02,160 Speaker 1: those were the results in October, just a fact, that's 449 00:25:02,160 --> 00:25:05,720 Speaker 1: how good it was then. But back to the concept 450 00:25:05,720 --> 00:25:08,800 Speaker 1: of AI years, two weeks is a long time. 451 00:25:10,000 --> 00:25:12,919 Speaker 2: Is that Are we in a steep part of the 452 00:25:12,960 --> 00:25:15,760 Speaker 2: model learning carve or do you expect this to continue 453 00:25:15,800 --> 00:25:17,520 Speaker 2: along this at this pace? 454 00:25:19,119 --> 00:25:23,359 Speaker 3: I think that is the big question and don't have 455 00:25:23,400 --> 00:25:24,080 Speaker 3: an answer yet. 456 00:25:24,480 --> 00:25:26,480 Speaker 1: By definition, at some point you would think it would 457 00:25:26,520 --> 00:25:29,000 Speaker 1: have to slow down a bit, but it's not obvious 458 00:25:29,040 --> 00:25:30,919 Speaker 1: that that is on the horizon. 459 00:25:31,000 --> 00:25:34,880 Speaker 2: Still speeding up. Yes, how fast. Can it get. 460 00:25:37,000 --> 00:25:40,159 Speaker 1: We've debated, can you actually have better results in the 461 00:25:40,200 --> 00:25:44,960 Speaker 1: afternoon than you did in the morning. Really it's nuts, Yeah, 462 00:25:44,960 --> 00:25:47,919 Speaker 1: I know, but that's why we came up with this term, 463 00:25:47,920 --> 00:25:50,000 Speaker 1: because I think you also have to think of like 464 00:25:50,600 --> 00:25:51,560 Speaker 1: concepts that. 465 00:25:53,680 --> 00:25:54,679 Speaker 3: Gets people's attention. 466 00:25:54,880 --> 00:25:58,040 Speaker 2: So you're basically turning into a bakery. You're like the 467 00:25:58,119 --> 00:26:00,359 Speaker 2: bread from yesterday. You know you can have it for 468 00:26:00,440 --> 00:26:04,040 Speaker 2: twenty five cents. But I mean you do proferential pricing. 469 00:26:04,080 --> 00:26:08,679 Speaker 2: You could say, we'll judge you x for yesterday's model, 470 00:26:09,119 --> 00:26:10,560 Speaker 2: two x for today's model. 471 00:26:12,160 --> 00:26:16,080 Speaker 1: I think that's dangerous as a merchandising strategy, but I 472 00:26:16,080 --> 00:26:16,680 Speaker 1: guess your point. 473 00:26:17,080 --> 00:26:20,199 Speaker 2: Yeah, but that's crazy. And this, by the way, so 474 00:26:20,240 --> 00:26:22,800 Speaker 2: this model is the same true for almost You're talking 475 00:26:22,840 --> 00:26:26,200 Speaker 2: specifically about a model that was created to help some 476 00:26:26,280 --> 00:26:30,560 Speaker 2: aspect of a financial services. So is that kind of 477 00:26:30,680 --> 00:26:33,720 Speaker 2: model accelerating faster and learning faster than other models for 478 00:26:33,800 --> 00:26:35,200 Speaker 2: other kinds of problems? 479 00:26:35,560 --> 00:26:37,680 Speaker 3: So this domain was code. 480 00:26:38,040 --> 00:26:41,840 Speaker 1: Yeah, and so by definition, if you're feeling feeding in 481 00:26:41,880 --> 00:26:45,400 Speaker 1: more data some more code, you get those kind of results. 482 00:26:46,359 --> 00:26:49,239 Speaker 1: It does depend on the model type. There's a lot 483 00:26:49,280 --> 00:26:52,080 Speaker 1: of code in the world, and so we can find 484 00:26:52,119 --> 00:26:55,520 Speaker 1: that we can create it. Like I said, there's other 485 00:26:55,640 --> 00:26:59,879 Speaker 1: aspects where there's probably less inputs available, which means you 486 00:27:00,000 --> 00:27:03,280 Speaker 1: probably won't get the same level of iteration. But for code, 487 00:27:03,280 --> 00:27:04,960 Speaker 1: that's certainly the cycle times that we're seeing. 