Speedrunning dead internet theory - an SMC listener weighs in.

Welcome to the Sanso Market Call. Before we get started, it's important we emphasize that the views and opinions expressed in this podcast are those of the hosts and guests and do not constitute investment advice or recommendations. All information provided is for educational and entertainment purposes only. Hey everyone, it is Thursday, 13th of August, 2026 and we have a pretty tepid bit of market action yesterday, mostly on the positive side if we look across the major indices in the US. Oils down a bit, the boil coming off there, even though there's no real improvement in the headlines coming from the Middle East. We had a US CPI that was in line with expectations, keeping the worries about the FMC and that next rate hike a bit on the back burner, in fact, the front end of the yield curve in the US falling a bit and the whole yield curve actually fell in reaction to this totally inline data, which does show that deceleration in the core year and year inflation. However, the long run sort of popped back up towards the highs and to mostly unchanged on the day, 10 and 30 year yields. There was a 10 year auction yesterday, not generating headlines, slightly softer demand there than the prior two auctions. We have a 30 year today, maybe only interesting, I would say, if the bidding metrics are out of whack, let's say people are reluctant to put their money to work at these levels or they're saying, hey, these are the highest levels in yields, 30 year yields since 2007, it's time to bid up on duration here. So an important option mainly because of where we are at these highs of the range. I want to say here, at the top of the podcast, that I'm going to share with you a very long email from a Mr. O, you know who you are, Mr. O. As you say, you're a regular listener, so you'll probably be listing this one on a read pretty much your entire email verbatim. Super interesting response to the discussion yesterday about the degrading quality of search and what comes next could essentially start killing off the internet or killing off the quality of search as AI takes over. This is from somebody clearly with experience at the coalface of IT operations. And thank you, Mr. O, for sending this. It really I think is going to enrich the perspectives of the listeners on this podcast as well as my own. So that'll be super interesting, just to tease that. All right. So let's just do a brief wrap on the market action yesterday and ask that 100 up three quarters of a percent, S and P, only a quarter of a percent and the S and P equal weight up a bit less than that. So the broader market just is very slightly positive on the day. Russell a bit more so the small caps up over half a percent and semiconductors up to and a half percent likely on some of those positive earnings that we saw that battery saw on the prior days action after the close. There was the whole vibe of some of the software as a service names coming off a bit. So we're still playing that little tune, it seems in markets. If we look over in Europe, the stocks 600 has been pretty darn sideways over the last week. There's tons of headlines about the Ryan is becoming unnavigable, the Ryan river and how much of Swiss GDP even is is relying on barch traffic coming up the Ryan and even German growth could be impacted quite heavily by this. I did look at the good old tropical tidbits dot com for the rain forecast and it looks like the accumulated rain is going to be pretty massive over the coming couple of weeks. So some relief might be coming on that front from Europe. Nikkei has been rebounding and rebounding from the very, very ugly sell off that it was seems very correlated with the general sell off and semi semis and for example NASDAQ 100 which has yet to reach those all time highs by the way again the 30,000 levels a bit of a sticking point. It seems on the NASDAQ 100 but back to the Nikkei it did reach a new high overnight but came off those high levels all right over to the macro picture which is not terribly interesting. I pretty much gave you the the big read there with the US Treasury yield curve steepening. That's not helpful for the Japanese yen bulls. The end continued to dribble to around the weaker side of the recent range here and that was despite a Bloomberg sourced with sources unnamed apparently close to Prime Minister Takaiichi in Japan that she is green lighting basically an imminent bank of Japan rate hike. There's been a word that there's pressure from the US Treasury secretary as a as a quick pro quo on this coordinated intervention that the bank of Japan needs to do. It's part of the heavy lifting even though the perhaps the Takaiji government doesn't want that growth impact from a more hawker should be a J so zero reaction long story short to this a news item from Bloomberg neither in the yen nor in bank of Japan expectations as we're still around 75% if I recall 80% maybe of the September bank of Japan meeting producing a rate hike across the FX space a bit of a carry trade vibe with those with all of this a reasonable risk on across markets and we have yours was actually posting marginal new highs for the year Swedish Krona lower low yielding at 1.75% and with a very, very tight inflation what are we at 0.6% year on year at the core for Sweden's inflation levels. That was out this morning but yesterday the Swedish Krona weakened sharply I think some of that technical as this 11 area in your stocky was was blasted through the parity level and knocky stocky so the Norwegian Krona versus the Swedish Krona also giving way some of that is backtracking this morning and we're seeing the Norwegian Krona a bit weaker as well the Norris bank was out this morning saying it was wishy washy way that the rate hike may still become necessary it's already got a very high policy rate about 4%. So it's just about showing up credibility around inflation really and those pricings are for a potential rate hike maybe 33% odds in September two thirds odds in November mostly priced for December for the Norris bank meetings I guess the it was a bit wishy year washy or relative to the recent price action I don't know also well coming down is a slight nook negative on the on the US event risk front we do have retail sales up tomorrow from the US but I think it really is this Jackson Hole conference symposium as they call it 27 through the 29th of of this month so two weeks basically it starts that is the next best chance for the dollar to do something to get a signal from the Fed there's all the talk about