WEBVTT - The FAA Halts Flights and More Apple Changes 0:00:13.400 --> 0:00:17.000 I'm Caroline Hyde, Bloomberg's world headquarters in New York. Ludlow 0:00:17.000 --> 0:00:19.959 were three days into our New York reunion. This is 0:00:19.960 --> 0:00:23.200 Bloomberg Technology coming up this hour. Ed we're grounded the 0:00:23.360 --> 0:00:26.480 f A a halted flight departures nationwide after a key 0:00:26.520 --> 0:00:31.720 pilot information system failed. What technology exactly failed us? We 0:00:31.800 --> 0:00:35.880 have the details and more Apple supply changes first chips, 0:00:36.040 --> 0:00:40.160 now screens, plus some plans for a touch screen mac book. 0:00:40.240 --> 0:00:44.120 Bloomberg's Mark German on all of this week's scoops, and well, 0:00:44.120 --> 0:00:46.839 where are the best companies to work for other than 0:00:47.080 --> 0:00:49.960 with us? Right here? Glass Door just came out with 0:00:50.040 --> 0:00:52.640 their annual list, but some of the Silicon Valley darlings 0:00:53.159 --> 0:00:56.040 they did not make the cut this time. The story 0:00:56.080 --> 0:01:00.279 of the day was flights grounded throughout the nation. Pre 0:01:00.360 --> 0:01:02.600 market we were lower. Most of the airlines actually, look 0:01:02.640 --> 0:01:04.440 at that screen now so much green. Many of the 0:01:04.480 --> 0:01:07.600 airlines recovering at the market open on Wednesday. Just a 0:01:07.640 --> 0:01:10.760 single name down in the road with Southwest. You know 0:01:10.880 --> 0:01:12.960 a lot of problems to Southwest coming out of that 0:01:13.280 --> 0:01:16.040 storm and busy holiday period, down seven tents of one percent. 0:01:16.080 --> 0:01:18.640 There's so much to discuss here. What on Earth went wrong. 0:01:18.800 --> 0:01:20.880 I mean, it's a technology story and it's very hard 0:01:21.040 --> 0:01:23.560 ed because the systems of Love, as the ticker is 0:01:23.600 --> 0:01:26.480 known Southwest, have been a key thorn in this side. 0:01:26.520 --> 0:01:30.440 But now the Federal Aviation Administration halting thousands of flights nationwide, 0:01:30.440 --> 0:01:33.120 as you just saying. It all happened early Wednesday, and 0:01:33.160 --> 0:01:36.039 it was a pilot notification system that exactly failed. But 0:01:36.400 --> 0:01:40.040 now we've got lawmakers worrying about investigating this and they 0:01:40.080 --> 0:01:42.080 want to see how this is going to influence course 0:01:42.160 --> 0:01:44.400 some major upcoming aviation bills that are going to be 0:01:44.440 --> 0:01:47.520 assessed by these lawmakers. It's bringing George Ferguson, who's our 0:01:47.560 --> 0:01:51.080 senior airline analyst over at Bloomberg Intelligence, and George just 0:01:51.320 --> 0:01:53.800 tell us exactly what failed, what went wrong from a 0:01:53.840 --> 0:01:57.120 tech perspective here, and so the no term system noticed. 0:01:57.160 --> 0:01:59.920 The Airman system is a system that does so many 0:02:00.000 --> 0:02:03.800 the information to flight crew, lets them know what what 0:02:03.960 --> 0:02:06.320 outages there might be at the airports they're arriving at, right, 0:02:06.360 --> 0:02:08.600 they might arrive at an airport that doesn't have an 0:02:08.600 --> 0:02:13.160 instrument approach that's working quite right, lets them know airspace restrictions. 0:02:13.200 --> 0:02:15.800 So it's I mean it's all about safety. And so 0:02:16.000 --> 0:02:19.359 that system seemed to be broke down earlier today. And 0:02:19.720 --> 0:02:21.720 when that happens, I think out of an abundance of caution, 0:02:21.760 --> 0:02:24.639 they stop all flights because they can't be sure they're 0:02:24.639 --> 0:02:27.639 feeding flight crews the right information to go into uh 0:02:27.680 --> 0:02:30.639 you know, airports all over the country. And so that 0:02:30.919 --> 0:02:33.320 was the breakdown, probably a technology breakdown. I don't think 0:02:33.320 --> 0:02:35.400 it's people. I think it's you know, they collect this 0:02:35.480 --> 0:02:39.000 information from flight service stations around the country, uh, and 0:02:39.040 --> 0:02:41.000 then they disseminate it back to the