WEBVTT - IBM CEO Arvind Krishna: Creating Smarter Business with AI and Quantum 0:00:00.120 --> 0:00:04.240 Welcome to Tech Stuff, a production from iHeartRadio. This season, 0:00:04.280 --> 0:00:07.880 non Smart Talks with IBM, Malcolm Glabwell is back, and 0:00:07.920 --> 0:00:10.600 this time he's taking the show on the road. Malcolm 0:00:10.640 --> 0:00:14.720 is stepping outside the studio to explore how IBM clients 0:00:14.760 --> 0:00:18.759 are using artificial intelligence to solve real world challenges and 0:00:18.840 --> 0:00:23.000 transform the way they do business. From accelerating scientific breakthroughs 0:00:23.079 --> 0:00:27.520 to reimagining education. It's a fresh look at innovation in action, 0:00:27.960 --> 0:00:31.720 where big ideas meet cutting edge solutions. You'll hear from 0:00:31.720 --> 0:00:36.000 industry leaders, creative thinkers, and of course Malcolm Glabwell himself 0:00:36.280 --> 0:00:39.879 as he guides you through each story. New episodes of 0:00:39.920 --> 0:00:43.280 Smart Talks with IBM drop every month on the iHeartRadio app, 0:00:43.440 --> 0:00:47.240 Apple Podcasts, or wherever you get your podcasts. Learn more 0:00:47.280 --> 0:01:05.560 at IBM dot com, slash smart Talks. 0:00:58.680 --> 0:01:03.640 Pushkin Hello. 0:01:03.680 --> 0:01:06.479 Hello, I'm Malcolm Gladwell and you're listening to Smart Talks 0:01:06.480 --> 0:01:09.600 with IBM. This season, we've been bringing you stories of 0:01:09.600 --> 0:01:13.000 how IBM works with its clients to solve complex problems, 0:01:13.720 --> 0:01:18.280 like helping Lareel reimagine how scientists approach cosmetic formulation, or 0:01:18.400 --> 0:01:22.399 enabling Scuderia FERRARIHP to connect with fans in new ways. 0:01:23.120 --> 0:01:25.399 But in this episode, we're going to zoom out and 0:01:25.440 --> 0:01:28.520 look at the bigger picture. Earlier this month, I had 0:01:28.520 --> 0:01:31.520 the chance to meet the person who's shaping IBM's future. 0:01:31.959 --> 0:01:36.240 It's CEO and Chairman Arvind Krishna. We sat down in 0:01:36.280 --> 0:01:38.520 front of an intimate live audience at IBM's New York 0:01:38.520 --> 0:01:42.400 City office and talked about his uncanny ability to anticipate 0:01:42.440 --> 0:01:46.240 where technology is heading, the future of AI, and his 0:01:46.360 --> 0:01:50.600 passion for quantum computing, which he says is as revolutionary 0:01:50.880 --> 0:01:55.559 as a semiconductor. Thank you everyone, Thank you to Arvin. 0:01:55.640 --> 0:01:58.280 You're a difficult man to schedule for one of these things, 0:01:58.320 --> 0:02:01.760 so we're enormously please you could join us. Let's start 0:02:01.760 --> 0:02:04.280 with a I have all these cousins, two cousins who 0:02:04.360 --> 0:02:06.720 work for IBM their entire career. I would ask them 0:02:07.360 --> 0:02:11.239 what does IBM do? And they would always give me different, confusing, 0:02:11.360 --> 0:02:13.200 complicated answers. 0:02:13.440 --> 0:02:15.800 What's your answer? What's your simple answer to that question. 0:02:16.639 --> 0:02:21.639 IBM's role is to help our clients improve their business 0:02:21.680 --> 0:02:26.480 by deploying technology. That means you're not ever gated to 0:02:26.560 --> 0:02:29.400 one product. It is what makes sense at that time, 0:02:29.880 --> 0:02:32.680 but it is about improving their business, not just giving 0:02:32.720 --> 0:02:33.880 them a commodity. 0:02:34.400 --> 0:02:35.240 Then to go to the. 0:02:35.200 --> 0:02:38.040 Next layer, I would say, we help them through a 0:02:38.120 --> 0:02:42.560 mixture of hybrid cloud and artificial intelligence and a taste 0:02:42.560 --> 0:02:44.680 of quantum coming down the road is kind of where 0:02:45.160 --> 0:02:45.760 I would take it. 0:02:45.919 --> 0:02:47.160 That's that's what IBM is. 0:02:47.760 --> 0:02:52.040 So you are technology agnostic in some sense. 0:02:52.320 --> 0:02:56.080 I'm product agnostic. Product I'm not technology agnostic. 0:02:56.200 --> 0:02:59.000 Yes, But if I twenty five years from now, IBM 0:02:59.040 --> 0:03:04.200 could be doing things that would be unrecognizable to contemporary IBM. 0:03:04.080 --> 0:03:05.280 It is completely possible. 0:03:05.480 --> 0:03:08.160 Yeah, it could be there in twenty five years from now. 0:03:08.360 --> 0:03:11.680 The only software IBM does is open source. It could 0:03:11.680 --> 0:03:14.280 be the only computing you do is quantum computers. And 0:03:14.320 --> 0:03:16.960 if I add those two people today, WLD say that's 0:03:17.000 --> 0:03:17.919 not the IEM of today. 0:03:18.200 --> 0:03:20.800 Is it even simpler to say you just IBM solves 0:03:20.880 --> 0:03:23.600 problems at the highest technical level. 0:03:24.080 --> 0:03:26.440 If you say highest technical level, yes, yeah. 0:03:26.560 --> 0:03:29.040 Like the guy who invented the bar code, he was 0:03:29.040 --> 0:03:33.480 solving a problem retailers wanted to scale. Many of you 0:03:33.520 --> 0:03:35.640 may not know it was an IBM or who invented 0:03:35.680 --> 0:03:38.680 the bar code, by the way, not somebody who was 0:03:38.720 --> 0:03:42.840 a PhD. Not somebody who was a deep researcher. I 0:03:42.840 --> 0:03:46.040 think it was actually a field engineer. Oh really, Yeah, 0:03:46.680 --> 0:03:50.200 And lasers were out and you could use lasers to 0:03:50.240 --> 0:03:53.120 scan things, but they could be upside down, they could 0:03:53.160 --> 0:03:56.760 be muddy, they could be partly scraped off. And he 0:03:56.840 --> 0:03:59.600 came up with the idea of the bar code. Yeah, 0:04:00.040 --> 0:04:03.320 and that changed inventory management forever. But the world needs 0:04:03.320 --> 0:04:05.760 to know that IBM invented the Parker. You guys should 0:04:05.760 --> 0:04:09.720 do a better job opicizing that. I am sure our 0:04:09.840 --> 0:04:14.240 CMO will listen to this podcast and we'll get that idea. 0:04:14.640 --> 0:04:18.880 Tell me you started at the Thomas Watson Research Center. 0:04:19.400 --> 0:04:21.520 What were you doing when you first started it, IBM. 0:04:22.120 --> 0:04:27.520 I started in nineteen ninety and that was the era 0:04:27.680 --> 0:04:31.599 in which computers and networking were beginning to converge. And 0:04:31.680 --> 0:04:35.320 for the first five years I was actually building networks. 0:04:35.720 --> 0:04:38.839 So let's remember this was pre laptops. Laptops came in 0:04:38.920 --> 0:04:41.760 ninety two or ninety three, but it was clear to 0:04:41.839 --> 0:04:45.159 us that they were going to come supportable computing, and 0:04:45.240 --> 0:04:48.240 I spent my first five years building what today you 0:04:48.240 --> 0:04:51.600 would call Wi Fi. We used to have these debates 0:04:51.640 --> 0:04:53.360 can be built it, it's got to be small enough. 0:04:53.440 --> 0:04:55.680 I mean like, it can't be more than one hundred 0:04:55.680 --> 0:04:58.320 grams was kind of our thought, because if it's more 0:04:58.360 --> 0:05:01.200 than that, you're on a three thousand grand laptop. 0:05:01.200 --> 0:05:02.560 Why would anybody put this on? 0:05:03.279 --> 0:05:05.480 And the debate used to be why would anybody want 0:05:05.520 --> 0:05:07.040 to walk around untethered? 0:05:07.480 --> 0:05:10.200 Won't you want to attach a big thick cable into 0:05:10.320 --> 0:05:10.920 it and sit down? 0:05:10.960 --> 0:05:13.640 Because that was the thought, that's how terminals worked. And 0:05:13.720 --> 0:05:16.640 I spent five years having a lot of fun building 0:05:16.720 --> 0:05:21.240 many iterations of those and making progress on that. 0:05:21.600 --> 0:05:24.920 If I had a conversation with your nineteen ninety self 0:05:26.120 --> 0:05:30.039 about what the next thirty years were going to look like, 0:05:30.760 --> 0:05:33.240 is it possible to reconstruct what your What were your 0:05:33.279 --> 0:05:35.560 predictions at that age about where the company, where the 0:05:35.600 --> 0:05:36.320 industry was going. 0:05:37.400 --> 0:05:39.919 It was more about where technology was going to go, 0:05:39.960 --> 0:05:42.640 I would say, than where industry would go. I would 0:05:42.680 --> 0:05:48.000 have told you that networking and computers would fuse in 0:05:48.120 --> 0:05:51.119 nineteen ninety That was a weird thought that some researchers held. 0:05:51.480 --> 0:05:54.440 By the late nineties, that was obvious that it became 0:05:54.440 --> 0:05:59.560 the Internet. I would have told you that I believe 0:05:59.640 --> 0:06:03.280 that streaming will be the primary way people will consume. 0:06:03.040 --> 0:06:05.080 Video you would have said that in nineteen ninety. 0:06:05.200 --> 0:06:08.359 Absolutely, Now, that didn't take five years. That took twenty. 0:06:08.920 --> 0:06:12.559 But it happened because you could do it technically, except 0:06:12.680 --> 0:06:15.600 it just too expensive and too cumbersome. And if you've 0:06:15.600 --> 0:06:19.039 been in technology, like in nineteen eighty five, I would 0:06:19.080 --> 0:06:23.120 have told you the internet is old because when I 0:06:23.200 --> 0:06:26.440 went to grad school, every one of us had a 0:06:26.560 --> 0:06:30.400 those days, an Apple, Mac or Lisa on our desks. 0:06:31.200 --> 0:06:34.080 They're all connected by a network. You're happily sending email 0:06:34.120 --> 0:06:37.760 to people all around the country. We were doing file transfers, 0:06:37.760 --> 0:06:39.280 so okay, you had to be a little bit more 0:06:39.320 --> 0:06:42.080 aware of the technology. And it didn't have a browser. 