488 00:27:05,000 --> 00:27:07,600 Speaker 2: Yeah, and how do you know that Let's stick with 489 00:27:07,640 --> 00:27:10,280 Speaker 2: this one example of this model you have. How do 490 00:27:10,320 --> 00:27:14,600 Speaker 2: you know that your model is better than big company 491 00:27:14,640 --> 00:27:17,640 Speaker 2: B down the street? The client asks you, why would 492 00:27:17,640 --> 00:27:20,639 Speaker 2: I go with IBM as opposed to some the some 493 00:27:20,840 --> 00:27:22,960 Speaker 2: firm in the valley that says, let's they have a 494 00:27:22,960 --> 00:27:27,240 Speaker 2: model on this, what's your how do you frame your advantage? 495 00:27:28,520 --> 00:27:31,679 Speaker 1: Well, we benchmark all of this, and I think the 496 00:27:31,680 --> 00:27:36,960 Speaker 1: most important is metric is price performance, not price, not performance, 497 00:27:36,960 --> 00:27:38,200 Speaker 1: but the combination of the two. 498 00:27:38,880 --> 00:27:40,960 Speaker 3: And we're super competitive there. 499 00:27:41,040 --> 00:27:44,240 Speaker 1: Well for what we just released, with what we've done 500 00:27:44,240 --> 00:27:46,719 Speaker 1: in open source, we know that nobody's close to us 501 00:27:46,760 --> 00:27:47,760 Speaker 1: right now on code. 502 00:27:47,920 --> 00:27:48,080 Speaker 3: Now. 503 00:27:48,119 --> 00:27:51,520 Speaker 1: To be clear, that will probably change because it's like leapfrog. 504 00:27:51,560 --> 00:27:53,960 Speaker 3: People will jump ahead, then we jump back ahead. 505 00:27:54,560 --> 00:27:59,040 Speaker 1: But we're very confident that with everything we've done in 506 00:27:59,080 --> 00:28:01,479 Speaker 1: the last few months taken a huge lead forward here. 507 00:28:01,800 --> 00:28:04,840 Speaker 2: Yeah, it's I mean, this goes back to the point 508 00:28:04,840 --> 00:28:07,600 Speaker 2: I was making in the beginning. So about the difference 509 00:28:07,640 --> 00:28:12,320 Speaker 2: between your twenty something self in ninety nine and yourself today. 510 00:28:12,640 --> 00:28:17,080 Speaker 2: But this time compression has to be a crazy adjustment. 511 00:28:17,520 --> 00:28:20,600 Speaker 2: So the concept of what you're working on and how 512 00:28:20,640 --> 00:28:23,959 Speaker 2: you make decisions internally and things has to undergo this 513 00:28:24,040 --> 00:28:27,760 Speaker 2: kind of revolution. If you're switching from I mean back 514 00:28:27,760 --> 00:28:31,720 Speaker 2: in the day, a model might be useful for how long. 515 00:28:31,960 --> 00:28:35,720 Speaker 1: Years years I think about you know, statistical models that 516 00:28:35,800 --> 00:28:40,200 Speaker 1: set inside things like SPSS, which is a product that 517 00:28:40,240 --> 00:28:40,600 Speaker 1: a lot of. 518 00:28:40,520 --> 00:28:41,600 Speaker 3: Students use around the world. 519 00:28:41,640 --> 00:28:43,600 Speaker 1: I mean, those have been the same models for twenty 520 00:28:43,680 --> 00:28:45,920 Speaker 1: years and they're still very good at what they do. 521 00:28:46,720 --> 00:28:50,600 Speaker 1: And so yes, it's a completely it's a completely different 522 00:28:51,480 --> 00:28:53,200 Speaker 1: moment for how fast this is moving. 