WEBVTT - Markets Wrap: Higher Yields, Higher Gold 0:00:02.720 --> 0:00:19.400 Bloomberg Audio Studios, Podcasts, Radio News. Welcome to the Maren 0:00:19.400 --> 0:00:21.759 Talks Money Market Wrap, where we talk about the biggest 0:00:21.800 --> 0:00:24.560 moves in the markets this week and what is driving them. 0:00:24.640 --> 0:00:27.960 I'm Maren Zum SAP Web. You gave Money editor at Large. 0:00:27.960 --> 0:00:30.080 And I'm joined step pick Ed. 0:00:31.680 --> 0:00:35.000 Who are you? John? Who are you? Everybody? 0:00:35.040 --> 0:00:37.800 Let's cut the man some slack. He was only you've 0:00:37.840 --> 0:00:39.479 been at work for two days or something. That's the 0:00:39.520 --> 0:00:41.720 longer stretch. He's done for months. For God's sake, You've 0:00:41.720 --> 0:00:42.800 got to cut on some flags. 0:00:43.360 --> 0:00:44.839 Who are you, John? Do me to help you. 0:00:44.840 --> 0:00:46.280 I'm going to do this for John right, so he 0:00:46.360 --> 0:00:49.159 can he can gather himself for the next bet. Okay, 0:00:49.479 --> 0:00:52.479 I'm John steppeck Holduly. I'm not his senior reporter. And 0:00:52.560 --> 0:00:54.960 over the Money Distilled newsletter. 0:00:54.920 --> 0:00:58.160 Award winning You'll never forget that bit. 0:00:58.240 --> 0:00:59.760 Forget your own name, forget your job. 0:01:00.080 --> 0:01:03.520 You don't get your awards, right, John, I think we 0:01:03.640 --> 0:01:06.840 have to start by talking about the bond market. Right, 0:01:06.920 --> 0:01:09.399 there's a lot going on. Let's have a quick look 0:01:09.480 --> 0:01:13.600 at yields today. We are talking on Thursday the twentieth 0:01:13.920 --> 0:01:16.200 about thirty year yields in the US and the UK 0:01:16.360 --> 0:01:19.040 are at a multi decade decade highs at this point 0:01:19.280 --> 0:01:21.959 tenure yields. The US is up near four point seven 0:01:22.040 --> 0:01:24.480 four point six eight, Japan is not far off three. 0:01:24.800 --> 0:01:27.800 Everywhere in the EU is knocking around three and four percent, 0:01:28.080 --> 0:01:31.400 and of course here in the UK we are firmly 0:01:31.840 --> 0:01:33.400 over five percent. 0:01:33.760 --> 0:01:34.720 Poorless Truss is. 0:01:34.720 --> 0:01:36.720 All I have to say to that, poor poor list 0:01:36.720 --> 0:01:38.960 crush lost a job for a bond dealed much lower 0:01:39.000 --> 0:01:39.320 than that. 0:01:40.400 --> 0:01:43.479 It's true, it's a double standard. It's almost like there's 0:01:43.520 --> 0:01:49.120 more than one tier involved, more than one tier, two tiers. 0:01:54.520 --> 0:01:59.960 All right, no more what's it called? No more dog whistling, John, 0:02:00.080 --> 0:02:03.559 That's right. What's going on in the bond market? 0:02:04.280 --> 0:02:08.000 We'll basically mean bond yields have gone up a lot, 0:02:09.040 --> 0:02:11.760 and they are, yes, a multi decade ties, particularly what 0:02:11.840 --> 0:02:14.120 it's known as the long end of the bond market, 0:02:14.240 --> 0:02:18.920 so you know, way out above ten years, thirty years. 0:02:19.520 --> 0:02:22.520 And I mean this is because I mean this is 0:02:22.520 --> 0:02:25.200 actually for pretty logical reasons. Governments have spent a lot 0:02:25.200 --> 0:02:26.920 of money, so there's a lot of points out there. 0:02:27.240 --> 0:02:29.320 EI companies are spending a lot of money, so there's 0:02:29.320 --> 0:02:31.400 even more points out there. So there's an awful lot 0:02:31.440 --> 0:02:35.080 of supply, and then on the other side, we've also 0:02:35.200 --> 0:02:41.160 got the warn Iran continues that keeps on pushing the 0:02:41.200 --> 0:02:44.200 oil price kind of up and down, but that also 0:02:44.240 --> 0:02:46.920 keeps feeding through the kind of diesel prices, petrol prices, 0:02:46.960 --> 0:02:50.360 et cetera, et cetera, And so markets are also still 0:02:50.360 --> 0:02:53.520 concerned about inflation, which also kind of pushes up bond yield. 