WEBVTT - Hubbard on Fed Cut Fallout, Open Source AI, Nuclear Bet, Department Store Revival 0:00:02.520 --> 0:00:17.200 Bloomberg Audio Studios, podcasts, radio news. 0:00:20.000 --> 0:00:22.760 This is Wall Street Week. I'm David Weston bringing you 0:00:22.840 --> 0:00:26.239 stories of capitalism. The race to an AI future takes 0:00:26.280 --> 0:00:29.600 us through the obscure land of open source code, and 0:00:29.640 --> 0:00:32.360 one of those leading the charge, Lisa Sue of AMD, 0:00:32.600 --> 0:00:36.360 explains what is at stake plus, however we get there, 0:00:36.440 --> 0:00:39.960 that AI future will require a whole lot more electricity, 0:00:40.280 --> 0:00:42.960 which we'll have to include a whole lot more nuclear 0:00:43.040 --> 0:00:46.199 power than we have ever seen before. And as the 0:00:46.240 --> 0:00:48.800 holiday season approaches, we take you on a trip down 0:00:48.800 --> 0:00:52.080 memory lane to the department stores of yesterday and the 0:00:52.120 --> 0:00:55.960 efforts of one of the remaining icons, Macy's, to transform 0:00:56.000 --> 0:00:58.880 itself for the future. But we start with the FED 0:00:58.920 --> 0:01:01.520 decision this week. What we learned about where the FED 0:01:01.640 --> 0:01:04.760 is going if it knows with Glenn Hubbard of the 0:01:04.760 --> 0:01:08.200 Columbia Business School, who headed the Council of Economic Advisors 0:01:08.360 --> 0:01:09.880 under President George W. 0:01:10.080 --> 0:01:15.360 Bush Well, the decision itself isn't a surprise. David, Why though, 0:01:15.480 --> 0:01:18.000 is a good question. I don't see really an argument 0:01:18.319 --> 0:01:21.399 for cutting the fund's rate, particularly given the Fed's own 0:01:21.600 --> 0:01:25.200 macroeconomic forecasts that would have real GDP growth now higher. 0:01:25.920 --> 0:01:29.560 Next year, not lower. There are reasonable minds on both 0:01:29.600 --> 0:01:32.520 sides who are debating the employment tradeoff and the inflation 0:01:32.640 --> 0:01:35.520 tradeoff were interestingly going forward, what could the Fed do? 0:01:35.880 --> 0:01:38.600 I don't see rate cuts. If you look at the 0:01:38.600 --> 0:01:41.760 FED funds rate now, it's roughly in the range of 0:01:41.840 --> 0:01:44.600 where it should be given views about what the real 0:01:44.680 --> 0:01:47.440 Federal funds rate is and the inflation forecast the Fed 0:01:47.480 --> 0:01:49.800 itself has put out. So I don't see much more 0:01:49.840 --> 0:01:50.840 room to cut. 0:01:50.720 --> 0:01:53.960 If you believe the Fed's projections next year. In twenty 0:01:54.000 --> 0:01:57.880 twenty six, they took up the GDP growth by half 0:01:57.880 --> 0:01:58.320 a percent. 0:01:59.000 --> 0:01:59.559 That's big. 0:02:00.000 --> 0:02:02.480 How do you cut rates into a growing economy? 0:02:02.880 --> 0:02:04.880 Well, I think they would have to argue that it's 0:02:04.920 --> 0:02:10.840 about employment, not just output, but that seems very tortured 0:02:10.880 --> 0:02:13.480 as an argument. They still have time to react to 0:02:14.080 --> 0:02:17.520 labor pressures should they worsen. Remember, the Fed also lowered 0:02:17.520 --> 0:02:20.680 the unemployment rate for next year as well. That doesn't 0:02:20.680 --> 0:02:23.600 look like an economy heading toward recession. I'm not saying 0:02:23.600 --> 0:02:25.840 that will come to pass, but that is the Fed's view. 0:02:26.080 --> 0:02:29.040 Artificial intelligence is all the talk of the town, including 0:02:29.600 --> 0:02:33.720 with the chair Pale in his decision this week. How 0:02:33.760 --> 0:02:37.320 can the FED understand AI well enough to know what 0:02:37.360 --> 0:02:40.440 that will do, for example, productivity, which the chair talked about. 0:02:40.760 --> 0:02:43.440 Well, I don't think any of us, if we're honest, No, 0:02:43.680 --> 0:02:48.959 and I talk to business people in California. They're extremely optimistic. 