523 00:28:53,680 --> 00:28:54,600 Speaker 3: And I think it just. 524 00:28:55,000 --> 00:28:59,160 Speaker 1: Raises the bar for everybody, whether you're a technology provider 525 00:28:59,240 --> 00:29:03,240 Speaker 1: like us, or you're a bank or an insurance company 526 00:29:03,600 --> 00:29:06,520 Speaker 1: or a shipping company, to say, how do you really 527 00:29:07,440 --> 00:29:11,840 Speaker 1: change your culture to be way more aggressive than you 528 00:29:11,960 --> 00:29:12,640 Speaker 1: normally would be? 529 00:29:14,680 --> 00:29:17,280 Speaker 2: Does this mean it's a weird question, but does this 530 00:29:17,320 --> 00:29:21,320 Speaker 2: mean a different set of kind of personality or character 531 00:29:21,360 --> 00:29:24,800 Speaker 2: traits are necessary for a decision maker in tech now 532 00:29:24,840 --> 00:29:26,440 Speaker 2: than twenty five years ago. 533 00:29:29,600 --> 00:29:33,360 Speaker 1: There's a book I saw recently, it's called The Geek Way, 534 00:29:33,680 --> 00:29:38,480 Speaker 1: which talked about how technology companies have started to operate 535 00:29:38,520 --> 00:29:45,600 Speaker 1: in different ways maybe than many traditional companies, and more 536 00:29:45,640 --> 00:29:51,440 Speaker 1: about being data driven, more about delegation. Are you willing 537 00:29:51,480 --> 00:29:55,200 Speaker 1: to have the smartest person in the room make decisions 538 00:29:55,280 --> 00:29:57,800 Speaker 1: opposed to the highest paid person in the room. I 539 00:29:57,840 --> 00:30:00,640 Speaker 1: think these are all different aspects that ever company is 540 00:30:00,680 --> 00:30:01,240 Speaker 1: going to face. 541 00:30:01,680 --> 00:30:06,480 Speaker 2: Yeah, yeah, next term, talk about open. When you use 542 00:30:06,520 --> 00:30:07,640 Speaker 2: that word open, what do you mean. 543 00:30:10,160 --> 00:30:12,920 Speaker 1: I think there's really only one definition of open, which 544 00:30:12,960 --> 00:30:18,000 Speaker 1: is for technology, is open source. An open source means 545 00:30:18,640 --> 00:30:24,520 Speaker 1: the code is freely available. Anybody can see it, access it, 546 00:30:25,440 --> 00:30:26,280 Speaker 1: contribute to it. 547 00:30:26,560 --> 00:30:29,960 Speaker 2: And what is Tell me about why that's an important principle. 548 00:30:32,720 --> 00:30:35,840 Speaker 1: When you take a topic like AI. I think it 549 00:30:35,880 --> 00:30:39,680 Speaker 1: would be really bad for the world if this was 550 00:30:39,720 --> 00:30:44,280 Speaker 1: in the hands of one or two companies, or three 551 00:30:44,360 --> 00:30:47,640 Speaker 1: or four, doesn't matter the number, some small number. Think 552 00:30:47,640 --> 00:30:52,240 Speaker 1: about like in history sometimes early nineteen hundreds, the Interstate 553 00:30:52,320 --> 00:30:55,959 Speaker 1: Commerce Commission was created, and the whole idea was to 554 00:30:56,080 --> 00:31:01,600 Speaker 1: protect farmers from railroads, meaning they wanted to allow free trade. 555 00:31:02,000 --> 00:31:04,400 Speaker 1: But they knew that well, there's only so many railroad tracks, 556 00:31:04,400 --> 00:31:07,720 Speaker 1: So we need to protect farmers from the shipping costs 557 00:31:07,760 --> 00:31:11,680 Speaker 1: that railroads could impose. So good idea, but over time 558 00:31:12,000 --> 00:31:15,760 Speaker 1: that got completely overtaken by the railroad lobby and then 559 00:31:15,760 --> 00:31:19,640 Speaker 1: they use that to basically just increase prices, and it 560 00:31:19,680 --> 00:31:23,760 Speaker 1: made the lives of farmers way more difficult. I think 561 00:31:23,760 --> 00:31:27,200 Speaker 1: you could play the same analogy through with AI. If 562 00:31:27,240 --> 00:31:31,480 Speaker 1: you allow a handful of companies to have the technology, 563 00:31:31,640 --> 00:31:34,719 Speaker 1: you regulate around the principles of those one or two companies, 564 00:31:34,760 --> 00:31:36,240 Speaker 1: then you've trapped the entire world. 