0:02:54.440 --> 0:02:58.720 So what happened this week is that yesterday Wednesday, US 0:02:58.800 --> 0:03:03.760 Treasury Secretary Being clearly has been feeling edgy about this 0:03:04.400 --> 0:03:07.639 because earlier in the year he basically said to Japan, look, 0:03:07.680 --> 0:03:10.240 if you want to strengthen the yen, I'm okay with that, 0:03:10.400 --> 0:03:13.720 but I don't want you selling US treasuries. Instead, you 0:03:13.760 --> 0:03:15.680 can borrow the money from us, and that's how you 0:03:15.720 --> 0:03:18.560 can raise the dollars to you know, we can strengthen 0:03:18.840 --> 0:03:21.680 the yen against the dollar. And now be the latest thing, 0:03:21.760 --> 0:03:26.040 what he's said is we're going to buy back double 0:03:26.200 --> 0:03:30.160 the number of long end US treasuries that we were 0:03:30.560 --> 0:03:34.880 planning to this quarter. And basically what that involves is 0:03:34.880 --> 0:03:38.320 issuing more short term debt to buy back in some 0:03:38.440 --> 0:03:40.960 of the thirty year points that have been, you know, 0:03:41.240 --> 0:03:43.960 going that are in the market, and they're where the 0:03:44.560 --> 0:03:47.000 kind of most strain has been felt in terms of 0:03:47.280 --> 0:03:51.360 bond yields going up, so that briefly sent bond yields 0:03:51.400 --> 0:03:55.280 back down. But of course, and I mean Scott Wessont 0:03:55.320 --> 0:03:57.000 used to be a hedge one manager, I'm sure he's 0:03:57.040 --> 0:03:59.520 aware of this. As soon as you draw a line 0:03:59.560 --> 0:04:02.840 in the sad and for the markets saying this is 0:04:02.880 --> 0:04:05.400 the thing that the government doesn't like and if you 0:04:05.440 --> 0:04:09.080 get to this number, we will intervene, they tend to 0:04:09.080 --> 0:04:14.280 push it. So today the youth has already bounced back 0:04:14.320 --> 0:04:16.840 to where it was before they said, right, we're going 0:04:16.880 --> 0:04:20.600 to do this intervention. So it's basically just gonna be 0:04:20.640 --> 0:04:22.080 interested to see what happens next. 0:04:22.520 --> 0:04:25.680 Yeah, I suppose, I think says that this doesn't fail 0:04:25.839 --> 0:04:27.840 like a Treasury sacruit job. 0:04:27.920 --> 0:04:31.159 This feels like a fed job. So this is the 0:04:31.200 --> 0:04:33.200 beginnings of what they call physical. 0:04:32.839 --> 0:04:36.599 Dominant right, Oh oh, yeah, the idea. I mean, no, 0:04:36.640 --> 0:04:41.360 you're absolutely right. I mean I don't think it's particularly new. 0:04:41.520 --> 0:04:46.400 I think obviously kind of the fiscal side of things 0:04:46.440 --> 0:04:50.400 has been intervening more and more and nudging central banks 0:04:50.400 --> 0:04:52.960 and the monetary policy element more and more, but it 0:04:53.000 --> 0:04:55.280 is getting more avert, and I do think it would 0:04:55.279 --> 0:04:58.280 be fair to describe this, and I mean George Saravelos 0:04:58.320 --> 0:05:01.360 over at Deutsche Bank did this is a soft form 0:05:01.360 --> 0:05:06.240 of financial repression because the desire is to cap yields 0:05:06.240 --> 0:05:09.160 in some way, and assuming that he does, my is 0:05:09.240 --> 0:05:12.360 to cap yields. The place where the strain will be 0:05:12.440 --> 0:05:16.960 felt is the foreign exchange market. So it's it's all 0:05:17.040 --> 0:05:20.920 else being equal. If he actually manages to put some 0:05:20.960 --> 0:05:23.560 sort of cap on long term yields, the dollar is 0:05:23.560 --> 0:05:26.479 probably where we would feel it. The dollar would go down, 0:05:27.160 --> 0:05:29.960 but I don't really see. I don't think he particularly 0:05:30.000 --> 0:05:31.560 would care about that at this point because if you 0:05:31.600 --> 0:05:34.240 look at where the dollar is, it's not ice particalarly weak, 0:05:34.800 --> 0:05:37.960 it's not mega strong, but you know it was basically 0:05:38.520 --> 0:05:41.440 the kind of dollar index, which is the dollar against 0:05:41.440 --> 0:05:46.440 all its partner currencies, is still a lot higher than 0:05:46.440 --> 0:05:48.839 it was for basically the whole of the two thousands. 