Speaker 1: All right, we should be live. Welcome to Thoughtful Money. I'm Thoughtful Money founder and your host. Thank you for joining me for this special report. Right at the end of last week, the world was rocked, I think it's fair to say, by the release of a new Chinese AI platform named DeepSeek. And it has thrown a lot into question about the future of the AI ecosystem, and then by association the financial markets, which are heavily revolving around all the market cap that's been priced into the AI stocks, the Magnificent Seven, if you will. So to help make sense of this for the layman, I thought I'd have this special report just sort of to explain what we know so far, and then to talk about what the potential implications of DeepSeek are both for the AI ecosystem, but also for the financial markets in general. And to do that, I have asked my good friend and mentor, Charles Hugh Smith, to join me. Charles, how are you doing?

Speaker 2: Good, Adam. Thank you very much for inviting me to talk about this super hot topic.

Speaker 1: Well, thank you. And folks, I mentioned Charles there is one of my mentors. He and I talk fairly regularly on the phone this weekend while I was taking my dogs out for a long two-hour walk. I spent all that time talking with Charles about the potential implications of DeepSeek. And so I thought he'd be a great guy to bring on and really just for us to sort of rehash the highlights of what we were talking about. Charles, I'll also let folks know, this has obviously caught our attention. If you've been on X at all over the weekend, you realize it's caught the world's attention. I reached out to Fred Hickey, who many regular viewers of this channel have seen me interview on technological matters. I have invited him to come on the program as well when he has something he wants to share with the world. I believe he's still just heads down making sense of all this and will first let his premium subscribers know. But then I'm hoping he may come on this channel. I will say he has gone public on X saying that he thinks this is a seminal event in the AI industry and shares a lot of the concerns that Charles and I are about to talk about with you here. All right, Charles, well, look, I guess first, let's just sort of recap for folks. Let's say somebody was away on a long weekend, hadn't heard about DeepSeek until they hopped on this call. Why is DeepSeek so noteworthy? Why is it such a potential game changer here?

Speaker 2: That's a great intro, Adam. And I want to just preface by saying that you're the one that saw the potential here for the classic black swan, you know, the disruptive event that no one anticipated. And so you alerted me and then I started digging around as a result of that. And then my readers started helping me, you know, find resources. And so it's a team effort to understand something this consequential. And so there's a lot of stuff out there already. And a lot of it is great stuff. And so what we're trying to do is summarize what's already out there. The basic idea here is that the U.S. approach to AI was sort of brute force, you know, use super fast, super costly, super energy consuming hardware, you know, processors to process vast amounts of data quickly and with high precision. And that was viewed as a monopolistic kind of moat. In other words, like if you didn't have billions of dollars to invest in this hardware, then you were out of the race, you know, you were going to be a loser in the AI global race. And so now what DeepSeek has shown, which shouldn't surprise anybody that knows much about software, and I don't claim to be an expert, but, you know, I'm an observer, is that they've relied on clever software rather than super fast hardware. And so this opens the door to getting the same results without the super costly, super energy consuming, super expensive hardware. And so that's where the revolution is. The point I would start with here is DeepSeek has issued their products as open source. So in other words, their software is available for people to examine. And so this is going to be used by many other smart people. They're going to start innovating along these same lines. And so this is not the end game here. It's the start of a whole new approach to AI.

