WEBVTT - TSMC's US Plant and the White House's Military AI Plan 0:00:01.639 --> 0:00:05.920 From Mahard where Innovation, Money and Power Collie in Silicon 0:00:06.040 --> 0:00:10.480 Vallet NBN. This is Bloomberg Technology with Caroline Hyde and 0:00:10.720 --> 0:00:11.800 Ed Ludlow. 0:00:24.720 --> 0:00:27.400 Live from New York and San Francisco. This is Bloomberg 0:00:27.440 --> 0:00:31.920 Technology coming up. TSMC's US plant yields surpassed that of 0:00:31.960 --> 0:00:33.199 similar factories in Taiwan. 0:00:33.479 --> 0:00:36.360 What that means for the US chip ambitions. 0:00:36.560 --> 0:00:39.839 And the Biden administration's planning a concrete strategy on the 0:00:39.960 --> 0:00:43.720 US military's use of AI in warfare. 0:00:43.520 --> 0:00:47.680 And Alphabet's autonomous driving unit. Weimo raises five point six 0:00:47.800 --> 0:00:50.920 billion dollars in its largest ever funding round. But first 0:00:51.320 --> 0:00:54.080 we check back on the public markets and new record 0:00:54.120 --> 0:00:56.920 high for the NASTAK. The Broader Technology indexes up one 0:00:56.920 --> 0:00:59.520 point three percent on the day. Consumer sentiment looks strong 0:00:59.560 --> 0:01:02.400 on the day, inject some risk on across the board, 0:01:02.520 --> 0:01:05.720 but most notably, we're still seeing the key magnificent seven 0:01:05.760 --> 0:01:08.800 players managing to drive up to a new record high end. 0:01:08.840 --> 0:01:10.120 What are you watching on the micro. 0:01:10.400 --> 0:01:11.800 Well, I want to get to that breaking news from 0:01:11.800 --> 0:01:15.000 this morning about WEIMO raising five point six billion dollars 0:01:15.040 --> 0:01:17.360 its largest ever round. It was a Series C and 0:01:17.480 --> 0:01:21.120 remember WAIMO is the autonomous driving unit of Alphabet, the 0:01:21.120 --> 0:01:24.320 parent of Google. Alphabet's leading the round, And actually I'm 0:01:24.360 --> 0:01:26.880 not drawing a direct causal link, but the stock did 0:01:26.920 --> 0:01:29.520 push higher after the open when that news hit. Later 0:01:29.560 --> 0:01:31.480 in the show will bring you the full details of 0:01:31.480 --> 0:01:33.679 the RAM because some really interesting names in the world 0:01:33.720 --> 0:01:36.600 of finance participating in it, and a lot of attention 0:01:36.760 --> 0:01:41.400 on autonomous driving right now. Elsewhere, our top story is TSMC. 0:01:41.760 --> 0:01:45.000 The basics of what Bloomberg's reporting is that its first 0:01:45.120 --> 0:01:49.680 US chip plant in Arizona has achieved yields that surpassed 0:01:49.680 --> 0:01:53.320 that of equivalent plants in its domestic home of Taiwan. 0:01:54.080 --> 0:01:57.080 An important political story and important industry story. These are 0:01:57.080 --> 0:01:59.919 the ADRs or US listed shares up almost three percent. 0:02:00.040 --> 0:02:01.760 Har more details, Yeah, let's. 0:02:01.600 --> 0:02:03.720 Get to it because, as you mentioned, just the nuance 0:02:03.800 --> 0:02:07.000 there of the yields in the first plant in Arizona 0:02:07.440 --> 0:02:09.280 that are surpassing similar. 0:02:08.960 --> 0:02:10.040 Factories back home. 0:02:10.040 --> 0:02:12.680 It's a significant breakthrough many feel for a US expansion 0:02:12.680 --> 0:02:16.600 project initially played by delays by worker strife in King 0:02:16.720 --> 0:02:18.920 joins US for some of the backstory on this because 0:02:18.960 --> 0:02:21.320 it was hard to get to this position that they're 0:02:21.320 --> 0:02:21.840 currently in. 0:02:23.040 --> 0:02:26.480 Yeah, there was several reports there was labor issues. Can 0:02:26.520 --> 0:02:29.639 they really do this outside of their home base? Is 0:02:29.720 --> 0:02:32.680 this big bet that the US government is making on 0:02:33.760 --> 0:02:36.200 a foreign company to bring them to the US where 0:02:36.200 --> 0:02:38.480 they really haven't done very much in the