0:06:43.040 --> 0:06:44.880 That took ten years to get the browser. That took 0:06:44.920 --> 0:06:47.320 five years to be a business. But when you see 0:06:47.320 --> 0:06:50.920 the speed and the pace of technology in usually ten 0:06:51.000 --> 0:06:56.520 or fifteen years, the cost point and the consumerization is 0:06:56.560 --> 0:06:59.760 at a scale that you couldn't imagine ten years ago 0:07:00.120 --> 0:07:01.640 until you've seen a few of those cycles. 0:07:02.320 --> 0:07:05.440 Wait, did you make the lead to sorry, this is fascinating, 0:07:05.640 --> 0:07:07.320 I'm curious, but how far did you take that? That's a 0:07:07.360 --> 0:07:10.720 really fundamental thing to have gotten right in nineteen ninety 0:07:10.920 --> 0:07:13.160 I think that we. 0:07:13.000 --> 0:07:17.120 Were pretty convinced that what we used to think of 0:07:17.160 --> 0:07:21.760 as linear television or broadcast would become digitized. That was 0:07:21.760 --> 0:07:26.560 a given two with cable already the preponderance of how 0:07:26.600 --> 0:07:31.640 people got it that if you put packet television over cable, 0:07:32.560 --> 0:07:36.559 then that becomes the way it will go. I fundamentally believe, 0:07:36.640 --> 0:07:41.000 actually way back eighty seven, that on demand movies would 0:07:41.000 --> 0:07:45.600 become the way people would consume movies. So those were 0:07:45.680 --> 0:07:48.360 all things that I could have predicted nine then personally 0:07:48.400 --> 0:07:51.080 work on all those. I mean, after networking, I moved 0:07:51.120 --> 0:07:55.000 on to doing other things, but those were easy to predict. 0:07:56.280 --> 0:07:59.400 If you had a conversation in those years with someone 0:07:59.440 --> 0:08:02.800 in the television industry and you gave them those predictions, 0:08:03.160 --> 0:08:03.880 did they see it? 0:08:03.960 --> 0:08:05.040 Were they convinced of this? 0:08:05.480 --> 0:08:07.600 I'm actually going to take it back to wireless networking. 0:08:08.760 --> 0:08:11.640 I think one of the reasons I do what I 0:08:11.680 --> 0:08:14.720 do today, which is at the intersection of business and technology, 0:08:15.360 --> 0:08:17.800 is because of what I saw happened with Wi Fi. 0:08:18.680 --> 0:08:22.240 So you built these wireless networks and then you say, hey, 0:08:22.240 --> 0:08:25.080 the market's going to be millions, tens of millions, billions 0:08:25.120 --> 0:08:27.840 of users, and the business looks at it and says 0:08:28.200 --> 0:08:33.200 we think the market is confined to warehouse workers doing inventory. 0:08:33.280 --> 0:08:35.600 You can look at them and say, why not people 0:08:35.600 --> 0:08:40.120 in their homes because they could imagine outside how people 0:08:40.160 --> 0:08:45.360 bought things at that time. And so I became convinced 0:08:45.400 --> 0:08:49.160 that I can't just help invent it. I got to 0:08:49.200 --> 0:08:52.520 think about, now, how do you market it? To whom 0:08:52.520 --> 0:08:54.600 do you market it? What are their routs? How do 0:08:54.640 --> 0:08:57.920 you make it easy enough? And that was probably I mean, 0:08:57.960 --> 0:09:00.559 I'm making it simple now. That is probably a five 0:09:00.640 --> 0:09:05.560 to ten year evolution of myself in those days. 0:09:05.679 --> 0:09:08.000 You know what this reminds me of when the telephone 0:09:08.040 --> 0:09:11.400 is invented in the eighteen seventies, it doesn't take off 0:09:11.440 --> 0:09:13.920 for forty years because the people running a telephone business 0:09:14.320 --> 0:09:16.679 they didn't want women using it because they were worried 0:09:16.679 --> 0:09:19.880 that women would gossip with their friends. They didn't understand 0:09:19.880 --> 0:09:23.440 that that's actually what telephone is, right, it's an exact parallel, 0:09:23.520 --> 0:09:24.760 Yes it is. 0:09:24.880 --> 0:09:26.040 You see it again and again. 0:09:26.400 --> 0:09:28.280 What is the source of that blindness? 0:09:28.280 --> 0:09:30.920 So there's a gap, in other words, between the invention, 0:09:31.120 --> 0:09:34.679 the technological achievement, and the social understanding of the technology. 0:09:35.040 --> 0:09:36.160 Why is there such a gap? 0:09:37.400 --> 0:09:41.760 I think that the gap is fundamental and rooted in 0:09:42.000 --> 0:09:47.480 a lot of academic disciplines. So even channeling some of 0:09:47.520 --> 0:09:49.480 your work, though you don't intend it to be used 0:09:49.480 --> 0:09:51.400 that way, you can say a lot of things that 0:09:51.520 --> 0:09:56.479 data driven. If it is data driven, then by definition, 0:09:56.600 --> 0:10:00.320 you're looking at history. If you're looking at history, that 0:10:00.360 --> 0:10:03.480 means you're looking at existing buying patterns. If you look 0:10:03.520 --> 0:10:07.800 at existing buying patterns, you forget. All of those who 0:10:07.840 --> 0:10:11.960 have created massive value in time have all created markets, 0:10:12.120 --> 0:10:13.760 meaning they've all created new markets. 0:10:14.600 --> 0:10:15.640 And I think that is why. 0:10:15.480 --> 0:10:19.040 The world is fascinated with people like Steve Jobs. For example, 0:10:19.120 --> 0:10:22.440 he imagined a market that didn't exist. So I think 0:10:22.480 --> 0:10:25.360 that is the gap. And then if you can get 0:10:25.800 --> 0:10:31.600 the technology the business acumen scaler company, and that imagination 0:10:31.640 --> 0:10:34.240 of making a market is how you create I think 0:10:34.640 --> 0:10:35.440 massive value. 0:10:35.480 --> 0:10:36.880 You got to get all three pieces going. 0:10:37.000 --> 0:10:37.640 It's not enough. 0:10:37.640 --> 0:10:39.240 In other words, you were thinking it's not enough to 0:10:39.280 --> 0:10:42.040 invent something new. I need to make a business case 0:10:42.080 --> 0:10:44.680 for it simultaneously, and that that's what gets you thinking 0:10:45.080 --> 0:10:46.800 along the path that leads you to this job. 0:10:46.880 --> 0:10:50.559 Oh yeah, I'll tell you if you had met or 0:10:50.600 --> 0:10:53.640 when the nineteen ninety four and you had talked about 0:10:53.760 --> 0:10:57.920 the stock market or about a balance sheet, or looked 0:10:57.920 --> 0:10:59.840 at you like, Okay, I got the those words are, 0:10:59.880 --> 0:11:00.839 I can parse them. 0:11:00.920 --> 0:11:02.959 I have no idea what they are. I have no 0:11:03.000 --> 0:11:04.280 intuition on what they are. 0:11:04.600 --> 0:11:07.719 I couldn't tell you why it's relevant or why it's not. 0:11:08.120 --> 0:11:11.080 But then you began to think, Okay, why do companies 0:11:11.120 --> 0:11:14.080 get higher values? Okay, that's the stock what does that 0:11:14.160 --> 0:11:17.800 capture if I have to spend working capital and that's 0:11:18.080 --> 0:11:21.280 the balance sheet? Well, so you learn. I mean, I figure, 0:11:21.320 --> 0:11:24.720 I'm willing to learn. I'm willing to read. However, the 0:11:24.720 --> 0:11:27.760 best way I read is to go to balance sheets. 0:11:28.080 --> 0:11:29.840 Here you can read the book. It's pretty damn dry. 0:11:30.559 --> 0:11:32.960 Much easier to go talk to a financial expert who's 0:11:33.000 --> 0:11:36.080 around the corner, and people are if you're curious about 0:11:36.120 --> 0:11:38.960 what they do, they're really happy to share their expertise, 0:11:39.400 --> 0:11:42.559 and over time you learn more and more and they 0:11:42.559 --> 0:11:45.679 actually become part of your network within the company. And 0:11:45.840 --> 0:11:48.920 that's how you can both learn and evolve yourself and 0:11:49.080 --> 0:11:50.880 actually gain the extra. 0:11:50.679 --> 0:11:54.120 Skills you are to be a successful business leader. 0:11:54.920 --> 0:11:58.880 Do you have to unlearn or deviate from some of 0:11:58.880 --> 0:12:01.720 the things that made you as successful scientist? 0:12:02.360 --> 0:12:03.760 I actually believe the exactly opposite. 0:12:04.000 --> 0:12:07.880 Yeah, but use what you're really good at as a foundation, 0:12:09.080 --> 0:12:11.480 but don't make it the only thing you use. So 0:12:11.520 --> 0:12:14.720 then how do you add the other skills? And there's 0:12:14.720 --> 0:12:17.600 many ways you can have people that you trust who 0:12:17.679 --> 0:12:21.240 help you out those skills. You can gain some intuition, 0:12:21.520 --> 0:12:24.120 maybe not the depth of expertise. I want to be 0:12:24.240 --> 0:12:27.760 deeper on certain areas of electrical engineering than I'm ever 0:12:27.840 --> 0:12:30.280 going to be, let's say, in finance or marketing. But 0:12:30.880 --> 0:12:32.959 I want to be curious about those. I don't want 0:12:33.000 --> 0:12:37.319 to dismiss them. So you build on your skills, and 0:12:37.360 --> 0:12:39.640 then you have to say, but I need a complete 0:12:39.800 --> 0:12:43.480 and holistic view. So I'm going to be a little deep, 0:12:43.559 --> 0:12:46.040 not very deep, in all of those. And you've also 0:12:46.080 --> 0:12:48.240 got to learn to trust your intuition a little bit. 0:12:48.600 --> 0:12:51.319 Yeah, but I forgot a question that I wanted to ask, 0:12:52.120 --> 0:12:56.040 But about the predictions of nineteen ninety arvand what did 0:12:56.080 --> 0:12:56.640 you get wrong? 