WEBVTT - DeepSeek, China AI Startup, Panics Markets 0:00:02.520 --> 0:00:13.760 Bloomberg Audio Studios, podcasts, radio news. This is the Bloomberg 0:00:13.840 --> 0:00:17.920 Surveillance Podcast. Catch us live weekdays at seven am Eastern 0:00:18.200 --> 0:00:22.000 on Apple CarPlay or Android Auto with the Bloomberg Business app. 0:00:22.360 --> 0:00:25.680 Listen on demand wherever you get your podcasts, or watch 0:00:25.760 --> 0:00:28.440 us live on YouTube right now. 0:00:28.280 --> 0:00:28.960 In joining us. 0:00:29.000 --> 0:00:31.120 And you know last night when this was blowing up, 0:00:31.200 --> 0:00:32.560 I said get me Mandeep saying. 0:00:32.440 --> 0:00:33.559 You'll join us in a bit here. 0:00:34.080 --> 0:00:37.080 But aniog Rana joins us now from Chicago with a 0:00:37.120 --> 0:00:40.520 little different take than Mandeep. More on the cloud, more 0:00:40.560 --> 0:00:43.760 on the systems analysis, if you will. Of these big companies, 0:00:44.280 --> 0:00:46.760 man Anerog, thank you so much for joining. 0:00:46.800 --> 0:00:48.120 You don't do biholed sell. 0:00:48.800 --> 0:00:54.040 But would you suggest the cell side will shift opinion today. 0:00:54.520 --> 0:00:55.560 For the cloud providers? 0:00:55.600 --> 0:00:58.160 I don't think so because, as you know Paul just 0:00:58.240 --> 0:01:01.960 mentioned about the tweet from my ACOFT CEO, he's probably 0:01:02.000 --> 0:01:04.560 a little bit happier because for him, if the cost 0:01:04.600 --> 0:01:07.200 curve goes down in the long run, and we don't 0:01:07.200 --> 0:01:09.520 know if that's going to happen, you know, his adoption 0:01:09.640 --> 0:01:11.679 rate goes up and his CAPEX goes down. 0:01:11.880 --> 0:01:14.360 So it's it's you know, in a roundabout way. 0:01:15.120 --> 0:01:16.560 Now, we don't think there is going to be an 0:01:16.560 --> 0:01:18.840 impact on CAPEX in the near term, but in the 0:01:18.920 --> 0:01:21.679 long run, if that just shifts, you know, it helps 0:01:21.680 --> 0:01:25.880 their profitability. But more important to that the enterprise you know, 0:01:26.160 --> 0:01:29.720 you could say take create on AI adoption goes up the. 0:01:29.520 --> 0:01:32.200 Theme this morning, Paul's been great about this is like 0:01:32.280 --> 0:01:36.000 drunken sailors from where in Anerich you're just world class 0:01:36.000 --> 0:01:40.600 at this. Are they drunken sailors in terms of CAPEX? 0:01:41.040 --> 0:01:43.840 Or is there a plan? Is there a sequence? Is 0:01:43.840 --> 0:01:45.839 there a pace to spending billions? 0:01:47.120 --> 0:01:48.760 So one of the things, if you would say, is 0:01:48.840 --> 0:01:53.360 if if this AI adoption curve, which is absolutely in 0:01:53.440 --> 0:01:55.640 infancy right now, the only thing we have seen is 0:01:55.640 --> 0:01:58.920 in the consumer world, if enterprises start to put this 0:01:59.080 --> 0:02:02.160 in their core app locations, then they would need a 0:02:02.240 --> 0:02:05.880 lot more you could say, firepower to run it. Most 0:02:05.880 --> 0:02:08.200 of the things that we have discussed is people would 0:02:08.240 --> 0:02:10.560 not want to do this on their own premise. They're 0:02:10.600 --> 0:02:13.520 bigger companies will but most of them will go to 0:02:13.639 --> 0:02:17.280 a handful of cloud providers, and that would be Microsoft, Aws, 0:02:17.360 --> 0:02:21.040 Google and then Oracle. So for all these companies to 0:02:21.080 --> 0:02:24.600 increase their capacity now they are only seeing the orders 0:02:24.639 --> 0:02:28.359 and based on that they are expanding their data center footwork. 0:02:28.680 --> 0:02:30.000 The question is whether do. 0:02:29.919 --> 0:02:32.120 They need the expensive chips to do with this or 0:02:32.120 --> 0:02:34.959 can they run this in with cheaper chips. And that's 0:02:35.000 --> 0:02:37.320 really I think the biggest discussion going to be over 0:02:37.360 --> 0:02:38.480 the next twelve months. 