0:05:49.920 --> 0:05:51.919 So I would think that if the trade off was 0:05:52.080 --> 0:05:56.680 a weeker dollar for interest rates looking better, I think 0:05:56.720 --> 0:05:59.120 it's being better because this will all feed into the 0:05:59.240 --> 0:06:02.440 US usage market as well. Then that's a trade off 0:06:02.440 --> 0:06:04.840 he would take, but I think he's going to have 0:06:04.920 --> 0:06:07.640 to do more, because otherwise the market is just going 0:06:07.720 --> 0:06:10.640 to say, well, you know, they'll push them to the 0:06:10.720 --> 0:06:13.680 yield that it feels uncomfortable with and then say, okay, 0:06:13.680 --> 0:06:15.320 what are you going to do now? You know, we 0:06:15.640 --> 0:06:17.479 need to see you pull out some big guns here 0:06:17.720 --> 0:06:20.320 is going to take more than just saying oh, we'll 0:06:20.360 --> 0:06:23.960 buy a few extra long entreasure needs to get rid 0:06:24.000 --> 0:06:25.720 of it, I think, hmm. 0:06:25.839 --> 0:06:28.120 And of course the line problem here is the thing 0:06:28.160 --> 0:06:29.880 that you and I now have been talking about for 0:06:30.080 --> 0:06:35.120 going on twenty years, which is the ongoing rises in 0:06:35.240 --> 0:06:38.440 government debt around the world that just keep going and 0:06:38.480 --> 0:06:41.240 going and going and going, and everyone feels like we're 0:06:41.240 --> 0:06:43.320 in a bit of an inflation re environment, and you 0:06:43.440 --> 0:06:46.599 keep seeing debt rising more than GDP over an avernomenal 0:06:46.680 --> 0:06:49.360 GDP in different countries, and't it just got to come 0:06:49.400 --> 0:06:52.280 a point when something breaks somewhere. 0:06:53.560 --> 0:06:58.720 I mean, that's what you would think. I only presumably 0:06:58.720 --> 0:07:01.479 that's what Scott basin as well, and that's why he's 0:07:01.480 --> 0:07:04.599 hoping to hold this down. And I mean I do 0:07:04.680 --> 0:07:05.160 think the. 0:07:05.360 --> 0:07:07.200 I mean US government debt, the number this week is 0:07:07.240 --> 0:07:11.200 going through forty trillion dollars, right, yeah, so much money. 0:07:11.640 --> 0:07:14.160 Yeah, And I mean it's also interesting that there's no 0:07:15.000 --> 0:07:19.239 there's not really like one difference between now and most 0:07:19.240 --> 0:07:21.320 of the time that we've ever been talking about this 0:07:21.880 --> 0:07:27.040 is there's usually at least one group of slightly obsessive 0:07:27.280 --> 0:07:31.040 people who are in politics and really want to see 0:07:31.160 --> 0:07:33.600 the debt and the deficit come down. When that was 0:07:33.640 --> 0:07:38.040 the case prior to the Great Financial Crisis, certainly the 0:07:38.040 --> 0:07:41.720 case after it. But now there is no political constituency 0:07:42.120 --> 0:07:50.640 for cutting spending really, certainly not in America, you know, 0:07:50.880 --> 0:07:53.480 and even here in the UK it's very hard to find. 0:07:53.880 --> 0:07:57.400 It doesn't matter, you know, where you sit on the spectrum, 0:07:58.280 --> 0:08:00.560 from you know, the furthest left to the father a 0:08:00.560 --> 0:08:02.559 straight everybody wants to spend more money. 