pilot. So it's 0:02:41.000 --> 0:02:44.880 a big sort of collection and dissemination uh, you know 0:02:45.000 --> 0:02:48.040 process system, right. Um, So I think that's what probably 0:02:48.080 --> 0:02:50.240 broke down. George, I'm one of those guys, you know, 0:02:50.280 --> 0:02:52.079 I walk out into the street, I look up in 0:02:52.120 --> 0:02:55.360 the sky, look at the airplanes overhead. I marvel about 0:02:55.440 --> 0:02:58.920 what's happening every day of the year, hundreds thousands of 0:02:58.960 --> 0:03:01.320 flights all around the world. Old. Is it just that 0:03:01.840 --> 0:03:05.000 we rely on an archaic system to manage all of this? 0:03:05.160 --> 0:03:07.920 I mean we're talking here about this technology as somebody 0:03:07.919 --> 0:03:11.440 that researches looks at the fundamentals of this industry. Are 0:03:11.480 --> 0:03:15.160 you really concerned was this just an unfortunate glitch. I 0:03:15.160 --> 0:03:17.680 think it was an unfortunate glitch. I think I think, Um, 0:03:17.720 --> 0:03:19.919 we won't continue to see problems like this, but we'll 0:03:19.919 --> 0:03:23.560 always see tech problems in this business, right because the 0:03:23.600 --> 0:03:26.120 tech is kind of always catching up with the business. 0:03:26.880 --> 0:03:28.919 I mean, we've we've been flying airplanes for over a 0:03:29.000 --> 0:03:32.760 hundred years now. Really commercial air travels come, you know, 0:03:33.000 --> 0:03:37.120 come into its golden age mid last century. You built systems. 0:03:37.560 --> 0:03:40.280 Those systems need to migrate over time. People that know 0:03:40.800 --> 0:03:42.840 I T systems know it's hard to plug the newest 0:03:42.880 --> 0:03:48.000 technology sometimes into the old technology. That migration is a challenge. Um. 0:03:48.040 --> 0:03:49.640 I've heard that there was some work being done in 0:03:49.640 --> 0:03:51.600 the no TAM system, so maybe that was part of 0:03:51.600 --> 0:03:54.640 the challenge here. But it's always a constant migration of 0:03:54.680 --> 0:03:58.200 bringing old technology to new bringing the best to the front. 0:03:58.720 --> 0:04:00.960 It takes time in the government doing it, and they 0:04:00.960 --> 0:04:02.800 don't always have the most minded to do it, so 0:04:03.200 --> 0:04:06.360 it probably moves on a slower scale than business. George 0:04:06.400 --> 0:04:09.320 ferguson bloom meg Intelligence, we thank you. Edit Do you 0:04:09.360 --> 0:04:11.720 actually do that? Do you actually like walk out and 0:04:11.760 --> 0:04:16.240 go wow? I do I'm a business traveler. I'm somebody 0:04:16.320 --> 0:04:19.560 that you know, I'm fascinated by the underlying technology of 0:04:19.640 --> 0:04:23.360 something that every single day you look up and when 0:04:23.400 --> 0:04:26.400 something like this happens, is a consumer is a technology? 0:04:26.480 --> 0:04:28.719 Does you ask what we just started? George? How on 0:04:28.760 --> 0:04:31.800 Earth or something like that happened? His answer, probably a glitch, 0:04:32.160 --> 0:04:34.359 you know, It's just it's incredible. Have they tried turning 0:04:34.360 --> 0:04:36.679 it on and off again? Right? Moved to the cloud? Whatever? 0:04:36.760 --> 0:04:39.440 I love her to Your Eneggie just brings me so 0:04:39.520 --> 0:04:42.120 much joy. Meanwhile, let's let's kick out a little bit 0:04:42.160 --> 0:04:44.200 more because we've got to talk Apple and it's planning 0:04:44.240 --> 0:04:47.560 to start using its own custom displays and mobile devices 0:04:47.560 --> 0:04:50.680 as early as and on top of that, maybe in 0:04:50.720 --> 0:04:53.719 the future those devices with those screens you can touch 0:04:53.760 --> 0:04:56.720 them even on a Mac computer. Let's bring in on 0:04:56.839 --> 0:05:00.280 the scoop Generator. That's Mark German and Mark, I mean 0:05:00.279 --> 0:05:03.040 you're out with another one. That start with the Macnews 0:05:03.080 --> 0:05:06.599 because I thought that Tim Cook was going to follow 0:05:06.600 --> 0:05:09.880 in his predecessors viewpoint on this and not want to 0:05:09.880 --> 0:05:13.600 have touch screens on computers. This is very