0:02:38.760 --> 0:02:41.760 An raglick Tom has often want to say, my head 0:02:41.800 --> 0:02:45.160 is spinning here. Just since Friday. I've got a big, 0:02:45.200 --> 0:02:48.320 big tech company called Meta I call it Facebook given 0:02:48.360 --> 0:02:51.160 the IPO. We spent a lot of time thinking about 0:02:51.200 --> 0:02:55.520 that name in the tech the ticker FB. They up 0:02:55.520 --> 0:02:57.640 their cap x from fifty billion to sixty five billion, 0:02:57.639 --> 0:03:00.800 a large part on AI. Now I've got this Deep 0:03:00.840 --> 0:03:02.600 Seek thinking about I can do this in. 0:03:02.720 --> 0:03:06.000 Much lower costs. How do we square those two things? 0:03:06.520 --> 0:03:08.880 Yeah, but one thing is for sure, this Deep Seek 0:03:08.919 --> 0:03:10.640 news of the paper has been out for a while, 0:03:10.680 --> 0:03:12.799 so it does got It did get publicized over the 0:03:12.840 --> 0:03:15.600 last few days. But I'd be very surprised if Meta 0:03:15.639 --> 0:03:18.840 AWS and Microsoft did not know about this. It has 0:03:18.880 --> 0:03:21.480 been around for a while, so both of them, both 0:03:21.600 --> 0:03:25.959 Meta and Microsoft talked about their capex last week after 0:03:26.040 --> 0:03:28.960 knowing this news, so you know they have moved. 0:03:29.000 --> 0:03:31.079 I mean, they're far smarter than I am when it 0:03:31.080 --> 0:03:31.520 comes to this. 0:03:31.639 --> 0:03:34.000 So if they're taking up their Capex, I mean there 0:03:34.040 --> 0:03:35.200 must be a real need for it. 0:03:35.720 --> 0:03:38.960 What's it mean for Coopertino there having an early morning? 0:03:39.160 --> 0:03:42.600 They don't have lotes at Coopertino? They got something, Lisa. 0:03:42.640 --> 0:03:45.080 What's the thing with the green powder in it? 0:03:45.280 --> 0:03:46.520 The machi? 0:03:46.640 --> 0:03:50.960 Yeah matter, they're having the machiat Coopertino some undrinkable bilge. 0:03:51.120 --> 0:03:53.720 What are they saying at Apple today? 0:03:53.600 --> 0:03:56.320 They're actually smiling and saying, you guys fight it out 0:03:56.360 --> 0:03:59.280 and tell me what's the best large language model out there. 0:03:59.400 --> 0:04:01.200 I'll just put it in my phone and I won't 0:04:01.200 --> 0:04:03.240 even pay for it. So they if you just look 0:04:03.280 --> 0:04:05.560 at you know, Apple right now, that's kind of the 0:04:05.600 --> 0:04:08.760 safety haven for people is they're not spending that capex. 0:04:08.800 --> 0:04:11.080 If you look at that Capex line, it hasn't moved 0:04:11.120 --> 0:04:13.080 at all in two years, and the free cash slow 0:04:13.200 --> 0:04:16.200 just keep on coming. Their issue, honestly, is more so 0:04:16.600 --> 0:04:19.360 on the you know, the growth side. They are having 0:04:19.440 --> 0:04:23.080 challenges in China that needs to you know, smoothened out 0:04:23.160 --> 0:04:25.440 over the next twelve months. Otherwise this is going to 0:04:25.440 --> 0:04:28.160 be another year for them. To grow only four five percent. 0:04:28.360 --> 0:04:30.760 That's the big issue for them. Nothing to do with AI. 0:04:30.920 --> 0:04:32.760 All right, Well, let's just stay on the Apple story. 0:04:32.800 --> 0:04:35.800 It stocks down eleven percent year to date this year, 0:04:35.800 --> 0:04:38.080 it's only a fifteen percent over the trailing twelve months, 0:04:38.120 --> 0:04:41.200 which is radical on the performance. And for a company 0:04:41.240 --> 0:04:45.640 like Apple, is there a bulkcase here short term? 0:04:45.640 --> 0:04:46.080 For Apple? 