past. Is 0:02:38.480 --> 0:02:42.880 that all going to work out? And what TSMC and 0:02:42.919 --> 0:02:45.400 our sour Thing is indicating is, well, yes, it is 0:02:45.440 --> 0:02:46.000 working out. 0:02:46.360 --> 0:02:49.440 I think it's worth explaining to the Bluemote Technology audience 0:02:49.520 --> 0:02:51.920 what yield is and why it's important. I've had to 0:02:51.960 --> 0:02:54.160 go at it and we'll bring the definition up on 0:02:54.200 --> 0:02:58.160 the screen. But it's basically the volume or number of 0:02:58.320 --> 0:03:01.240 usable chips in a batch, and that's super important when 0:03:01.240 --> 0:03:03.079 you think about profit and efficiency. 0:03:03.200 --> 0:03:06.120 Yeah, I mean the key phrase here is production yields. 0:03:06.200 --> 0:03:08.640 That means basically, if you don't have production yields, you 0:03:08.680 --> 0:03:11.640 shouldn't be in production. You are losing money on what 0:03:11.680 --> 0:03:15.000 you're doing. How many of these things do you have 0:03:15.040 --> 0:03:17.200 to throw away? Proper production backs because you have to 0:03:17.200 --> 0:03:19.480 throw away a lot in the early stages. What they're 0:03:19.520 --> 0:03:21.840 saying here is, hey, we've got through those early stages, 0:03:22.040 --> 0:03:27.000 perhaps quicker than had been expected, and we're in a. 0:03:26.960 --> 0:03:31.160 Good position, in a good position, What then, is the 0:03:31.240 --> 0:03:35.080 ramifications for the US focus on bringing production to the 0:03:35.200 --> 0:03:38.760 United States. Are we going to see yet more TSMC production? 0:03:38.920 --> 0:03:40.760 Are we going to see yet more money coming from 0:03:40.800 --> 0:03:41.720 the federal government. 0:03:42.680 --> 0:03:45.760 I mean, there's no doubt that, given the landscape, that 0:03:46.200 --> 0:03:49.800 that is what the US government would want from TSMC. 0:03:50.280 --> 0:03:53.080 We've got Intel, as we know, having problems. We've got 0:03:53.120 --> 0:03:57.200 Samsung as we know, having some issues with its foundary business. 0:03:57.240 --> 0:04:00.680 So if we want this production in this country, TSMC 0:04:00.840 --> 0:04:02.480 has to be the leading vehicle for that. 0:04:03.360 --> 0:04:06.240 I wouldn't read too much into it. I would say 0:04:06.240 --> 0:04:09.200 that the timing and manner by which we came to 0:04:09.320 --> 0:04:11.400 this information is interesting. 0:04:11.520 --> 0:04:13.440 Could you help me explain why. 0:04:14.200 --> 0:04:15.760 The US needs a win? 0:04:16.040 --> 0:04:16.240 Right? 0:04:16.760 --> 0:04:18.719 The Biden administration needs a win. 0:04:18.920 --> 0:04:19.080 Right. 0:04:19.080 --> 0:04:23.440 There's a lot of negative newsflow around what's going on 0:04:23.560 --> 0:04:26.800 with this initiative. At the same time, if you're TSMC, 0:04:27.120 --> 0:04:30.039 you've been criticized for what you've done here, for not 0:04:30.160 --> 0:04:32.920 going as big as possible, for coming in with an 0:04:32.960 --> 0:04:36.440 older production technique. So perhaps you want to remind the 0:04:36.440 --> 0:04:38.240 world that actually, no, we are in the lead. 0:04:38.279 --> 0:04:39.800 We can do this. We know what we're doing. 0:04:40.680 --> 0:04:41.840 On shoring is one part. 0:04:41.920 --> 0:04:44.000 The other part is the export control which we can 0:04:44.040 --> 0:04:46.760 talk about a little bit more bloombergsy and King thank 0:04:46.800 --> 0:04:48.679 you very much. I want to bring an Annya Manuel 0:04:49.040 --> 0:04:51.760 for more on this exact point. Executive director of the 0:04:51.800 --> 0:04:55.080 Aspen Strategy Group and former diplomat. Let's just start with 0:04:55.120 --> 0:04:58.880 the basics of this whole situation. TSMC the sort of 0:04:58.880 --> 0:05:03.280 crown jaw of Taiwan technology industry having success early in 0:05:03.320 --> 0:05:06.240 the US and the support of the US to do 0:05:06.360 --> 0:05:09.920 it in the context of China. That's a difficult one. 0:05:10.880 --> 0:05:11.320 That's right. 