Speaker 1: All right. So I think that last point I was going to raise, if you didn't, but let me just sort of knock three things about it off for people, again, if you didn't know what DeepSeek was before. Essentially, it's a, you know, generative AI, large language model platform that is apparently as good, if not in some cases, better than the leading ones that we've all been hearing about so far. Open AI, Gemini, BARD, whatever, the ones that were being built by these big AI behemoths. And what's notable about that is it was created for a pittance versus what the billions and billions that have been sunk into these existing big models. So basically, you get as good, if not even better results for way, way cheaper. It's way, way cheaper to build, way, way cheaper to run in terms of energy, as you talked about. You don't need, and we'll talk a little bit more specifically about this in a bit, you don't need as expensive hardware to run this. In fact, you can run it on fairly cheap hardware. And to your point, Charles, and very importantly, the AI models that we've known of before last week were all closed, meaning they were owned by the companies that were developing them, whereas this is open source code. So anybody can copy this, run their own version of it, make whatever tweaks they want, make whatever improvements they want. So it really kind of has changed the game. So one of the, it raises some questions about technology, which is, wow, is AI going to be a lot cheaper? Is this going to be an accelerant to the future of AI because it's going to be so ubiquitous and cheap and all that stuff and better. But also it's going to throw into question a lot about the current market value that's in the Magnificent Seven AI-related stocks, most of which are the biggest companies in the world. And a lot of their market cap over the past couple of years, a lot of the growth in their market cap has been granted on the expectation that these guys were essentially going to have a corner on the market on AI. And DeepSeek has maybe just blown up those expectations and said, well, maybe those guys might not have much of a corner on anything going forward. You're nodding a bit as I'm saying this, but are those, you know, kind of the seminal parts of DeepSeek that the average layman should understand at this point in time?

Speaker 2: Yeah, I think that's an excellent summary, Adam. And we have to look at the current space, the Mag Seven, as quasi-monopoly or monopolistic. And that was the value proposition here. In other words, was these companies have a lock and they're going to have 20 years of growth or 25 years or 50 years or whatever. And that, of course, is extremely reminiscent of the dot-com era bubble in the around 2000, which was looking at the Internet as a permanent growth story with super high margins. You know, famously, NVIDIA has like 75 percent to 90 percent margins on their hardware. So everyone thought that was a forever story. And so now that's not a forever story. That's already a past story. So now it's a new game. And so we can't rely on valuing stuff as if it's going to grow for decades at its current pace. And so I think there's a couple of other quick points here, which is I've and again, this all the information that's coming out is like, you know, trying to drink out of a firehose. Right. We're all limited by what we can consume and understand. But apparently very much in real time here, DeepSeek's product can be used on a phone. Like in other words, it fits on a smartphone, so you don't need anything gigantic. And it can also be downloaded and taken offline. So in other words, you're not relying on an Internet connection to, you know, the owner or the issuer of the software for anything. And so those are powerful features. And a couple of other points I want to make here is I've been observing the AI space for like 40 years. And if anybody remembers Lisp, you know, the language of AI in the 80s and that kind of thing, we've seen this space just disrupted and then crash again and again and again. And so this is not actually unusual to say that the future that everyone thought was hammered out and rock solid has already dissolved and it's dissolved entirely. In other words, there's a lot of people going, oh, well, NVIDIA chips are now going to be needed in robotics and blah, blah, blah. Well, all that may very well be true. But the bottom line is software has eaten the monopoly world that existed before DeepSeek. And that's a good thing because monopoly is inherently strangling, right? I mean, what it does is it limits innovation and creates barriers to widespread use and so on and so forth. So in a way, the stock market has been overvaluing the Mag7 because they thought they had a monopolistic lock on their markets. And so now that it's a free for all, then what's the value of those platforms and so on? And the answer is considerably less than when they were valued as monopolies. And again, I can't stress this point hard enough is it's actually way better for the economy that these monopolies have just been dissolved. And in other words, that it's more of a free for all now because the core of productivity is two things, transparency, like everybody's on the same level playing field, and two, that there's real competition. And so now with DeepSeek, there's going to be a lot of different people coming out into the market with software based products that will run on smartphones, they'll run offline if you need to. In other words, it's going to be a wide open space here again, which is good for any economy that's open.