0:08:03.920 --> 0:08:07.760 So I think absolutely nobody out there saying to get 0:08:07.800 --> 0:08:11.080 off this road to future missure, we have to reduce 0:08:11.120 --> 0:08:14.720 the size of government relative to GDP and we have 0:08:14.800 --> 0:08:17.320 to see some massive rise in GDP. And I saw 0:08:17.360 --> 0:08:20.840 from macro Strategy Partnership of Noteubt this morning saying the 0:08:20.920 --> 0:08:24.160 US is getting thirty four percent of nominal GDP growth 0:08:24.200 --> 0:08:27.120 per dollar of additional debt China is getting eighteen point 0:08:27.200 --> 0:08:31.600 eight cents. Yeah, I mean, you're just borrowing from the 0:08:31.600 --> 0:08:36.240 future in a completely absurd and unsustainable way. So if 0:08:36.240 --> 0:08:40.280 I with the bond marketsily pushing it yelled to regardless 0:08:40.280 --> 0:08:41.760 of all the short term stuff. 0:08:42.040 --> 0:08:44.480 And I mean, this may be what we need to see, 0:08:44.520 --> 0:08:49.560 but I think again the resk here is that well, 0:08:50.600 --> 0:08:52.600 the US, for example, is able to get away with 0:08:52.679 --> 0:08:56.079 us for death and that and other countries can get 0:08:56.080 --> 0:08:59.480 away with it to an extent. The issue will be 0:08:59.600 --> 0:09:02.960 the presumably if they MCEE take cap boind yields in 0:09:03.000 --> 0:09:07.920 some way, then inflation will make itself felt more strongly 0:09:07.960 --> 0:09:11.040 because obviously if you get a weaker currency, that will 0:09:11.040 --> 0:09:14.840 translate into imported inflation or you know, more kind of 0:09:14.880 --> 0:09:19.600 costly goods coming in. But yeah, I mean, I guess 0:09:19.600 --> 0:09:22.120 the fight can go on for quite a bit longer, 0:09:22.320 --> 0:09:28.640 assuming that that this is that they take the steps 0:09:28.679 --> 0:09:31.760 that we think they would take to suppress bond yields. 0:09:49.600 --> 0:09:52.400 Charlie Morris, our old podcast friend Charlie Morris, broke today 0:09:52.440 --> 0:09:55.120 that gold smells Q. 0:09:57.360 --> 0:09:58.439 And we see we did. 0:09:58.280 --> 0:10:00.360 See a big jump in the gold briad yesterday four 0:10:00.360 --> 0:10:04.520 per centers, and obviously in an environment like this, the 0:10:04.559 --> 0:10:06.320 gold price probably is going to rise. 0:10:06.720 --> 0:10:09.760 Yeah, I mean I I see no reason to be 0:10:09.960 --> 0:10:12.880 perish on gold right now. And you know the other 0:10:12.960 --> 0:10:15.559 thing that caught my eye, which I mean and also 0:10:15.600 --> 0:10:18.080 caught Charlie's eye because he likes his crypto as well. 0:10:18.760 --> 0:10:23.160 Bitcoin really took off in a way that it hasn't 0:10:23.400 --> 0:10:26.439 for a while. And actually bitcoin gold have not been 0:10:26.480 --> 0:10:31.199 doing the same thing for some time. But I definitely 0:10:31.200 --> 0:10:34.480 think this is because people suddenly thought, oh, look, this 0:10:34.559 --> 0:10:37.679 is it. The debasement trade is absolutely back. This is 0:10:37.720 --> 0:10:41.400 all about the weaker US dollar. Buy all the things 0:10:41.440 --> 0:10:43.760 that you buy if you think the dollar is going 0:10:43.840 --> 0:10:48.920 to slide. And I think that's again, I don't think 0:10:48.960 --> 0:10:50.680 that's an unreasonable conclusion. 0:10:51.720 --> 0:10:53.880 I asked a tech guy yesterday on one of the 0:10:53.920 --> 0:10:56.720 podcasts coming out next week, actually Topless and do listen 0:10:56.760 --> 0:10:59.440 on quantum computing and all sorts of fabulous things like that. 0:11:00.400 --> 0:11:01.079 I asked him. 0:11:01.080 --> 0:11:02.840 I haven't asked at the end of our podcast for 0:11:02.880 --> 0:11:05.640 a while, bitcoin or gold, but given that we talk 0:11:05.720 --> 0:11:08.840 so much about about quantum computing in this podcast, I. 0:11:08.880 --> 0:11:11.160 Thought I would ask him bitcoin or gold. 0:11:11.679 --> 0:11:14.760 You know, Dick tech guy, right, And I wasn't entirely 0:11:14.760 --> 0:11:16.640 sure which way he would go, but he was very 0:11:16.880 --> 0:11:17.439 very clear. 