WEBVTT - Former IBM CEO Sam Palmisano Talks AI 0:00:02.400 --> 0:00:06.760 Bloomberg Audio Studios, Podcasts, radio news. 0:00:06.960 --> 0:00:09.240 All Right, well, the biggest headline, of course, out of 0:00:09.240 --> 0:00:13.600 the WWDC was Tim Cook talking about artificial intelligence. 0:00:15.120 --> 0:00:19.640 Recent developments and generative intelligence and large language models offer 0:00:19.880 --> 0:00:24.439 powerful capabilities that provide the opportunity to take the experience 0:00:24.480 --> 0:00:29.680 of using Apple products to new heights. Introducing Apple Intelligence, 0:00:30.240 --> 0:00:33.839 the new personal intelligence system that makes your most personal 0:00:33.960 --> 0:00:36.680 products even more useful and delightful. 0:00:37.880 --> 0:00:40.199 Here in the studio with more is Sam Palmisano. He's 0:00:40.240 --> 0:00:43.280 former CEO of IBM. He is also currently the board 0:00:43.320 --> 0:00:46.360 chair of America's Frontier Fund, and nonprofit developed to support 0:00:46.360 --> 0:00:50.200 critical tech to ensure US competitiveness. Sam, it's so great 0:00:50.200 --> 0:00:52.199 to get you on set, your first guest on set for. 0:00:52.159 --> 0:00:54.040 The new show. I'm honor, thank you so much. 0:00:54.080 --> 0:00:55.560 We're going to get you a hatter ache show that 0:00:55.560 --> 0:00:57.920 it's something okay, perfect, We're gonna work my hair, little do, 0:00:58.160 --> 0:00:59.840 We're gonna work with shoot someone. 0:01:01.200 --> 0:01:01.480 Okay. 0:01:01.880 --> 0:01:03.840 So let's take a broader view. Then when it comes 0:01:03.920 --> 0:01:07.520 to AI, how does the US take the appropriate amount 0:01:07.600 --> 0:01:10.200 of risk to make sure that it can stay ahead. 0:01:10.800 --> 0:01:13.959 That's a great question, quite honestly, and I'll step back 0:01:14.040 --> 0:01:16.200 and say, look, at the models today and why we 0:01:16.319 --> 0:01:19.839 created the American Frontier Fund. The models today are heavily 0:01:19.959 --> 0:01:23.680 oriented around software. You heard that today or the Apple announcements. 0:01:24.080 --> 0:01:25.720 To stay ahead in tech, it has to. 0:01:25.760 --> 0:01:28.160 Be much more than just software, So you need long 0:01:28.240 --> 0:01:31.800 term investments in science and research. US is far ahead 0:01:31.840 --> 0:01:35.160 in science and research, but we're losing track on commercialization 0:01:35.600 --> 0:01:38.840 that builds the new technologies that changes the future, and 0:01:38.959 --> 0:01:41.880 not just consumer convenience. I mean how healthcare is done, 0:01:42.200 --> 0:01:44.880 how banking and our financial systems are run, how the 0:01:44.959 --> 0:01:48.520 national security, national defense occurs. All those types of things 0:01:48.600 --> 0:01:51.640 require much longer cycles of investment, and that's why we 0:01:51.760 --> 0:01:53.800 created this thing called the American Frontier Fund. 0:01:53.920 --> 0:01:55.720 We want to keep the West ahead of the East. 0:01:56.280 --> 0:01:57.560 So we have the Chips Act. 0:01:57.680 --> 0:02:00.840 Right, Yes, we clearly have big tech innovating. Like where 0:02:00.880 --> 0:02:02.960 are we falling short? Like where do you think the money, 0:02:03.040 --> 0:02:05.960 in the talent and the time needs to go into That's. 0:02:05.840 --> 0:02:09.160 A great question because obviously the issues telling me that, Well, okay, 0:02:09.200 --> 0:02:11.119 then I'll tell you that's an okay question. 0:02:11.240 --> 0:02:12.919 I asked Roman for a good question. 0:02:13.760 --> 0:02:17.480 But seriously, what it is that the point that the 0:02:17.600 --> 0:02:20.959 cycles that we're talking about for these very advanced applications 0:02:21.160 --> 0:02:23.639 to require I'll say, let's say ten to fifteen years, 0:02:23.680 --> 0:02:25.679 we call it the value of death. To take it 0:02:25.760 --> 0:02:28.600 from science and research through the value death to get 0:02:28.639 --> 0:02:31.639 it to commercialization takes a long time. So to do 0:02:31.760 --> 0:02:34.919 that you have to create a partnerships or an ecosystem 0:02:34.960 --> 0:02:38.720 with the research laboratories, with the academic institutions, with the 0:02:38.919 --> 0:02:41.360 private sector as well, and you need government involved because 0:02:41.360 --> 0:02:44.519 of the risk associated of getting through the value of death. 0:02:44.680 --> 0:02:45.960 And that's what we are about. 