0:05:11.520 --> 0:05:14.279 This is really very good news, and it means that 0:05:14.320 --> 0:05:18.839 the Biden policy of both having some tough export controls 0:05:18.880 --> 0:05:24.560 on certain very small critical technologies, but also these subsidies. 0:05:25.000 --> 0:05:29.960 I think TSMC got six billion in subsidies, five billion loans. 0:05:30.120 --> 0:05:34.600 Other companies are getting similar That project of pushing our 0:05:34.640 --> 0:05:37.040 semiconductor industry forward is actually working. 0:05:37.839 --> 0:05:40.440 We come to you because of your rich experience when 0:05:40.480 --> 0:05:42.839 it comes to foreign policy, the work that you do 0:05:43.120 --> 0:05:45.400 with the Aspens Strategy Group, the work that you've done 0:05:45.640 --> 0:05:49.040 with the US government thus far. What do you grade 0:05:49.200 --> 0:05:54.080 this focus of the United States on shoring of chip 0:05:54.120 --> 0:05:55.840 supply from an A to a D. 0:05:55.920 --> 0:05:57.400 Are they surpassing. 0:05:56.880 --> 0:06:00.200 Expectations it's a good question I would give it about 0:06:01.240 --> 0:06:03.719 So the important thing is to keep our eye on 0:06:03.760 --> 0:06:07.280 the long term ball. There are certain key technologies where 0:06:07.320 --> 0:06:10.600 the US and our friends and allies should stay in 0:06:10.640 --> 0:06:13.440 the lead. We should be more advanced than the Chinese, 0:06:13.480 --> 0:06:15.920 who are becoming more and more competitors in that space. 0:06:16.279 --> 0:06:19.080 We do that in two ways, by slowing them down 0:06:19.400 --> 0:06:23.719 that's the export controls, and by speeding ourselves up. My 0:06:23.880 --> 0:06:27.160 personal view is that we've done pretty close to what 0:06:27.240 --> 0:06:29.320 we can do to slow the Chinese down. 0:06:29.360 --> 0:06:33.320 They're great entrepreneurs themselves. 0:06:33.400 --> 0:06:36.320 We have some very tough export controls on the most 0:06:36.360 --> 0:06:40.159 advanced chips and also on the tools that makes the chips. 0:06:40.839 --> 0:06:42.840 Making those controls even tougher. 0:06:43.200 --> 0:06:47.039 Nilatterly on the US side won't slow China down. It 0:06:47.200 --> 0:06:50.840 just means our companies won't have the revenue that they 0:06:50.880 --> 0:06:53.839 need to fund the research and development to keep ourselves 0:06:53.839 --> 0:06:54.240 in the lead. 0:06:55.040 --> 0:06:58.080 Let's just set the context of this week as well, 0:06:58.440 --> 0:07:01.520 because there will some cups, shall we say, to the 0:07:01.640 --> 0:07:07.720 US focus on containing semiconductor expertise, limiting it from going 0:07:07.760 --> 0:07:10.920 into Chinese hands. In particular, we hear that TSMC has 0:07:10.960 --> 0:07:13.320 had to admit that somehow it's chips are working its 0:07:13.360 --> 0:07:17.440 way into Huawei's products, even though Huawei has been cut 0:07:17.440 --> 0:07:20.680 off in theory since twenty twenty. News today that al 0:07:20.800 --> 0:07:23.520 Chip Technology said it has no director and direct business 0:07:23.520 --> 0:07:26.080 dealings with Huawei. Of course, many trying to work out 0:07:26.360 --> 0:07:30.400 how and which client of TSMC's allowed the chips to 0:07:30.400 --> 0:07:34.400 get into the hand of Huawei. But how effective can 0:07:34.440 --> 0:07:35.880 these limitations really be on you? 0:07:36.960 --> 0:07:40.240 That's a perfect question, and this example shows exactly how 0:07:40.280 --> 0:07:42.760 difficult it is. You know, there are lots of well 0:07:42.800 --> 0:07:46.239 meaning people in the US government in European governments trying 0:07:46.280 --> 0:07:48.400 to get this right and trying to make the controls 0:07:48.480 --> 0:07:52.560 narrowly focused and effective, very very hard to do. 0:07:53.040 --> 0:07:54.840 I think they're doing quite a good job. 0:07:55.280 --> 0:07:59.200 I worry, however, that, for example, you see some folks 0:07:59.320 --> 0:08:02.440 in the might go into a Trump administration very tough 0:08:02.480 --> 0:08:04.520 on China, China hawks, saying well, we're just going to 0:08:04.560 --> 0:08:07.720 control more, We're going to put more companies on the 0:08:07.840 --> 0:08:11.240 entity list. And as you're seeing with this TSMC example, 0:08:11.360 --> 0:08:14.600 even when people are well meaning, it's just very very. 0:08:14.400 --> 0:08:17.440 Difficult to do things with controls alone. You've got to run. 0:08:17.360 --> 0:08:19.840 Faster, which is more powerful. 0:08:19.880 --> 0:08:23.920 Then we've covered the export controls part of this story 0:08:24.040 --> 0:08:28.520 deeply on this program, but they almost work in parallel. 0:08:28.600 --> 0:08:30.440 You sort of can't have one way out the other. 0:08:30.480 --> 0:08:32.480 If your end goal is to make your own nation 0:08:33.240 --> 0:08:36.040 more competitive than any given field or domain. 0:08:37.040 --> 0:08:38.200 That's absolutely right. 0:08:38.280 --> 0:08:41.439 I'm firmly on the side of running faster is better. 0:08:41.640 --> 0:08:45.000 We have an amazing innovation engine. I live out here 0:08:45.040 --> 0:08:48.080 in California. You all report on it all the time. 0:08:48.440 --> 0:08:50.720 We're doing great, so we need to make sure that 0:08:50.840 --> 0:08:54.040 engine keeps running and keeps running fast. One of the 0:08:54.080 --> 0:08:57.160 things that's underappreciated. Everyone always talks about the CHIPSAC the 0:08:57.200 --> 0:09:01.240 subsidies to semiconductor fabs. What they don't talk about so 0:09:01.320 --> 0:09:03.000 much is that there was supposed to be one hundred 0:09:03.000 --> 0:09:08.480 and seventy billion dollars for going into basic science R. 0:09:08.440 --> 0:09:09.160 And D funding. 0:09:09.480 --> 0:09:12.920 This is really important. This creates all of the technologies 0:09:12.920 --> 0:09:16.600 of the future, you know, okay, Serie, GPS, and so 0:09:16.720 --> 0:09:18.600 those are the things we should be doubling down on. 0:09:19.559 --> 0:09:19.880 ANYA. 0:09:19.960 --> 0:09:22.480 At the same time, China is taking its own action. 0:09:22.800 --> 0:09:27.000 How closely do you follow and assess their response to 0:09:27.080 --> 0:09:28.400 the TSMC situation. 0:09:30.880 --> 0:09:34.280 China is great. They're doing a lot of innovating. Their 0:09:34.280 --> 0:09:37.000 companies are excellent. If you look at, for example, what 0:09:37.000 --> 0:09:40.760 they're doing on artificial intelligence, they're large language models. 0:09:41.080 --> 0:09:42.640 The biggest models. 0:09:42.240 --> 0:09:45.880 In AI are depending who you ask, three six, nine 0:09:45.920 --> 0:09:49.000 months behind because of our chip's controls. But on certain 0:09:49.120 --> 0:09:55.280 other parts of artificial intelligence, computer visions, surveillance, drones, they're 0:09:55.280 --> 0:09:56.120 really first rate. 0:09:56.160 --> 0:09:57.960 So the Chinese are going to keep running fast. 0:09:58.240 --> 0:10:01.160 And I would just say one thing this, this policy 0:10:01.400 --> 0:10:04.880 needs to be nuanced. It's not like we are competing 0:10:04.920 --> 0:10:07.479 with China on absolutely every technology. 0:10:07.760 --> 0:10:09.160 There's certain areas where we. 0:10:09.080 --> 0:10:12.400 Should still collaborate on healthcare, tech, on other things. It's 0:10:12.559 --> 0:10:16.400 just those things that touch national security that are really 0:10:16.440 --> 0:10:19.600 sensitive that can be dual use, where we need to be. 0:10:19.559 --> 0:10:21.360 More careful push us forward. 