significant. Now 0:05:13.600 --> 0:05:16.200 they won't fly, so ED won't be able to marvel 0:05:16.240 --> 0:05:18.560 at them in the air, but they will get touch 0:05:18.600 --> 0:05:21.720 screens right. Apple has a project. They have multiple teams 0:05:21.800 --> 0:05:24.360 working on effort to bring touch displays to the Mac 0:05:24.720 --> 0:05:26.560 for the first time. This is the first time where 0:05:26.560 --> 0:05:29.279 they're actually serious about this. They've explored in the past, 0:05:29.640 --> 0:05:32.520 but work has ramped up on these new machines and 0:05:32.720 --> 0:05:35.600 this is a big u turn from what Apple has 0:05:35.640 --> 0:05:38.120 done in the past. They've talked about how the iPad 0:05:38.240 --> 0:05:40.720 is the best experience for touch. They talked about how 0:05:40.760 --> 0:05:43.520 mac os, the Mac operating system doesn't work well with touch. 0:05:43.800 --> 0:05:46.479 They talked about the ergonomic concerns about that, but the 0:05:46.560 --> 0:05:49.680 reality is is that consumers want touch. The Mac can 0:05:49.760 --> 0:05:51.920 run iPhone and iPad apps now, but they don't have 0:05:51.960 --> 0:05:54.279 a touch screen, so they don't work very well. Kids 0:05:54.320 --> 0:05:57.640 these days have grown up on touch screens, iPhones, iPads, 0:05:57.680 --> 0:06:01.240 Apple watches, Amazon, Kindled table, It's you name it right. 0:06:01.720 --> 0:06:04.360 The next generation of people who are going to get 0:06:04.440 --> 0:06:07.640 laptops and desktops, get computers, get max they are used 0:06:07.680 --> 0:06:10.120 to touch. So Apple has to think about that new 0:06:10.160 --> 0:06:13.200 generation and that learning curve and what this new generation 0:06:13.200 --> 0:06:16.400 of potential buyers is used to. So touch displays are 0:06:16.440 --> 0:06:18.480 important to the future of the Mac, and that's why 0:06:18.480 --> 0:06:20.080 Apple is going to go down this road in a 0:06:20.080 --> 0:06:22.839 few years. Let's pivots your other big scoop, which is 0:06:22.920 --> 0:06:26.320 the starting with high end Apple Watch and later the 0:06:26.400 --> 0:06:30.719 iPhone handsets, Apple wants to move to in house technology 0:06:30.839 --> 0:06:33.880 for the displays. Uh, you know the currently l G 0:06:34.200 --> 0:06:36.919 and Samsung are the big supplies when it comes to 0:06:36.960 --> 0:06:38.760 those displays. What have you learned in the course of 0:06:38.760 --> 0:06:41.120 your reporting? So right now, the iPhone and the Apple Watch, 0:06:41.200 --> 0:06:43.440 both you use what is known as an ol end display. 0:06:43.760 --> 0:06:46.560 Those are supplied primarily by Samsung. They get some from 0:06:46.640 --> 0:06:50.039 Japan Display LG. They're also working with a company called 0:06:50.040 --> 0:06:52.760 BOE in China and source displays from there too. But 0:06:52.880 --> 0:06:56.279 that underlying technology that's made by those manufacturing companies that 0:06:56.279 --> 0:07:00.000 make those displays, right, Apple can tweak and customize them. 0:07:00.440 --> 0:07:04.080 The underlying technology has not been developed by Apple. Now 0:07:04.160 --> 0:07:08.320 they're developing the underlying technology and that manufacturing process. So 0:07:08.360 --> 0:07:11.680 now no longer will Apple be competing with Samsung but 0:07:11.840 --> 0:07:14.600 also sourcing parts from Samsung for one of the most 0:07:14.640 --> 0:07:18.640 important components inside of their products. All right, bloombergs Mark German, 0:07:18.680 --> 0:07:21.280 thank you very much. Well, let's talk about some progress 0:07:21.280 --> 0:07:23.640 being made when it comes to executive searches, Walt Disney 0:07:23.640 --> 0:07:27.520 in fact electing a new independent director, Mark Parker, as 0:07:27.560 --> 0:07:29.960 the chairman and the board. Now. Parker already been on 0:07:30.000 --> 0:07:32.840 the board for seven years now and he was, of 0:07:32.840 --> 0:07:35.960 course still executive chairman of Nike. Says some big roles 0:07:36.000 --> 0:07:39.240 he's got under his belt. He's succeeding Susan Arnold, who 0:07:39.240 --> 0:07:41.640 will not be