0:11:17.600 --> 0:11:23.400 Gold gold a quantum guy M destroying. 0:11:24.440 --> 0:11:28.360 Yeah, QD coming up, and then you know, can anyone 0:11:28.400 --> 0:11:30.080 be bothered to go around the place where you know, 0:11:30.080 --> 0:11:32.080 once you've got a quantum computer, you can probably use 0:11:32.120 --> 0:11:34.040 it for more exciting things than going around the place 0:11:34.120 --> 0:11:36.480 breaking into people's wallets and stealing their bitcoin. You know, 0:11:36.559 --> 0:11:38.199 probably not going to get it's not going to happen. 0:11:38.280 --> 0:11:40.120 It's not going to happen in ourria. It's way down 0:11:40.160 --> 0:11:42.000 the list of what happens when people gets their hands 0:11:42.040 --> 0:11:44.200 on a nice quantum computer. To the extent that they 0:11:44.240 --> 0:11:47.160 will in the short to medium term, but nonetheless the 0:11:47.160 --> 0:11:50.840 fact that they could is surely enough to make it 0:11:51.720 --> 0:11:52.720 to make it worthless. 0:11:53.120 --> 0:11:55.240 Yeah, I mean to be for Chris, it's Chris Woods 0:11:55.280 --> 0:11:58.120 all at the c l S that excess c l S. 0:11:58.320 --> 0:12:02.400 I think he was a bitcoin but quite recently sold 0:12:02.400 --> 0:12:04.119 out entirely, didn't it because. 0:12:04.080 --> 0:12:07.559 On that basis, Yeah. 0:12:06.840 --> 0:12:10.040 I mean that is interesting and I sually, I mean 0:12:10.040 --> 0:12:11.840 you're really I mean, at the end of the day, 0:12:11.840 --> 0:12:13.640 you don't need to wrap your head round that if 0:12:13.760 --> 0:12:15.880 even if bitcoin is digital gold, you may as well 0:12:15.880 --> 0:12:18.840 stick with the physical one that can't be disrupted by 0:12:19.200 --> 0:12:23.040 quantum ill Presumably gold could be screwed up if we 0:12:23.080 --> 0:12:27.559 get fusion energy. Yes, because you probably make gold in 0:12:27.679 --> 0:12:28.760 a fusion reactor. 0:12:29.520 --> 0:12:33.680 Yes, I mean alchemy is currently possible, just it takes 0:12:33.720 --> 0:12:37.640 so much energy that's not worth thinking about. Yeah, but 0:12:37.760 --> 0:12:45.640 one day anyway, moving on bonds gold help it, you know, 0:12:45.679 --> 0:12:49.240 the age of abundance the fusion power combined with the 0:12:49.320 --> 0:12:51.320 AI and quantum will bring. The mean is that we 0:12:51.320 --> 0:12:53.199 we won't need to worry about the boll gold buried 0:12:53.240 --> 0:12:55.120 under our house or not anymore. Everything's can be so 0:12:55.360 --> 0:12:56.640 mawfless nothing. 0:12:56.360 --> 0:12:56.880 To worry about. 0:12:57.120 --> 0:13:02.480 But on the matter of disruption, I wanted to talk 0:13:02.520 --> 0:13:08.000 briefly again about an academic report from Stanford University which 0:13:08.679 --> 0:13:13.600 someone at PAMA Liberian pointed us to. And we've talked 0:13:13.640 --> 0:13:18.040 about this before. We've talked about whether the buildout of 0:13:18.400 --> 0:13:22.320 data centers is the right way to go, whether the 0:13:22.360 --> 0:13:27.280 future of a future of AI really is all about 0:13:27.480 --> 0:13:31.040 large language models. Go back and listen to our podcast 0:13:31.120 --> 0:13:34.040 with doctor Jenny's Mareki anyone who wants them. Was the 0:13:34.120 --> 0:13:37.120 granular detail on this, but one of the things that 0:13:37.240 --> 0:13:40.960 he said was, Look, you don't have to use LMS 0:13:40.960 --> 0:13:44.400 anymore to get the same efficacy. 