0:04:47.040 --> 0:04:48.880 In the short term, it's very difficult to come up 0:04:48.880 --> 0:04:51.680 with the bulkcase because, as we talked about it, even 0:04:51.800 --> 0:04:54.600 you know a few months ago, Apple Intelligence is not 0:04:54.640 --> 0:04:56.440 going to be out in other parts of the world. 0:04:56.480 --> 0:04:58.680 It's not out in China, and China is where they 0:04:58.720 --> 0:05:01.520 are struggling honestly at this and they need to come 0:05:01.560 --> 0:05:04.279 out and pacify investors that you know, we have a 0:05:04.279 --> 0:05:08.040 strategy in China, we will gain market share back. Otherwise 0:05:08.040 --> 0:05:10.080 this is going to be another mut the year for them. 0:05:10.640 --> 0:05:12.680 We think the next phone, which is the one that's 0:05:12.680 --> 0:05:15.360 coming out in September, I know that's fa I think 0:05:15.400 --> 0:05:18.039 that's the big catalyst, but that's like, you know, that's 0:05:18.080 --> 0:05:19.839 eight months out, nine months out. 0:05:19.880 --> 0:05:23.280 Right, and Agrana with us in Chicago right now. We 0:05:23.279 --> 0:05:25.760 welcome all of you on your commute across the Nation. 0:05:26.200 --> 0:05:29.240 For those of you in the office at home on YouTube, 0:05:29.320 --> 0:05:31.880 thank you for being on youtubday. Just brilliant live chat 0:05:31.960 --> 0:05:34.760 of learning a lot from people smarter than me on 0:05:34.800 --> 0:05:39.480 the YouTube live chat this morning. Subscribe at Bloomberg Podcast. 0:05:39.600 --> 0:05:42.520 That's the best way to get to the show out there. 0:05:42.720 --> 0:05:45.880 This is a rare privilege. And for senior management listening 0:05:45.880 --> 0:05:49.200 on radio right now. We adhere to their policy which 0:05:49.200 --> 0:05:51.760 is Ana Agarana and man Deep Sing can never be 0:05:51.800 --> 0:05:54.719 in the same ruct together. It's too much worse to 0:05:54.760 --> 0:05:57.800 the to the franchise. But Paul, why don't you bring 0:05:57.800 --> 0:05:58.560 in man Deep Sing? 0:05:58.600 --> 0:06:00.800 As we talked to both and Ruana and. 0:06:00.800 --> 0:06:03.760 Mandy man Deep Seeing covers the technology space along with 0:06:03.800 --> 0:06:07.440 Ana rog Rana, there are two leaders globally running our 0:06:07.480 --> 0:06:11.440 technology franchise and it is a global franchise analyst in Europe, 0:06:11.520 --> 0:06:15.719 Asia and North America. Man Deep, But I guess what 0:06:15.760 --> 0:06:20.520 we're hearing such Antela from Microsoft put out something that hey, 0:06:20.839 --> 0:06:24.520 cost coming down is pretty natural for the technology space. 0:06:24.560 --> 0:06:27.839 That happens, and it's a good thing because it helps adoption. 0:06:28.800 --> 0:06:31.880 Does that make sense to you, Well, clearly, I think 0:06:32.400 --> 0:06:35.600 the investor expectations were different, which is why you see 0:06:35.640 --> 0:06:37.039 this sort of stock reaction. 0:06:37.200 --> 0:06:38.320 So yes, we. 0:06:38.360 --> 0:06:41.080 Have Moore's law and we have other laws in terms 0:06:41.080 --> 0:06:44.200 of figuring out how the cost curve may look like 0:06:44.320 --> 0:06:47.200 over time. But clearly this is a step change in 0:06:47.279 --> 0:06:49.800 terms of, you know, what you can do with the 0:06:50.640 --> 0:06:53.240 existing chips that are out there, as well as what 0:06:53.640 --> 0:06:56.919 you will need going forward. So I think this is 0:06:57.000 --> 0:07:00.960 a big moment in terms of just that we can 0:07:01.000 --> 0:07:03.640 see with these chips and the kind of scale that 0:07:03.760 --> 0:07:07.760 is required for the next versions of these foundational models. 0:07:08.000 --> 0:07:11.200 But there is no doubt that this makes generative AI 0:07:11.440 --> 0:07:14.320 more accessible to software companies as well as you know, 0:07:14.360 --> 0:07:15.200 overall users. 