standing for re election. Student too, what 0:07:41.840 --> 0:07:44.520 is a fifteen year term in it under Disney's board 0:07:44.560 --> 0:07:48.960 tenure policy, shares likinginess. I mean he had a cool, 0:07:49.000 --> 0:07:52.200 storied career over at Disney, I mean over at Nike now, 0:07:52.280 --> 0:07:55.640 says a chair. But his main priority, ed, he says, 0:07:56.040 --> 0:07:57.800 is to be getting a new CEO. Who is that 0:07:57.840 --> 0:07:59.600 going to be? He says? Already the searchers upon it. 0:07:59.640 --> 0:08:01.760 There is a so a deeper story here, which is 0:08:01.840 --> 0:08:06.280 that Nelson Pelts and Tree and an activist investor we 0:08:06.360 --> 0:08:08.720 found out in November eight million dollars state they were 0:08:08.720 --> 0:08:11.080 trying to get a board seat and the board rejecting 0:08:11.440 --> 0:08:14.559 that idea. But the reason why is that Nelson Pelts 0:08:14.560 --> 0:08:16.880 and Tree and they did not want Bob Iger to 0:08:17.000 --> 0:08:19.920 return as CEO, which again was a shock in itself 0:08:19.960 --> 0:08:23.480 when those headlines broke. So there's things happening behind the 0:08:23.480 --> 0:08:25.280 scenes here at Disney, but at least they have some 0:08:25.320 --> 0:08:37.200 certainty now about the new board chair chat GPT. Yes, 0:08:37.240 --> 0:08:39.480 we're talking in again and there's a lot of fanfare 0:08:39.520 --> 0:08:41.960 around the AI powered chat pot that's gone viral in 0:08:42.040 --> 0:08:46.240 recent weeks. But there's some negative connotations here potentially posing 0:08:46.320 --> 0:08:49.400 propaganda hacking risks. Now, this is all according to a 0:08:49.440 --> 0:08:52.960 report published by researchers on Wednesday. Here's the thing. Two 0:08:53.000 --> 0:08:56.520 of those authored the study actually worked at open ai, 0:08:56.960 --> 0:08:59.040 and one of them, however, has since left there. Not 0:08:59.080 --> 0:09:03.000 for profit and blog accompanying the report stated quote, Our 0:09:03.040 --> 0:09:06.280 bottom line judgment is that language models will be useful 0:09:06.320 --> 0:09:11.600 for propagandists and will likely transform online influence operations. Joining 0:09:11.679 --> 0:09:14.600 us now to discuss the risks as well as the positives, 0:09:14.600 --> 0:09:17.920 the use cases, the future of the technology. Street Or Ramaswamy, 0:09:18.280 --> 0:09:22.320 co founder CEO of the ad free private search engine Neiva, 0:09:22.400 --> 0:09:24.400 and we're going to go into the what Neiva does 0:09:24.440 --> 0:09:26.559 and the way in which you use natural language in 0:09:26.600 --> 0:09:28.560 a moment, you know, but just talk to us first 0:09:28.640 --> 0:09:32.800 and foremost about what you made of this discussion about 0:09:32.840 --> 0:09:37.840 the fear of propaganda about misinformation being used within AI 0:09:38.040 --> 0:09:39.800 because also on the flip side, there's a lot of 0:09:39.840 --> 0:09:41.960 positives the way in which you can actually uncover who's 0:09:42.000 --> 0:09:46.280 behind certain cyber attacks at least two. So large language 0:09:46.280 --> 0:09:48.719 models are one of the most exciting developments in the 0:09:48.840 --> 0:09:51.240 last few years. These are models that have been trained 0:09:51.280 --> 0:09:54.840 pretty much on everything that's been written down, everything that's 0:09:54.920 --> 0:09:57.600 on that page. Um. The results of that is that 0:09:57.640 --> 0:10:00.199 they have enormous fluency. Can ask chat DPD too write 0:10:00.240 --> 0:10:02.160 a look form or a story and it will do 0:10:02.360 --> 0:10:05.440 an amazing job. Um. It doesn't know right from wrong 0:10:05.520 --> 0:10:08.840 or truth from falsehood. So you know, the word that 0:10:08.880 --> 0:10:12.880 we use for that is halstening um. But it is 0:10:12.920 --> 0:10:17.600 incredibly prolific. I think of these large language models as savants. Um. 0:10:17.640 --> 0:10:20.559 They can say very convincing things. They can say things 0:10:20.600 --> 0:10:22.800 in the voice of somebody like you're writing if you 0:10:22.880 --> 0:10:25.200 give it enough of a sample. Um. So it's not 0:10:25.240 --> 0:10:29.520 surprising obviously that they can be used for propaganda or misinformation. 