0:02:49.040 --> 0:02:54.560 Every CEO is optimistic. Some others are more dour among economists, honestly, 0:02:54.760 --> 0:02:57.680 I think there's every reason to believe generative AI could 0:02:57.680 --> 0:02:59.960 be a big productivity boom. But there lies the rub 0:03:00.080 --> 0:03:04.079 for the FED. If you think AI is the future 0:03:04.160 --> 0:03:08.920 of productivity, that's a story for higher, not lower interest rates. 0:03:09.160 --> 0:03:12.480 The real interest rate will go up and inflation is 0:03:12.520 --> 0:03:16.320 still stuck at a higher level. If we're wrong about AI, 0:03:16.440 --> 0:03:18.280 then we've got many things to worry about because we 0:03:18.280 --> 0:03:20.480 have a lot of cap X going into AI data 0:03:20.480 --> 0:03:22.720 centers and things like that. But I don't think the 0:03:22.760 --> 0:03:26.960 AI productivity story gets the FED to low rates. This 0:03:27.000 --> 0:03:30.240 is not a replay of green Span in the nineteen nineties. 0:03:30.960 --> 0:03:35.080 What sorts of pressures could AI in success with increased 0:03:35.120 --> 0:03:38.960 productivity increase investment, as you say, put on the dual mandate, 0:03:39.280 --> 0:03:42.000 because you could have on the one said growth economic 0:03:42.080 --> 0:03:47.119 growth increased productivity radically with lower employment, so you could 0:03:47.200 --> 0:03:48.680 be on both sides of that mandate. 0:03:48.840 --> 0:03:49.440 You well could. 0:03:49.520 --> 0:03:51.720 At the moment, I'm not so sure. So we have 0:03:52.000 --> 0:03:56.160 seen obviously a weakening labor market, but I'm not sure 0:03:56.200 --> 0:03:58.520 AI is associated with it. When you look at all 0:03:58.600 --> 0:04:01.200 of the sectors that are weakening, to me, it looks 0:04:01.240 --> 0:04:04.480 more like an unwind of excess hiring during COVID. So 0:04:04.520 --> 0:04:07.480 we don't yet know, and we don't know going forward. 0:04:07.560 --> 0:04:09.800 Is AI going to be more of a compliment to you, 0:04:09.920 --> 0:04:11.960 me and everybody else in terms of the skills we 0:04:12.040 --> 0:04:15.440 have or a substitute. But if it were the case 0:04:15.520 --> 0:04:18.320 that AI displaced a lot of jobs, to me, that's 0:04:18.400 --> 0:04:20.560 less of a federal reserve challenge and more of a 0:04:20.680 --> 0:04:23.640 government challenge that the government needs to do something, have 0:04:23.680 --> 0:04:26.880 a policy to prepare people better than we did in 0:04:26.920 --> 0:04:29.039 previous waves of technological change. 0:04:29.120 --> 0:04:32.000 One of the things that Cherpeal addressed was tariff's When 0:04:32.000 --> 0:04:35.160 he was asked about the elevated well above the two 0:04:35.200 --> 0:04:38.200 percent number and inflation, he said, well, that's really a 0:04:38.200 --> 0:04:40.800 tariff thing, and we think that will dissipate. Is he 0:04:40.920 --> 0:04:41.600 right about that? 0:04:42.600 --> 0:04:45.800 In principle, yes, a tariff is if you had a 0:04:45.880 --> 0:04:51.280 once in four all tariff that should raise prices, not inflation. 0:04:51.440 --> 0:04:54.080 It only raises inflation for a period of time. The 0:04:54.160 --> 0:04:57.320 problem is we've seen tariffs come, tariffs go, rates go up, 0:04:57.440 --> 0:05:00.479 rates go down, so it becomes a little harder. My 0:05:00.760 --> 0:05:04.760 worry about the tariffs is less about inflation somewhat for 0:05:04.800 --> 0:05:08.360 the reasons Chair Powell has suggested, but more for their 0:05:08.440 --> 0:05:13.840 long run corrosive effect on the economy in productivity, in 0:05:13.960 --> 0:05:18.360 supply chains. You know again, America's imports are more than 0:05:18.400 --> 0:05:21.480 half of them, or intermediate goods. We're shooting ourselves in 0:05:21.520 --> 0:05:24.640 the foot whenever we do large across the board tariffs. 0:05:24.839 --> 0:05:26.080 What do you make of the fact we're now up 0:05:26.080 --> 0:05:28.560 to three descents, two going one way, one going the 0:05:28.600 --> 0:05:30.800 other way. We went for a long time actually with 0:05:31.279 --> 0:05:33.800 essentially no dessence in the FED. Is this a good thing, 0:05:33.800 --> 0:05:35.760 a bad thing, or something in between? 