Speaker 1: Right. And let me just note real quick, as far as we know so far, and I'm already seeing some questions here about, hey, how do we know this thing's even for real? It's from China. It could just be a big hoax. I do think that we've had enough people banging on it since this was announced, I think on Friday, that real programmers are saying, no, this thing's actually really pretty good. So as far as we know, Charles, what you've said is right, but folks, there are probably still going to be some curveballs to come out around this over the next couple of days. So take everything we say with a little bit of a, you know, knowing that we're peering through the fog of war here. And speaking of war, we were talking about this yesterday, Charles, that, you know, what China appears to have done here is the AI agapoli, right? These magnificent seven companies said, look, you know, we're going to be, this is a winner take all game. It's an arms race. We're going to win it like the US did versus the Soviets, where we're just going to spend our competitors into bankruptcy, right? And China basically said, all right, well, we're not going to be able to win that war necessarily. So let's try to fight a different type of war. Not unlike, you know, how oftentimes when there's a big military power, the opposition resorts to guerrilla tactics, right? And China said, okay, look, we're just probably not going to be able to win or win on the timeline we want on a hardware basis. So let's try to win on a software basis. And there are two things that you mentioned that stuck in my mind around this. One you said, this is kind of not unlike a Sputnik moment for those that were alive during the time where the Russians launched Sputnik, which was famously the first satellite ever to go into orbit. And it was really where, you know, the West had its confidence ripped away from its eyes to realize, wow, someone's just made a technological leap that we didn't think possible. And we're now perhaps behind in this race and have to catch up. So one, I'd love to hear you opine a little bit on that. And secondly, you were making a lot of direct comparisons to the Voyager satellite that was sent out in the seventies, where they had so little available processing capability in that satellite that they had to get really, really good with the little bit of code they were able to write for it. And it has performed remarkably well. So if you can just sort of elaborate on both those two points of yours, because I think they're really interesting in sort of explaining why deep sea could be so bigger.

Speaker 2: Well, thank you, Adam. And I noticed on today's scan of the web that famous investor and, you know, innovator in the browser space, Mark Andreessen, also used the exact same phrase, the Sputnik moment. So I'm not sure whether that's a meme that came to us at the same moment or not, but I'm in good company referring to that.

Speaker 1: I saw Goldman mentioned it too, but again, that was after I'd heard you say it. So I want to give you original props.

Speaker 2: Yeah, I might have. I might have been first. Who knows? Anyways, it's what I think attracts Goldman and Mark Andreessen and myself to this idea is that it was a cultural shock as well as a geopolitical shock that our dominance was not guaranteed. And I think that's kind of the vibe behind this use of this analogy in AI, that our so-called dominance is not guaranteed. And to Adam's point, asymmetric advances can undermine brute force kinds of advances. And again, I don't claim to be a software guru, but I have observed the space for like 40 years. And so I refer back to the Voyager mission and you can go further back and say Pioneer and so on. But the Voyager missions launched in 1977, you know, 47 years ago, are interesting because they're still out there. They're 15 billion miles away and we're still communicating with them and they're still providing data. And they carry, remember, in the 1970s, you know, the integrated circuit chip was new and they were very expensive and modest by today's standards. So the Voyager 1 spacecraft carries three little computers and with a total memory of 69 kilobytes, which is about the size of a low resolution JPEG photo. So how do you play with something that's that tiny? Well, you have to be really clever. And so the Voyager spacecraft, Voyager 1, it had a failure, it lost contact last year. And so the mission controllers played around with it and realized there was one component that had failed in one of the three computers on board. And so they used software to work around this by chopping it up into little pieces and then finding little bits of memory space where they could put it. And what I think the point here is that software, as Marc Andreessen famously said in 2011, is eating the world because it allows you to bypass a lot of hardware barriers, if you will. And to Adam's point, necessity is the mother of invention. And we've seen this throughout history where when some sort of abundance goes away and scarcity is the reality, then people get a lot smarter about how they're using the resource. And so I think that the Sputnik moment is also powerful because we're talking about an energy kind of picture here, too, where people are talking about bringing online, you know, thousands or hundreds of thousands of more kilowatts of energy to fuel these vast data farms that AI supposedly needed. And so it's going to change the energy picture, because if we don't need those vast new energy sources, then that frees up capital, if you will, for other better uses.