0:07:15.480 --> 0:07:18.440 Anrak, what do you think this means to you know, 0:07:18.480 --> 0:07:21.320 I know you guys at Bloomberg Intelligence you have access 0:07:21.360 --> 0:07:23.200 to this I d C data, which is some of 0:07:23.240 --> 0:07:26.600 the best data on technology spending in the marketplace, and 0:07:26.640 --> 0:07:29.280 you guys feature that in your research. Is this going 0:07:29.320 --> 0:07:32.920 to change i DC's forecast of tech spending? 0:07:34.600 --> 0:07:37.080 AI spending? Do you think so? 0:07:37.160 --> 0:07:39.880 AI spending has been going up, Paul, We've talked about 0:07:39.880 --> 0:07:42.400 this a few times, but it's the other non AI 0:07:42.480 --> 0:07:45.160 spending that has actually not done so well in the 0:07:45.200 --> 0:07:48.120 last two years and over the next I would say month, 0:07:48.200 --> 0:07:50.240 month and a half, we will get a much clearer 0:07:50.280 --> 0:07:54.440 picture if corporations or enterprises are moving up there over 0:07:54.480 --> 0:07:57.680 our tech budgets or not, because right now AI spending 0:07:57.760 --> 0:08:00.880 is coming at the expense of areas you know, normal 0:08:00.960 --> 0:08:05.360 software or you know, just bording hardware upgrades or consulting 0:08:06.240 --> 0:08:08.640 that really needs to pick up because a large portion 0:08:08.760 --> 0:08:11.000 of the tech ecosystem is some of those companies. 0:08:12.000 --> 0:08:12.280 Mande. 0:08:12.280 --> 0:08:14.680 If I think when I think of this news today again, 0:08:14.720 --> 0:08:17.120 I'm like everybody else. Our listeners and viewers are learning 0:08:17.120 --> 0:08:19.800 about deep Seek and what it means. I can't help 0:08:19.840 --> 0:08:23.480 but think about TikTok at some point. Again, I feel 0:08:23.480 --> 0:08:25.280 like by the end of business today, we're gonna have 0:08:25.280 --> 0:08:27.240 a story coming into Washington. There's gonna be an opinion 0:08:27.280 --> 0:08:31.080 coming from Washington as it relates to deep Seek. Does 0:08:31.120 --> 0:08:33.800 oppose a similar risks? Do you think to TikTok from 0:08:33.800 --> 0:08:36.080 a regulator's perspective, Well. 0:08:35.840 --> 0:08:39.600 It's still the app is still you know, early in 0:08:39.679 --> 0:08:43.720 terms of adoption, and look, I'm sure it's got millions 0:08:43.760 --> 0:08:49.000 of users in East Asia and you know, China and whatnot. 0:08:49.080 --> 0:08:54.120 But I think when it comes to AI consumers, just 0:08:54.240 --> 0:08:56.960 look at the functionality, and you know the fact that 0:08:57.000 --> 0:09:00.800 they have an app wage is accessible for free as 0:09:00.800 --> 0:09:03.320 opposed to paying two hundred dollars a month, they'll end 0:09:03.400 --> 0:09:05.959 up using it. So it's the same argument with TikTok. 0:09:06.000 --> 0:09:07.840 You know they will go with the functionality. They don't 0:09:07.880 --> 0:09:09.120 care about the privacy and. 0:09:09.160 --> 0:09:11.920 The data man keep saying with us and ana ug Runa, 0:09:11.960 --> 0:09:15.640 we continue anerog, let's go to earnings. Let's like remove 0:09:15.679 --> 0:09:19.800 ourselves from deep seek. What is the character of the 0:09:19.880 --> 0:09:23.679 cloud mystery that you see in the earning season, Aniog. 