0:05:36.000 --> 0:05:38.239 I think it's It could be any of you bobs. 0:05:38.240 --> 0:05:40.279 So it's a good thing in the sense that in 0:05:40.320 --> 0:05:42.600 a healthy organization, if you and I don't have the 0:05:42.640 --> 0:05:45.880 same view, we should speak it. If we're not going 0:05:45.920 --> 0:05:50.760 to be reconciled. That said, you've got to ultimately govern 0:05:50.800 --> 0:05:52.880 the FED. We're about to have a new FED chair. 0:05:52.960 --> 0:05:54.960 The President will tell us who it is sometime in 0:05:55.000 --> 0:05:58.719 the near future, and that gentleman or lady is going 0:05:58.760 --> 0:06:01.080 to have to bring together a group of people with 0:06:01.200 --> 0:06:05.120 disparate opinions. That's where the challenge seems to hold. Your 0:06:05.160 --> 0:06:07.440 Powell has done a very good job, I think, in 0:06:07.520 --> 0:06:12.239 managing through a situation where people could reasonably hold different 0:06:12.360 --> 0:06:16.040 views about the economy and obvious political cross currents. He's 0:06:16.080 --> 0:06:18.839 been able to get through that. The new FED chair, 0:06:18.960 --> 0:06:21.200 whoever it is, will face that challenge. 0:06:21.360 --> 0:06:25.400 It's one thing to have disagreements over policy, over theory, 0:06:26.600 --> 0:06:28.480 even over what the data are telling. You don't think 0:06:28.480 --> 0:06:31.240 they have political disagreements. We tend to think that the 0:06:31.480 --> 0:06:34.719 FED should stay away from political disagreements and disagree over 0:06:34.880 --> 0:06:35.920 the merits. 0:06:35.480 --> 0:06:38.839 As it Well, that's absolutely right. You can make a 0:06:38.920 --> 0:06:41.760 case right now in the economy for different points of 0:06:41.839 --> 0:06:43.719 view about the path of interest rates. The fact that 0:06:43.760 --> 0:06:46.720 I have one point of view does not mean, for example, 0:06:46.720 --> 0:06:49.159 that Governor Waller's wrong because he is a different point 0:06:49.200 --> 0:06:50.640 of view. This is a different way of looking at 0:06:50.680 --> 0:06:53.760 the world. I respect that if your argument is based 0:06:53.800 --> 0:06:58.280 on politics, then that's not a technocratic argument for the FED. 0:06:58.920 --> 0:07:02.640 Political arguments are fine when you're talking about fiscal policy 0:07:02.720 --> 0:07:05.039 or things that the Congress does. But the FED was 0:07:05.120 --> 0:07:09.400 set up to provide independent judgment about monetary policy, and 0:07:09.440 --> 0:07:10.320 that's what it should do. 0:07:10.640 --> 0:07:14.280 To provide independent judgments and to be perceived as providing 0:07:14.320 --> 0:07:15.360 independent judgments. 0:07:15.400 --> 0:07:19.960 It's critical, David, because if people don't believe that you're independent, 0:07:20.040 --> 0:07:23.760 even if you yourself think you are, that doesn't work, 0:07:23.920 --> 0:07:26.040 and you have to ask yourself why does it not work. 0:07:26.520 --> 0:07:30.400 The bond market and capital markets depend on the perception 0:07:30.720 --> 0:07:33.160 that the FED is an independent agency. 0:07:33.600 --> 0:07:35.280 One of the things that came up more than once 0:07:35.320 --> 0:07:39.840 in the news conference, although Cherpal really refused to address it, 0:07:39.880 --> 0:07:43.280 is the question in Governor Cook's case depending before the 0:07:43.280 --> 0:07:46.400 Supreme Court, if you had to choose which is more 0:07:46.400 --> 0:07:48.680 important who the FED chair is or how the Supreme 0:07:48.680 --> 0:07:49.440 Court rules. 