0:02:46.000 --> 0:02:49.600 And what we've created is a new model for commercialization 0:02:49.800 --> 0:02:52.040 of this tech that it's required to keep the US 0:02:52.080 --> 0:02:54.040 ahead of the rest of the world, but also the 0:02:54.120 --> 0:02:56.919 funding models that make it attractive for private investment to 0:02:57.000 --> 0:02:58.720 come in to these opportunities. 0:02:58.880 --> 0:03:00.600 Do we have that right now? I mean, when you 0:03:00.680 --> 0:03:02.919 look at sort of all the movement we see seen 0:03:02.919 --> 0:03:04.800 an AI as of lately, so much of it seems 0:03:04.800 --> 0:03:06.600 to be driven by the private sector. And even when 0:03:06.639 --> 0:03:09.000 the government got involved, who do they turn to first? 0:03:09.040 --> 0:03:11.440 They said, basically they brought the private sector and it 0:03:11.480 --> 0:03:14.120 said give us a solution to regulating this, which I 0:03:14.160 --> 0:03:15.880 think caused everyone to kind of scratch their heads. 0:03:15.960 --> 0:03:19.560 Well, they're t two different questions regulations different than actually 0:03:19.600 --> 0:03:22.600 the creation of the tech and the creation of the tech. 0:03:22.720 --> 0:03:25.120 The reason why go back to the early days of 0:03:25.360 --> 0:03:28.000 your INCU tell the CEO of the organization's give them 0:03:28.240 --> 0:03:29.280 who created incutel. 0:03:29.520 --> 0:03:31.360 That was for national security and defense. 0:03:31.840 --> 0:03:33.400 But it was the same sorts of things that what 0:03:33.480 --> 0:03:36.480 could you do long term in technology to make the 0:03:36.520 --> 0:03:40.320 country more secure? And sometimes you need governments involvement because 0:03:40.320 --> 0:03:43.040 of their risk factors, no different than the Chips Act. 0:03:43.120 --> 0:03:46.440 As far as you're talking about billions of dollars to 0:03:46.520 --> 0:03:49.000 create these facilities which are necessary that we control the 0:03:49.040 --> 0:03:50.320 resilience of the supply chain. 0:03:50.720 --> 0:03:52.400 But who's going to take the risk of forty or 0:03:52.480 --> 0:03:54.280 fifty billion in the private sector? 0:03:54.600 --> 0:03:57.120 I mean, Intel might invest twenty billion, but you're doubling 0:03:57.200 --> 0:03:59.920 it up now for these kinds of technologies going into 0:03:59.920 --> 0:04:00.520 the future. 0:04:00.840 --> 0:04:02.680 So this is why we've come up with this. 0:04:02.840 --> 0:04:06.440 Model that says, Okay, we're going to collaborate in the 0:04:06.880 --> 0:04:11.240 science with the national laboratories and everybody else, and then 0:04:11.760 --> 0:04:15.120 when you have commercialization, we'll take it to the private sector. 0:04:15.160 --> 0:04:17.920 But whether there's high risk involved, we're going to ask 0:04:18.000 --> 0:04:19.680 for We're going to try to seek which we've done 0:04:19.720 --> 0:04:22.719 in our first fund, there's government involvement, there's private sector 0:04:22.800 --> 0:04:25.360 involvement as well. The combined are going to create the 0:04:25.400 --> 0:04:27.040 first fund we're about to announce in two or three. 0:04:27.279 --> 0:04:29.280 Are there parallels between what you're trying to do and 0:04:29.360 --> 0:04:31.640 what we're seeing now and maybe what we saw coming 0:04:31.680 --> 0:04:34.360 out of the sixties seventies with some of the Defense 0:04:34.400 --> 0:04:36.599 Department projects and now there's some of the science based 0:04:36.640 --> 0:04:40.040 projects that were government based that eventually created the foundations 0:04:40.080 --> 0:04:41.400 for what really is our tech industry. 0:04:41.440 --> 0:04:43.400 Well, you go through a dark both, right, is the 0:04:43.480 --> 0:04:45.840 in there? I mean, you go through those initials. I'm 0:04:45.880 --> 0:04:47.920 so old I was involved with these things. I mean, 0:04:48.200 --> 0:04:49.560 so I can take you through the history of hell 0:04:49.640 --> 0:04:52.760 was all was done. But fundamentally, it's those kinds of 0:04:53.040 --> 0:04:55.440 core technologies. Did you ever think of when you were 0:04:55.480 --> 0:04:58.280 exchanging technical documents it would lead to what you do 0:04:58.400 --> 0:05:00.440 today on your Apple phone, which you just heard about. No, 0:05:00.560 --> 0:05:02.880 of course not now, this is like what thirty forty 0:05:02.960 --> 0:05:06.240 years later, But nonetheless it's the same sort of thing here. 0:05:06.720 --> 0:05:09.280 But a lot of this stuff does originate quite honestly, 0:05:09.400 --> 0:05:12.560 in the science organizations within the government and the academics 0:05:12.600 --> 0:05:16.800 research organizations. But you still need the private sector to commercialize, 0:05:17.279 --> 0:05:19.680 not just the fund but also to commercialize because the 0:05:19.800 --> 0:05:21.880 talent and the skills that are required to build these 0:05:21.960 --> 0:05:25.400 companies don't exist in government. In fact, I would never 0:05:25.520 --> 0:05:27.480 let government do it if I had an opinion on 0:05:27.560 --> 0:05:29.960 that subject, because that's not what they do. So we 0:05:30.080 --> 0:05:32.960 created these things we call road weather laboratories that actually 0:05:33.000 --> 0:05:35.440 will take what we're not talking about and drive the 0:05:35.480 --> 0:05:36.760 commercialization phase. 