0:12:58.120 --> 0:13:01.000 All lots of things. I think that people were thinking 0:13:01.160 --> 0:13:05.840 that in those days, and it started my phrase, but 0:13:05.880 --> 0:13:08.400 I'll come back to it. I think most people thought 0:13:08.520 --> 0:13:12.040 that the communication companies would turn out to be the 0:13:12.080 --> 0:13:15.520 winners of how networking got carried. If you all think 0:13:15.559 --> 0:13:18.000 through the nineties of the investments that were being done 0:13:18.520 --> 0:13:20.920 by let's not take the names of all of the 0:13:21.160 --> 0:13:25.920 telecom carriers, didn't turn out to be the case. Actually, 0:13:25.920 --> 0:13:28.840 I think that's the business model case. The reason is 0:13:29.320 --> 0:13:31.400 they all had in their heads that you can charge 0:13:31.440 --> 0:13:32.679 people by the minute. 0:13:34.280 --> 0:13:36.000 Because they had been doing that already. 0:13:35.720 --> 0:13:37.320 Because they'd been doing that for one hundred years. 0:13:37.360 --> 0:13:41.720 Yeah, and in the end, the winners the networking were 0:13:41.720 --> 0:13:44.480 those who sat flat price thirty bucks a month or 0:13:44.520 --> 0:13:47.640 fifty bucks a month or whatever, and that was just 0:13:47.760 --> 0:13:48.439 too much of. 0:13:48.400 --> 0:13:49.920 A leap for them. 0:13:50.040 --> 0:13:53.559 You think it's as simples, That is the most parsimonious 0:13:53.559 --> 0:13:55.800 explanation for why you think they failed. 0:13:56.080 --> 0:13:58.240 No, there were a couple of other more technical things. 0:13:58.320 --> 0:14:01.199 The one was written by somebody who was inside one 0:14:01.200 --> 0:14:04.559 of these telecom companies, and he labeled his article the 0:14:04.640 --> 0:14:08.439 Rise of the Stupid Network. So telephone people believe that 0:14:08.480 --> 0:14:11.680 the network should be really smart. The end device is dumb. 0:14:12.760 --> 0:14:14.559 If you think about the telephone, telephone is dumb. It 0:14:14.559 --> 0:14:17.040 doesn't actually do anything. It's just about the relays. And 0:14:17.080 --> 0:14:19.520 the network is smart. It routes you, it figures out 0:14:19.560 --> 0:14:24.880 where to send it. It does echocaculation backwards, and the 0:14:24.920 --> 0:14:27.360 current Internet is completely dumb with the inside. It just 0:14:27.440 --> 0:14:29.360 takes the bits and shoves them out the other end. 0:14:29.760 --> 0:14:32.440 All the intelligence is the computer at the end. That's 0:14:32.760 --> 0:14:35.720 probably a bit more of a found explanation. But business 0:14:35.760 --> 0:14:37.160 model didn't help them either. 0:14:37.400 --> 0:14:37.680 Yeah. 0:14:37.960 --> 0:14:41.720 Wait, did nineteen ninety r end think that the network 0:14:41.760 --> 0:14:43.840 should be dumb or smart? 0:14:45.240 --> 0:14:47.920 I'm not sure I thought about it deeply, but everything 0:14:47.960 --> 0:14:49.400 I worked on the network was dumb. 0:14:50.120 --> 0:14:51.800 The network movements, That's all I did. 0:14:52.440 --> 0:14:56.320 Yeah, because even I in those days understood I can't 0:14:56.320 --> 0:14:57.720 imagine all the applications. 0:14:58.120 --> 0:15:00.320 So if all you do is voice me, maybe the 0:15:00.320 --> 0:15:01.280 network can be smart. 0:15:01.440 --> 0:15:02.960 But if you're doing all those other things, how could 0:15:02.960 --> 0:15:05.760 the network possibly know all those things and be smart 0:15:05.760 --> 0:15:06.040 for it? 0:15:06.240 --> 0:15:10.200 Yeah, so you've been COO for five years. 0:15:10.320 --> 0:15:10.960 Five years? 0:15:11.200 --> 0:15:13.920 Wait, so in your five year increment, what was your 0:15:14.000 --> 0:15:15.920 most misunderstood decision? 0:15:16.680 --> 0:15:18.480 Well, you ended up being right, but everyone thought you 0:15:18.480 --> 0:15:18.960 were crazy. 0:15:19.600 --> 0:15:23.560 Twenty and eighteen, I proposed to our boat that we 0:15:23.560 --> 0:15:28.000 should buy a company called rad Hat. IBM does proprietary 0:15:28.000 --> 0:15:31.640 but that was open source. The stock twelve fifteen percent 0:15:31.680 --> 0:15:35.040 of the day we announced it, and today most people 0:15:35.080 --> 0:15:37.040 will turn around and say, this is the most successful 0:15:37.080 --> 0:15:40.760 acquisition that IBM has done in all time, and probably 0:15:40.840 --> 0:15:45.520 the most successful software acquisition in history. So it was 0:15:45.560 --> 0:15:49.120 completely misunderstood because people didn't see that you actually did 0:15:49.200 --> 0:15:53.400 need a platform that could make it agnostic across multiple 0:15:53.800 --> 0:15:57.720 cloud platforms, across on premise environments. So you've got to 0:15:57.760 --> 0:16:00.160 have a view of what it could be, and we 0:16:00.240 --> 0:16:02.880 drove it to a place where I think today it 0:16:03.000 --> 0:16:04.760 stands as the leader in its space. 0:16:05.520 --> 0:16:08.560 So how did you come to believe this heretical notion? 0:16:10.720 --> 0:16:15.160 So Cloud was happening, you could ask yourself the question, 0:16:15.640 --> 0:16:20.160 should we spend a lot of capital and chase Cloud? Okay, 0:16:20.400 --> 0:16:24.560 you're five years to be generous, maybe longer behind at 0:16:24.560 --> 0:16:27.960 that point the two leaders, So you could spend maybe 0:16:28.000 --> 0:16:31.960 ten billion a year, and a lot of businesses tend 0:16:32.000 --> 0:16:34.240 to do that. Okay, it's so important, it's going to 0:16:34.280 --> 0:16:37.760 be half the market. I can't not. My view was 0:16:38.200 --> 0:16:40.240 we'll always be five years behind. They're not dumb and 0:16:40.240 --> 0:16:42.880 they're not slow. So if you're going to be there, 0:16:43.000 --> 0:16:46.400 you're going to be best case, a distant third, worst 0:16:46.400 --> 0:16:48.440 case maybe a fourth or a fifth. Because there's Chinese 0:16:48.480 --> 0:16:52.080 also in the mix. Why would you do that instead? 0:16:52.440 --> 0:16:55.000 Is there a different space you can occupy instead of 0:16:55.080 --> 0:16:58.440 competing with them? Can you become their best partner? In 0:16:58.480 --> 0:17:02.120 which case you right there success. If I want to 0:17:02.120 --> 0:17:04.520 be the best partner, then what are the set of 0:17:04.560 --> 0:17:08.040 technologies that would be useful? So you can flip The 0:17:08.080 --> 0:17:11.640 problem is how I thought about it. 0:17:12.480 --> 0:17:16.119 How hard was it to convince people needed convincing before 0:17:16.160 --> 0:17:17.879 that acquisition. 0:17:18.359 --> 0:17:20.840 Probably six to nine months of. 0:17:22.359 --> 0:17:26.920 Breaking my head with no success, and then six months 0:17:26.960 --> 0:17:33.160 of building the momentum once a couple of people began 0:17:33.240 --> 0:17:33.679 to see it. 0:17:34.119 --> 0:17:35.920 Yeah, you're very persistent. 0:17:36.440 --> 0:17:41.840 Oh yes, very Would you describe that as you defining trade? 0:17:43.240 --> 0:17:45.880 I am very persistent and I'm very patient. 0:17:46.600 --> 0:17:50.080 I'm also probably very impatient, But I'm not a yeller 0:17:50.119 --> 0:17:50.720 and screamer. 0:17:50.840 --> 0:17:52.480 I don't rant and rave. 0:17:53.080 --> 0:17:55.639 But as I say, if I think we're going to 0:17:55.680 --> 0:18:00.439 do something, I can be remarkably stop about it. 0:18:00.480 --> 0:18:01.120 We will do it. 0:18:01.280 --> 0:18:03.600 If I got your family, put them up on stage 0:18:03.640 --> 0:18:06.239 and ask them this exact question, is this how they 0:18:06.240 --> 0:18:06.760 would answer? 0:18:06.800 --> 0:18:07.320 As well? 0:18:07.480 --> 0:18:13.239 They will tell you I'm very stubborn. They might not 0:18:13.280 --> 0:18:15.160 agree that I don't rant and drave. 0:18:17.320 --> 0:18:20.920 Well, you know, one of the principal observations of psychology 0:18:21.000 --> 0:18:24.400 is that our home self and our work self are uncorrelated. 0:18:25.040 --> 0:18:26.400 Once you know that, you know everything. 0:18:26.720 --> 0:18:29.760 Wait, I'm curious one last question about that. How long 0:18:29.800 --> 0:18:33.160 does it take for you to be vindicated with red Hat? 0:18:34.200 --> 0:18:39.119 Probably took five maybe four years, I think by twenty 0:18:39.280 --> 0:18:41.880 twenty three, So twenty eighteen we announced it, We took 0:18:41.880 --> 0:18:43.760 the big shot croft. It took a year to close 0:18:43.840 --> 0:18:48.199 twenty nineteen, so if I count, not that I'm counting 0:18:48.240 --> 0:18:57.600 that much, but July ninth, twenty nineteen as the day 0:18:57.640 --> 0:19:00.000 that we got all the approvals. Took another few weeks 0:19:00.200 --> 0:19:04.280 to actually transfer the money. But from there, probably twenty 0:19:04.359 --> 0:19:06.560 twenty three, the world woke up and said, hey, you 0:19:06.600 --> 0:19:08.440 guys deserve credit for this was actually. 0:19:08.160 --> 0:19:10.080 A great move, not a bad move. 0:19:10.240 --> 0:19:14.439 Yeah, but this is it's interesting because this is a 0:19:14.480 --> 0:19:18.520 real gamble. If it doesn't work, you're not sitting in 0:19:18.520 --> 0:19:19.200 this chair right now. 0:19:19.320 --> 0:19:19.560 Right. 0:19:19.680 --> 0:19:21.440 Oh, for sure. There were two steps. 0:19:21.440 --> 0:19:24.159 One if it was obviously not going to work, I 0:19:24.160 --> 0:19:26.720 wouldn't have been selected, and two if it hadn't worked 0:19:26.720 --> 0:19:29.520 after that, that's why CEOs can be short lived. 0:19:30.040 --> 0:19:31.480 Can I ask you a sort of a personal question. 0:19:31.720 --> 0:19:33.439 How much sleep did you lose over this? 0:19:36.920 --> 0:19:40.680 Once we had made the decision, none. 0:19:43.000 --> 0:19:44.719 Can you give me pointers on how you do this? 0:19:44.800 --> 0:19:48.000 Because I wake up at two am every morning and 0:19:48.040 --> 0:19:50.159 I over much more trivial things than this. 0:19:50.640 --> 0:19:52.639 Once a week, I'll probably wake up at two or 0:19:52.640 --> 0:19:53.320 three in the morning. 