0:10:21.360 --> 0:10:25.840 Then, no matter what the administration ofter November fifth, sixth, 0:10:26.600 --> 0:10:32.360 will there forever be this idea of tension and ultimately 0:10:32.800 --> 0:10:36.000 opposing nations between China and US when it comes to AI. 0:10:36.400 --> 0:10:41.040 Do you see any grounds for a more dubbsh approach? 0:10:42.480 --> 0:10:42.800 Yeah? 0:10:43.200 --> 0:10:46.440 I do think we are entering a phase where the 0:10:46.559 --> 0:10:50.920 Chinese government, led by Si Jinping specifically, has said they 0:10:50.960 --> 0:10:55.960 want to advance in certain technologies AI semiconductors, five G, 0:10:56.160 --> 0:10:59.880 six G, certain others. They've set this out in the plan, 0:11:00.720 --> 0:11:04.520 and they want to have a very dominant military, and 0:11:04.600 --> 0:11:08.400 those two things are quite worrying. Understandably so to the US, 0:11:08.520 --> 0:11:11.720 Europe are friends and allies, and so I do think 0:11:11.760 --> 0:11:15.040 no matter who wins our election, this part of the 0:11:15.120 --> 0:11:17.480 technology race, and I don't call it a war, I 0:11:17.520 --> 0:11:19.880 call it a tech race, is going to continue. This 0:11:19.920 --> 0:11:23.239 is a big, long term push that's going to continue. 0:11:23.360 --> 0:11:26.600 That doesn't mean, however, that the US and China can't 0:11:26.640 --> 0:11:27.520 do business. 0:11:27.200 --> 0:11:29.000 Together and lots of other areas. 0:11:29.120 --> 0:11:32.520 When I see companies in the consumer sector and agriculture 0:11:32.520 --> 0:11:34.920 and other places still doing a lot of work with China, 0:11:35.000 --> 0:11:36.800 it doesn't mean that there has to be a completely 0:11:36.880 --> 0:11:38.520 coupling nuance. 0:11:38.600 --> 0:11:39.120 We love it. 0:11:39.240 --> 0:11:42.600 Anna Manuel, thank you, executive director of the Aspen Strategy Group. 0:11:43.200 --> 0:11:45.800 Coming up, how the UK Prime. 0:11:45.640 --> 0:11:48.280 Minister is looking to win over skeptics with they plan 0:11:48.400 --> 0:11:49.640 to lead on. 0:11:49.640 --> 0:11:51.640 Your guest at artificial intelligence. 0:11:51.640 --> 0:12:02.080 That's next, there's a bring back technology. 0:12:04.800 --> 0:12:08.480 After some early confusion around the UK Labor Party's commitment 0:12:08.559 --> 0:12:12.280 to AI, Prime Minister Kiss starm is now positioning himself 0:12:12.559 --> 0:12:16.840 as a champion for AI. His party's inaugural investment summit 0:12:16.920 --> 0:12:20.320 last week. He even appeared alongside former Google CEO Eric 0:12:20.320 --> 0:12:24.400 Schmidt and called AI an opportunity rather than something to 0:12:24.480 --> 0:12:26.880 be scared of. Let's bring in Bloomberg's Matt Bergen, who's 0:12:26.920 --> 0:12:30.240 out in London for us, and let's ask the simple question, 0:12:30.800 --> 0:12:33.240 what is the UK's AI plan. 0:12:35.280 --> 0:12:36.880 Well, we don't know it yet. 0:12:36.960 --> 0:12:40.199 Wear our story today we talked to Secretary Cumpeter Kyle, 0:12:40.280 --> 0:12:44.600 who's the Secretary for Science and Technology runs the department here. 0:12:44.920 --> 0:12:46.959 He said, they're coming out with two false strategy and 0:12:47.040 --> 0:12:50.000 one is this opportunities plan where they're going to be 0:12:50.080 --> 0:12:52.839 using AI both for they talk about for improvements in 0:12:52.880 --> 0:12:55.439 the public sector and for the private sector, and we've 0:12:55.480 --> 0:12:57.720 seen a little bit of that in the public sector 0:12:57.800 --> 0:13:01.080 so far. The second part is the prom a legislation 0:13:01.200 --> 0:13:04.040 around AI, and what he told us very explicitly is 0:13:04.040 --> 0:13:05.720 we are not going to go as far as the 0:13:05.720 --> 0:13:08.839 EU have. They basically said the EU has gone it's 0:13:08.880 --> 0:13:12.120 kind of overreached on AI, overregulated in a way. This 0:13:12.160 --> 0:13:15.280 is an opportunity, one of the silver silver linings for Brexit, 0:13:15.320 --> 0:13:17.040 if you will. For a Labor Party, which which stood 0:13:17.080 --> 0:13:21.559 against that obviously, and they've given kind of two rough 0:13:21.840 --> 0:13:24.720 commitments that AI bill. One is that the AI labs 0:13:25.080 --> 0:13:27.480 are going to have to make their voluntary commitments they 0:13:27.559 --> 0:13:30.360 made last year, made those into law, and he's going 0:13:30.400 --> 0:13:33.440 to have a new agency to kind of monitor and 0:13:33.559 --> 0:13:38.240 enforce the safety and criteria for these frontier rols like 0:13:38.280 --> 0:13:39.680 open AI and Google. 