0:10:29.559 --> 0:10:32.280 That's a problem that we face online anyway. I know 0:10:32.280 --> 0:10:34.199 what I see something I always look to see, well, 0:10:34.320 --> 0:10:37.160 is this from Boomberg? Is this reputable? So I think 0:10:37.600 --> 0:10:39.320 those kinds of risks are going to go up, but 0:10:39.360 --> 0:10:41.800 I think there's also a lot of positive potential that 0:10:41.880 --> 0:10:44.559 come from these models. Do you know what's interesting is 0:10:44.600 --> 0:10:46.880 we went to our own audience earlier in the day 0:10:47.000 --> 0:10:49.840 using Twitter a pole, and we asked them, look to 0:10:49.880 --> 0:10:53.160 these sorts of reports about the ability to incite propaganda 0:10:53.360 --> 0:10:58.160 put you off from using chat gypt Interestingly, most still say, look, 0:10:58.240 --> 0:11:01.280 they're going to be continuing to use it despite these reports. 0:11:01.320 --> 0:11:03.560 But about said, no way, I'm not going to be 0:11:03.640 --> 0:11:06.440 using it from your perspective. Is this where you step 0:11:06.480 --> 0:11:10.200 in Because one of the fixes perhaps that neither has 0:11:10.320 --> 0:11:13.200 for chat GPT is by allowing us to know what 0:11:13.320 --> 0:11:16.120 the sourcing of the information is that we're reading. Yeah, 0:11:16.160 --> 0:11:18.440 that was one of our big goals. We wanted to 0:11:18.480 --> 0:11:22.400 integrate the fluency of these large language models into search. 0:11:22.880 --> 0:11:25.200 Right now, search is pretty hard to put in a keyword. 0:11:25.200 --> 0:11:26.920 You have to go click on a link, figured out 0:11:27.000 --> 0:11:29.360 what's there, um, and then come back maybe it goes 0:11:29.360 --> 0:11:32.480 somewhere at else um. So we wanted the fluent answers 0:11:32.520 --> 0:11:34.960 of chat GPT, but we wanted to be done in 0:11:35.360 --> 0:11:39.480 an authentic way. So we set out to make sure 0:11:39.640 --> 0:11:43.880 that every sentence he wrote had its source clearly put there, 0:11:43.920 --> 0:11:46.440 so there's a citation. UM. We also wanted to make 0:11:46.440 --> 0:11:48.920 sure that we could take into our count real time information. 0:11:49.000 --> 0:11:52.480 This is what we launched earlier UM last week, and 0:11:52.520 --> 0:11:55.720 it's been exceptionally well received because two of the biggest 0:11:55.720 --> 0:11:59.679 problems that GPT has old information it's sort of hallucinate, 0:11:59.760 --> 0:12:02.199 sort of makes a fact or addressed to a large 0:12:02.240 --> 0:12:06.160 degree by our ability to combine a search engine, which 0:12:06.240 --> 0:12:10.400 is all about finding you authentic, believable information UM with 0:12:10.559 --> 0:12:13.839 the fluency of a large language model. So we are 0:12:13.880 --> 0:12:16.640 at the very much the beginning inning of how large 0:12:16.720 --> 0:12:19.760 language models are going to be integrated into various functions, 0:12:19.760 --> 0:12:23.120 including search engines, so they can be put to very 0:12:23.200 --> 0:12:25.480 very positive views. Just like the Internet. The Internet is 0:12:25.520 --> 0:12:28.640 great for finding things, but obviously a lot of misinformation 0:12:28.640 --> 0:12:31.040 and bad things also happened, so we should think of 0:12:31.080 --> 0:12:34.360 these language models as being the same. But it's a tool. Hey, Sea, 0:12:34.440 --> 0:12:37.200 we're so grateful for you joining us from Munich, Germany 0:12:37.320 --> 0:12:39.439 or attending d l D. I know you're surrounded by 0:12:39.520 --> 0:12:43.320 people who understand the underlying technology, but there are so 0:12:43.360 --> 0:12:45.600 many people out there that do not, And I know 0:12:45.720 --> 0:12:48.560 that it's not straightforward, but could you try and explain 0:12:48.600 --> 0:12:53.160 to our audience what large language models are in Layman's terms, 0:12:53.200 --> 0:12:57.719 how do they work? Yeah, so language models understand the 0:12:57.960 --> 0:13:01.920 structure of how we write and speak um, and you 0:13:01.960 --> 