565 00:31:36,480 --> 00:31:40,600 Speaker 3: I think that would be very bad. So the danger 566 00:31:40,600 --> 00:31:42,320 Speaker 3: of that app for sure. 567 00:31:42,440 --> 00:31:45,600 Speaker 1: I mean there's companies in Watson in Washington every week 568 00:31:46,080 --> 00:31:48,680 Speaker 1: trying to achieve that outcome. 569 00:31:49,600 --> 00:31:50,080 Speaker 3: And so the. 570 00:31:50,040 --> 00:31:51,960 Speaker 1: Opposite of that is to say it's going to be 571 00:31:51,960 --> 00:31:56,840 Speaker 1: an open source because nobody could dispute open source because 572 00:31:56,840 --> 00:32:00,960 Speaker 1: it's right there, everybody can see it. So I'm a 573 00:32:00,960 --> 00:32:03,239 Speaker 1: strong believer that open source will win for AI. It 574 00:32:03,280 --> 00:32:06,120 Speaker 1: has to win. It's not just important for business, but 575 00:32:06,160 --> 00:32:09,680 Speaker 1: it's important for humans. 576 00:32:10,440 --> 00:32:13,560 Speaker 2: On the I'm curious about on the list of things 577 00:32:13,600 --> 00:32:17,360 Speaker 2: you worry about, Actually, let me before I ask, let 578 00:32:17,400 --> 00:32:19,840 Speaker 2: me ask this question very generally, what is the list 579 00:32:19,880 --> 00:32:22,719 Speaker 2: of things you worry about. What's your top five business 580 00:32:22,720 --> 00:32:23,880 Speaker 2: related worries right now? 581 00:32:25,320 --> 00:32:27,680 Speaker 3: Tops from those are the first question. We could be 582 00:32:27,720 --> 00:32:28,960 Speaker 3: here for hours for me to answer. 583 00:32:30,720 --> 00:32:32,640 Speaker 2: I did say business related. We could leave. You know, 584 00:32:33,920 --> 00:32:36,200 Speaker 2: your kids' haircuts got it out of. 585 00:32:36,080 --> 00:32:41,040 Speaker 1: The Number one is always it's the thing that's probably 586 00:32:41,080 --> 00:32:46,200 Speaker 1: always been true, which is just people. Do we have 587 00:32:46,240 --> 00:32:48,400 Speaker 1: the right skills? Are we doing a good job of 588 00:32:48,440 --> 00:32:51,880 Speaker 1: training our people? Are our people doing a good job 589 00:32:51,920 --> 00:32:55,880 Speaker 1: of working with clients like that's number one? Number two 590 00:32:56,000 --> 00:33:02,240 Speaker 1: is innovation? Are we pushing the envelope enough? Are are 591 00:33:02,240 --> 00:33:06,880 Speaker 1: we staying ahead? Number three is which kind of feeds 592 00:33:06,920 --> 00:33:09,640 Speaker 1: into the innovation one is risk taking? Are we taking 593 00:33:09,760 --> 00:33:13,800 Speaker 1: enough risk? Without risk, there is no growth. And I 594 00:33:13,800 --> 00:33:18,680 Speaker 1: think the trap that every larger company inevitably falls into 595 00:33:18,840 --> 00:33:24,360 Speaker 1: is conservatism. Things are good enough, and so it's are 596 00:33:24,360 --> 00:33:27,800 Speaker 1: we pushing the envelope? Are we taking enough risk to 597 00:33:27,880 --> 00:33:30,000 Speaker 1: really have an impact? I'd say those are probably the 598 00:33:30,040 --> 00:33:32,600 Speaker 1: top three that I spend talk about. 