0:19:53.400 --> 0:19:56.880 I acknowledge it because I wake up and my brain 0:19:56.960 --> 0:19:58.879 is running, and once it's running, I don't even. 0:19:58.760 --> 0:19:59.560 Try to go back to sleep. 0:19:59.520 --> 0:20:02.560 I mean, go get up and do work and make 0:20:02.600 --> 0:20:04.800 yourself productive. You're gonna be tired before in the afternoon, 0:20:04.800 --> 0:20:08.080 that's fine, you'll sleep well that night. I have actually 0:20:08.119 --> 0:20:11.240 learned a long time back. You can't do it across. 0:20:11.240 --> 0:20:14.800 You can't do it early morning, through the day and 0:20:14.920 --> 0:20:18.080 late at night. So an hour before I think I 0:20:18.119 --> 0:20:20.800 want to go to bed, I will actually change what 0:20:20.840 --> 0:20:25.280 I'm doing, meaning I will start reading something interesting to 0:20:25.359 --> 0:20:28.920 me but completely outside the scope of work. I may 0:20:28.960 --> 0:20:34.320 read a biography, I might read somebody who's spontefecating on 0:20:34.400 --> 0:20:38.120 demographics and population. But I won't read it on leadership, 0:20:38.119 --> 0:20:40.720 because that's too close now. Twenty years ago I might 0:20:40.760 --> 0:20:43.800 have that would have been different. I won't read it 0:20:44.080 --> 0:20:46.560 on deep signs because that's too close to what we 0:20:46.640 --> 0:20:48.920 do for a living. So it's got to be outside 0:20:49.800 --> 0:20:53.199 the things that will make my brain churn about work. 0:20:53.640 --> 0:20:56.639 But it's got to be something that is dense enough 0:20:56.680 --> 0:20:59.560 to occupy your brain, so it shift gears. 0:21:00.119 --> 0:21:02.320 Sorry, I want to dwell on this just for a moment. 0:21:02.359 --> 0:21:03.040 The red hat thing. 0:21:03.800 --> 0:21:06.320 Was there someone or is there someone who you went 0:21:06.400 --> 0:21:10.600 to and explained the logic of this and they saw 0:21:10.600 --> 0:21:12.800 the logic of this, and that made a big difference 0:21:12.800 --> 0:21:13.000 to you. 0:21:14.720 --> 0:21:18.359 Getting their support made a big difference. You'd be surprised. 0:21:18.520 --> 0:21:25.320 I'm remarkably open inside. I mean when I have are 0:21:25.359 --> 0:21:28.360 there probably a half dozen to a dozen people inside 0:21:28.359 --> 0:21:31.040 that I'll talk to and I'll be completely open about Hey, 0:21:31.040 --> 0:21:33.359 this is what I'm thinking. I don't know. Here are 0:21:33.400 --> 0:21:35.359 the risks. I'm open about those. Also, it's not just 0:21:35.400 --> 0:21:38.440 the benefits. I think the other risks, but I think 0:21:38.480 --> 0:21:41.280 the benefits outweigh the risks. I talk about that to 0:21:41.320 --> 0:21:46.440 people all the time. So whether for example, I mean 0:21:46.520 --> 0:21:50.120 i'll take names. I think our current chro Nicol who 0:21:50.160 --> 0:21:53.000 introduced us, she has been in that loop since at 0:21:53.080 --> 0:21:56.800 least twenty fifteen. For me, if I look at our CFO, 0:21:57.400 --> 0:21:58.120 Jim Cavanaugh. 0:21:58.200 --> 0:21:58.880 He's been in that. 0:21:58.840 --> 0:22:02.720 Loop probably since twenty thirteen, and the IBM's will probably wonder, 0:22:02.960 --> 0:22:05.240 what the hell intersection did you guys have? It didn't 0:22:05.280 --> 0:22:08.200 when I talked about learning finance. I will go to 0:22:08.280 --> 0:22:09.960 him and say, hey, explain this to me. I don't 0:22:10.040 --> 0:22:13.639 understand why it's like this, and to me it's okay. 0:22:14.359 --> 0:22:15.480 A patient you go learn. 0:22:16.160 --> 0:22:18.480 If I think about many of the people in the 0:22:18.520 --> 0:22:22.160 software business, they've been having these discussions with me for always. 0:22:22.920 --> 0:22:27.399 I mean, now I'll acknowledge I can get probably impatient 0:22:27.600 --> 0:22:30.920 and a serbic, but it's meant to be a discussion. 0:22:31.000 --> 0:22:33.040 I mean, like, let's have the discussion. If you have 0:22:33.080 --> 0:22:35.199 a strong point of view, I got it. Nobody has 0:22:35.240 --> 0:22:38.040 a going to be perfectly correct, but I always look 0:22:38.119 --> 0:22:41.280 for if you have a strong point of view, that 0:22:41.320 --> 0:22:44.320 means it's from a different perspective than mine. So what 0:22:44.400 --> 0:22:47.520 do I learn from that is the question which helps 0:22:47.560 --> 0:22:48.880 to improve my point of view? 0:22:48.960 --> 0:22:50.320 That makes sense. 0:22:51.520 --> 0:22:53.400 I actually think that each person should try to build 0:22:53.400 --> 0:22:56.280 a community of a one hundred people inside your enterprise 0:22:56.680 --> 0:23:00.159 and a hundred outside that you can call up. I 0:23:00.160 --> 0:23:04.520 have no hesitation. Somebody introduced me to a long time 0:23:04.560 --> 0:23:07.600 back to a CEO on the outside. I called them 0:23:07.640 --> 0:23:08.760 up all the time and say, hey, do you have 0:23:08.800 --> 0:23:11.960 five minutes. I'm just thinking about something this way, the 0:23:12.040 --> 0:23:14.440 CEO of red At who left IBM in twenty twenty one, 0:23:15.080 --> 0:23:17.800 we probably talk every two or three months on a 0:23:17.880 --> 0:23:18.600 random topic. 0:23:18.920 --> 0:23:21.320 One way, it becomes mutual. He'll asks me my opinion 0:23:21.400 --> 0:23:21.960 on some things. 0:23:22.520 --> 0:23:24.400 Now by the way, three or four times he might 0:23:24.440 --> 0:23:28.399 do something different, but he wants my opinion to the 0:23:28.400 --> 0:23:29.040 other way around. 0:23:29.240 --> 0:23:30.960 If I gave you my phone number, can I be 0:23:31.040 --> 0:23:33.840 on that list? I would just be fascinating. I don't 0:23:33.840 --> 0:23:35.000 know if I can help you, but it would be 0:23:35.000 --> 0:23:35.880 really fun to get the call. 0:23:36.000 --> 0:23:36.439 Sure you can. 0:23:36.560 --> 0:23:39.320 Do you think that we can ever succeed unless people 0:23:39.359 --> 0:23:44.720 who influence opinions say things about us. So you may 0:23:44.720 --> 0:23:48.120 not think deeply about maybe the physics of one of computing, 0:23:48.600 --> 0:23:52.040 But would you think deeply about why. 0:23:51.720 --> 0:23:54.040 And what moment may make it much more. 0:23:53.920 --> 0:23:57.119 Attractive to a large audience. Sure you would. You'd be 0:23:57.160 --> 0:24:00.520 far better as a thinker on that topic. Probably most 0:24:00.520 --> 0:24:01.360 of the people. 0:24:02.080 --> 0:24:03.479 I was thinking, you know, when you were making your 0:24:03.480 --> 0:24:09.320 comments about your nineteen ninety self and streaming that the 0:24:10.040 --> 0:24:13.359 rational thing would have been for there have been a 0:24:13.480 --> 0:24:19.000 reserved board seat for every television network from someone from 0:24:19.000 --> 0:24:22.320 the world of technology, which I one hundred percent sure 0:24:22.320 --> 0:24:24.600 they did not have that in nineteen ninety but they 0:24:25.040 --> 0:24:28.120 they're board was probably composed of people like them. Let's 0:24:28.119 --> 0:24:32.480 talk a little bit about technology now. There's so much, 0:24:32.560 --> 0:24:34.879 so much of the changes going on right now are 0:24:35.560 --> 0:24:40.000 accompanied by a great deal of hype. What are we overestimating? 0:24:40.000 --> 0:24:41.200 What are we underestimating? 0:24:41.600 --> 0:24:44.400 Okay, let's go back to ninety ninety five the Internet, 0:24:44.440 --> 0:24:46.320 because I think that the current moment is very much 0:24:46.400 --> 0:24:48.680 like the Internet moment. Actually, all the moments in the 0:24:48.720 --> 0:24:51.679 middle were much smaller. I think mobile streaming were much smaller. 0:24:51.720 --> 0:24:54.440 Internet was the major moment. If you remember back to 0:24:54.520 --> 0:24:56.520 ninety nine and two thousand, people claimed there was a 0:24:56.520 --> 0:24:59.359 lot of hype. Would we say that the Internet of 0:24:59.359 --> 0:25:02.800 today has more than fulfilled all those expectations and more? 0:25:03.359 --> 0:25:03.639 Yes. 0:25:04.760 --> 0:25:08.280 Along the way, they'd eight out of ten of the 0:25:08.320 --> 0:25:11.440 companies that were invested in heavily go bankrupt. 0:25:11.520 --> 0:25:11.800 Yes. 0:25:12.800 --> 0:25:15.639 I actually think of that as being the huge positive 0:25:16.400 --> 0:25:20.400 of the United States capital system that that investment happened. 0:25:21.400 --> 0:25:23.640 Eight out of ten went broke. By the way, those 0:25:23.680 --> 0:25:26.600 acids didn't go away. They got consumed at ten cents 0:25:26.600 --> 0:25:28.679 in the dollar by somebody else who could then make 0:25:28.680 --> 0:25:30.840 a lot of money. But the two out of ten, 0:25:31.640 --> 0:25:34.639 just take two, it probably has paid for all the capitals. 0:25:34.800 --> 0:25:35.440 If you just. 0:25:35.480 --> 0:25:40.280 Take Amazon and Alphabet aka Google, just those two have 0:25:40.320 --> 0:25:43.679 probably paid for all the capital of that time. So 0:25:44.320 --> 0:25:47.240 that's what's going to happen this time. There will be 0:25:47.280 --> 0:25:50.560 a lot of tears, but in aggregate, there will be 0:25:50.560 --> 0:25:53.000 a lot of success. And I think that's the fundamental 0:25:53.000 --> 0:25:55.840 difference between the US model and almost. 0:25:55.600 --> 0:25:56.520 All other countries. 0:25:57.080 --> 0:25:59.439 On all other countries, they're desperate to keep all the 0:25:59.440 --> 0:26:00.000 companies alone. 