The Dollar Is Toast (Why a Cyberattack Could Collapse the AI Trade)

0:00Welcome back to the Dengayla Cumbona Show, lots of ground to cover today. David Wu is here with me and he's already said the quiet part out loud. If the AI bubble burst, well, the dollar is toast. On Wednesday, the 10-year retreated on cooler inflation. Stocks like that. The question for David today is whether that dip and yield saves the AI trade or just delays the moment one of them has to give. David of David Wu and Bound, always good to be with you. Welcome back to the show. Thanks for having me. You're closing out the week here. Thoughts on all these moving news items and what you're paying attention to the most. I mean, I mentioned the AI trade. We'll talk that. But you're also closely tracking the price of oil here, David. Yeah, I think you know, listen, I was on the phone this week with lots of clients. Most of them are some of the biggest institutional, whatever macro hitters, if you like. And I would say the general consensus is that the only two things to pay attention to right now.

1:03This AI is oil. That's it. Like everything else is endogenous. Like, wow, you say, wow, what about the Fed? Well, you know, listen, what the Fed is going to do depends on what's going to happen to oil. It's going to happen to AI. So whatever is. So from now point of view, we can see here and talk about the dollar rates, whatever it is. But in the end, the two main drivers right now are AI and oil. And to some extent, they're related. But I think what generally exhaustion is. And then both both are actually, you know, in a way, the biggest themes of 2020, you know, six so far. So the really question is in the final stretch into the US midterm elections, you know, whether these things are, you know, how they're going to basically play out. I think that's sort of the big crucial question I'm focusing on the moment. Okay, so let's break it down into chunks here because there's a lot to digest. You say if AI bursts, the dollar is toast, but yields just eased. Does that buy the AI trade more time?

2:03Or is the dollar already living on borrowed demand? No, I think I think you know, if you actually look at the second quarter, for example, which has got the latest gone some payment data, what he showed is that there was a search in foreign buying of basically US equities. Okay, now I'm just telling you right now foreigners, I think people have the other way around. People think, oh, wow, higher rates is a problem for the AI bubble. The way I look at this, that's quite different. The only reason why what interest rates are going up is because of the AI bubble. And the reason is because, you know, people, I mean, you have a lot of people out there who are forecasting these extraordinary expected returns on the AI trade, the AI capics and so on. Well, you know, a couple of weeks ago Morgan Stanley goes out with a report forecasting a 25 to 50% return on capics in AI by 2030. Let me just put it this way, if you believe that number, then you literally would not be buying anything else other than, you know, basically AI stocks.

3:05If you believe in that, you wouldn't be investing any other countries up for the United States. If you believe in that, you would not be buying treasuries of 5.2%. Because if you think you're going to be able to Microsoft Amazon is going to be able to make 25 to 50% return on their capics in three years time, it's a totally different ballgame. So this is what we're talking about here. I mean, there is this right now, you know, there are a lot of, and I did a whole report explaining what that actually means. I mean, Morgan Stanley just put out this report. What does that actually mean? And what I basically, so what I did was that I actually, um, backed out, okay, the sort of the macro assumption, okay, behind their 25 to 50% forecast. And what I came to the conclusion is that for them to be right, SMB, vibrant companies will have to be able to cut the operating cost by at least 25% by 2030. And given most of the operating costs involved wage and salaries, yeah, if you believe that 20%

4:07Americans are going to get laid off, okay, in three years time, then yeah, then you could basically make the case that AI is going to generate that kind of returns. But other than that, I think it's very unlikely. So I think from now point of view, the point here is that right now, you know, you've got Trump who's pushing, you've Trump said this two weeks ago, it was probably the most important thing he said all year, which is that if we lose the AI war, then we've lost everything else. So for this administration, it's like AI or nothing. So from now point of view, he is putting the entire US federal government essentially behind the whole AI tray, AI bubble, AI, whatever you want to call it. And he's basically most powerful, countless at this point is Jensen Wong, who is the most, I mean, the CEO of the biggest company in the world. And that guy is willing to put whatever much money, you know, I don't have to tell you every time the AI tray running to a little issue, Jensen Wong comes here and throw another 50 billion here, 100 billion there just yesterday, he said he's going to buy back 100 billion dollars of stocks. This is after having said that he's

5:11going to put 10 billion dollars into the anthropic IPO. And what I'm saying is that this is what's really going on, which is that you've got, I mean, I think the valuation is totally crazy. But right now, you're up against the president of the United States, the most powerful in the world, who believes that there's nothing more important than this AI. And then, you know, and then you are up against basically Jensen Wong, you know, biggest company in the world. And they have every reason to keep the music going for the music of this party going fast, as long as possible. And that's what you up against. And you know, now, this is why like, sure sellers have been frustrated. And I've been frustrated, but either as it may, that is the reality that we are up against right now. But that's the sort of the, that's why rates are up here, by the way, because nobody wants to pay out now. You have heard economic guru Mitch Vaxler say time and time again on my show, we are a razor blade away from a crash of the entire financial system. Now it's your turn to ask the questions.