0:13:05.160 should really think of them as this intermediarly. Just like 0:13:05.200 --> 0:13:08.520 the keyboard takes the strokes that you put on it 0:13:08.600 --> 0:13:13.199 and produces sentences, these language models can take your input, um, 0:13:13.320 --> 0:13:16.440 but they're able to respond back to you um with 0:13:16.679 --> 0:13:19.480 writings off their own because they've been trained on so 0:13:19.559 --> 0:13:23.440 much text. UM. But these writings by themselves, like don't 0:13:23.480 --> 0:13:25.960 always have meaning. If you ask you to write a poem, 0:13:26.200 --> 0:13:27.600 you know, should it will write a poem. It will 0:13:27.600 --> 0:13:30.120 be fun. On the other hand, it might not know 0:13:31.040 --> 0:13:33.440 true from false, or like what is a fact and 0:13:33.480 --> 0:13:36.360 what is believable and things like that. So this technology 0:13:36.440 --> 0:13:39.520 is very much developing, but right now you should think 0:13:39.520 --> 0:13:41.880 of this the same way, um you do. You think 0:13:41.880 --> 0:13:45.199 of like any interface or you know, like an interpreter 0:13:45.600 --> 0:13:47.840 um that can listen to you in English and translate 0:13:47.920 --> 0:13:51.040 it into into German. They have that kind of a skill, 0:13:51.520 --> 0:13:53.680 but they need to be augmented with you know, other 0:13:53.760 --> 0:13:55.840 things from real world to be truly useful and those 0:13:55.840 --> 0:13:58.320 things are going to come up sew. We're talking about 0:13:58.360 --> 0:14:00.679 chat GPT for a reason, right because of the news 0:14:00.679 --> 0:14:05.520 cycle Bloomberg reporting that Microsoft is looking at investing maybe 0:14:05.520 --> 0:14:08.800 ten billion dollars into open Ai and chat GPT over 0:14:08.840 --> 0:14:11.320 a series of years. And it seems like, you know, 0:14:11.360 --> 0:14:14.520 the common sense conclusion is that they would use chat 0:14:14.559 --> 0:14:18.280 GPT to improve being as a search engine, make it 0:14:18.320 --> 0:14:22.600 more competitive against Google as somebody that operates in search. 0:14:22.680 --> 0:14:26.000 You know, what's your read on that piece of news. Absolutely, 0:14:26.080 --> 0:14:27.920 you know, being as clearly a number two in the 0:14:28.000 --> 0:14:31.480 search space, and they want to leave pragu over, you know, 0:14:31.640 --> 0:14:34.760 just like we want to leave prob over and create 0:14:34.800 --> 0:14:37.960 a better search experience. It absolutely makes sense that Microsoft 0:14:38.040 --> 0:14:41.120 would want to do that. Um. But we should remember 0:14:41.160 --> 0:14:45.239 that Microsoft actually a juggernaut in things like work communication, 0:14:45.360 --> 0:14:49.040 personal communication. Um. And these language models are going to 0:14:49.080 --> 0:14:52.360 be helping any time we write an email, any time 0:14:52.440 --> 0:14:55.680 we want to consume a document, auto web page. So 0:14:55.760 --> 0:14:58.760 the use cases for the language models are truly massive 0:14:59.160 --> 0:15:01.640 and clearly microsof us being very smart by making a 0:15:01.680 --> 0:15:05.000 big early bet in the space. Should you make bets 0:15:05.160 --> 0:15:07.440 for living to your venture partner at Greylock as well 0:15:07.480 --> 0:15:10.280 as of course co founder of this business. Neither talk 0:15:10.360 --> 0:15:12.640 to us about where the money is being made in 0:15:12.680 --> 0:15:14.760 your business if you're not going to have ads, and 0:15:15.000 --> 0:15:19.080 and if we're trying to visualize what a language format 0:15:19.080 --> 0:15:20.720 would be in terms of search that gives you a 0:15:20.720 --> 0:15:22.920 perfect answer and not only the links to click on 0:15:23.160 --> 0:15:25.960 what what's the what's the money's been ahead? Well, so 0:15:26.080 --> 0:15:28.280 this is real. It actually gets really exciting for Neiva 0:15:28.320 --> 0:15:30.480 because we have a subscription search engine, be said, or 0:15:30.560 --> 0:15:32.880 to create a search engine that was all about you, 0:15:33.120 --> 0:15:36.000 the user. UM, And we have a premium model. You know, 0:15:36.040 --> 