0:13:40.320 --> 0:13:42.880 Piece of Carl though the UK Science of Technology Secretory 0:13:43.640 --> 0:13:47.760 rather upended everyone though earlier by having to it seems 0:13:47.960 --> 0:13:51.800 nix a couple of very important investment areas in to 0:13:52.400 --> 0:13:57.040 ultimately compute and I'm interested as to how he's tried 0:13:57.040 --> 0:14:00.000 to navigate that a little bit more and educate us 0:14:00.120 --> 0:14:00.760 why he did it. 0:14:01.960 --> 0:14:03.679 Yeah, that's correct. I was kind of shocking to a 0:14:03.720 --> 0:14:07.040 lot of investors here. And so that was a project 0:14:07.040 --> 0:14:09.000 in Scotland, one of the countries very few kind of 0:14:09.240 --> 0:14:13.120 these supercomputing projects. His explanation from that, and the Labor 0:14:13.160 --> 0:14:16.559 Party's explanation is that the prior government promised this money 0:14:16.559 --> 0:14:19.760 without allocating the funds. They are in a big fiscal 0:14:19.760 --> 0:14:22.760 hole that they're trying to fill, among other things, discussing 0:14:22.800 --> 0:14:25.680 the possibility of capital gains tax and some measures that 0:14:25.720 --> 0:14:28.800 the tech community is very afraid of and doesn't want 0:14:28.840 --> 0:14:31.720 to happen. I think they certain people we've talked to 0:14:32.360 --> 0:14:34.760 around the Labor government who advised them admit that that 0:14:34.840 --> 0:14:38.560 was sort of clumsy handling of that, But he told 0:14:38.640 --> 0:14:41.640 us that they're reevaluating the computing budget for the country. 0:14:41.920 --> 0:14:45.680 Mob you cover bench capital in the UK and Europe. 0:14:45.920 --> 0:14:47.680 What is the VC community to say right now? What 0:14:47.760 --> 0:14:50.320 is the amount of money going into AI and AI 0:14:50.440 --> 0:14:52.840 innovation compared to the US and globally. 0:14:54.520 --> 0:14:57.960 Yeah, I mean it's it's small, and there's a concern 0:14:58.080 --> 0:15:01.800 that the UK will is is losing its footing to France, 0:15:02.160 --> 0:15:04.680 which has invested quite a bit and the government has 0:15:04.840 --> 0:15:07.600 been like working the prop up the industry. There's a 0:15:07.640 --> 0:15:11.280 concern that the efforts not just of say pension funds 0:15:11.320 --> 0:15:14.720 moving into VC, but but government research project. I think 0:15:14.720 --> 0:15:17.200 the other side of that is, you know that they're 0:15:17.240 --> 0:15:19.880 defenders of the government that are saying they're much more realistic. 0:15:20.120 --> 0:15:21.000 We had Rishi. 0:15:20.840 --> 0:15:24.360 Sunak, the Prime Minister, talking about the UK being a 0:15:24.440 --> 0:15:27.560 scigned superpower, having Elon Musk on stage. Some of those 0:15:27.560 --> 0:15:29.800 things that maybe weren't realistic for the UK to be. 0:15:30.840 --> 0:15:33.440 Caroline and I have some experience in reporting on the 0:15:33.520 --> 0:15:36.200 UK government and in the context of things like AI, 0:15:36.680 --> 0:15:40.720 they often add in many buzzwords and political words like efficiency, 0:15:41.280 --> 0:15:44.160 some of which we've mentioned in this conversation. But are 0:15:44.160 --> 0:15:47.480 they actually acting with the private sector, not just the vcs, 0:15:47.520 --> 0:15:50.400 but you cover the startups too, right, They need kind 0:15:50.400 --> 0:15:52.600 of some relationship there to move forward. 