the insanely unsustainable rate of of the growth of the deficit in the US let's recall though the treasury issuance in terms of the coupons it has not been growing even though some some the banks involved in auctions and so on are wondering why not or saying it should it's all being absorbed in bill issuance so kind of a curious setup there you know how does the system digest all these exploding supply of bills essentially as we've just crossed the psychological 40 trillion in national debt in the US so yeah that's that's about it for the macro picture there I noticed across metals we've got gold coming off a bit not helpful for silver either that copper prices have rolled over a bit after testing those all time highs at least for the front the front futures contract in the US so awaiting for technical signals there and whether we get a follow up move higher and gold after it's exploded out of that prior range couple of and I forgot on the equity wrap to go through a couple of single stock stories I'll do so now a nebius was 34% after their earnings report yesterday feels like it may be a short squeeze vibe to a degree with that kind of reaction but in any case for whatever reason the stock was that much higher seemed to be very confident in their forecast for growth and talking about adding a gigawatt per year of capacity that seems to be the unit even for a company and I'll get to the minutes like Sarah Brostes it's making basically chips talking about the numbers of hundreds of megawatts and gigawatts it's going to be delivering into so that was that story but interestingly Michael Berry of the big short fame has been out I can't remember the exact quote something along the lines of I'm glad it's gone higher that means it's a much better level to short it from so color Michael Berry skeptical on the outlook for nebius by the way just because I've on the subject he was also out with some comments on Palantir saying it's worth under a dollar that sounds a bit excessive to me on the bearishness what is it trading 140 something like that now but he did say that there are some key things we need to consider when looking at Palantir and that he's bearish the article that I'll send a link to in the podcast episode description talks about him buying March 2027 puts in the low to mid 100 range whatever the heck that means I don't know if that means 105 or 110 it says that the market isn't valuing this company correctly because it's not pricing in things like 31 million dollar 31 million shares of a stock based compensation to staff that's a that's a chunk of change right there and off balance sheet commitments is another one infrastructure purchase obligations that are set to triple it's at least what I've noticed noted here from that article and so color him skeptical on both Palantir and nebius cerebrus the The maker of these wildly large, basically the entire, but I can print on a single wafer chip or occupying the entire wafer. It was down 17,17% after hours yesterday and after reporting. The revenue growth was not quite living up to expectations, but you see some relatively positive looking headlines here and there in this earnings report like tripling its revenue in 2027. That's what the company said. We'll do a better hope so. I mean, this company is incredibly richly valued, incredibly pricing in an incredible amount of growth going forward. And it did see a near quadrupling of its so-called fast inference cloud business in Q2. Cisco was down around 4%, after its earnings report, AI outlook, apparently disappointing, and coherent. I didn't see any headlines. I thought that it was up around 8% or more during the session yesterday before reporting earnings. A lot of that in sympathy with Luminthum's very strong report, but it then fell around 3.5% after the close. Let's see. Another, I'll send a link to, I'm a bit reluctant to do so. I find this guy a bit annoying, but Bill Ackman out with the report is getting a lot of coverage. Some of his shifts in his Pershings holdings. We might find it interesting reading. Some of this is going to be stuff he's already done and things have already moved and who knows the guy could have sold it and moved on. That's what we don't know about these reports when they come after the fact. But a couple of interesting names he's added are things like S&P Global saying the market is overpricing AI disruption, similar for Visa and MasterCards saying that those companies could actually gain from stablecoin adoption, not find that stablecoin or stablecoin infrastructure, whatever, for payments will be in any way disruptive. QSR was another name in there. The restaurant brand that has Burger King under its umbrella, Burger King apparently out with some solid growth numbers relative to what's going on McDonald's, et cetera. Netflix was another interesting name if I didn't say that already. We have also in the links, a cool one to pluralistic. I love just putting that in there. It's a totally left-leaning guy, but the guy writes so well. He talks about sophisticated technical matters and technical subjects as well. In this case, he's talking about model collapse and how an AI model will try to, especially in predictions, let's say something in trying to predict, I think the example, whether it was him or the link he sends us to in this post, to this Lauren leak post, looks super interesting. I've only read a part of it talking about model predictions. Let's say it's an AI that's trying to predict what people would like to eat, it finds the mode. So in a mode in a distribution is the one, the dish that people pick the most. It starts to, as Lauren leak says, it starts to, the model starts to eat its own tail. The model prediction starts to eat its own tail, starts throwing out choices that most people don't make, but some people may really prefer. And everything just starts to take on a sameness across the board. Super interesting subject, a little bit along the lines with the subject yesterday, and very much a good segue and dovetailing with the comments from Mr. Oh, thank you very much. Mr. Oh, for sending in this long email, and I'm going to read through it now, and there's a link within this email that I will post in today's podcast description as well. So here we go. Listener, Mr. Oh, writes just some thoughts on your recent AI commentary from someone who has 20 plus years background in IT operations instead of being in finance. I love this because it's somebody dealing, you know, at the coalface of, and seeing, in real time, what is, what is happening with, with AI. It might