599 00:33:32,600 --> 00:33:35,920 Speaker 2: The vast trend to define productivity paradox something I know 600 00:33:35,960 --> 00:33:37,920 Speaker 2: you've thought a lot about what does that mean? 601 00:33:39,360 --> 00:33:42,080 Speaker 1: So I started thinking hard about this because all I 602 00:33:42,120 --> 00:33:47,920 Speaker 1: saw and read every day was fear about AI, and 603 00:33:48,960 --> 00:33:52,600 Speaker 1: I studied economics, and so I kind of went back 604 00:33:52,600 --> 00:33:56,479 Speaker 1: to like basic economics, and there's been like a macro 605 00:33:56,560 --> 00:34:00,560 Speaker 1: investing formula I guess I would say it's been around 606 00:34:00,600 --> 00:34:08,640 Speaker 1: forever that says growth comes from productivity growth plus population 607 00:34:08,760 --> 00:34:14,640 Speaker 1: growth plus debt growth. So if those three things are working, 608 00:34:15,080 --> 00:34:18,479 Speaker 1: you'll get GDP growth. And so then you think about 609 00:34:18,480 --> 00:34:22,279 Speaker 1: that and you say, well, debt growth, we're probably not 610 00:34:22,320 --> 00:34:25,560 Speaker 1: going back to zero percent interest rates, so to some 611 00:34:25,600 --> 00:34:28,680 Speaker 1: extent there's going to be a ceiling on that. And 612 00:34:28,719 --> 00:34:32,920 Speaker 1: then you look at population growth. There are shockingly few 613 00:34:33,080 --> 00:34:35,840 Speaker 1: countries or places in the world that will see population 614 00:34:35,920 --> 00:34:38,600 Speaker 1: growth over the next thirty to fifty years. In fact, 615 00:34:38,640 --> 00:34:43,279 Speaker 1: most places are not even at replacement rates. And so 616 00:34:43,320 --> 00:34:45,040 Speaker 1: I'm like, all right, so population growth is not going 617 00:34:45,040 --> 00:34:45,560 Speaker 1: to be there. 618 00:34:46,880 --> 00:34:48,800 Speaker 3: So that would mean if you just take. 619 00:34:48,640 --> 00:34:53,600 Speaker 1: It to the extreme, the only chance of continued GDP 620 00:34:53,760 --> 00:35:02,600 Speaker 1: growth is productivity. And the best way to solve productivity 621 00:35:02,600 --> 00:35:03,000 Speaker 1: as AI. 622 00:35:03,840 --> 00:35:05,000 Speaker 3: That's why I say it's a paradox. 623 00:35:05,120 --> 00:35:09,600 Speaker 1: On one hand, everybody's scared after death it's going to 624 00:35:09,600 --> 00:35:12,960 Speaker 1: take over the world, take all of our jobs, ruin us, 625 00:35:14,440 --> 00:35:16,560 Speaker 1: But in reality, maybe it's the other way, which is 626 00:35:16,640 --> 00:35:18,240 Speaker 1: it's the only thing that can save us. 627 00:35:18,560 --> 00:35:20,840 Speaker 3: Yeah, and if you believe. 628 00:35:20,600 --> 00:35:23,560 Speaker 1: That economic equation, which I think has proven quite true 629 00:35:23,600 --> 00:35:26,040 Speaker 1: over hundreds of years, I do think it's probably the 630 00:35:26,120 --> 00:35:27,480 Speaker 1: only thing that can save us. 631 00:35:28,520 --> 00:35:31,680 Speaker 2: Actually looked at the numbers yesterday for total random reason 632 00:35:31,960 --> 00:35:35,120 Speaker 2: on population growth in Europe and receive this is a 633 00:35:35,160 --> 00:35:38,280 Speaker 2: special bonus question. See how smart you are? Which country 634 00:35:38,560 --> 00:35:42,120 Speaker 2: in Europe continentally Europe has the highest population growth? 635 00:35:43,840 --> 00:35:49,400 Speaker 1: It's small continental Europe, probably one of the Nordics, I 636 00:35:49,440 --> 00:35:50,440 Speaker 1: would guess. 