0:05:37.000 --> 0:05:39.680 So to that point, I've been wondering what the cycle, 0:05:40.240 --> 0:05:41.640 I don't know what to call it, like the cycle 0:05:41.680 --> 0:05:43.920 of AI or the cycle of AI chips, Like, what 0:05:44.360 --> 0:05:45.280 what does that look like? 0:05:45.520 --> 0:05:46.680 Is it short? Is a long? 0:05:46.800 --> 0:05:49.520 How does that affect the private the private money coming in? 0:05:49.800 --> 0:05:51.040 Well, give you a sense. 0:05:51.240 --> 0:05:54.200 Historically, that's technology cycle. If I take you back fifty 0:05:54.279 --> 0:05:57.320 years is fifteen to twenty years. So go back to 0:05:57.400 --> 0:06:00.400 the mainframe mainframe, to the client server PC. Any clients 0:06:00.440 --> 0:06:03.000 ever precede the cloud, they were fifteen to twenty year cycles. 0:06:03.520 --> 0:06:08.720 They started slower, Televisions, phones all started slower. Today with 0:06:08.880 --> 0:06:12.520 Jenner the AI, it is like on steroids relative to 0:06:12.600 --> 0:06:15.559 the pace of it. However, a lot of people argue 0:06:15.560 --> 0:06:17.560 that it's at the peak, even there's research out there 0:06:17.600 --> 0:06:18.680 that says we're peaking. 0:06:18.440 --> 0:06:19.000 In the cycle. 0:06:19.400 --> 0:06:21.200 So then what happens next and you hear it in 0:06:21.400 --> 0:06:24.800 the enterprise talking about we can't use the stuff it hallucinates. 0:06:25.080 --> 0:06:27.080 That's an indication that this thing is peaking. 0:06:27.560 --> 0:06:30.680 And maybe it's okay for music and poetry or marketing material, 0:06:31.000 --> 0:06:33.359 but you're not going to go defend the United States 0:06:34.480 --> 0:06:37.480 infrastructure with that kind of technology, or you're not going 0:06:37.520 --> 0:06:39.000 to have a bank and not deal a deal with 0:06:39.080 --> 0:06:41.680 fraud or cyber with that kind of technology. So my 0:06:41.800 --> 0:06:44.400 point being is this is where we are, but it's 0:06:44.480 --> 0:06:48.040 going really really fast. Now the key, and we believe 0:06:48.120 --> 0:06:50.560 the key for it to really be adopted in enterprise, 0:06:51.000 --> 0:06:52.839 it's basically you have to deal with we will call 0:06:52.839 --> 0:06:56.360 them guardrails versus regulations room and we would argue that, 0:06:56.560 --> 0:06:59.480 and we do argue this that the right source for 0:06:59.520 --> 0:07:02.920 the regular is not the tech industry itself, because there 0:07:02.960 --> 0:07:04.880 are leaders in the tech industry and they're going to 0:07:05.080 --> 0:07:07.240 show up with an opinion which is a valid opinion, 0:07:07.560 --> 0:07:11.400 but also it serves the hyperscalers. We would say that 0:07:11.600 --> 0:07:13.200 you really need to do who the users of the 0:07:13.280 --> 0:07:16.560 technology are to take the businesses the CEOs of these 0:07:16.640 --> 0:07:20.120 companies in healthcare, banking, government, whatever it happens to be, 0:07:20.600 --> 0:07:23.960 and let them come up with let them innovative the technology, 0:07:24.040 --> 0:07:28.040 but regulate the use cases the application, regulate the use 0:07:28.080 --> 0:07:31.360 of the data, regulate the privacy, establish the guard rails 0:07:31.800 --> 0:07:34.280 so that people are protected in their privacy and their 0:07:34.360 --> 0:07:37.120 data is protected, both the consumer as well as the enterprise. 0:07:37.560 --> 0:07:39.320 That should be done by the enterprises. 0:07:39.720 --> 0:07:42.080 Look, I was in the tech industry for forty years, 0:07:42.560 --> 0:07:44.520 not by me or my counterparts. 0:07:44.800 --> 0:07:47.320 All right, Sam, great conversation. I really appreciate you being 0:07:47.360 --> 0:07:47.840 here today. 0:07:47.920 --> 0:07:51.360 Sam pal Masano, former CEO over at IBM, and of 0:07:51.400 --> 0:07:54.160 course now helping to shepherd that effort of that public 0:07:54.240 --> 0:07:58.440 private partnership to further our technological advancement here in the US.