be a different angle from what you usually hear, similar to what you said about the degrading quality of search, that was yesterday's podcast, by the way, parenthetically, AI critics have for a long time pointed out that we're simply speed running dead internet theory, and unfortunately, they've been proven more right with every passing week. So dead internet theory, I had to do a little search on that is basically the idea that there's just so much out there is becoming sort of auto-generated content and not really produced by real people, but it's just, you know, essentially fake or dead content. Back to the mail. The amount of bot traffic on pretty much any kind of internet exposes, internet exposes site or service is getting completely ridiculous. I'm sure if you've talked to your IT folks, they'll have quite a few stories of their own. In my personal experience, some of the worst offenders are open AI into the topic themselves who have bots that regularly act in a completely brain dead manner that would make even junior IT engineers back in the 90s when nobody knew what they were doing, blush in embarrassment. I've seen requests from these frontier labs on multiple customers systems trying to crawl everything access stuff. They really should not or even they really should not or even brute force credentials over and over and over again millions of times after being unmistakably denied to a point where you really can't tell anymore if it's sheer incompetence on their part. Their bots just going haywire like LLNs do every now and then or if it's in a malicious attack be it on purpose or by accident. So that's a pretty damning accusation there. If you haven't seen it yet, it's a bit technical but well presented and really worth of time. Security researchers from open AI a few days ago gave a talk about what happened behind the scenes with the hugging face incident and why it is so significant. It's on YouTube here and I will provide the link in the podcast episode description by the way with a short summary ad here. I imagine that at some point people will start asking about the AI labs quote polluting the internet unquote beyond recognition. How come everyone else has to pay for it be it an increased expenses for cyber security, storage, compute, bandwidth, etc. just to defend against this new threat at exponentially rising costs, especially now that the same hardware needed to cope with a flood of bots is becoming increasingly unaffordable thanks to the very same crowd that is the root cause of the bottlenecks and well almost everything really. They could very well be killing one of mankind's greatest achievements by polluting it beyond recognition. The whole internet is turning into the prime example of the tragedy of the comments I'm afraid. Not to mention the question, how come it's seemingly totally fine for AI labs to quote accidentally unquote do things that would be considered an open and shut criminal case for anyone else in most parts of the developed world. I don't even bother talking about civil issues anymore because concepts like IP or copyright seemingly nobody cares about anymore anyway. Now we're just doing crime because it's okay if a machine did it a little bit confused by this next sentence, but I'll just get through it. I could go on and on about the technical and legal nightmares that running agentic harnesses entails if people just thought about it for more than a few minutes or had any idea of what is really going on when they send off a prompt and watch quote, their agent unquote start quote, doing things, doing it's thing unquote, but I don't want to waste too much of your time. That would be a whole lengthy essay on it's own. Well, thank you very much for wasting my time, Mr. William, you're certainly not doing so. And I'm quote unquote wasting the time of our listeners too with this wonderful with your wonderful thoughts. Back to the email, don't get me wrong. I think those tools can do amazing things if done right. It's just that the complexity of what's going on is so off the charts that the chances of these last couple of years will turn into the next quote, what were they thinking unquote in our history books is quite significant. And that's just the engineering side, not even thinking of all the financial engineering I have much less understanding of, but it sure sounds like a similar mess the closer you look at it. Then again, social media still has not had its big tobacco moment either and might never have one because now we've just moved on to the next thing. So what do I know? And then he goes on to think, I think the Saxon market call. Thank you, Mr. Oh, I think some super worthy thoughts for everyone to think, sink their teeth into and think about. And it's just amazing the speed with which this technology has just been sort of, it's like unleashing the crack in the chaos, if you will, hand over the box. Some might say on the world and the lack of controls and speed, you know, speed limiters or whatever on how this all unfolds. Because in some cases, the regulators are always slow in terms relative to the march of technology. But then on the other hand, they think, well, AI is innovative, it's good. We must have it more quicker than the others can have it because it's a national security risk, but not thinking about the or all the basically unintended consequences that can come downstream of what is going on. So anyway, I thought I'd finish on that very good note. And yeah, we'll see where things are taking us in the days ahead. We are into the last couple of weeks, plus really of summer. And on occasion, August has proven to provide some interesting market pivots, even though it's normally supposed to be just a humdrum summer month. And September, October, classically, the times of year when or classically the time of year when volatility can pick up, at least it has in a couple of very notable instances. Let's see what's going on. You know, if you think the volatility is exceptionally low, you can express the view through options were implied volatility is at least an extra. are quite low in some cases. I'm going to bet you for course you understand what you're doing before doing any such thing. Alright, stay careful out there everybody. I'll be back tomorrow with the next Saxo Market Call. This has been the Saxo Market Call podcast. Thanks for joining today's episode. We're always happy for your feedback and questions of all kinds. To reach out you can drop us an email at [email protected]. That's [email protected]. Saxo, serious trading worldwide. (upbeat music)