0:13:43.800 --> 0:13:44.600 From your AI. 0:13:44.720 --> 0:13:48.840 You can use an SLM, a small language model which 0:13:48.840 --> 0:13:50.960 you can run on your laptop or on your phone, 0:13:51.040 --> 0:13:53.800 and you get pretty much the same performance. And not 0:13:53.880 --> 0:13:56.319 only that, but you hang on to your own data. 0:13:56.360 --> 0:14:00.520 You maintain your privacy by using an SLM over an LLLM. Cheaper, 0:14:00.840 --> 0:14:07.679 uses pulerless energy and is contained. So the paper on 0:14:07.760 --> 0:14:14.120 this looks at how effective and how accurate SLMs are 0:14:14.320 --> 0:14:17.240 against lms, and I'll put the link to it in 0:14:17.400 --> 0:14:19.080 the show notes, and I'm going to write about it 0:14:19.080 --> 0:14:21.720 in the Merintalks Money and Use letter this this. 0:14:21.600 --> 0:14:22.240 Week as well. 0:14:22.240 --> 0:14:26.880 But the basic point here is that if you look 0:14:26.920 --> 0:14:31.120 at all the tasks on reasoning tasks, SLMs are nearly 0:14:31.160 --> 0:14:33.760 as good as llms. They provide a better or at 0:14:33.840 --> 0:14:35.280 least as good an answer as LMS. 0:14:35.320 --> 0:14:35.880 And while it was. 0:14:35.880 --> 0:14:40.120 Sixty percent of the cases, look at other task, real 0:14:40.160 --> 0:14:44.960 life tasks, this mixture of chat and reasoning performance, and 0:14:45.080 --> 0:14:49.360 on average SLMs are as good, if not better than 0:14:49.640 --> 0:14:53.080 lllms in eighty one point two percent of the cases, 0:14:53.360 --> 0:14:57.160 with lllms having significant advantae. I'm bringing aloud here a 0:14:57.160 --> 0:14:59.960 significant advance only an area like I think you mean, 0:15:00.160 --> 0:15:02.480 such as, let's stick with a grammar correctly, guys, that's 0:15:02.520 --> 0:15:06.080 what AI is for for your grammar. Significant run only 0:15:06.120 --> 0:15:09.520 in areas such as engineering, life science, as transport or transportation, 0:15:09.640 --> 0:15:13.840 and computer science. It's not just accuracy. SLMs achieved this 0:15:13.960 --> 0:15:18.080 performance and energy and compute costs that are between fifty 0:15:18.120 --> 0:15:21.880 and eighty five percent lower than an LLM. 0:15:22.720 --> 0:15:27.680 So at slm's like the SA engine versions. Basically like 0:15:27.720 --> 0:15:30.040 if you're just googling or whatever, use that if you're 0:15:30.080 --> 0:15:34.360 trying to cure kinds of us and ALLM. 0:15:33.080 --> 0:15:35.080 Well, I mean maybe maybe not. 0:15:35.960 --> 0:15:38.920 By October twenty five, the SLMs achieved ninety nine percent 0:15:38.960 --> 0:15:41.280 success for the easiest reasoning task in level one and 0:15:41.360 --> 0:15:44.200 two eighty five to ninety two percent in harder tasks, 0:15:44.320 --> 0:15:46.880 and only lagged in the hardest tasks of all that 0:15:46.960 --> 0:15:50.000 b aqre concept where there's a fifty one point five 0:15:50.040 --> 0:15:51.040 percent success right, So. 0:15:51.000 --> 0:15:51.840 They're exactly right. 0:15:51.880 --> 0:15:55.520 There's are only a very small number of very complex 0:15:55.640 --> 0:15:57.440 and vital tasks that you might want to use in 0:15:57.640 --> 0:15:59.840 LLLM for and the rest perhaps if what you're just 0:16:00.200 --> 0:16:02.000 to do is hang onto your data and hang onto 0:16:02.000 --> 0:16:04.360 your privacy and do things at a low cost, you 0:16:04.400 --> 0:16:05.760 would use an slum. 0:16:06.040 --> 0:16:07.560 And if that is the case. 0:16:07.920 --> 0:16:10.600 If that is the case, and again neither you or 0:16:10.680 --> 0:16:12.600 I are experts on this stuff. 0:16:13.320 --> 0:16:14.680 If that is the case, and. 0:16:14.720 --> 0:16:19.200 The the head that the title of this a little 0:16:19.200 --> 0:16:21.800 bit of research is if this is true, the hyperscalers 0:16:21.880 --> 0:16:23.040 are toast. 