Speaker 1: Right. And energy for other better uses. Right. Right. And can you talk about that for just a moment? And again, you do a good job of putting this in layman's terms. You helped me understand it yesterday. Can you just talk really briefly about why DeepSeek, you know, can deliver equal value at much lower cost? It's really sort of a difference in approach. Right. It basically has traded off a little bit of accuracy for a lot of efficiency.

Speaker 2: Correct. Yeah, I think that's a good a good way of putting it. And again, this is my interpretation. All right. So I could be wrong, you know, so but my interpretation of their approach is that precision in getting an answer, you know, in the AI process, right, where you're taking samples of stuff that you've learned and then you're projecting based on those examples. Right. And so that process needs to be guided. And so that's guided by a system that we call like rewards, like the system is rewarded for increasing the probability that the answer is correct. Yeah. So if the open AI kind of brute force hardware based approaches, we're going to go in with extremely high precision. And you can kind of think about this as like, OK, how many how many decimal points are we working with in a number? Right. And so if you're going to and then at the end of that process, after it's concluded, it's its answer, then it's compressed, right, like everything else in the digital world. It's like your photo on your iPhone is two megabytes. And then, you know, you want to send it in an email, then you compress it down to 200 kilobytes. So it's the same kind of thing. You lose a lot of that precision in the compression. And so what the software approach of DeepSeek does is it says we're not going to go in seeking 95 percent precision. We're going to go in at 85 percent at the start and we're going to we're going to tweak the rewards of elements of the software so that the software will if the probability of a correct answer is declining, the software goes, well, wait a minute, let's just start over here. And so you see the huge advantage of that instead of pressing through, like grinding through this really high precision kind of stuff, seeking some kind of perfect answer. Instead, the software is starting out at a lower precision and it's looking to see if it's if it's on the wrong path. And so you can see the huge advantages in that and how it could use 30 to 40 percent less processing power, which is just staggering, right? I mean, that's that's a that's an order of magnitude decline in the use of of energy and processing power. So but you can see how these sort of tricks or clever programming could compensate for just brute force processing.

Speaker 1: Yeah. And so, you know, that that's why it's so much cheaper. And I think I've heard you said 30 to 40, I think I heard 45 times cheaper or something like that, you know, something crazy or 45 times less energy usage or whatever. But when you look at what companies have committed to CapEx so far, it's in the tens of billions, right? I mean, it's some of the big companies that's approaching a hundred billion that they've put into these massive farms of NVIDIA GPUs and stuff like that. What are they saying was the total capital cost of developing DeepSeek?

Speaker 2: Well, everybody's extremely skeptical of this five million dollars they put out there. And so but you know what? Let's just add two orders of magnitude to that and say actually they spent five hundred million. But even that's not that that's still modest, as you say, compared to like 80

Speaker 1: billion. Yeah.

Speaker 2: And so and so that's going to change the valuation of these companies. And it's also going to free up a lot of capital for them to invest in something else. But again, the whole technology space is always about agility and adaptability. And, you know, a lot of those things are cliches, but but they're real. And so I think I want to I want to touch on this geopolitical thing, too. A lot of people are saying, oh, wait a minute. Does this mean China's got the lead in AI? I don't think that it's a state or government run option here to say, oh, we're going to control the world's AI. I think what we really the dominance here is which economy is more open, more transparent, more open to sharing information and is going to allow the greatest number of people to start innovating with these new tools. And so that's really where what it boils down to.

Speaker 1: How weird would it be if that country were China? It's just not historically, culturally, those are not adjectives that are used around

Speaker 2: China. And so what's interesting is the whole world really is has an equal opportunity here. It's like so it's like there's systemic issues here. It's not just the software or the hardware or whatever. It's which economy as a system is more open, more transparent, more adaptable, more flexible, and is going to give rewards to somebody working in their garage, famously, that comes up with a product that that people love. Now, the other thing is, what's the pricing of of this of these tools? In other words, the the monopoly idea or the quasi monopoly idea was we're going to charge a couple hundred bucks subscription rate for access to our tools. And now if there's a tool that's free and can be downloaded on any smartphone, that subscription model is dead in the water.