0:26:00.520 --> 0:26:03.960 So that means you're dialuting. But that's a horrible thing. 0:26:04.440 --> 0:26:08.199 So to me let the system works worked really effectively, 0:26:08.400 --> 0:26:10.119 by the way, not just now, I mean all the 0:26:10.160 --> 0:26:14.840 way back to railways and electrification, and you mentioned telephone system. 0:26:15.320 --> 0:26:19.919 You can keep going on oil, I mean consumer goods. 0:26:20.280 --> 0:26:22.240 It goes on and on. I think this system is 0:26:22.359 --> 0:26:26.000 very effective. It deploys capital. Its census is a big 0:26:26.040 --> 0:26:29.880 market is completely willing to over deploy capital in the short. 0:26:29.720 --> 0:26:31.240 Term, not the long term. 0:26:31.600 --> 0:26:35.440 That results in more competition, so it actually improves a 0:26:35.520 --> 0:26:38.840 rate of innovation. That means what might have taken twenty 0:26:38.880 --> 0:26:42.920 years takes five and the winners emerge exactly the same 0:26:43.000 --> 0:26:44.240 is going to happen this time. 0:26:44.400 --> 0:26:45.800 Yeah, I saw that. 0:26:46.040 --> 0:26:50.120 I grew up in Waterloo and BlackBerry closest from Waterloo. 0:26:50.200 --> 0:26:52.119 Yep, everyone used to work for BlackBerry. 0:26:52.240 --> 0:26:52.320 Ye. 0:26:52.440 --> 0:26:55.679 BlackBerry goes into its dive and that's the best thing 0:26:55.720 --> 0:26:57.760 that happened to Waterloo because it was not just capital 0:26:57.800 --> 0:26:58.280 but talent. 0:26:58.600 --> 0:27:00.800 Yep. Talent is way any other companies. 0:27:00.840 --> 0:27:03.359 So all these smart people went on the next really 0:27:03.440 --> 0:27:07.240 more interesting thing. And yeah, the wait, you haven't answered, 0:27:08.240 --> 0:27:11.120 So what is you an idea that we are underestimating 0:27:11.119 --> 0:27:13.800 at the moment that's in the current kind of suite 0:27:13.880 --> 0:27:14.919 of innovations. 0:27:15.119 --> 0:27:17.800 So I don't think AI is being underestimated because when 0:27:17.800 --> 0:27:19.320 you look at the amount of capital and the amount 0:27:19.359 --> 0:27:23.000 of things chasing it, I think it's incredible. I do 0:27:23.040 --> 0:27:25.520 think that a lot of enterprises are deploying it in 0:27:25.560 --> 0:27:29.119 the wrong place. They're running after shiny experiments. There's a 0:27:29.200 --> 0:27:32.359 lot of basic things you can do to use AI 0:27:32.440 --> 0:27:33.800 to improve the business today. 0:27:34.200 --> 0:27:36.280 So that's really just my one advice to them. 0:27:36.600 --> 0:27:40.160 Pick areas you can scale, don't pick the shiny little 0:27:40.400 --> 0:27:41.399 toys on the side. 0:27:42.240 --> 0:27:44.480 Then I think, for example, that. 0:27:46.320 --> 0:27:51.679 If anybody has more than ten percent of what they 0:27:51.720 --> 0:27:56.439 had for customer service ten years ago, they're already five 0:27:56.520 --> 0:28:02.480 years behind. If anybody is not using AI to make 0:28:02.560 --> 0:28:07.199 their developers who write software thirty percent more productive today 0:28:07.560 --> 0:28:11.280 with the goal of being seventy percent more productive, that's 0:28:11.280 --> 0:28:13.240 not to say you will need less, you'll just get 0:28:13.280 --> 0:28:14.160 more software done. 0:28:15.560 --> 0:28:16.760 Then they're not. 0:28:17.080 --> 0:28:18.560 And I would turn around and tell you I think 0:28:18.600 --> 0:28:22.560 only maybe five percent of the enterprises on both. 0:28:22.400 --> 0:28:23.320 Those metrics today. 0:28:23.560 --> 0:28:28.280 Yeah, yeah, and the one that is completely underestimated. I 0:28:28.359 --> 0:28:32.440 kind of put it like this, Quantum today is where 0:28:32.600 --> 0:28:36.200 GPUs a AI war in twenty fifteen, and I bet 0:28:36.320 --> 0:28:39.960 you every AI person is thinking and hoping. I wish 0:28:40.000 --> 0:28:43.400 I had started doing more in twenty fifteen as opposed 0:28:43.400 --> 0:28:46.600 to wait until twenty twenty two. Quantum today is there, 0:28:46.880 --> 0:28:49.240 So it's not good enough that you can get a 0:28:49.320 --> 0:28:51.960 big advantage, But if you learn how to use it, 0:28:52.400 --> 0:28:55.840 then in five years you'll be ready to exploit what comes. 0:28:55.960 --> 0:28:58.160 Yeah, we're gonna get to quantum in a moment. But 0:28:58.200 --> 0:29:01.400 I have a couple other AI questions you know I 0:29:01.520 --> 0:29:04.160 as you know where This conversation is part of this 0:29:04.200 --> 0:29:07.400 thing that we do with IBM Smart Talks, and I've 0:29:07.400 --> 0:29:11.640 been The last episode I did was on Kenya, which 0:29:11.640 --> 0:29:15.360 has a massive deforestation problem, and they got together IBM 0:29:15.600 --> 0:29:19.600 took all the NASA satellite data, ran it through an LLM, 0:29:19.880 --> 0:29:23.600 and gave them this incredibly precise ten meter by ten 0:29:23.680 --> 0:29:27.000 meter analysis of what trees to plant, where to plant 0:29:27.000 --> 0:29:30.400 them exactly where the you know, an astonishing kind of 0:29:30.400 --> 0:29:33.880 blueprint about how to fix their country ecologically. And it 0:29:33.920 --> 0:29:38.360 made me think, when we analyze the potential of AI, 0:29:39.160 --> 0:29:41.400 are we making a mistake by spending too much thinking 0:29:41.400 --> 0:29:44.320 about the developed world when it's actually the developing world 0:29:44.640 --> 0:29:47.040 where the greatest ROI for this is to me. 0:29:47.160 --> 0:29:50.680 Look, software technologies are wonderful and the sense they can 0:29:50.680 --> 0:29:52.880 scale and they can be an ad so you don't 0:29:52.920 --> 0:29:56.520 have to do one or the other. You use deforestation. 0:29:57.080 --> 0:30:01.200 How about the use of pesticides and fertilizers, We overuse it. 0:30:01.640 --> 0:30:04.840 We tend to for irrigation, we tend to just flood everything, 0:30:04.920 --> 0:30:07.280 as opposed to say, okay, only that one needs it. 0:30:07.680 --> 0:30:09.240 You could do all those things to get a ten 0:30:09.280 --> 0:30:13.000 times effectiveness and that all would apply to the developing world. 0:30:13.480 --> 0:30:16.560 How about remote healthcare or telehealth using an AI agent. 0:30:16.880 --> 0:30:21.560 So the examples are numerous. In the developed world, I 0:30:21.720 --> 0:30:25.360 believe we are running out of people. I know that 0:30:25.600 --> 0:30:29.160 nobody likes to hear it. Most of the Far East 0:30:29.520 --> 0:30:31.400 is going to have half the number of people by 0:30:31.480 --> 0:30:33.960 twenty seventy compared to today. That's not that far away. 0:30:34.680 --> 0:30:38.000 If I look at Europe, birth rates are far under 0:30:38.440 --> 0:30:43.040 sustaining or keeping population flat. How the US, also, depending 0:30:43.040 --> 0:30:44.600 on which number you want to look at, is either 0:30:44.640 --> 0:30:46.560 one point six births per women. 0:30:46.960 --> 0:30:50.480 Or two or two point one. Why are the three numbers? 0:30:50.480 --> 0:30:52.240 One point six is to women who were born in 0:30:52.240 --> 0:30:55.440 the United States, It becomes two point two if you 0:30:55.520 --> 0:30:59.000 include immigrant women. It becomes two point one if you 0:30:59.000 --> 0:31:00.520 include children who to make getting in. 0:31:01.120 --> 0:31:03.560 So you got to decide where the trend is obvious. 0:31:04.000 --> 0:31:04.760 This is going down. 0:31:05.120 --> 0:31:08.120 So AI in the developed world is going to be 0:31:08.320 --> 0:31:11.800 essential because to keep our current quality of life, you 0:31:11.880 --> 0:31:14.600 need more work done or what's going to do the 0:31:14.640 --> 0:31:16.320 work if they're people to do the work. 0:31:16.560 --> 0:31:19.200 So the problems are different in the places. 0:31:19.280 --> 0:31:22.360 Yeah, it gives you In the developing world, you get 0:31:22.400 --> 0:31:26.479 access to a suite of technologies and things at a 0:31:26.600 --> 0:31:28.880 price that you could never been able to afford. 0:31:29.080 --> 0:31:29.440 Correct. 0:31:29.720 --> 0:31:31.600 That was my in talking to the Kenyon thing. 0:31:31.600 --> 0:31:34.320 It was like the whole it's it's maybe one of 0:31:34.320 --> 0:31:38.520 the largest ecological projects of its kind at fifteen billion 0:31:38.560 --> 0:31:39.600 trees they want to plan. 0:31:39.520 --> 0:31:41.240 And that is one country that might get it done 0:31:41.280 --> 0:31:43.520 because they do take a lot of pride in their 0:31:43.920 --> 0:31:47.680 ecology and the sort of returning to the land and 0:31:47.720 --> 0:31:48.280 giving back. 0:31:48.520 --> 0:31:48.760 Yeah. 0:31:48.960 --> 0:31:54.200 Yeah, what's different about IBM's version of AI versus some of. 0:31:54.160 --> 0:31:56.760 Your So we are not a consumer company, so we 0:31:56.800 --> 0:31:59.960 have no focus on a B two C chat pot. 0:32:00.880 --> 0:32:03.440 And the reason I say that is, if you're making 0:32:03.520 --> 0:32:06.280 a B two C chat bot, does it help you 0:32:06.360 --> 0:32:09.920 to make it even bigger and more computationally inefficient? And 0:32:09.960 --> 0:32:12.200 the short answer is yes, because you have a certain 0:32:12.240 --> 0:32:14.920 number of users, and you kind of say, I kind 0:32:14.920 --> 0:32:18.880 of say this jokingly. If I add finished to French capabilities, 0:32:19.120 --> 0:32:22.360 I can probably add five million users. If I add 0:32:23.040 --> 0:32:25.000 writing a high coup I might be able to add 0:32:25.000 --> 0:32:29.000 another five million users. If I add writing an email 0:32:29.080 --> 0:32:31.160 in the voice of Steinbeck, I can probably add another 0:32:31.160 --> 0:32:34.800 five million users. Do all those things if my goal 0:32:35.040 --> 0:32:38.440 is to get help a company summarize the legal documents 0:32:38.480 --> 0:32:39.720 in English. 