637 00:35:50,560 --> 00:35:57,640 Speaker 2: Close Luxembourg. Okay, something that's going on in Luxembourg. I 638 00:35:57,680 --> 00:36:00,239 Speaker 2: feel like, well, all of this need to investigate. There're 639 00:36:00,239 --> 00:36:02,080 Speaker 2: at one point four nine, which in the day, by 640 00:36:02,120 --> 00:36:06,000 Speaker 2: the way, would be a relatively that's the best performing country. 641 00:36:06,400 --> 00:36:08,839 Speaker 2: I mean in the day, you'd countries had routinely had 642 00:36:08,880 --> 00:36:12,520 Speaker 2: two points something, you know, percent growth in a given year. 643 00:36:13,840 --> 00:36:16,200 Speaker 2: Last question, you're writing a book. Now, we were talking 644 00:36:16,239 --> 00:36:20,400 Speaker 2: chatting about it backstage, and now I appreciate the paradox 645 00:36:20,440 --> 00:36:24,160 Speaker 2: of this book, which is universe with a model is 646 00:36:24,160 --> 00:36:25,960 Speaker 2: better in the afternoon than it is in the morning. 647 00:36:26,440 --> 00:36:28,760 Speaker 2: How do you write a book that's like printed on paper? 648 00:36:29,320 --> 00:36:30,640 Speaker 2: I expected to reuse Aul. 649 00:36:34,360 --> 00:36:38,280 Speaker 1: This is the challenge. And I am an incredible author 650 00:36:38,320 --> 00:36:41,439 Speaker 1: of useless books. I mean most of what I've spent 651 00:36:41,480 --> 00:36:44,760 Speaker 1: time on in the last decade of stuff that's completely useless, 652 00:36:44,840 --> 00:36:49,120 Speaker 1: like a year after it's written. And so when we 653 00:36:49,120 --> 00:36:50,520 Speaker 1: were talking about it, I was like, I would like 654 00:36:50,560 --> 00:36:54,919 Speaker 1: to do something around AI that's timeless. Yeah, that would 655 00:36:54,920 --> 00:36:59,160 Speaker 1: be useful ten or twenty years from now. But then 656 00:36:59,520 --> 00:37:04,440 Speaker 1: to your so, how is that even remotely possible if 657 00:37:04,760 --> 00:37:06,800 Speaker 1: the model is better in the afternoon and in the morning. 658 00:37:07,400 --> 00:37:09,120 Speaker 3: So that's the challenge in front of us. 659 00:37:09,400 --> 00:37:12,520 Speaker 1: But the book is around AI value creation, so kind 660 00:37:12,520 --> 00:37:15,360 Speaker 1: of links to this productivity paradox, and how do you 661 00:37:16,120 --> 00:37:22,640 Speaker 1: actually get sustained value out of AI, out of automation, 662 00:37:23,600 --> 00:37:27,200 Speaker 1: out of data science. And so the biggest challenge in 663 00:37:27,200 --> 00:37:29,120 Speaker 1: front of us is can we make this relevant? 664 00:37:30,360 --> 00:37:31,680 Speaker 3: How's the day that it's published? 665 00:37:31,760 --> 00:37:33,000 Speaker 2: How are you setting out to do that? 666 00:37:35,160 --> 00:37:38,120 Speaker 1: I think you have to to some extent level it 667 00:37:38,239 --> 00:37:40,840 Speaker 1: up to bigger concepts, which is kind of why I 668 00:37:40,880 --> 00:37:46,520 Speaker 1: go to things like macroeconomics, population geography as opposed to 669 00:37:46,560 --> 00:37:49,960 Speaker 1: going into the weeds of the technology itself. If you 670 00:37:50,040 --> 00:37:52,719 Speaker 1: write about this is how you get better performance out 671 00:37:52,719 --> 00:37:56,640 Speaker 1: of a model we can agree that will be completely 672 00:37:56,719 --> 00:37:59,200 Speaker 1: useless two years from now, but