0:07:49.680 --> 0:07:52.120 Well, I'm going to give you a non lawyer's answer 0:07:52.160 --> 0:07:54.880 because I honestly I don't know. But to me, the 0:07:55.000 --> 0:07:59.240 Cook case is very important, and not just because it's 0:07:59.240 --> 0:08:02.280 about Governor Cook. Per sees call it about X. Whoever 0:08:02.480 --> 0:08:07.120 X is if the President can fire a member of 0:08:07.160 --> 0:08:11.160 the Federal Reserve Board for what doesn't appear to be caused, 0:08:11.200 --> 0:08:14.840 at least as an outsider. Then that speaks volumes about 0:08:14.840 --> 0:08:17.840 the Fed's independent So I think that case is one 0:08:17.880 --> 0:08:19.960 to watch. You know, I'm not a lawyer. I don't 0:08:20.000 --> 0:08:23.360 know what the outcome will be, but I wish the 0:08:23.400 --> 0:08:25.720 Supreme more well in deciding. 0:08:25.320 --> 0:08:27.640 It looking forward into twenty twenty six, which is very, 0:08:27.720 --> 0:08:30.440 very difficult obviously. Do you think there's more risks to 0:08:30.480 --> 0:08:32.760 the upside or the downside in the economy? 0:08:33.280 --> 0:08:37.080 You know, I see risks in both directions. The AI 0:08:37.240 --> 0:08:40.400 boom that we spoke about is clearly an upside risk 0:08:40.800 --> 0:08:43.320 to me. The resilience of the economy is also an 0:08:43.400 --> 0:08:45.839 upside risk. If we were having this conversation a year 0:08:45.840 --> 0:08:49.920 ago and you had told me about Liberation Day, I 0:08:50.000 --> 0:08:53.520 alone got to know that. I would have then thought 0:08:53.559 --> 0:08:57.080 the macroeconomy would do worse than it did. I would 0:08:57.080 --> 0:08:59.760 have predicted market effects, which did happen, but not the 0:08:59.800 --> 0:09:01.600 mac for economic effects. And I think part of that 0:09:01.720 --> 0:09:05.640 is the overwhelming resilience of the American economy combined with 0:09:05.720 --> 0:09:09.240 the productivity book. I think those are real upsides. The 0:09:09.280 --> 0:09:12.520 downsides I would worry about is that the labor market 0:09:12.720 --> 0:09:16.679 could weaken gradually and suddenly, and we already have a 0:09:16.720 --> 0:09:22.000 credit cycle that looks relatively mature to me, and so 0:09:22.040 --> 0:09:25.520 I would start to worry about credit accidents which could 0:09:25.520 --> 0:09:28.320 affect the bod market, So that would worry me as 0:09:28.320 --> 0:09:30.280 a downside. And then of course there are always geopolitical 0:09:30.360 --> 0:09:33.000 risks that I'm certainly not an expert in and are 0:09:33.040 --> 0:09:35.319 hard to predict. So I think there are risks on 0:09:35.840 --> 0:09:37.880 both sides. It's going to be a very interesting year 0:09:37.920 --> 0:09:41.600 for the Fed for fiscal policy, and obviously because it's 0:09:41.640 --> 0:09:43.920 an election year, a highly charged. 0:09:43.679 --> 0:09:46.839 Year coming up. 0:09:47.040 --> 0:09:49.320 Getting to an AI future, it's not just a matter 0:09:49.320 --> 0:09:51.800 of how much money is spent, it's also what the 0:09:51.840 --> 0:09:55.559 basic structure of the system looks like. AMD Lisa Sue 0:09:55.679 --> 0:09:58.680 on the critical bets being made on open source versus 0:09:58.679 --> 0:10:15.320 proprietary code. This is a story about frontemies, when to 0:10:15.400 --> 0:10:17.800 try and beat your competition, and when to share with 0:10:17.840 --> 0:10:21.760 those competitors to build a bigger overall business and compete 0:10:21.800 --> 0:10:25.240 in other ways. The competition to develop AI models could 0:10:25.240 --> 0:10:28.840 not be more fierce, with big tech companies investing record 0:10:28.880 --> 0:10:31.400 amounts of money on chips and data centers. 0:10:31.960 --> 0:10:35.480 These companies investing trillions of dollars in CAPEX. 0:10:35.040 --> 0:10:38.280 Trillions of dollars in the raise to build artificial intelligence 0:10:38.320 --> 0:10:39.800 support systems. 0:10:39.360 --> 0:10:43.200 Trillions of dollars of our tech companies investing in building 0:10:43.280 --> 0:10:44.440 data centers in America. 