6:13Join us live on October 15th at 12 p.m. Eastern for an exclusive one hour Q&A with Mitch Vaxler. Bring your questions about property taxes, the bomb market, real estate, gold, silver, and the entire financial system. One hour live unfiltered. Your questions answered by Mitch Vaxler. Don't miss it. Register now with the QR code on the screen or the link below in the description. See you there. Okay, so let's play out that scenario, right? Because you just made the case for it. I mean, yeah, they're serious. They're even naming an AI SAR. You know, so basically if the AI bubble keeps growing and growing and growing, does that mean that the dollar is safe? I guess the question is what's safe? What's not? Sure. I think the way I think about this is that, you know, there are two major risks to the AI bubble at this point. Okay, it's not it's not about interest rates, not even

7:14about oil price. There are two major risks that even Trump cannot deal with and that's what I'm sort of like focus on. And when that happens and whenever that happens, that's when everything is going to turn around. Now, then you know, as we know, I mean, the biggest problem, okay, the biggest risk that I've been talking about, many people have been talking about for the last 12 months about the whole entire AI trade is the iris on the assumption that these AI companies are going to make a lot of money. In other words, that they're going to be able to monetize this technology. Let's put it this way. I have no problem with the technology. So I'm a huge user of AI. My problem is how anthropic and opening up money from it, when you've got the Chinese basically essentially the catching up rapidly, okay, with cheaper models and models that can deliver the same level performance. Now, that's the problem. My problem is about the profit margin of this business. And we already see that planning out in open AI, right? Because open AI, the consecutive quarterly revenue growth already slowed to 18% last quarter, which is not that amazing. What I'm saying to you is that the

8:17issue is China. Now, there's another problem related to China, which is the US. And Trump is increasing or that China is going to somehow reach, you know, you know, artificial general intelligence before the US. Okay, that's why he came out and said we lose the AI war is all over for us. Okay, and this is also the reason why he decided not to regulate. He decided he doesn't want to take the risk of slowing AI development because he sees the risk of allowing China to essentially surpass the US as being much greater. Now, I think this is where it gets very interesting. So there are two things that could happen from here because the Chinese, you know, basically issue is not going away. And therefore the question now is what happens next? The first major risk therefore comes out of the fact that since the US decided not to regulate AI because they feel like they cannot afford to regulate AI because the Chinese are going to get ahead. I can tell you most likely what Trump basically offered Xi Jinping was to slow it down on both sides. The Chinese didn't

9:20believe him. So this is the reason why they probably go told him forget it. That was not going to happen. That's why there was no deal. In my humble opinion, what this basically means is very simple. Number one risk is that, you know, what you're going to see a cyber attack. You're going to see any kind of attack whether it is by a malicious actor or by rogue agents like we saw in Huggingface is going to be all over for the AI trick. I mean, Jason Huang pretty much told us in so many words last week in interview. He's rightly said that if there was an attack, okay, we're not talking about a model that's under testing that they release a new model opening in traffic and attacks something big. Guess what? The legal liability's loan will be hundreds of billions. This is why a third of anthropics perspective is devoted to warning of human extinction risk. And let me just tell you this. This is why an act in I don't know people even wear this last week for me. One of the most interesting piece of news that caught my ear was the fact that a group of hackers said that they

10:24using AI presumably they found a vulnerability in the FBI storage system and they downloaded all the personal information of 35,000 FBI officers. If you don't think that is serious, I don't know what is serious if there's a more secure place. You know, that is scary. So the first attack, first risk is that it isn't attacked on cyber cyber attack of any kind. You're going to see a collapse in the AI trade because there's going to be a massive basically a wave support to slow down AI topics, which will be a big problem. The other, of course, the other risk is that Trump might just simply decide. And I think this could very well happen before year end. I think it's at least 50, 50 and I just say 50, 50 is actually very hard. I would say Trump is going to ban Chinese AI models. Okay. And maybe even ahead of the anthropic IPO, in order to sort of like keep that, but if they do that, China is going to retaliate very likely by restricting rare earth exports. And then we see an end of the fragile truth between the US and China. And then the stock market goes down. I think