Speaker 1: It's gone. So I was going to ask you about that. So so I have the memory of it's funny. I've told folks here that I might try to get Brad Garlinghouse, the CEO of Ripple, on this program at some point, because I knew Brad when I worked at Yahoo way back in the day. He used to we ran several departments, but I knew him best when he was running Yahoo Mail. And I remember the day when Yahoo was, you know, you got kind of the base mail for free. But if you wanted like some substantial storage so you didn't have to keep deleting emails, you had to upgrade the Yahoo Mail Plus. And Yahoo was really trying to build a premium revenue stream off of that because it had so many users. And I remember the day where Gmail came out with unlimited storage. Right. And so basically they just declared email is going to be free. Unlimited email is going to be free to everybody forever. Right. And I just remember the sense of panic going on around Yahoo that day where they just saw, you know, this this revenue stream that they had and all their pro forma models just poof, go up in smoke. Is this kind of that moment for a lot of the Mag7 AI players, the open AIs of the world who had a business model they were counting on? And all of a sudden, you know, China said, look, guys, price of AI to the consumers

Speaker 2: going to zero. I think so. And let's look at the successful subscription models, which are basically supported by monopolistic moats. In other words, Microsoft can charge a subscription rate for Office 365 because they control 80 percent of the operating systems on the planet. And so if you don't have a monopoly moat that forces people to subscribe to your service, well, then the fact is digital material is basically free to copy. And so, you know, the Web is famously a giant copy machine. So I don't see any scarcity value that anybody can use to charge any kind of premium for these tools. And again, we can come back to the precision element of it, which is if something works 85 percent as good as something that's super expensive, it's probably going to be good enough for 85 percent of the users. Right. And so I don't see the use value in AI in a lot of cases, like a lot of the big Mag7 are thinking it's going to they're using it to improve their search. But we're not paying a premium for that search. It's just costing them a fortune to run it. You know, we're not offered some platform where we have to pay money for a better search engine. So I think a lot of the AI tools are in that category that the big players are thinking that there's going to be a revenue stream somewhere down the road, per your example of email, which doesn't materialize because there's no scarcity value in this.

Speaker 1: All right, so again, folks, that's one of the reasons why this DeepSeek launch has just caught so many people's attention that literally overnight changed people's expectations of the technology. But I think it also changed their expectations of the economics of this space. OK, so there are some pretty big negatives that I want to talk about with you, big potential negatives coming out of this. But before we get to them, let me just ask you on the positive side. You mentioned this earlier that you said, look, this is great for the consumer, right? It's great for commerce, that this new tool is going to be perhaps much more ubiquitous, much more quickly, much more cheaply. It's going to require less energy inputs, all that stuff. Do you see this as a potential catalyst for really unleashing an economic boom, not necessarily tomorrow, but maybe over the next five, 10 years? Is this something we'll look back as sort of a watershed moment of when AI became

Speaker 2: democratized? Well, that's a great phrase, Adam, democratizing AI. And it's unknown because as you and I discussed privately yesterday, there's a downside to the efficiencies of AI, which is replacing human labor that can be automated. Now, not all human labor can be automated, but whatever can be automated will be automated. And so there's some excitement around using AI, say, to optimize small business supply chains. And if that means being able to have fewer employees or ideally a one person operation that used to require five or 10 people, then that's the way that it's going to go. But then the question that you raised, which is, well, what do people who have been displaced by AI automation, what are they going to do for a living? And a lot of the sort of standard response is, oh, well, they're going to find jobs in AI. And it's like, no, I'm sorry. This is not that kind of revolution.

Speaker 1: By the way, this is one of the negatives that I have on my list.