0:10:45.280 --> 0:10:48.520 How far AI will take us and how fast may 0:10:48.559 --> 0:10:51.520 depend in part on a basic choice about the overall 0:10:51.600 --> 0:10:56.280 approach to sharing for withholding information, a choice often mentioned 0:10:56.280 --> 0:10:59.640 in passing, but one that investors may not have identified 0:10:59.760 --> 0:11:00.400 as key. 0:11:01.200 --> 0:11:03.160 I want to see AI everywhere. You know we were 0:11:03.320 --> 0:11:07.840 such so early in the real usage of the technology. 0:11:08.760 --> 0:11:11.439 Lisa Sue is chair and CEO of a m D, 0:11:11.800 --> 0:11:14.440 the chip maker with hundreds of billions of dollars at 0:11:14.520 --> 0:11:17.679 stake in the development of AI. When it comes to 0:11:17.720 --> 0:11:20.839 the software that runs on their products, she and her 0:11:20.880 --> 0:11:24.160 company have made their choice in favor of open source 0:11:24.280 --> 0:11:26.480 technology sharing where they can. 0:11:27.000 --> 0:11:30.280 The difference with an open source or open standards is 0:11:30.320 --> 0:11:34.959 the idea that you set out a set of standards 0:11:35.000 --> 0:11:38.640 that people can follow, and then different companies can choose 0:11:38.679 --> 0:11:41.079 to implement in a different way. And you know, the 0:11:41.120 --> 0:11:46.800 whole idea is to provide platforms that the entire ecosystem, 0:11:46.880 --> 0:11:51.400 lots and lots of developers can develop, and that that 0:11:51.440 --> 0:11:54.040 you can be able to interchange sort of the best 0:11:54.040 --> 0:11:55.679 of breed from different sources. 0:11:56.040 --> 0:11:58.720 Why is there so much passion around open I mean, 0:11:58.800 --> 0:12:00.360 is it a matter of business or is there a 0:12:00.400 --> 0:12:01.280 matter of philosophy? 0:12:01.440 --> 0:12:05.319 The philosophy is, you know, do you believe that there's 0:12:05.520 --> 0:12:08.760 any one company or any one group that will have 0:12:09.000 --> 0:12:11.520 all of the best ideas in the world, or do 0:12:11.600 --> 0:12:14.680 you believe that if you have an open ecosystem where 0:12:14.800 --> 0:12:18.160 different researchers and different groups can contribute, that you will 0:12:18.280 --> 0:12:23.320 end up with a better product overall? And philosophically, I think, 0:12:23.400 --> 0:12:25.720 you know, from an am D standpoint. You know, actually, 0:12:25.760 --> 0:12:28.040 if I think about throughout my career, I look at, 0:12:28.320 --> 0:12:30.880 you know, sort of the different inflection points on technology. 0:12:31.800 --> 0:12:36.000 Things often start new ideas often start in a closed environment. 0:12:36.360 --> 0:12:38.720 But when you actually think about when they get big 0:12:38.760 --> 0:12:41.520 and when you know a lot of people adopt, you 0:12:41.640 --> 0:12:45.560 like to be an open environment. And you know that's philosophically. Now, 0:12:45.600 --> 0:12:47.520 let's talk about the business aspects of it. You know, 0:12:47.600 --> 0:12:49.880 the business aspects of it are you know, we all 0:12:49.920 --> 0:12:52.920 have to make choices of where we invest and do 0:12:52.960 --> 0:12:55.080 you want to trust you know, all of your crown 0:12:55.160 --> 0:12:58.520 jewels and all of your data in a closed ecosystem 0:12:58.640 --> 0:13:01.200 that you know, may or may not be the most 0:13:01.200 --> 0:13:03.720 competitive at any given time, or you know, may have 0:13:03.880 --> 0:13:06.720 you know, a flaw or any of those things. Whereas 0:13:07.240 --> 0:13:09.840 you know, in an open ecosystem, you have choices. You 0:13:09.880 --> 0:13:12.400 can decide who is the best at any given time. 0:13:12.679 --> 0:13:15.840 You can decide if I have my data in a 0:13:15.840 --> 0:13:18.880 cloud environment, I can easily move it to another environment. 