11:27it's going to be at this point, that's what we're looking at as two risks that could potentially centrally break this whole momentum of the AI trade. And then, you know, and what what what let's park that let's park data second. What about the other narrative, the other big narrative of the billion deaths warning? I mean, Bill Gates came out on record. He was on Meet the Press. He said that unchecked AI could cause a billion deaths if weaponized by bad actors. Now, President Donald Trump dismissed the warnings. And I guess to your point, he wants he wants, you know, he doesn't want to stop the AI train. But he dismissed the warnings from the Microsoft co-founder that unregulated artificial intelligence can cause billions at deaths. And instead, he announced a voluntary self-regulation according with top tech execs. So your thoughts on this narrative of, you know, AI and could it wipe out the planet basically? Terminator type situation. Like, I mean, I think you don't even have to be a co-founder of Microsoft to reach that conclusion

12:32because at this point, you know, I think the this view is now the consensus view in Silicon Valley among the senior basically software engineers who are developing this thing. I mean, in fact, you know, I mean, I don't know if you be I mean, the chatters that I've been hearing lately is that a lot of these so-called basically founders or senior engineers at these leading AI firms all rushing to basically by farmland and New Zealand just in the event we get a human extinction event, they can go go there and basically buy their grow their own food. You know, what I'm saying to you is, I mean, I want to say something to every audience would understand just so that what we're talking about. Two days ago, as you know, Nvidia launched this new piece of software called Sentinel, which Jason Wong said it could have prevented, okay, the attack on hugging face. I didn't been in place as though that this was going to be like a, you know, with this, like, you know, we don't even have to worry about it anymore. Well, let me tell you this, that what this Sentinel does,

13:34what it does is to simply limit the access of AI models, you know, from basically network. So because for these for AI to do an even damage, they need to have access to network, right? I mean, in other words to the internet, right? The reason why basically that, you know, you know, open AI models attacked hugging face was because even though they were on the testing, they were supposed to be operating in a sandbox. Basically, they were, there was actually, they had limited and permitted access to the internet and exploited that and gain access to were able to do that serious damage. Now, what I'm saying to you is that once you launch a model, I'm just think about like, Jason Wong's argument that Sentinel could actually prevent this from happening is to say, yeah, if, if, if, if, if, if, if these agents have absolutely no access to the internet, then they won't be doing any damage. But the whole point about AI is that if they don't have access

14:35to internet, they're useless. Their whole entire basically premise is based on being able to access this and that and do search and and basically do all that. So what I'm saying to you is that that's why, you know, there is no now. I mean, if you, if you, anybody who's read through the post post mortems of the hugging face attack, okay? You know, which only came out two, three weeks ago. I mean, was horrified. And because you're talking about these thousands of basically rogue agents who were lying through their teeth, they were collaborating, they were even some even went into sacrifice in order for others to succeed. I mean, it's crazy. What you're dealing with here is something that's completely unpredictable. And, and in a way, this is what I think it's a very worrisome development. And I think from now point of view, this is the problem. We're now gotten to that point that, you know, these things are becoming more and more unpredictable, which means the pressure to, you know, send should the, you know, to regulate them is only going to grow. And then

15:40if we get to that point that you can basically develop more advanced models because they're just, they're going to have smart as then the Chinese going to basically catch up right away. And then how is any of this AI catpex is going to and to produce a return for investors. You know, I think that is for me the biggest. It sounds crazy. But I think it's, this is not like a 1% risk anymore. This is like a 10% maybe even 20% risk any day, any week, any month, any year. So that, I think that that's the way we're looking, we should be looking at this. I want to shift gears here a second and talk about the pressure on on gold prices, David. What do you see happening on the gold front? Why the pressure? I think, I think for me, I mean, it's, it's pretty obvious. I mean, I think actually gold has actually done not badly actually relative to the bond market, right? I mean, you because we know that, you know, going bond yields tend to be inversely correlated. Given bond yields at these level, you know, I would have thought that gold might have been hit even harder. But you know,

16:44the fact that gold has been holding up reasonably okay, that's not a bad thing in itself. I think it just tells you that people understand that this AI bubble cannot last forever, that, you know, the rates in a way have overshot. And that if this anything happens with this AI bubble gold, it's just going to go like 10,000, whatever it is because I mean, there's no doubt if the AI bubble burst rates were collapsed, okay, dollar were collapsed and gold, it's just going to go through the roof. So in a way, only goal is the biggest protection against basically the bursting of the AI bubble. So I think a lot of people understand that now, of course, you know, like there is the sort of the day to day that, you know, whatever it is. But I think there's no doubt that that's sort of like that's the way I think a lot of people are looking at it and rightly so. And like is it safe to say David, like it's inevitable that that bubble has to burst at one point? I personally think so because again, you know, like, you know, like it's, you know, unless,