0:16:25.520 --> 0:16:29.280 I mean link in the show notes two. 0:16:31.360 --> 0:16:33.680 The thing is, I agree, I mean, I guess My 0:16:33.760 --> 0:16:36.280 problem is I'm trying to work why I'm wrong, if 0:16:36.320 --> 0:16:38.040 you see what I mean, or why this is wrong, 0:16:38.240 --> 0:16:42.160 because on the one hand, you know, we'll alvisos get 0:16:42.160 --> 0:16:46.720 a natural skepticism lean against it in these cases because 0:16:46.760 --> 0:16:48.560 clearly EI is a big deal. 0:16:51.280 --> 0:16:52.160 It's not a big deal. 0:16:52.240 --> 0:16:54.080 It's just to say it doesn't have to be done 0:16:54.200 --> 0:16:55.400 in this expensive way. 0:16:55.760 --> 0:16:57.800 Well, this is it. And also the even if you 0:16:57.960 --> 0:17:01.040 discount the s l A M L l M thing, 0:17:01.840 --> 0:17:06.240 I mean, the Chinese models are just as good, or 0:17:06.240 --> 0:17:08.760 certainly seem to be nearly just as good as the 0:17:08.800 --> 0:17:12.080 most divined ones in the US, but they are not 0:17:12.480 --> 0:17:16.400 spending the same amount of money or in the infrastructure. 0:17:17.200 --> 0:17:22.160 And I do yeah, I would need some I would 0:17:22.240 --> 0:17:24.040 like somebody to kind of make the kind of steal 0:17:24.160 --> 0:17:28.040 mind case for why it's a good idea to be 0:17:28.119 --> 0:17:31.359 building all of these data centers and what they're going 0:17:31.400 --> 0:17:34.359 to be used for if this all turns out to 0:17:34.520 --> 0:17:37.920 be essentially unnecessary. 0:17:38.040 --> 0:17:41.439 Or unnecessary in this volume, it's a little bit. It's 0:17:41.480 --> 0:17:46.960 a little bit like it's like wind power. Yeah, and 0:17:47.000 --> 0:17:50.560 then on a really windy day in the UK, for 0:17:50.760 --> 0:17:56.040 short periods, we can support the entire electricity network from wind. 0:17:56.240 --> 0:17:57.280 So why are we building more? 0:17:58.080 --> 0:18:02.920 Yeah, yeah, it's I mean, I guess I suppose. The 0:18:02.960 --> 0:18:06.119 other thing with this is that all points to the 0:18:06.280 --> 0:18:09.480 data centers being in a funny kind of way that 0:18:09.680 --> 0:18:14.159 kind of like peacock tails. It's like, you know, I mean, 0:18:14.160 --> 0:18:17.920 you get these whatever kind of like for hyperscalers kind 0:18:17.920 --> 0:18:22.040 of like building these massive ornaments because they can and 0:18:22.040 --> 0:18:26.200 it's almost like the last one standing is the winner 0:18:27.000 --> 0:18:31.840 because they kind of out spent each other. I don't know, it's. 0:18:32.760 --> 0:18:39.080 Your animal metaphors are getting kind of kind of interesting. 0:18:39.320 --> 0:18:40.320 I don't know if you. 0:18:40.320 --> 0:18:46.359 Remember that squirrel squirrels standards, And here we are with 0:18:46.520 --> 0:18:52.080 data centers as peacocks, data center data center infrastructure builders 0:18:52.119 --> 0:18:53.159 as peacocks. 0:18:54.880 --> 0:18:59.399 Look at my mighty edifice is of no value, but 0:18:59.480 --> 0:18:59.960 can be some. 0:19:01.680 --> 0:19:02.120 Exactly. 0:19:02.200 --> 0:19:07.359 So let's hold that in your head, this whole SLM 0:19:07.920 --> 0:19:11.040 LLM Chinese LLM dynamic. 0:19:11.720 --> 0:19:12.560 And then let's go. 0:19:12.600 --> 0:19:14.760 Back to what we were talking about at the beginning, 0:19:14.800 --> 0:19:17.440 when we were talking about the bond market. I mean, 0:19:17.680 --> 0:19:22.040 almost always you think, you know, yields up, equity market down. 0:19:22.800 --> 0:19:26.639 Yeah, but that isn't really happening at the moment. 