0:13:19.400 --> 0:13:22.839 If I have my applications that are built on one chip, 0:13:23.200 --> 0:13:25.880 if there's a better chip, you know, a few years later, 0:13:26.040 --> 0:13:27.959 you can move you know, quickly. That's that's kind of 0:13:27.960 --> 0:13:29.160 the idea of both. 0:13:29.440 --> 0:13:31.959 What broadly speaking of the advantages of open and let 0:13:31.960 --> 0:13:34.439 me throw all out For example, in a new technology, 0:13:35.080 --> 0:13:37.520 is it more likely, all of the things being equal, 0:13:38.120 --> 0:13:41.280 that the technology will develop more quickly if it's open. 0:13:41.520 --> 0:13:44.840 I would say that the main advantage of an open 0:13:44.880 --> 0:13:48.880 ecosystem is that you can get many, many more developers 0:13:49.040 --> 0:13:52.679 on that ecosystem, and they will contribute, you know, sort 0:13:52.720 --> 0:13:56.079 of in their you know, special way. So you know, 0:13:56.120 --> 0:13:58.000 an example I can give you is, you know, when 0:13:58.040 --> 0:14:01.760 we think about supercomputing, like the largest supercomputers in the world, 0:14:02.280 --> 0:14:05.880 we usually like to develop them on open software stacks, 0:14:06.240 --> 0:14:10.040 so that researchers from all different labs can actually contribute 0:14:10.400 --> 0:14:13.800 to those applications and those learnings. We see that a 0:14:13.840 --> 0:14:17.920 lot in open ecosystems, which is the notion that we 0:14:17.960 --> 0:14:20.360 want more developers. Like Linux is a great example, right. 0:14:20.400 --> 0:14:25.240 Linux is one of the examples where with an open 0:14:25.280 --> 0:14:28.040 operating system, you can see that it's now really become 0:14:28.360 --> 0:14:32.000 the standard for a lot of computing going forward. 0:14:32.600 --> 0:14:35.520 For all the advantages of the open source approach to AI, 0:14:36.160 --> 0:14:39.960 many of the most important players, including most recently Meta 0:14:40.040 --> 0:14:43.600 with a new model expected to debut next year, have 0:14:43.720 --> 0:14:46.880 gone the other way, keeping much of their generative AI 0:14:47.000 --> 0:14:50.680 tech proprietary. And there are some advantages to that approach, 0:14:51.120 --> 0:14:54.480 including preventing bad actors from having access to source code 0:14:54.760 --> 0:14:58.200 that they can manipulate. That's something Sue believes can be 0:14:58.320 --> 0:14:59.680 remedied by community. 0:15:00.800 --> 0:15:04.880 I think there is a definite view that there's a 0:15:04.920 --> 0:15:07.840 place for open and then there's a place for proprietary. 0:15:08.120 --> 0:15:10.840 And when I say that, I mean, look, there are 0:15:11.840 --> 0:15:14.720 amazing groups of researchers that are working on the largest 0:15:14.720 --> 0:15:18.320 foundational models. When you think about you know, what's happening 0:15:18.320 --> 0:15:21.000 at open AI and what's happening at Google and what's 0:15:21.040 --> 0:15:25.000 happening at Anthropic, what's happening at Meta. These are phenomenal 0:15:25.080 --> 0:15:28.880 labs that are doing a tremendous research in AI. There 0:15:28.920 --> 0:15:31.120 are also a number of open models. You know open 0:15:31.160 --> 0:15:34.240 Ai has an open model, you know Meta, their Lama 0:15:34.400 --> 0:15:37.680 series as have had open models. And what you find 0:15:37.680 --> 0:15:41.119 when those models are open is there are more researchers 0:15:41.160 --> 0:15:43.280 that are able to build on top of that, and 0:15:43.600 --> 0:15:46.760 I think there's significant advantages there. So you know, our 0:15:46.840 --> 0:15:49.840 thought process in how this evolves over time is you're 0:15:49.880 --> 0:15:53.840 going to have both in the ecosystem And just like 0:15:53.880 --> 0:15:57.720 my very early example of sort of the Apple iOS 0:15:57.760 --> 0:16:02.000 ecosystem and the Google Android ecosystem, it's not like you're 0:16:02.000 --> 0:16:04.640 going to have all of one takeover all of the 0:16:04.680 --> 0:16:08.400 other takeover. But what you do see is there's value 0:16:08.440 --> 0:16:11.280 in richness in having let's call it a little bit 0:16:11.280 --> 0:16:14.920 of competition between the ecosystems. And most importantly, what we 0:16:14.960 --> 0:16:18.160 want is what's best for the consumer is to get 0:16:18.200 --> 0:16:22.800 the best overall product experience and the best overall capabilities. 