0:15:38.960 people pay for premium features like being able to connect 0:15:39.000 --> 0:15:42.440 their email or works like to them, so our model 0:15:42.520 --> 0:15:46.600 is perfectly aligned. UM. I can imagine a scenario in which, yes, 0:15:46.640 --> 0:15:48.720 there is a chat like interface but some of the 0:15:48.840 --> 0:15:52.960 sentences are sponsored. Maybe, UM, but I think the one 0:15:53.080 --> 0:15:56.760 answer format it doesn't really jive all that well with 0:15:56.760 --> 0:15:59.280 with advertising. But there will be other cases. You know, 0:15:59.320 --> 0:16:01.520 people are going to use these models to create like 0:16:01.600 --> 0:16:04.640 personalized advertising for each and every one of us. So 0:16:04.800 --> 0:16:07.160 you're going to lose some there are other places where 0:16:07.200 --> 0:16:09.560 you're going to live in as well. In advertisement. All right, 0:16:09.600 --> 0:16:13.320 Suda Ramaswami really smart analysis of this. Nascent Space co 0:16:13.480 --> 0:16:16.720 founder and CEO of Neiva, thank you so much for joining. 0:16:16.960 --> 0:16:20.000 Now coming up, what's new in well Elon Musk's world, 0:16:20.240 --> 0:16:23.560 from new potential deals to being grand favorite at Goldman Sacks. 0:16:23.560 --> 0:16:26.040 Will have the latest and everything Tesla everything you need 0:16:26.080 --> 0:16:39.600 to know. This is Bloomberg. Now let's get to the 0:16:39.600 --> 0:16:42.400 news in the world of Elon Musk because sources say 0:16:42.440 --> 0:16:44.800 Tesla is close to a prelim deal to set up 0:16:44.840 --> 0:16:48.240 a factory in Indonesia, a nation where reserves of key 0:16:48.280 --> 0:16:51.680 battery metals. Multiple facilities in the country could be serving 0:16:51.720 --> 0:16:55.560 different functions like production across the supply chain. That's according 0:16:55.600 --> 0:16:57.880 to one source, and the plant could produce as many 0:16:57.880 --> 0:17:00.640 as one million cars a year in him with Tessa's 0:17:00.640 --> 0:17:03.760 ambitions for all of its factories globally to eventually reach 0:17:03.840 --> 0:17:06.040 that capacity. But the deal is not signed it could 0:17:06.080 --> 0:17:10.359 still fall through, according to sources. Meanwhile, Tesla remains of 0:17:10.359 --> 0:17:13.879 technology leader in EV's and three top pick in the 0:17:13.920 --> 0:17:17.600 auto industry for Goldman Sacks, with the firm emphasizing that 0:17:17.840 --> 0:17:20.879 the Inflation Reduction Act will have a positive impact on 0:17:20.920 --> 0:17:23.960 the EV maker. The brokerage maintains Tesla at a bye, 0:17:24.040 --> 0:17:26.520 though it cut its price target for the second time 0:17:26.520 --> 0:17:29.000 in less than a month last week to two five 0:17:29.000 --> 0:17:31.120 dollars per share. That's down from a target of four 0:17:31.200 --> 0:17:33.520 hundred dollars a share a year ago, when the firm 0:17:33.560 --> 0:17:36.720 also called Tesla a top pick and to boot Caroline 0:17:36.960 --> 0:17:40.960 Elon Musk is the subject of Wednesday's Big Take, and 0:17:41.000 --> 0:17:43.359 what a big take? Was? The big take all about 0:17:43.600 --> 0:17:46.800 money or Musk Casa CEO. And he called himself, of 0:17:46.800 --> 0:17:49.359 course chief twit. What else has he called himself across 0:17:49.359 --> 0:17:52.919 the years? But he might never reclaim perhaps the title 0:17:52.960 --> 0:17:57.040 as the world's richest person overall aid This entire take 0:17:57.440 --> 0:17:59.879 from Bloomberg is about the way which he paid a 0:18:00.119 --> 0:18:03.560 self basically, and the way he's being taken to task 0:18:03.680 --> 0:18:06.119 in the courts an example, the way in which he 0:18:06.240 --> 0:18:09.439 was compensated. And the whole point was he was compensated 0:18:09.560 --> 0:18:12.399 by options and certain ways in which he had to 0:18:12.440 --> 0:18:15.560 meet certain targets. Huge growth in the share price. But 0:18:15.600 --> 0:18:18.119 for many a shareholder, they're basically saying, look, the idea 0:18:18.200 --> 0:18:20.160 is to lock you in and keep your attention on Tesla, 0:18:20.240 --> 0:18:24.200 but his