maybe even two months 673 00:37:59,200 --> 00:38:03,120 Speaker 1: from now, and so it will be less in the 674 00:38:03,280 --> 00:38:07,920 Speaker 1: technical detail and more of what is sustained value creation 675 00:38:08,080 --> 00:38:11,560 Speaker 1: for AI, which if you think on what is hopefully 676 00:38:11,600 --> 00:38:14,719 Speaker 1: a ten or twenty year period, it's probably we're kind 677 00:38:14,760 --> 00:38:18,200 Speaker 1: of substituting AI for technology. Now I've realized, because I 678 00:38:18,239 --> 00:38:20,880 Speaker 1: think this has always been true for technology. It's just 679 00:38:20,960 --> 00:38:24,120 Speaker 1: now AI is I think that everybody wants to talk about. 680 00:38:25,280 --> 00:38:27,560 Speaker 1: But let's see if we can do it. Time will tell. 681 00:38:28,400 --> 00:38:31,040 Speaker 2: Did you get any inkling that the pace that this 682 00:38:31,200 --> 00:38:34,879 Speaker 2: AI year's phenomenon was gonna that things with the pace 683 00:38:34,920 --> 00:38:37,440 Speaker 2: of change was going to accelerate so much? Because you 684 00:38:37,520 --> 00:38:40,239 Speaker 2: had More's law, right, You had a model in the 685 00:38:40,239 --> 00:38:45,560 Speaker 2: technology world for this kind of exponential increase in so 686 00:38:45,640 --> 00:38:50,040 Speaker 2: we you were you thinking about that kind of accelerate 687 00:38:50,280 --> 00:38:52,560 Speaker 2: similar kind of acceleration in the. 688 00:38:55,120 --> 00:38:57,680 Speaker 1: I think anybody that said they expected what we're seeing 689 00:38:57,719 --> 00:39:03,439 Speaker 1: today is probably exactly. I think it's way faster than 690 00:39:03,560 --> 00:39:08,880 Speaker 1: anybody expected. Yeah, but technologies, back to your point at 691 00:39:08,880 --> 00:39:13,000 Speaker 1: More's law has always accelerated through the years, so I 692 00:39:13,000 --> 00:39:16,280 Speaker 1: wouldn't say it's a shock, but it is surprising. 693 00:39:16,880 --> 00:39:22,400 Speaker 2: Yeah, you've had a kind of extraordinary privileged position to 694 00:39:22,640 --> 00:39:25,319 Speaker 2: watch and participate in this revolution, right, I mean, how 695 00:39:25,360 --> 00:39:29,920 Speaker 2: many other people have been in that have ridden this 696 00:39:30,360 --> 00:39:31,120 Speaker 2: wave like you have? 697 00:39:32,480 --> 00:39:35,640 Speaker 1: I do wonder is this really that much different or 698 00:39:35,680 --> 00:39:37,439 Speaker 1: does it feel different just because we're here? 699 00:39:38,480 --> 00:39:40,120 Speaker 3: I mean, I do think on one level. 700 00:39:40,200 --> 00:39:43,640 Speaker 1: Yes, So in the time I've been an IBM, Internet happened, 701 00:39:45,200 --> 00:39:51,400 Speaker 1: Mobile happened, social network happened, blockchain happened. 702 00:39:51,960 --> 00:39:53,360 Speaker 3: AI, So a lot has happened. 703 00:39:53,680 --> 00:39:55,040 Speaker 1: But then you go back and say, well, but if 704 00:39:55,080 --> 00:40:00,719 Speaker 1: I'd been here between nineteen seventy and ninety five, there 705 00:40:00,719 --> 00:40:03,600 Speaker 1: were a lot of things that are pretty fundamental then too, say, 706 00:40:03,600 --> 00:40:06,920 Speaker 1: I wondered, almost do we always exaggerate the. 707 00:40:06,840 --> 00:40:13,359 Speaker 3: Timeframe that we're in. I don't know. Yeah, but it's 708 00:40:13,360 --> 00:40:14,080 Speaker 3: a good idea though. 