0:09:24.800 --> 0:09:27.760 Yeah, we are anticipating in improvement in growth rates for 0:09:27.800 --> 0:09:31.040 both Microsoft and AWS going in. In fact, Microsoft, we 0:09:31.080 --> 0:09:34.040 are hoping they come out and increase their guidance for 0:09:34.200 --> 0:09:36.920 Azure numbers for the next two quarters. They have been 0:09:36.960 --> 0:09:40.560 struggling with supply challenges where they are not able to 0:09:40.600 --> 0:09:43.679 fulfill the demand that's coming in in their data centers, 0:09:43.679 --> 0:09:46.760 and if they give that signal that you know, that 0:09:47.000 --> 0:09:50.760 should revive or pacify some of the fears that we 0:09:50.800 --> 0:09:54.200 are seeing right now in the market. Same thing for 0:09:54.240 --> 0:09:57.160 a WS AWS has been talking about a recovery in 0:09:57.200 --> 0:10:01.319 consumption for their core business looking forward to some comments 0:10:01.360 --> 0:10:03.120 sometimes as well from them. 0:10:03.240 --> 0:10:04.319 So I'm hopeful that. 0:10:04.280 --> 0:10:06.760 Both companies will come up with good numbers over the 0:10:06.760 --> 0:10:08.120 next you know, ten days. 0:10:08.200 --> 0:10:10.560 But let's say it's not like you know, Procter and Gamble. 0:10:10.600 --> 0:10:13.840 It's not a unit and price store hit the revenue line. 0:10:14.080 --> 0:10:17.600 But where is the fear in the earning season analog 0:10:18.080 --> 0:10:20.920 on the income statement. Is it a margin fear, a 0:10:21.040 --> 0:10:23.599 revenue fear, or something I don't understand. 0:10:24.679 --> 0:10:26.840 I think it's going to be a backlock slash revenue 0:10:26.840 --> 0:10:30.320 fear more than the margins, because everybody understands this is 0:10:30.360 --> 0:10:32.760 a very you know, unique business. When you have a 0:10:32.760 --> 0:10:35.560 fixed cost business like this, you can scale it up 0:10:35.559 --> 0:10:38.400 in the near town, but once you reach a mature face, 0:10:38.640 --> 0:10:40.760 you know, all the revenue comes back, or you can 0:10:40.760 --> 0:10:42.920 distribute that over a very large platform. 0:10:43.080 --> 0:10:45.040 That has been the case for Cloud over the last 0:10:45.160 --> 0:10:45.960 you know, ten years. 0:10:46.520 --> 0:10:49.320 So in an expansion of capacity, people don't really you know, 0:10:49.480 --> 0:10:51.240 see that as a big issue in the near town 0:10:51.920 --> 0:10:52.400 man Deep. 0:10:52.520 --> 0:10:56.760 Last Friday, Meta took their capex guidance from fifty billion, 0:10:57.000 --> 0:11:00.560 which was consensus, which was kind of guidance, to sixty 0:11:00.600 --> 0:11:04.960 five billion. I've never seen that before. In order of magnitude, 0:11:05.400 --> 0:11:06.679 what's going on there. 0:11:06.840 --> 0:11:11.199 Well, and I guess they probably knew this deep Seek advancement, 0:11:11.320 --> 0:11:13.080 so it's not as if, you know, they were cut 0:11:13.080 --> 0:11:16.000 off guard with the deep Seek release, so they did 0:11:16.000 --> 0:11:20.000 it despite the you know, the whole thing around deep Seek. 0:11:20.080 --> 0:11:23.040 So to my mind, you know, every company is doubling 0:11:23.080 --> 0:11:26.160 down on AI right now, even though we are talking 0:11:26.160 --> 0:11:29.440 about pricing, compression and whatnot. But this is the moment 0:11:29.520 --> 0:11:34.480 where generative AI will be incorporated in all software, all applications, 0:11:34.520 --> 0:11:36.600 and every company wants a piece of it, so they 0:11:36.600 --> 0:11:38.839 don't mind spending more on County in. 0:11:38.760 --> 0:11:41.480 The room a ROD does a public one AI. Is 0:11:41.480 --> 0:11:45.480 there any interest that Lisa Mateo is waiting and waiting 0:11:45.559 --> 0:11:46.800 and waiting for AI. 0:11:48.320 --> 0:11:51.000 See, the thing is that the consumers have already embraced 0:11:51.040 --> 0:11:55.640 it through chat, GPT, through Gemini, through Perplexity. But when 0:11:55.640 --> 0:11:58.040 you look at the enterprise use case, we are still 0:11:58.040 --> 0:12:00.880 in the building phase. And that's probably this year is 0:12:00.920 --> 0:12:03.360 going to be like next year, we should see a 0:12:03.440 --> 0:12:06.880 much fostered adoption by companies in terms of and that 0:12:07.160 --> 0:12:09.839 you said will show up in cloud numbers right now. 0:12:09.880 --> 0:12:12.480 Even when you look for somebody like a Microsoft, most 0:12:12.480 --> 0:12:15.080 of their AI revenue is from consumer app, which is 0:12:15.320 --> 0:12:17.000 basically open AI is shock GPT. 0:12:18.000 --> 0:12:20.240 So all right, Tom, we like to talk to anog. 0:12:20.480 --> 0:12:22.440 Do you like Ludlow this morning? 