0:15:54.400 --> 0:15:56.480 Yeah, I think you know, the term we heard a 0:15:56.520 --> 0:15:59.080 lot was it's a wait and see. I think right 0:15:59.120 --> 0:16:02.640 now they're they're certainly having conversations with the startup industry. 0:16:02.520 --> 0:16:06.600 They are the choice few sort of AI startups that 0:16:06.640 --> 0:16:09.040 are promising that are involved in conversations. But I don't 0:16:09.080 --> 0:16:11.120 think we've seen a lot of evidence of them either 0:16:11.160 --> 0:16:14.880 being drawn in as government contracts or partners. But that's 0:16:14.920 --> 0:16:16.920 something that at least in LIB service they're talking about 0:16:16.920 --> 0:16:17.160 a lot. 0:16:18.200 --> 0:16:19.400 We want to thank you, Mart Bergen. 0:16:28.000 --> 0:16:30.920 It's time for talking tech and first start. Four astronauts 0:16:30.960 --> 0:16:34.560 have made touchdown back on Earth. SpaceX's Crew eight capsule 0:16:34.760 --> 0:16:38.240 landed earlier this morning after undocking from the International Space 0:16:38.280 --> 0:16:41.560 Station on Wednesday. The crew arrived at the ISS on 0:16:41.680 --> 0:16:44.880 March fifth, but could not return due to delays and 0:16:45.040 --> 0:16:49.480 issues with Boeing Starliner two Boeing Starliner test pilots are 0:16:49.520 --> 0:16:53.720 still stuck in space for the return set for February plus. 0:16:53.760 --> 0:16:54.960 Elon Musk is said to have. 0:16:54.920 --> 0:16:59.760 Had secret conversations with Russian President Vladimir Putin. That's according 0:16:59.800 --> 0:17:02.240 to The Street Journal, who say the two have been 0:17:02.520 --> 0:17:06.399 chatting regularly since twenty twenty two on a wide ranging 0:17:06.480 --> 0:17:10.920 number of topics, including business and geopolitical issues. And Tesla 0:17:11.240 --> 0:17:14.800 is drawing up an eclectic menu as it prepares to 0:17:14.880 --> 0:17:18.640 open a fifties diner been charging station in the heart 0:17:18.640 --> 0:17:19.359 of La where. 0:17:19.240 --> 0:17:21.320 Ev drivers will be able to grab a bite. 0:17:21.320 --> 0:17:24.000 Or catch a movie or topping off their battery assign 0:17:24.040 --> 0:17:28.960 the industry's moving away from the utesarian utilitarian. 0:17:28.240 --> 0:17:30.840 Gas station model. It's something I thought i'd never read. 0:17:30.880 --> 0:17:33.439 Character I say it, that's quite cool. 0:17:33.840 --> 0:17:36.480 Meanwhile, let's talk about the Biden administration saying that the 0:17:36.560 --> 0:17:39.760 US must accelerate its adoption of AI for military and 0:17:39.800 --> 0:17:42.840 intelligence use. In a new national security memorandum, the White 0:17:42.840 --> 0:17:46.440 House outlined efforts to maintain an edge of arrivals and 0:17:46.520 --> 0:17:49.840 increase safety. Courtney rosen imanly by government joins US now, 0:17:50.000 --> 0:17:52.920 and you've really been focusing in on why the US 0:17:53.040 --> 0:17:55.480 government is suddenly ramping up this narrative. 0:17:55.680 --> 0:17:56.719 What's behind it? Corney. 0:17:58.440 --> 0:18:02.760 Well, the primary partomle of this is that US adversaries 0:18:02.800 --> 0:18:06.520 are already deploying AI in their militaries and their intelligence uses. 0:18:07.040 --> 0:18:10.840 And that's what President Biden's National security advisor said yesterday. 0:18:11.240 --> 0:18:12.040 He said that. 0:18:12.240 --> 0:18:15.600 These things are advancing without American values, and so the 0:18:15.720 --> 0:18:18.479 US has to keep up in order to keep up 0:18:18.480 --> 0:18:19.280 with the adversaries. 0:18:20.280 --> 0:18:20.639 Courney. 