0:16:22.920 --> 0:16:24.720 And you do that when you. 0:16:26.200 --> 0:16:30.520 Allow a playing field that allows good competition across the board. 0:16:33.400 --> 0:16:36.600 The debate between open source and proprietary is not a 0:16:36.600 --> 0:16:40.560 new one. This IBM seven oh five. The computer industry 0:16:40.600 --> 0:16:43.600 had to confront a similar choice long before the current 0:16:43.720 --> 0:16:47.840 rush to artificial intelligence models. Sam Palmasana was the CEO 0:16:47.920 --> 0:16:51.080 of IBM from two thousand and three through twenty eleven. 0:16:51.680 --> 0:16:54.080 How this whole thing began was basically in the late 0:16:54.160 --> 0:16:57.920 nineties and there were two different model There were proprietary 0:16:57.960 --> 0:17:00.880 models that exist at IVM and the main framed Microsoft 0:17:01.320 --> 0:17:05.040 Client Server, et cetera. They're all proprietary, and we believed 0:17:05.160 --> 0:17:07.920 that IBM that a better model was one that would 0:17:07.960 --> 0:17:11.439 drive more innovation work growth if it was open source 0:17:11.520 --> 0:17:16.320 so everybody could participate. So that led to this Linux Initiative, 0:17:16.400 --> 0:17:18.679 which was the open source in this institute that was 0:17:18.720 --> 0:17:21.440 created by the industry, which we were part of. And 0:17:22.040 --> 0:17:23.600 obviously what happened over time. 0:17:23.640 --> 0:17:25.439 Now we put a lot behind it. 0:17:25.480 --> 0:17:28.480 We made a billion dollar bet that Linux would become 0:17:28.560 --> 0:17:31.719 commercial and that in the future Linux in the patche 0:17:31.760 --> 0:17:33.080 would be kind of the operating the. 0:17:33.119 --> 0:17:34.240 Environment of the Internet. 0:17:34.760 --> 0:17:36.840 It worked out, but it could have been a mistake, 0:17:36.920 --> 0:17:38.920 but it worked out, so that gave it a lot 0:17:38.920 --> 0:17:42.240 of momentum. But the whole point was open innovation is 0:17:42.280 --> 0:17:46.560 going to win over single company proprietary approaches because they 0:17:46.600 --> 0:17:49.920 have a limited amount of resource and unlimited market. Even 0:17:49.960 --> 0:17:51.679 if they have a large share position, it's still a 0:17:51.680 --> 0:17:54.320 smaller market than the entire industry. 0:17:54.880 --> 0:17:57.480 When you made the decision to go open source with Linux, 0:17:58.640 --> 0:18:01.240 was there controversy about it? The pluses and the minuses 0:18:01.240 --> 0:18:02.280 of that decision. 0:18:01.960 --> 0:18:04.800 Huge internal debate. Think about this stuff, IBM inventative. With 0:18:04.840 --> 0:18:07.879 the history of the industry, it was all proprietary, whether 0:18:07.920 --> 0:18:10.639 that was not just in mainframe or storage and database 0:18:10.680 --> 0:18:13.520 and all those things were all invented by IBM. So 0:18:13.600 --> 0:18:15.920 there's the camp that said, hey, you guys, this is crazy. 0:18:15.960 --> 0:18:19.640 We have this proprietary model. It's very profitable. Slower growth, 0:18:19.680 --> 0:18:23.200 but it's very profitable, and it's from a business perspective. 0:18:23.240 --> 0:18:25.919 The alternative was that, no, if we can open up 0:18:25.960 --> 0:18:29.119 the opportunity and then participate in a bigger opportunity than 0:18:29.160 --> 0:18:32.400 the one we're participated in, IBM will be better off 0:18:32.480 --> 0:18:34.280 long term, and that was the decision. 