0:12:22.640 --> 0:12:22.720 No? 0:12:22.800 --> 0:12:26.120 You just I mean we're likes will be Bloomberg Tech exactly. 0:12:26.200 --> 0:12:28.960 I mean, can you get the shoes with the White Souls? 0:12:29.040 --> 0:12:32.040 Yes, that's what David shoes. 0:12:33.040 --> 0:12:36.000 No, you weren't shoes of the White Souls. 0:12:36.520 --> 0:12:39.600 No, No, old school. 0:12:39.400 --> 0:12:41.880 And Tom, you know Man Deep and Hon they're pretty good. 0:12:41.920 --> 0:12:45.319 But the research, the real investment research on this whole 0:12:46.200 --> 0:12:48.520 Deep seek Chinese AI. 0:12:48.559 --> 0:12:51.760 Robert Lee. These are Bloomberg intelligence analysts in Hong Kong. 0:12:51.800 --> 0:12:53.680 So folks who have access to the Bloomberg terminal go 0:12:53.720 --> 0:12:57.679 there because Man Deep and they're just he was out 0:12:57.679 --> 0:12:59.440 with research a couple of months ago on this and 0:12:59.480 --> 0:13:01.000 he says, hey, this stuff is there. 0:13:01.080 --> 0:13:02.160 Okay. 0:13:02.440 --> 0:13:06.840 You know Lea in Uh in Hong Kong was way 0:13:06.840 --> 0:13:08.960 out front of this. Why are we stunned this morning? 0:13:09.520 --> 0:13:12.199 Well because he didn't found the table in terms of saying, 0:13:12.240 --> 0:13:15.560 my god, the table we're negative? 0:13:15.679 --> 0:13:18.240 Where are we negative? Three point five percent? In YAC 0:13:18.360 --> 0:13:21.880 right now? What I mean, let's bring it back to 0:13:25.080 --> 0:13:28.280 what happened. Why were we surprised. 0:13:27.720 --> 0:13:30.160 By this, No, I don't think you are surprised because 0:13:30.240 --> 0:13:32.760 open source has been talked about a lot. We have 0:13:32.880 --> 0:13:35.679 been talking about how companies are looking to you know, 0:13:35.840 --> 0:13:40.280 drive scaling uh at inference time, and how pricing was 0:13:40.320 --> 0:13:43.520 a constraint, and so look, this is the way the 0:13:43.559 --> 0:13:46.440 industry was headed. It's just that Deep Sea came out 0:13:46.440 --> 0:13:49.839 with a you know, a long paper that describes what 0:13:49.880 --> 0:13:52.480 they did in detail, and everyone else is kind of 0:13:52.679 --> 0:13:56.160 closed source. So being open source has its advantages, and 0:13:56.200 --> 0:13:57.840 that's what we're seeing with deep Seat here. 0:13:58.600 --> 0:14:02.440 So the baniers to entry here on ROD just in general, 0:14:02.840 --> 0:14:06.240 on AI are the banished entry low for AI. 0:14:08.120 --> 0:14:11.079 I think that's exactly what this paper that Mandep is 0:14:11.080 --> 0:14:13.360 alluding to is talking about. That you and I can 0:14:13.400 --> 0:14:15.880 go out and build some of this thing without much capital. 0:14:16.120 --> 0:14:18.160 Now the question is how are you going to distribute it? 0:14:18.400 --> 0:14:19.960 And which is why I go back to the cloud 0:14:19.960 --> 0:14:22.400 providers that those are the guys who are going to 0:14:22.440 --> 0:14:25.600 work with enterprises to embed some of this technology. If 0:14:25.640 --> 0:14:28.000 you look at somebody like an AWS, they work with 0:14:28.040 --> 0:14:31.080 all these providers, they work with Meta Islama model or 0:14:31.080 --> 0:14:35.600 Anthropics cloud models. So it's really up to the customer 0:14:35.680 --> 0:14:39.200 which model they choose or sometimes they would write the 0:14:39.240 --> 0:14:43.160 reroute the inquiry based on what's better for that particular question. 