0:32:40.040 --> 0:32:41.280 That can be a model. 0:32:41.000 --> 0:32:45.920 That's one hundred size as effective, probably higher quality, but 0:32:46.040 --> 0:32:49.360 I don't need to go wide. So if you're focusing 0:32:49.360 --> 0:32:52.840 on the enterprise, that actually takes away the focus of 0:32:52.920 --> 0:32:56.920 having to go to extremely large models, which but definition 0:32:56.960 --> 0:33:01.320 are going to be computationally expensive, power hungry, and demand 0:33:01.360 --> 0:33:02.480 lots and lots of data. 0:33:02.920 --> 0:33:04.480 So I can turn ontell the enterprise. 0:33:04.560 --> 0:33:06.719 You don't need to worry about copyright issues, about all 0:33:06.720 --> 0:33:09.240 those because you can train on a much smaller amount 0:33:09.240 --> 0:33:09.680 of data. 0:33:10.080 --> 0:33:11.240 And now, by. 0:33:11.080 --> 0:33:14.160 The way, turning it for you yourself is a weekend exercise, 0:33:14.400 --> 0:33:17.720 it's not a six month on a big super computer 0:33:17.800 --> 0:33:19.360 cluster somewhere out there. 0:33:19.520 --> 0:33:21.200 That's one big difference of what we do. 0:33:21.920 --> 0:33:27.400 Second, we are very focused on helping those problems that 0:33:27.440 --> 0:33:29.680 can give people immediate benefit. 0:33:29.320 --> 0:33:30.720 Where we have domain knowledge. 0:33:30.880 --> 0:33:36.560 So our domain knowledge is around operations, is around programming, encoding, 0:33:37.280 --> 0:33:43.400 is around customer service, is around customer experience, logistics, procurement. 0:33:43.880 --> 0:33:47.200 Let's change the areas where we have a lot of expertise. 0:33:47.880 --> 0:33:51.120 And then three we kind of apply it to ourselves 0:33:51.800 --> 0:33:54.200 and so we are not asking our clients to be 0:33:54.320 --> 0:33:55.680 the first experiment on it. 0:33:55.760 --> 0:33:59.040 We say you can leverage what we did. We're happy to. 0:33:59.000 --> 0:34:01.840 Bring out all our learnings, including what needs to change 0:34:02.320 --> 0:34:04.719 in the process, because the biggest change is not technology, 0:34:05.120 --> 0:34:08.000 is getting people to accept that there's a different way 0:34:08.040 --> 0:34:08.720 to do things. 0:34:09.640 --> 0:34:14.080 Other challenges to explaining what makes you different to potential customers. 0:34:14.080 --> 0:34:17.040 For sure, the shining object is always attractive. Oh, I 0:34:17.120 --> 0:34:20.480 can go and try chat GPT. Why don't you have 0:34:20.520 --> 0:34:21.560 your GPT version? 0:34:23.080 --> 0:34:24.240 Do you use chat gipt? 0:34:25.200 --> 0:34:26.040 I have used it. 0:34:27.440 --> 0:34:29.440 I asked it a question recently which I thought was 0:34:29.480 --> 0:34:32.800 really simple, and it made up about ten people. 0:34:34.239 --> 0:34:35.440 Anyway, I had a bad experience. 0:34:35.480 --> 0:34:38.719 I should think that that's the fundamental issue with all 0:34:39.040 --> 0:34:41.480 lms as they get larger. Yeah, because you had to 0:34:41.520 --> 0:34:45.560 ask what was the original insight that led to these 0:34:46.719 --> 0:34:50.960 It was a reward function with intent. So if it 0:34:51.040 --> 0:34:55.000 has learned by using a reward function, it's reward function 0:34:55.200 --> 0:34:59.160 comes from giving an answer that satisfies you. So if 0:34:59.160 --> 0:35:01.279 it thinks that if it makes up an answer that 0:35:01.320 --> 0:35:04.600 will satisfy you, how will you stop it? Why do 0:35:04.680 --> 0:35:08.520 we think this is different than the clever college kid 0:35:08.520 --> 0:35:09.439 who doesn't know an answer. 0:35:09.480 --> 0:35:12.840 What bullshits the way to an answer. Well, it's exactly 0:35:12.920 --> 0:35:13.239 the same. 0:35:13.280 --> 0:35:16.240 It's like the example of clever hands in that story 0:35:16.120 --> 0:35:18.920 the horse that they thought could speak, then all it. 0:35:19.000 --> 0:35:21.880 Was doing was pleasing it. It's master. Yes, it is 0:35:21.880 --> 0:35:22.959 a little bit of clever hands. 0:35:23.040 --> 0:35:25.240 Yeah, it's like dogs kind of imitating and looking. 0:35:26.520 --> 0:35:31.640 What would you identify as the most significant bottleneck in 0:35:31.719 --> 0:35:34.200 the development of AI? What's slowing us down right now? 0:35:36.080 --> 0:35:40.600 I am not convinced that LLMS is the way to 0:35:40.719 --> 0:35:45.880 get much beyond where we'll get incremental improvements, But I, 0:35:46.360 --> 0:35:49.319 for one, don't believe that LMS are going to get 0:35:49.400 --> 0:35:55.480 us to super intelligence or AGI. So I'll park that 0:35:55.600 --> 0:35:58.120 on the side and simply say, we have to find 0:35:58.120 --> 0:36:02.680 a way to fuse and how do you represent knowledge 0:36:03.239 --> 0:36:06.000 as opposed to have to statistically rediscover it each time 0:36:06.000 --> 0:36:09.240 I ask a question? And how do we fuse knowledge 0:36:09.320 --> 0:36:15.760 with LLM? Maybe then we'll get to leaves and beyond 0:36:15.800 --> 0:36:20.360 today on LMS alone, my view is, I think we 0:36:20.440 --> 0:36:24.520 can get a thousand x efficiency in power and cost 0:36:24.640 --> 0:36:28.040 and compute from today. So if you make something a 0:36:28.120 --> 0:36:31.359 thousand times cheaper, would people use a lot more of it? 0:36:32.520 --> 0:36:32.880 Yes? 0:36:33.719 --> 0:36:37.040 And I think those answers lie as is usually in compute. 0:36:37.239 --> 0:36:42.080 So advances in semiconductors, advances in software, and advances in 0:36:42.120 --> 0:36:45.160 agorthic techniques all three. But how come we're not working 0:36:45.160 --> 0:36:46.600 in any of those three. We're just taking the current 0:36:46.600 --> 0:36:51.280 sevic conductor and going more. We're taking the current algorithmic 0:36:51.360 --> 0:36:53.319 techniques and not really trying to invent new ones. 0:36:53.440 --> 0:36:56.440 So I think those are all happen less than five years. 0:36:58.200 --> 0:37:01.640 But why say there is a we're in a moment 0:37:01.640 --> 0:37:06.719 where people are not pursuing the the optimal strategy for 0:37:06.800 --> 0:37:08.040 exploiting this technology. 0:37:08.360 --> 0:37:13.120 Why because when you see a few people running really 0:37:13.160 --> 0:37:16.200 hard and they're willing to invest any amount of money, 0:37:16.200 --> 0:37:20.239 so efficiency is not the focus. People feel if you 0:37:20.280 --> 0:37:21.880 don't do the same, you'll get left behind. 0:37:23.120 --> 0:37:25.879 So is this a case where there's too much money 0:37:26.040 --> 0:37:27.160 ones have never had for more? 0:37:27.239 --> 0:37:27.640 Right? Ever? 0:37:28.200 --> 0:37:33.480 Yeah, but this is this a consequence of overinvestment in 0:37:33.520 --> 0:37:34.560 the in the field. 0:37:34.680 --> 0:37:36.960 Going back to my internet allology, if two out of 0:37:37.040 --> 0:37:40.440 ten are going to succeed, how do you guarantee or 0:37:40.440 --> 0:37:42.120 how do you improve the orders that you are one 0:37:42.120 --> 0:37:45.080 of those two? So if you pause to say I 0:37:45.120 --> 0:37:47.200 want to become more efficient, that's not the way to win. 0:37:47.520 --> 0:37:49.640 So first you win, then you become efficient. 0:37:50.719 --> 0:37:54.560 Yeah, let's talk about what is I was told your 0:37:54.560 --> 0:37:59.759 favorite topic, it's quantum. It is what boy even go 0:37:59.800 --> 0:38:00.279 to further? 0:38:00.400 --> 0:38:01.920 Why is quantum your favorite topic? 0:38:03.160 --> 0:38:06.080 We've only had two kinds of compute in the history, 0:38:06.960 --> 0:38:10.319 so nineteen forty five was to use that year for 0:38:10.400 --> 0:38:13.160 anyac all the way till twenty twenty, we had one 0:38:13.200 --> 0:38:16.040 kind of computerlassical what today you would call a classical computer. 0:38:17.600 --> 0:38:22.080 Then GPUs and AI came around, so you would say 0:38:21.880 --> 0:38:25.080 the intuition there is you went from sort of bits, 0:38:25.120 --> 0:38:29.200 which is algebra or high school algebra, to including neurons, 0:38:29.320 --> 0:38:32.719 which is captured in linear algebra. But that gives you 0:38:32.719 --> 0:38:36.200 a different kind But it can do problems that are 0:38:36.280 --> 0:38:38.359 really hard to do. I don't say impossible, just hard 0:38:38.400 --> 0:38:43.000 to do on normal computers. Quantum adds a third kind 0:38:43.000 --> 0:38:49.560 of math. Yes, the physics properties which really get people 0:38:49.719 --> 0:38:52.640 energized and the imagination going. And we use all these 0:38:52.640 --> 0:38:57.160 words about entanglement and silver position. But maybe because I'm 0:38:57.160 --> 0:38:59.800 a bit of a math guy. The real thing is 0:39:00.200 --> 0:39:03.120 a third kind of math to make it really simple, 0:39:03.600 --> 0:39:06.919 a third kind of math that comes from the field 0:39:06.960 --> 0:39:08.120 of abstract algebra. 