17:47unless somehow the Chinese just, you know, they just basically fall away, they some of this appear that the competition from China somehow just basically disappears unless basically we don't have any major attacks and somehow AI become super productive and then super well behave and then somehow like, you know, we squeeze a lot of productivity gains without losing all the jobs. I mean, all these ifs are my humble opinions together amount to a very, very slim chance of the kind of path that we are that could basically put us in a sort of soft landing. I think it's last question. I know your boots on the ground, you live in Israel. You have elections coming up October 28th at Netanyahu warned of, you know, possible attack ahead of the election. You know, we obviously saw the news of what happened with this flight, but your thoughts and what you're tracking. Yeah, I think this is where you're at now. I mean, I think the market, I mean, a lot

18:50of my clients and I mean, I think, you know, and then I think they're very representative of the of the market. I think a lot of people think, oh, well, Trump is going to talk, oh, because you've got the midterm coming up and so on and so forth. I've been saying it's the opposite. Okay, you know, Trump doesn't care about them. I think one thing I really believe, okay, you know, is what what Trump says when he says he doesn't care about the midterm because guess what, you know, I like 2018. He's not going to be seeking a reelection. And I think in that respect, especially given the Republicans are sentiently polling so far behind the Democrats, he could talk on Iran tomorrow and oil price dropped 30% and then still basically go on to lose the midterm election. So why the hell? Okay, would he basically talk? Would he want to compromise on the arguably the defining objective policy of his presidency, which is this war against Iran? I mean, if Trump was going to talk, he would have talked already. If he talks now and then go on to lose the election,

19:52it would be like a terrible thing, I think, for his legacy in general. So I think Trump is not going to talk. Now, of course, what it's also interesting to me is the fact that Trump, when he says, you know what he said, you know, last year and that he's going to he's considering resuming bombing Iran after the midterm election, if you recall, right? That's what he said. But one thing we know about Trump, he said that so many times in the past, which is that if he's going to attack you, the last thing he's going to do is to give you a advanced warning, right? So he says that he's going to wait until after the midterm to basically attack Iran. For all I know, he could very well do it before. Now, this is where it gets very interesting because, you know, two days ago, BB Nadeyev, who actually flew to UAE, where he gave a presentation to representatives of 10 countries about the kind of threat, okay? That, you know, that the whole entire region faces for me, Iran. I'm talking about nuclear threats. And then what is interesting to me is that two days later, you match me today, the vice president of UAE got on the plane,

20:52and flew to guess where Saudi Arabia. This is the first, basically, visit by a senior of, you know, UAE official in Saudi Arabia since the rift between these two countries. What I'm saying to you is right now, I think Trump is trying to rally the region around him, you know, essentially on a hardline approach to Iran. And that I think, you know, something could very well happen very soon. And I think that the Iranians must be waking up to this. And because also, Iran is in a much weaker position economically speaking now than there were six weeks ago. Because what a lot of people don't realize is that from February all the way to August, Iran was exporting a lot of oil, despite the so-called blockade. Because the US never really fully enforced the blockade because Trump was worried about high oil price. So they never actually, so US Navy was actually turning a blind eye when Iranian, you know, oil tankers were selling onto China. But in the last two months, Iranians have not exported any oil. So from now point of view, economically, they're in

21:55dio streets right now. Therefore, they cannot afford to wait around, okay, for Trump to basically buy his time to talk about negotiate that for them, they're running out of time. In my humble opinion, the only reason why oil prices not much higher is because Iran is allowing oil traffic to get through the straight up hormones because they're trying to, you know, basically bring around the Saudis UAE to their side to make the case for Iranian control of the straight. When they decide that that is never going to happen because Trump is putting his foot down, then I think they're going to shut down the traffic in one go. And then you can see it potentially fairly not only in your movie and oil, that will be a big problem for the stock market. So this is something else, I think, you know, before between now, the midterm election, I think even before that, we're just talking about next week, even. I think that probably, the address is very, very, very, very, very real. It might be. David Wu, thank you always for your latest thoughts. I will catch up with you soon and

22:56stay safe. Okay. That's great. Great talking to you. Always great talking to you, David. David Wu, you could find more of him on David Wu on bound. We'll see you all soon.