0:19:26.800 --> 0:19:29.199 And if you read all the endless broker reports that 0:19:29.280 --> 0:19:30.800 you and I both look at all the time, what 0:19:30.840 --> 0:19:33.440 they do is they tell us that that dynamic isn't 0:19:33.480 --> 0:19:36.240 working at the moment, or isn't working as it normally 0:19:36.280 --> 0:19:44.440 does because AI because it's simply not sensitive to yield 0:19:44.680 --> 0:19:47.960 in the same way as historically major parts of the 0:19:48.000 --> 0:19:50.960 market has been. Because the demand is so great, the 0:19:51.160 --> 0:19:54.439 need is so strong, the revenues are so high, and 0:19:54.480 --> 0:19:57.720 the future profits are so ginormous that a couple of 0:19:57.760 --> 0:19:59.920 percentage points here or there on the ten and third 0:20:00.040 --> 0:20:01.200 to you yield Yeah, who. 0:20:01.119 --> 0:20:05.440 Case, Yeah, that's not very convincing. 0:20:06.560 --> 0:20:09.280 Well, it's not, and it's particularly not if it's if 0:20:09.520 --> 0:20:12.840 if there is any truth at all in the idea 0:20:12.880 --> 0:20:15.720 that data centers are I would beat toast in the 0:20:16.000 --> 0:20:17.560 in the face of. 0:20:17.840 --> 0:20:21.560 S l MS. Yeah, I mean it's going to be 0:20:21.640 --> 0:20:23.800 it's gonna be fascinating to watch how it falls out. 0:20:23.840 --> 0:20:27.199 And I suppose the wine the wine sort of like 0:20:28.480 --> 0:20:31.520 it's a potential positive in a kind of negative way 0:20:31.560 --> 0:20:34.280 for the sovereign there, or at least for US sovereign. 0:20:34.400 --> 0:20:37.919 There is that if everyone is in a rush to 0:20:37.960 --> 0:20:41.800 go Ei boinds at some point, then that you would 0:20:41.880 --> 0:20:45.080 sort of expect lump me dash for the safety of 0:20:45.200 --> 0:20:51.000 US treasuries if it got that bad. I don't know, 0:20:52.240 --> 0:20:55.600 is this is one of those words. I just think 0:20:55.640 --> 0:20:57.359 it's gonna be fascinating to watch because there's all this 0:20:57.480 --> 0:21:00.320 both balance sheet stuff that they've got as well stuff. 0:21:00.400 --> 0:21:01.400 Yeah. 0:21:01.520 --> 0:21:04.040 Yeah, I think John that we should probably end it 0:21:04.080 --> 0:21:07.080 there because we don't know the answers, but we can 0:21:07.160 --> 0:21:09.800 just say to our listeners, watch out for peacocks. 0:21:09.960 --> 0:21:12.760 I'd say, one tiny thing. This is another reason why 0:21:12.840 --> 0:21:16.399 you should watch what's in your pension Again. If you 0:21:16.440 --> 0:21:20.320 don't know what's in your auto enrollment pension, go and 0:21:20.359 --> 0:21:23.679 find out. Because there is more and more pressure to 0:21:23.760 --> 0:21:29.480 stick more and more money into private credit, private assets, 0:21:29.600 --> 0:21:31.640 all of the sort of stuff that will end up 0:21:31.640 --> 0:21:37.160 being exposed to the worst of the excesses of this 0:21:37.200 --> 0:21:41.680 particular build out, I suspect, and the pressure to do 0:21:41.720 --> 0:21:44.760 that is going to keep mounting the more of this 0:21:44.800 --> 0:21:48.199 stuff that ends up spreading into the wider financial system. 0:21:48.840 --> 0:21:51.639 So yeah, just know where your money is and make 0:21:51.680 --> 0:21:53.720 sure it's not anywhere you don't want it to be. 0:21:54.720 --> 0:21:57.280 Another little personal finance heads up on that one, by 0:21:57.280 --> 0:21:59.080 the way, I was listing earlier today. 0:21:59.320 --> 0:22:02.080 Let's talk about it. This is a UK specific. 0:22:01.560 --> 0:22:05.080 Point, but uh, you know, you get these whopping great 0:22:05.080 --> 0:22:08.920 inheritance tax bills and you can, occasionally, for very very 0:22:08.960 --> 0:22:11.600 non liquid acids, spread those bills out over a long 0:22:11.640 --> 0:22:13.879 time if you have you know, maybe if the the 0:22:14.000 --> 0:22:16.399 acid in question is land or whatever, you can arrange 0:22:16.440 --> 0:22:20.280 to pay the bill in bits. But I don't think 0:22:20.320 --> 0:22:23.399