0:39:08.560 --> 0:39:09.360 It does the math. 0:39:10.320 --> 0:39:13.440 You can use Hamiltonians for those who like physics, or 0:39:13.480 --> 0:39:15.640 you can use the word Lee algebras for those who 0:39:15.800 --> 0:39:19.759 like abstract mathematics. If you can do a third kind 0:39:19.760 --> 0:39:22.759 of math, which algorithms are suited to that third kind 0:39:22.760 --> 0:39:26.120 of math? So it excites me because we can now 0:39:26.400 --> 0:39:29.360 approach algorithms that you just could never do on the 0:39:29.400 --> 0:39:32.319 other two it's impossible. Now it's different than AI. It's 0:39:32.320 --> 0:39:35.600 not data intensive, it's compute intensive. So we kind of 0:39:35.640 --> 0:39:38.560 had compute and supercomputers. Then we went to data which 0:39:38.600 --> 0:39:40.840 is AI. And now if you say there's another class 0:39:40.840 --> 0:39:42.520 of problems, it require lots of compute. 0:39:42.680 --> 0:39:44.440 That's quantum. 0:39:44.480 --> 0:39:47.120 A couple months ago was at the to watch some 0:39:47.160 --> 0:39:49.240 research center and they have you know, on the ground floor, 0:39:49.239 --> 0:39:50.680 they have those behind the glass. 0:39:51.239 --> 0:39:55.279 There's incredibly exciting looking machines. But where are we in 0:39:55.320 --> 0:39:56.399 the timeline of this. 0:39:57.960 --> 0:40:01.560 Three to five years away from shocking people? 0:40:02.520 --> 0:40:03.839 What does shocking people mean? 0:40:04.200 --> 0:40:07.000 Do something that nobody thought was possible in that timeline? 0:40:07.280 --> 0:40:08.720 Does an example come to mind? 0:40:09.560 --> 0:40:14.080 I was actually pleasantly surprised. So one of our clients, HSBC, 0:40:15.160 --> 0:40:19.280 last week published a result that using a quantum computer. 0:40:20.520 --> 0:40:24.680 Bond trading was thirty four percent more accurate than their 0:40:24.719 --> 0:40:25.520 prior technique. 0:40:26.440 --> 0:40:27.640 Thirty four percent. 0:40:27.760 --> 0:40:28.560 Thirty four percent. 0:40:29.120 --> 0:40:32.480 This is an industry that's used to one percent correct, 0:40:32.760 --> 0:40:34.000 zero point five percent. 0:40:34.480 --> 0:40:36.880 Yes, that's astonishing. 0:40:37.160 --> 0:40:39.960 Now that was not at a scale when they could 0:40:40.000 --> 0:40:42.600 turn it into production today, but that was sort of 0:40:43.040 --> 0:40:45.520 their original thought experiment, and that's what they did. 0:40:46.040 --> 0:40:51.239 Now. Can you imagine when will somebody so you. 0:40:50.960 --> 0:40:53.920 Were correct, you talk about an industry where one basis point, 0:40:54.239 --> 0:40:57.360 if I remember I may be wrong, like thirteen trillion 0:40:57.440 --> 0:40:59.720 dollars of money kind of moves around in the financial 0:40:59.719 --> 0:41:07.080 industr each day, right, So basis point would be thirteen 0:41:07.120 --> 0:41:11.600 billion something like that, right, one over ten thousand. So 0:41:11.920 --> 0:41:14.280 when you think about the kind of profit that people 0:41:14.280 --> 0:41:16.799 can make, if you can tell somebody that you can 0:41:16.840 --> 0:41:21.680 come up with a better price than your competition by 0:41:21.800 --> 0:41:24.759 just one basis point, they would actually gain the dire 0:41:24.840 --> 0:41:30.040 market share. Yeah, so I think something around there or 0:41:30.080 --> 0:41:31.680 something in the world of materials. 0:41:31.960 --> 0:41:35.160 Can we make a better battery? Could we make a 0:41:35.200 --> 0:41:36.200 solid state battery? 0:41:37.520 --> 0:41:42.239 Which means your risk of fires heating decrease dramatically. 0:41:42.280 --> 0:41:44.560 And the reason, sorry to ask a really nice question. 0:41:45.920 --> 0:41:48.279 Why is it that a quantum computer would be better 0:41:48.320 --> 0:41:52.840 at solving a battery problem than our existing methods of computing. 0:41:52.920 --> 0:41:57.080 So the equations of quantum mechanics and chemistry and how 0:41:57.640 --> 0:42:02.000 things interact are well known. To solve them, that are 0:42:02.000 --> 0:42:04.279 no known techniques. So these are not like closed form, 0:42:04.320 --> 0:42:06.160 you know, it's not like the square root of a 0:42:06.239 --> 0:42:09.480 qualtertic equation. So the only way to solve them is 0:42:09.520 --> 0:42:12.480 to explore the state space. So if you have a 0:42:12.480 --> 0:42:17.399 few hundred electrons, you need two to the one hundred states. Well, 0:42:17.480 --> 0:42:19.719 I'm sorry, you don't have that much memory. It's impossible. 0:42:20.160 --> 0:42:23.000 So it takes a really really long time on a 0:42:23.040 --> 0:42:26.160 normal computer to solve those problems. Right, But that's simple 0:42:26.200 --> 0:42:31.280 a problem. If a quantum computer operates in the equation domain, 0:42:31.760 --> 0:42:33.759 it doesn't need to explore the state space. It can 0:42:33.800 --> 0:42:36.080 actually solve it. That's why I call it a different 0:42:36.200 --> 0:42:38.600 kind of math. That's the kind of math it does. 0:42:39.239 --> 0:42:42.000 So in a couple of seconds, it can tell you 0:42:42.400 --> 0:42:44.080 this is how that material will be here. 0:42:45.560 --> 0:42:46.960 Oh, I see, so you've. 0:42:46.760 --> 0:42:50.000 Taken what could take years to a few seconds. Yeah, 0:42:50.080 --> 0:42:51.640 that's a pretty big change. 0:42:51.760 --> 0:42:54.080 Yeah. Yeah, it's speaking a different language. 0:42:54.360 --> 0:42:54.640 Different. 0:42:54.880 --> 0:42:57.120 So any kind of problem that comes along. 0:42:56.960 --> 0:43:00.000 That's specific to that language correctly, which is not all problems. 0:43:00.080 --> 0:43:02.040 Yeah, just as I call it. 0:43:02.040 --> 0:43:03.120 It's one more kind of math. 0:43:03.440 --> 0:43:07.719 Yeah, what's an example? So so many questions. L give 0:43:07.719 --> 0:43:11.359 me another example of a of a of a kind 0:43:11.400 --> 0:43:13.439 of problem that a quantum computer would love. 0:43:15.239 --> 0:43:18.560 This one is a bit more speculative, and I'm going 0:43:18.600 --> 0:43:21.839 to use a little bit of poetic license. So let's 0:43:21.840 --> 0:43:25.319 take a post office in a mid size country. They 0:43:25.400 --> 0:43:28.120 probably burn a billion gallons of fuel per year delivering 0:43:28.120 --> 0:43:31.640 packages and letters. Because most posts and advanced country says 0:43:31.960 --> 0:43:37.440 every house, every address, each day. The way to optimize 0:43:37.480 --> 0:43:39.840 this is we can formulate the problem. It's called the 0:43:39.880 --> 0:43:44.040 proveling sales and problem solving. It is really hard, so 0:43:44.040 --> 0:43:48.239 people have heuristics. Let's suppose today our heuristics get us 0:43:48.280 --> 0:43:51.880 to within twenty percent of the optimal answer. Let's suppose 0:43:51.880 --> 0:43:55.879 a quantum computer can get you the next ten percent. Well, 0:43:55.920 --> 0:43:58.160 if I can get ten percent of a billion gallons 0:43:58.160 --> 0:43:59.800 that I think is one hundred million gallons of my 0:43:59.880 --> 0:44:02.840 man others right, and in the country I'm thinking about, 0:44:03.080 --> 0:44:06.360 that could be eight hundred million pounds of saving to 0:44:06.440 --> 0:44:12.040 one entity in one year, and the associated carbon footprint 0:44:12.160 --> 0:44:16.000 climate change were in't less mileage on vague. I'm not 0:44:16.000 --> 0:44:19.719 even counting all that. These are pretty attractive problems to 0:44:19.760 --> 0:44:22.560 go after. So if I look at the interest recently, 0:44:23.800 --> 0:44:28.239 New York started a whole program in some places. Illinois 0:44:28.680 --> 0:44:32.400 stood up a quantum algorithm center between a number of 0:44:32.400 --> 0:44:37.319 the universities. The governor there was heavily behind it, etc. 0:44:38.040 --> 0:44:41.080 So I wouldn't say that this is widespread. This is 0:44:41.080 --> 0:44:42.960 why I'm saying three to four years for that moment. 0:44:43.600 --> 0:44:47.200 But there's enough people who are deeply cognizant who are saying, 0:44:47.200 --> 0:44:49.279 wait a moment, we kind of get it. 0:44:49.320 --> 0:44:51.000 This is a new kind of man. What are the 0:44:51.040 --> 0:44:52.120 new problems we can solve? 0:44:52.840 --> 0:44:55.280 And the fact that we have about roughly two hundred 0:44:55.280 --> 0:44:58.920 clients who worked with us very early stage, small experiments 0:44:59.239 --> 0:45:01.600 is because they're in is I could do something here 0:45:01.600 --> 0:45:04.040 that I couldn't do in other places. 0:45:04.320 --> 0:45:05.960 Three to four years is not a long time. 0:45:06.400 --> 0:45:06.600 No. 0:45:08.520 --> 0:45:10.280 But if I'm in the battery business. 0:45:10.520 --> 0:45:14.040 And I don't have a line out to a quantum 0:45:14.239 --> 0:45:16.319 computing experiment. 0:45:17.480 --> 0:45:19.359 I have a problem. Don't have a problem. 0:45:19.800 --> 0:45:23.839 Yeah, you'll probably be out of business in ten years. Well, 0:45:23.840 --> 0:45:25.680 maybe you could write a big check and buy the 0:45:25.719 --> 0:45:27.040 technology from somebody else. 0:45:27.719 --> 0:45:30.600 You had to what is quantum of rank in the 0:45:30.680 --> 0:45:33.840 kind of great inventions of the last one hundred and 0:45:33.880 --> 0:45:34.719 fifty years. 0:45:34.480 --> 0:45:40.000 Equal to something conductor? And I think that if sem 0:45:40.080 --> 0:45:42.160 Conductor's vanished, modern life. 0:45:42.000 --> 0:45:43.840 Would stop, like just stop. 0:45:44.560 --> 0:45:53.040 Yeah, no electricity, no automobile, no streaming. You can imagine 0:45:53.080 --> 0:45:54.759 the yells from all the kids who ever hear that 0:45:54.840 --> 0:45:55.800 no streaming? 0:45:57.680 --> 0:46:01.719 The and is that it's funny because don't. As someone 0:46:01.760 --> 0:46:05.480 who's outside this world, I feel like quantum is underdiscussed 0:46:05.520 --> 0:46:09.520 relative to its potential for transforming society. 