0:14:43.680 --> 0:14:45.520 So it's really the distribution. 0:14:45.880 --> 0:14:48.360 What I'm saying is is the value here which takes 0:14:48.400 --> 0:14:50.600 us back to the cloud providers and also do Apple. 0:14:50.840 --> 0:14:52.960 This has been fun erragran and your trooper to get 0:14:53.040 --> 0:14:56.560 up this early in Chicago aner Agroana with Bloomberg Intelligence 0:14:56.560 --> 0:14:58.680 at Man Deep saying, can we say on behalf of 0:14:58.680 --> 0:14:59.160 all of us? 0:14:59.160 --> 0:15:00.440 Mandep, thank you much. 0:15:01.040 --> 0:15:03.680 Okay, we'll try not to do it three times tomorrow, 0:15:04.080 --> 0:15:05.680 maybe once, or maybe we'll see. 0:15:05.480 --> 0:15:06.480 A Wednesday or whatever. 0:15:06.840 --> 0:15:09.920 Man Deep sing and enter Agrana. Is what the show 0:15:09.960 --> 0:15:13.480 is about, just world class on tech analogy. 0:15:18.360 --> 0:15:21.960 You're listening to the Bloomberg Surveillance Podcast. Catch us live 0:15:22.000 --> 0:15:25.200 weekday afternoons from seven to ten am Eastern Listen on 0:15:25.240 --> 0:15:28.920 Applecarplay and Android Otto with the Bloomberg Business app, or 0:15:29.080 --> 0:15:30.720 watch us live on YouTube. 0:15:31.200 --> 0:15:31.880 We Get Lucky. 0:15:31.960 --> 0:15:35.200 Joe Wisenthal has done so much for Bloomberg, I should say, 0:15:35.200 --> 0:15:37.560 for business journalism, and of course it's work here at 0:15:37.560 --> 0:15:41.520 Bloomberg with Tracy Alloway. But one of the great things 0:15:41.840 --> 0:15:44.400 is he did Paul, he you know, I'm like a 0:15:44.640 --> 0:15:48.400 I E I E I O. Joe's actually doing it, 0:15:48.480 --> 0:15:51.320 he said, no, but but nobody. He looked at the 0:15:51.360 --> 0:15:54.480 summer tour of Light Sweet Crude and said, I can't 0:15:54.560 --> 0:15:59.240 do the Light Sweet Crude analysis without AI. You three 0:15:59.320 --> 0:16:04.680 days ago you nailed deep Seek. You're using it at home. Yeah, 0:16:04.920 --> 0:16:07.160 what's the experience of deep Sea? 0:16:07.240 --> 0:16:09.600 So I'll just see the main There are two main 0:16:09.640 --> 0:16:14.560 takeaways for me. One is I have never experienced any 0:16:14.600 --> 0:16:18.160 technology in the history of any tech, computers, apps, whatever, 0:16:18.400 --> 0:16:20.040 that have zero switching. 0:16:19.720 --> 0:16:21.400 Costs the way AI chatbots have. 0:16:21.840 --> 0:16:25.720 I used Chad GBT for a long time. Claude came out, 0:16:26.360 --> 0:16:29.280 the one from Anthropic, I was like, oh, this works Gemini. 0:16:29.600 --> 0:16:31.600 I used Gemini now that for some reason, that one 0:16:31.600 --> 0:16:33.600 hasn't sucking as much then I. But then deep Sea 0:16:33.680 --> 0:16:35.320 came out and I said, or I saw the chap 0:16:35.400 --> 0:16:37.760 out last week online and I was like, I'm gonna 0:16:37.800 --> 0:16:40.360 try this out. It took me thirty seconds to register, 0:16:40.720 --> 0:16:42.600 and I was like, this is just as good as 0:16:42.680 --> 0:16:44.800 what I'm getting from Chad. You know, I ask at 0:16:44.840 --> 0:16:46.480 random things during the day and. 0:16:47.680 --> 0:16:52.400 It's just zerop radio. It's all, you know. 0:16:52.520 --> 0:16:55.440 I asked a random historical questions or linguistic I'm interested 0:16:55.440 --> 0:16:56.040 in Lenco, like. 0:16:55.960 --> 0:16:58.680 Why did Texas lose to Ohio? No, I can't. That's 0:16:58.720 --> 0:17:00.640 too painful. I don't ask that. 0:17:00.920 --> 0:17:04.240 But then the other thing is I made a twenty 0:17:04.280 --> 0:17:07.760 twenty five one of my New Year's resolutions. Yep, I've 0:17:07.800 --> 0:17:09.800 never I said I'm going to try and code something 0:17:09.920 --> 0:17:12.159 with AI one of my goals. Yeah, And so I 0:17:12.200 --> 0:17:14.280 started like building