0:18:34.960 --> 0:18:38.320 The IBM moved to open source for operating systems twenty 0:18:38.359 --> 0:18:41.840 five years ago helped spur what became a software revolution 0:18:42.359 --> 0:18:45.920 as firms adopted the system and built their own proprietary 0:18:45.960 --> 0:18:49.240 applications on top of it. Looking back on it now, 0:18:49.320 --> 0:18:51.520 it's hard not to conclude that it was good for 0:18:51.560 --> 0:18:55.600 the business overall and ultimately good for companies like IBM. 0:18:56.400 --> 0:18:59.280 But is history likely to repeat itself this time as 0:18:59.320 --> 0:19:03.119 the world experien ements with various generative AI approaches said, the. 0:19:03.160 --> 0:19:05.399 Chip is actually quite small. What you really we want 0:19:05.440 --> 0:19:07.840 to present a framework where we're going to get the 0:19:07.880 --> 0:19:10.760 best ideas coming out of this, and from an AMD standpoint, 0:19:10.960 --> 0:19:14.080 we'd love you to run on our chips, but more importantly, 0:19:14.800 --> 0:19:17.080 we want you to run on an open ecosystem so 0:19:17.160 --> 0:19:20.280 you can decide, you know, two years from now or 0:19:20.320 --> 0:19:23.280 four years from now, if you made a choice of AMD, 0:19:23.400 --> 0:19:25.800 that doesn't mean you're locked into our chips for the 0:19:25.840 --> 0:19:29.400 next five years. It means that you have an open 0:19:29.440 --> 0:19:33.160 ecosystem where you can benefit from all of the competitive 0:19:33.880 --> 0:19:36.800 capabilities that come on board over the next five to ten. 0:19:36.760 --> 0:19:38.600 Years, which sounds like it should be good for the 0:19:38.640 --> 0:19:41.760 business long term overall the business. Is it also good 0:19:41.800 --> 0:19:43.520 for AMD in the sense that in that world where 0:19:43.520 --> 0:19:45.120 you can switch and you don't have to be committed, 0:19:45.320 --> 0:19:47.399 you'll sell more chips over the long term. 0:19:47.480 --> 0:19:50.439 I believe we will because we are giving people the 0:19:50.520 --> 0:19:54.119 opportunity to choose, and look, we always have to be 0:19:54.160 --> 0:19:58.760 best in class, right, there's no question that it's there's 0:19:58.880 --> 0:20:02.240 so many innovations to come on board. But the idea 0:20:02.320 --> 0:20:05.160 that if you develop on AMD, you have choice and 0:20:05.240 --> 0:20:08.720 you also have the ability to work closely with us 0:20:09.280 --> 0:20:12.760 in terms of, you know, how we develop this ecosystem together, 0:20:13.119 --> 0:20:15.439 that this is a case where one plus one is 0:20:15.440 --> 0:20:19.359 going to be greater than three because you know, we're 0:20:19.440 --> 0:20:24.760 getting the smartest people from all of the ecosystem versus 0:20:24.920 --> 0:20:27.040 just you know, we're going to figure out everything on 0:20:27.080 --> 0:20:27.480 our own. 0:20:28.200 --> 0:20:30.680 And it's not just the big US tech companies making 0:20:30.720 --> 0:20:34.399 their bets on open versus a proprietary. China is casting 0:20:34.440 --> 0:20:38.520 its vote as well and perhaps ironically siding with the 0:20:38.600 --> 0:20:39.480 open approach. 0:20:40.080 --> 0:20:43.120 Where we started, it was mostly proprietary, and the winners 0:20:43.160 --> 0:20:47.800 at that point in time were Meta, Google, Open Ai, Microsoft, 0:20:47.840 --> 0:20:48.360 you know, et cetera. 0:20:48.440 --> 0:20:49.240 They were the winners. 0:20:49.480 --> 0:20:52.520 And we're growing like crazy with these large language models. 0:20:52.920 --> 0:20:54.919 Then all of a sudden, this thing happens in China. 0:20:55.600 --> 0:20:57.399 And there's two things that happen in China that are 0:20:57.440 --> 0:21:01.560 open sourced and they're based on this view that innovation 0:21:01.680 --> 0:21:04.280 will scale faster if we have an open source approach. 0:21:05.040 --> 0:21:07.120 Deep Seek it's got a lot of press, a lot 0:21:07.119 --> 0:21:09.439 of coverage. The other one's called huggy Face, which you 0:21:09.440 --> 0:21:12.040 hardly hear about. But if you look at the numbers 0:21:12.080 --> 0:21:15.359 compared to the growth before twenty twenty four, it was 0:21:15.400 --> 0:21:19.080 heavily dominated by the three companies, the hyperscalers I talked about, 0:21:19.640 --> 0:21:23.080 and then since then it's been these other companies and 0:21:23.119 --> 0:21:26.520 the China approach. So since twenty twenty four, China actually 0:21:26.640 --> 0:21:30.200