0:18:20.640 --> 0:18:22.440 I want to get into kind of the specifics of 0:18:22.480 --> 0:18:24.399 what a strategy might look like, but I think to 0:18:24.400 --> 0:18:27.359 be fair to our audience on Bloomberg Technology around the world, 0:18:27.560 --> 0:18:29.440 we're all kind of conscious that we have a US 0:18:29.480 --> 0:18:34.000 presidential election in less than two weeks. Therefore, the existing 0:18:34.000 --> 0:18:37.119 administration has a time limit on what it can or 0:18:37.160 --> 0:18:39.639 can't do while in office. Just kind of take that 0:18:39.680 --> 0:18:40.919 and explain that bit to us. 0:18:40.960 --> 0:18:41.280 Please. 0:18:43.119 --> 0:18:47.120 This memo that President Biden's team published yesterday lays out 0:18:47.160 --> 0:18:50.080 a whole bunch of to do lists for military officials, 0:18:50.119 --> 0:18:53.240 intelligence officials. The President's not going to be in power 0:18:53.400 --> 0:18:56.160 long enough to see a lot of that through, so 0:18:56.280 --> 0:18:58.840 we need to be thinking about what the next president 0:18:58.920 --> 0:19:02.560 would do on these issues now. The Republican platform says 0:19:02.640 --> 0:19:06.240 that President Trump would repeal a lot of the Biden 0:19:06.280 --> 0:19:09.520 administration's consumer policies on AI. It doesn't really have a 0:19:09.560 --> 0:19:13.680 take on military and intelligence uses. So that's one particular 0:19:13.760 --> 0:19:16.240 vision forward. And then you've got Vice President Kamala Harris, 0:19:16.280 --> 0:19:19.520 who has really been the face of AI for this 0:19:19.600 --> 0:19:22.760 administration on the international stage and had a hand in 0:19:22.840 --> 0:19:26.360 developing a lot of these policies. So two very different visions. 0:19:27.000 --> 0:19:29.960 But regardless, it's not going to be President Biden who's 0:19:30.000 --> 0:19:31.119 executing on these things. 0:19:32.080 --> 0:19:34.560 Accorney, on this program, we talked to companies all the time, 0:19:34.640 --> 0:19:39.159 like Palenteer and Anderill, private sector, publicly traded companies or 0:19:39.240 --> 0:19:42.199 startups that tell us they're working more and more with 0:19:42.280 --> 0:19:44.960 the government. What about from the government's side, is this 0:19:45.000 --> 0:19:48.600 strategy they plan based on working with the private sector. 0:19:49.800 --> 0:19:50.119 Yes. 0:19:50.720 --> 0:19:54.800 Jake Sullivan, who's President Biden's National security advisor, yesterday made 0:19:54.800 --> 0:19:58.600 a speech about this Artificial intelligence and national security strategy, 0:19:58.920 --> 0:20:01.520 and he devoted a lot time to talking about how 0:20:01.560 --> 0:20:05.600 AI innovation is largely happening in the private sector and 0:20:05.640 --> 0:20:08.399 not in government. There was a lot of talk about 0:20:08.440 --> 0:20:10.679 how the US government's going to have to work with 0:20:10.800 --> 0:20:14.320 tech companies to make that happen in a way that 0:20:14.600 --> 0:20:17.360 maybe they hadn't in different eras of technology. 0:20:17.720 --> 0:20:17.920 Now. 0:20:17.960 --> 0:20:20.320 I was sitting in a room here in Washington a 0:20:20.359 --> 0:20:24.280 couple months ago during NATO when Oracle Saffricats said something 0:20:24.400 --> 0:20:27.800 very similar that the tech is existing in the companies, 0:20:28.320 --> 0:20:31.240 the government is behind, and the two have to get 0:20:31.280 --> 0:20:32.560 along to make this work. 0:20:33.920 --> 0:20:38.040 What's interesting is the ethics around it and any limitations 0:20:38.040 --> 0:20:41.320 coming from the United Nations. There is this tension around 0:20:41.359 --> 0:20:45.920 autonomous weapons and AI, and ultimately they're used in the battlefield. 0:20:45.920 --> 0:20:47.959 How is that something that perhaps Jake Sullivan is speaking 0:20:47.960 --> 0:20:49.080 to or trying to navigate. 0:20:50.960 --> 0:20:53.440 They put together what they call or what I would 0:20:53.480 --> 0:20:56.399 call a risk based approach. Right, the riskiest uses of 0:20:56.440 --> 0:21:00.000 AI in military and intelligence settings, for example, use it 0:21:00.280 --> 0:21:02.600