0:46:09.480 --> 0:46:13.279 Because I use my Internet example. Ninety five was the 0:46:13.280 --> 0:46:18.200 moment with Netscape that Internet came on people's consciousness. I said, 0:46:18.239 --> 0:46:20.360 when eighty five I considered it to be this is 0:46:20.400 --> 0:46:24.000 a solve problem because it needs something that makes it 0:46:24.080 --> 0:46:25.760 accessible easy. 0:46:26.080 --> 0:46:28.680 That was the browser. The Netscape browser. 0:46:28.400 --> 0:46:32.640 Is what brought it easy to understand. We have probably, 0:46:32.680 --> 0:46:35.600 as I said, about four to five years from that moment. 0:46:35.640 --> 0:46:38.560 That's why it's under discussed, because the moment I say, 0:46:38.560 --> 0:46:40.480 and you got a math, I've probably lost ninety nine 0:46:40.520 --> 0:46:43.280 percent of the audience. If I go to quantum mechanics. 0:46:43.320 --> 0:46:46.680 I've probably lost nine percent of the audience. 0:46:48.480 --> 0:46:51.640 So you, as c over the last five years, have 0:46:51.800 --> 0:46:54.680 been really the birth mother for a lot of the 0:46:54.760 --> 0:46:59.319 quantum computing work. I'm curious, so you come in When 0:46:59.360 --> 0:47:03.800 you started this, CEO, was this your first priority. 0:47:04.320 --> 0:47:06.840 I had already started investing in it back in twenty 0:47:06.880 --> 0:47:12.520 fifteen when I was leading IBM research. So let me 0:47:12.560 --> 0:47:14.360 acknowledge and like nobody should try to copy it. And 0:47:14.400 --> 0:47:16.960 I've had a i'll call it a weird career. 0:47:17.400 --> 0:47:19.359 I was a researcher at some point. 0:47:19.360 --> 0:47:20.640 If he had asked me out, I said, I'm one 0:47:20.680 --> 0:47:22.239 of those people, you know, throw a pizza under our 0:47:22.280 --> 0:47:23.560 door and like leave me alone. 0:47:23.719 --> 0:47:24.839 I don't want to talk to people. 0:47:25.360 --> 0:47:27.719 Then I decided I was interested in the business. Then 0:47:27.760 --> 0:47:30.759 I went and started acquiring companies and doing that. Then 0:47:30.800 --> 0:47:32.840 somebody told me, hey, why did you start doing some 0:47:32.880 --> 0:47:37.560 business strategy. Then I went back to research and led 0:47:37.560 --> 0:47:40.359 our research division for a couple of years. And when 0:47:40.400 --> 0:47:44.160 the people described it to me, I asked some questions. 0:47:44.239 --> 0:47:46.480 So it wasn't a big investment. 0:47:46.000 --> 0:47:46.399 At that time. 0:47:46.440 --> 0:47:48.719 It was hey, can we make a computer not just 0:47:48.760 --> 0:47:52.239 a science experiment? Can it run by itself all night? 0:47:52.760 --> 0:47:55.839 Can you think about software so that even people who 0:47:55.840 --> 0:47:58.440 are not deeper quantum mechanics can begin to use it? 0:47:59.360 --> 0:48:00.760 And they began to do those things. 0:48:01.040 --> 0:48:05.640 So over three four years, did I get enough confidence, Yeah, okay, 0:48:05.640 --> 0:48:07.320 this is something that can really work. 0:48:07.840 --> 0:48:09.279 And then you've got to nurture it to. 0:48:10.000 --> 0:48:12.759 Where it gets bigger and bigger until you get the 0:48:12.800 --> 0:48:14.479 confidence that okay, now it's a big bet. 0:48:15.280 --> 0:48:18.120 And what was the moment when you when you realize now. 0:48:17.960 --> 0:48:22.240 It's a big bet, probably two or three years ago. 0:48:22.760 --> 0:48:25.280 How do you decide, as the head of a company 0:48:25.360 --> 0:48:29.319 like this, how much money, how many resources, and how 0:48:29.360 --> 0:48:31.399 many people and how what kind of prominence to give 0:48:31.440 --> 0:48:32.239 to an idea like that? 0:48:33.000 --> 0:48:37.040 So three layers the set of people who actually have 0:48:37.200 --> 0:48:41.680 the knowledge and the intensity to fundamentally advance the technology. 0:48:42.400 --> 0:48:44.080 If I could find more out higher then. 0:48:44.080 --> 0:48:47.359 So I'm constrained on people on that one because normally 0:48:47.400 --> 0:48:49.600 there's only so many people who can do these things. 0:48:50.320 --> 0:48:52.960 Two, you got to be careful. 0:48:53.239 --> 0:48:56.000 If you push too hard on timing, you will get 0:48:56.000 --> 0:48:58.719 people to take so much risk that actually the thing 0:48:58.760 --> 0:49:02.560 will fail. So that's the art of between the leadership 0:49:02.600 --> 0:49:06.279 on the project and me to say, Okay, how hard 0:49:06.320 --> 0:49:08.959 can you push? But not so hard that you cause 0:49:09.000 --> 0:49:12.160 it to fail because then they get compelled to commit 0:49:12.200 --> 0:49:14.239 timelines that are just impossible. 0:49:14.800 --> 0:49:17.480 Yeah, how do you This is fascinating. 0:49:17.600 --> 0:49:20.759 So it's ultimately a question of judgment trying to figure 0:49:20.760 --> 0:49:23.399 out what's the sweet spot between enough pressure to keep 0:49:23.440 --> 0:49:25.960 them ahead of the pack, but not too much pressure 0:49:26.360 --> 0:49:29.759 so that they start taking risks. How do you calibrate 0:49:30.040 --> 0:49:32.839 whether you're hitting that sweet spot? I mean, do you 0:49:33.040 --> 0:49:36.239 reassess every few months and say I think I'm over 0:49:36.280 --> 0:49:37.960 correcting or undercorrecting at this moment. 0:49:38.800 --> 0:49:42.160 So one you got to have what I call and 0:49:42.239 --> 0:49:44.920 this is channeling a word from one of my favorite books, 0:49:44.960 --> 0:49:48.879 The Geekway, How open can you be? So I want 0:49:48.920 --> 0:49:52.399 to press hard, but the team knows that they're allowed 0:49:52.440 --> 0:49:56.600 to push back and really argue back hard. That means 0:49:56.600 --> 0:50:02.360 you'll get to probably that correct goldilocks pressure do the 0:50:02.480 --> 0:50:05.240 people themselves should want to go as hard as possible, 0:50:05.760 --> 0:50:09.320 but not harder than possible. So that is then personality 0:50:09.719 --> 0:50:12.000 of leadership that makes sense. 0:50:12.239 --> 0:50:14.800 But you have to be someone who people feel comfortable 0:50:14.840 --> 0:50:15.440 being honest with. 0:50:15.640 --> 0:50:19.200 Yes, absolutely, and people feel. 0:50:18.960 --> 0:50:20.040 Comfortable being honest with you. 0:50:20.239 --> 0:50:20.880 I believe so. 0:50:21.239 --> 0:50:25.719 Yeah. When has there been a moment in this path 0:50:25.760 --> 0:50:28.560 with quantum where you did think you were pushing too hard? 0:50:30.360 --> 0:50:35.000 No, because I think that the leadership there will argue 0:50:35.000 --> 0:50:37.959 back with me any day of the week. I don't 0:50:37.960 --> 0:50:40.400 think that they feel that they have to forward. 0:50:41.480 --> 0:50:44.600 Do you drop by at sort of Saturday night at 0:50:44.600 --> 0:50:46.040 ten pm to see if people are working? 0:50:47.120 --> 0:50:53.080 I tend to text people and ask questions and like, 0:50:53.160 --> 0:50:55.880 I'll read something and say, hey, are these people doing this? 0:50:56.440 --> 0:51:00.400 And if they can answer me in reasonable terms, I 0:51:00.440 --> 0:51:03.640 actually don't say great. They're already watching the competition, they're 0:51:03.680 --> 0:51:06.920 watching the literature, they're watching the science. I don't need 0:51:06.960 --> 0:51:09.200 to push hard. If they are already ahead of it, 0:51:09.239 --> 0:51:12.280 then me I can answer my question. I'll say thoughtfully, 0:51:12.600 --> 0:51:16.719 not always completely accurately. You're thinking about it on their own. 0:51:16.719 --> 0:51:17.919 I don't need to push Yeah. 0:51:18.760 --> 0:51:21.239 One last question I wanted to ask you, do you 0:51:21.320 --> 0:51:23.200 have the most interesting job in America? 0:51:23.360 --> 0:51:25.719 I believe that it's the most interesting job, which I 0:51:25.719 --> 0:51:26.880 won't give up anything. 0:51:27.880 --> 0:51:30.760 It also sounds like you're enjoying yourself. 0:51:31.320 --> 0:51:34.600 I enjoy it as long as look my role and 0:51:34.719 --> 0:51:38.719 goal should be to make the enterprise thrive as long 0:51:38.760 --> 0:51:42.040 as than making the enterprise thrive and are clients delighted? 0:51:43.640 --> 0:51:46.520 I love it. If I don't, somebody else should do it. 0:51:47.520 --> 0:51:50.840 Harvin, this has been so much fun. Thank you so 0:51:50.960 --> 0:51:56.440 much taking the time and a fascinating, completely fascinating conversation. 0:51:56.480 --> 0:51:58.120 I wish I was one of those people who could 0:51:58.160 --> 0:52:00.440 help you out with quantum, but I'm afraid I'm not 0:52:01.080 --> 0:52:01.759 in a few years. 0:52:03.200 --> 0:52:04.680 Good. Thank you so much. 0:52:14.719 --> 0:52:18.400 Smart Talks with IBM is produced by Matt Ramano, Amy Gaines, McQuaid, 0:52:18.880 --> 0:52:23.080 Trina Menino, and Jake Harper. Mastering by Sarah Bugerer, music 0:52:23.080 --> 0:52:28.880 by Gramoscope, Strategy by Tatiana Lieberman, Cassidy Meyer and Sophia Derlong. 0:52:29.600 --> 0:52:32.239 Smart Talks with IBM is a production of Pushkin Industries 0:52:32.440 --> 0:52:36.840 and Ruby Studio at iHeartMedia. To find more Pushkin podcasts, 0:52:37.239 --> 0:52:41.160 listen on the iHeartRadio app, Apple Podcasts, or wherever you 0:52:41.239 --> 0:52:45.160 listen to podcasts. I'm Malcolm Glabo. This is a paid 0:52:45.160 --> 0:52:49.719 advertisement from IBM. The conversations on this podcast don't necessarily 0:52:49.800 --> 0:53:03.800 represent IBM's positions, strategies, or opinions.