attention isn't on Tesla. So the Moonshot awards 0:18:24.240 --> 0:18:26.960 so much money, lots of options, but he borrowed against 0:18:26.960 --> 0:18:30.320 it margin loans to invest in other things. When Tesla 0:18:30.320 --> 0:18:32.679 stopped was buoyant. Well it's not now, So now we're saying, well, 0:18:32.720 --> 0:18:35.080 what about margin calls? Will he ever get back to 0:18:35.080 --> 0:18:37.480 top spot? A really fascinating read from our team at 0:18:37.520 --> 0:18:40.000 Bloomberg News. I mean, I think the only person who 0:18:40.000 --> 0:18:42.320 got to the wealthy heights that he did and then 0:18:42.560 --> 0:18:44.040 the only person who have lost it. What is it? 0:18:44.080 --> 0:18:56.520 Two hundred two hundred billion dollars? Welcome back to bloembog Technology. 0:18:56.560 --> 0:18:58.479 I'm Caroline Hyde in New York and I'm Ed Ludlow. 0:18:58.560 --> 0:19:01.800 Also in you special tree this week only for one 0:19:01.840 --> 0:19:04.560 week only though. Meanwhile, well the more updates aid at 0:19:04.560 --> 0:19:07.520 the moment on FTX is bankruptcy case advisors. Well they 0:19:07.560 --> 0:19:10.200 found some money, indeed, about five billion dollars worth in 0:19:10.240 --> 0:19:13.680 casual crypto assets that may help repay creditors RESUS saving 0:19:13.720 --> 0:19:15.880 emotion anti us is here to explain, Well, how far 0:19:15.960 --> 0:19:19.119 this is actually going to go? Yeah, exactly five billion 0:19:19.119 --> 0:19:21.760 dollars is probably not enough for the nine million accounts 0:19:21.800 --> 0:19:23.600 that are owed money. At the end of the day, 0:19:23.600 --> 0:19:27.200 and the bankruptcy advisors did advise as much. Caroline. Now 0:19:27.200 --> 0:19:30.280 remember that number, that sheer number of customers that they're 0:19:30.320 --> 0:19:32.680 dealing with here as they go through this bankruptcy process 0:19:33.160 --> 0:19:36.240 is important because of course there are of all different sizes. 0:19:36.320 --> 0:19:39.000 We know that they have asked earlier on to keep 0:19:39.040 --> 0:19:43.359 the names silent of especially the fifty largest which have 0:19:43.520 --> 0:19:45.600 much more at stake. But at the end of the day, 0:19:45.800 --> 0:19:47.960 once you're spreading the love around the money back to 0:19:48.119 --> 0:19:51.040 the clients, it's not as much as you would get 0:19:51.080 --> 0:19:53.080 back as if you're going to be paid and whole. 0:19:53.800 --> 0:19:56.520 Remember also, they are also saying that above and beyond 0:19:56.560 --> 0:19:59.760 that five billion dollars, there are a lot of illiquid 0:20:00.040 --> 0:20:03.040 sets here that they're dealing with. So in addition to 0:20:03.119 --> 0:20:05.320 the assets that they have said about five billion dollars 0:20:05.359 --> 0:20:07.320 or so, the question is how much is in crypto, 0:20:07.359 --> 0:20:08.959 how much is in cash, and how much of that 0:20:09.080 --> 0:20:11.040 is still worth five billion dollars at the end of 0:20:11.040 --> 0:20:13.680 the day. O pationally one of the more obscure headlines 0:20:13.720 --> 0:20:17.919 across the Bloomberg terminal. Salana based bunk in neu n 0:20:17.960 --> 0:20:23.200 f T surge tenfold after mint, but listing attracts criticism. 0:20:23.320 --> 0:20:28.159 Very simple question, what is the controversy around bunk? A 0:20:28.240 --> 0:20:30.960 simple question without a very simple answer, because at the 0:20:31.040 --> 0:20:33.520 end of the day, Bonk, what was interesting about it 0:20:33.560 --> 0:20:37.280 is really when it made this n f T mint, 0:20:37.359 --> 0:20:38.960 if you will, And a lot of it happened on 0:20:39.040 --> 0:20:42.879 Magic Eden, which remember has been closely tied to Salana. 0:20:43.359 --> 0:20:46.199 They had recently raised money here. A lot of it 0:20:46.240 --> 0:20:50.359 comes back to the purpose of the bonk itself. Of course, 0:20:50.400 --> 0:20:53.600 you have this idea here that the meme stock or 0:20:53.760 --> 0:20:56.639 the mean crypto crazes back in some ways the idea 0:20:56.640 --> 0:20:59.000 of having fun in the industry again, which even Mark 0:20:59.040 --> 0:21:01.000