a little software. But I just asking 0:17:14.320 --> 0:17:17.480 to create this code, and uh, I was able to 0:17:17.520 --> 0:17:19.840 plug my app that I'm building on my home. It's 0:17:19.840 --> 0:17:24.920 a little linguistics app into the Deepseek API, and it's 0:17:24.920 --> 0:17:26.840 like a fraction of the cost of what I use 0:17:26.880 --> 0:17:28.920 for the open AI API. So, like I just played 0:17:28.920 --> 0:17:30.920 around with it. I was like, Wow, here's this thing 0:17:30.960 --> 0:17:35.520 that seems just as good as changchipyt. Sorry, here's this thing. 0:17:35.359 --> 0:17:37.879 That seems just good changing formosa. 0:17:39.040 --> 0:17:41.960 So yes, I'm not surprised that, Like, all right, the 0:17:42.080 --> 0:17:44.639 dumb question there, Yeah, are we not using Google anymore 0:17:44.640 --> 0:17:45.080 because of this? 0:17:45.200 --> 0:17:45.280 No? 0:17:45.320 --> 0:17:47.719 I still well, look it is you know, like I 0:17:47.760 --> 0:17:49.800 still do. I still use Google all the time. I 0:17:49.840 --> 0:17:53.000 still use Wikipedia all the time. They're just different. 0:17:53.240 --> 0:17:55.679 But what's your initial go to, oh, I need to 0:17:55.760 --> 0:17:56.440 know about this. 0:17:56.720 --> 0:18:01.000 I think it really depends on the context. I say, 0:18:01.080 --> 0:18:05.399 I really like using AI for things like understanding a 0:18:05.520 --> 0:18:08.639 term within a specific context or something like that or something. 0:18:08.880 --> 0:18:10.480 But you know, like you don't know if it's true, 0:18:10.520 --> 0:18:12.640 So you know, I always do a backup. I see 0:18:12.640 --> 0:18:14.720 something and then I say, oh, look at Google is 0:18:14.760 --> 0:18:15.600 like is this actually true? 0:18:16.160 --> 0:18:21.120 Morgan Brown, Dropbox Imperial w He's out with the smartest 0:18:21.160 --> 0:18:23.359 note I've seen where he's talking about decimal points. I 0:18:23.359 --> 0:18:25.880 don't want to go into that. You and Tracy at 0:18:25.880 --> 0:18:30.520 audience talk a lot more to techy people about this. 0:18:31.000 --> 0:18:36.600 What's their message on the future of American AI versus 0:18:36.640 --> 0:18:37.720 these shock threats. 0:18:39.400 --> 0:18:41.840 I mean, you know what, here's what my my take, 0:18:42.600 --> 0:18:46.000 which is that they're already I mean, we saw the 0:18:46.000 --> 0:18:49.000 Trump announcement last week with Stargate right half a trillion dollar. 0:18:49.200 --> 0:18:51.800 I thought that was I thought that was the movie 0:18:51.880 --> 0:18:54.320 with what's happening now? 0:18:54.440 --> 0:18:56.680 I think because of the arrival of this very compelling 0:18:56.920 --> 0:18:59.680 Chinese model, is it just cements the notion that there 0:18:59.720 --> 0:19:02.119 is a global arms race allah the bomb And I 0:19:02.119 --> 0:19:04.920 don't know if achieving AGI or whatever people talk about 0:19:05.080 --> 0:19:08.760 is a breakthrough on par or is strategic as a 0:19:08.960 --> 0:19:12.159 nuclear bomb. But now everyone is just going to be 0:19:12.480 --> 0:19:13.959 we have to beat them, we have to beat them, 0:19:13.960 --> 0:19:16.480 and there's gonna be so much money flowing to tech 0:19:16.520 --> 0:19:19.760 companies now in this idea that there is a global race, Joe. 0:19:19.560 --> 0:19:21.840 I feel like by the end of business today, the 0:19:21.840 --> 0:19:24.719 top star on the Bloomberg terournal will be US government 0:19:24.760 --> 0:19:28.840 looking at deep Seek as it's looking at TikTok TikTok. 0:19:29.119 --> 0:19:30.560 Well, you know what, so this is interesting. 0:19:30.560 --> 0:19:32.240 The other thing, and thanks for reminding me of this, 0:19:32.440 --> 0:19:35.520 which is that deep Seek is a piece of open, 0:19:35.920 --> 0:19:38.960