709 00:40:16,000 --> 00:40:19,040 Speaker 2: I think the ending with the phrase, I don't know 710 00:40:19,760 --> 00:40:23,520 Speaker 2: it's a good idea though. Comd great way to wrap 711 00:40:23,560 --> 00:40:23,920 Speaker 2: this up. 712 00:40:24,320 --> 00:40:25,680 Speaker 3: Thank you so much, Thank you, Malcolm. 713 00:40:29,719 --> 00:40:32,920 Speaker 2: In a field that is evolving as quickly as artificial intelligence, 714 00:40:33,280 --> 00:40:36,120 Speaker 2: it was inspiring to see how adaptable Rob has been 715 00:40:36,200 --> 00:40:39,959 Speaker 2: over his career. The takeaways from my conversation with Rob 716 00:40:40,239 --> 00:40:44,120 Speaker 2: had been echoing in my head ever since. He emphasized 717 00:40:44,239 --> 00:40:47,920 Speaker 2: how open source models allow AI technology to be developed 718 00:40:47,920 --> 00:40:53,359 Speaker 2: by many players. Openness also allows for transparency. Rob told 719 00:40:53,400 --> 00:40:58,360 Speaker 2: me about AI use cases like IBM's collaboration with Sevilla's 720 00:40:58,360 --> 00:41:02,040 Speaker 2: football club. That exam really brought home for me how 721 00:41:02,120 --> 00:41:07,320 Speaker 2: AI technology will touch every industry. Despite the potential benefits 722 00:41:07,320 --> 00:41:12,280 Speaker 2: of AI, challenges exist in its widespread adoption. Rob discussed 723 00:41:12,480 --> 00:41:17,520 Speaker 2: how resistance to change, concerns about job security and organizational 724 00:41:17,560 --> 00:41:23,439 Speaker 2: inertia can slow down implementation of AI solutions. The paradox, though, 725 00:41:23,480 --> 00:41:26,040 Speaker 2: according to Rob, is that rather than being afraid of 726 00:41:26,040 --> 00:41:29,640 Speaker 2: a world with AI, people should actually be more afraid 727 00:41:29,719 --> 00:41:33,440 Speaker 2: of a world without it. AI, he believes, has the 728 00:41:33,480 --> 00:41:36,400 Speaker 2: potential to make the world a better place in a 729 00:41:36,400 --> 00:41:40,879 Speaker 2: way that no other technology can. Rob painted an optimistic 730 00:41:40,960 --> 00:41:44,480 Speaker 2: version of the future, one in which AI technology will 731 00:41:44,520 --> 00:41:48,799 Speaker 2: continue to improve at an exponential rate. This will free 732 00:41:48,880 --> 00:41:52,960 Speaker 2: up workers to dedicate their energy to more creative tasks. 733 00:41:53,560 --> 00:41:58,080 Speaker 2: I for one am on board Smart Talks with IBM 734 00:41:58,200 --> 00:42:02,320 Speaker 2: is produced by Matt Romano, Joey Fishground, and Jacob Goldstein. 735 00:42:02,760 --> 00:42:06,479 Speaker 2: We're edited by Lydia gene kott Our engineers are Sarah 736 00:42:06,560 --> 00:42:11,680 Speaker 2: Bruguer and Ben Tolliday. Theme song by Gramscow. Special thanks 737 00:42:11,680 --> 00:42:14,200 Speaker 2: to the eight Bar and IBM teams, as well as 738 00:42:14,200 --> 00:42:17,759 Speaker 2: the Pushkin marketing team. Smart Talks with IBM is a 739 00:42:17,800 --> 00:42:22,560 Speaker 2: production of Pushkin Industries and Ruby Studio at iHeartMedia. To 740 00:42:22,600 --> 00:42:28,240 Speaker 2: find more Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts, 741 00:42:28,360 --> 00:42:33,080 Speaker 2: or wherever you listen to podcasts. I'm Malcolm Gladwell. This 742 00:42:33,239 --> 00:42:36,840 Speaker 2: is a paid advertisement from IBM. The conversations on this 743 00:42:36,960 --> 00:42:49,680 Speaker 2: podcast don't necessarily represent IBM's positions, strategies, or opinions.