Speaker 2: Yeah. Yeah. Anyways, it's it's it's like AI will just get better at doing its own work. And so we're not going to we're not going to fill that employment void with 10 million programmers of AI. And so that's that is an issue. But on the other side of that coin is a lot of people are assuming that virtually all human work can be automated or done by robots or something. And so there's but actually there's a lot of human labor that requires intuition, tacit knowledge. And therefore, it's not really that easy to automate it. You know, you can you can have some tools that might help the human do their job quicker, faster, better. But the human is still going to be cheaper and better than than than the robot. So there's upsides to that. And so we can talk about that almost for another couple hours. Like, well, how do you deal with the employment, unemployment issue here?

Speaker 1: OK, all right, so but if I if I if I took from what you said there, Charles, there are lots of reasons to have some excitement about what the future holds, both commercially and also just, you know, hopefully we're going to see all sorts of breakthroughs in medical science and all sorts of things that will hopefully enrich our lives and stuff going forward. OK, now let's get to my list of worries. And AI as a job killer is really high on that list, as folks have heard me talk about in the past. I don't want to start there, though. I want to start here. And regular viewers of this channel have seen me put up this image before. But, you know, this is a meme for those that are listening on the podcast. It's a meme of an elephant balancing itself on a beach ball. And the elephant represents global capital. The beach ball represents U.S. equities. So we've never had so much of global capital inside the U.S. equity market. There's just been huge inflows to the U.S. stock market from the rest of the world over the past several years. Yet we've never had the U.S. equity market so concentrated in so few stocks. So in this image here, the beach ball that the elephant is balancing on is itself being held up by a couple of ants, which I've labeled the magnificent seven stocks. I think it's something like they represented just those seven stocks themselves represent like 38 percent of market cap, or at least they did at the time I made this meme here. So, you know, I've been sharing this because I've been very concerned about how dependent the entire global, you know, equity market was on so few stocks. And then to ask the question, hey, if if one or more of those stocks start to stumble, you know, that could will that bring down, you know, the global stock market? I fear we may be about to find out the answer to that question. And as we're talking here, let's say that so this is Monday. It's these are the first couple of trading hours that the market's been open since Deep Seek was released. S&P is down two percent. The Nasdaq is down three and a half percent. I mean, honestly, Charles, we've seen markets down much greater previous corrections. It is funny that people are kind of running around with their hair on fire at the moment when the indices aren't down by all that much yet. But Nvidia is down 17 percent as we're talking. I mean, that's that's a lot. That's a few percentage points away from a fifth of its value being gone. And, you know, a single trading open. I'm going to see if I can share this on my screen. And if I can't, I will stop trying to do this. But bear with me one second. If you can share screen, share screen. Zero hedge. All right. So hopefully, folks, you can see some of these headlines here. But Nvidia, hot money turns dead money. Nvidia is crashing below its 50 day, 100 day and 200 day moving average. This is a rare exclamation point. Two are the stories down is cheap AI a black swan event? So, you know, the media obviously is breathing pretty hard right now about the potential that, you know, this new Deep Seek release could have in terms of its threat on the market value of these Mag 7 companies that are keeping up the entire global equity market. So I guess where I'm going with you on this, Charles, is let's assume that Deep Seek, you know, it's assumed for a moment it is as disruptive as we think it's going to be here. How big do you think the impact is going to be on the financial markets?

Speaker 2: Well, I think you've outlined the core issue here, Adam, which is the incredible and historic extremes of concentration in these seven stocks and not just in the U.S. markets, but globally. And so that that leverage or concentration is, of course, makes it makes these issues extremely vulnerable and the whole market vulnerable, because, of course, as you've posted before, Adam, the number of global stock market concentration in the U.S. is something like 65 percent of all global equity wealth is in the U.S. markets, despite the U.S. having 4 percent of the world's population and maybe 18 percent of the GDP. So this kind of these kinds of extremes are just sort of begging for some sort of disruption and return to some sort of baseline. So I think that is a great. Potential risk, and I don't really see it as as a as a possibility, I think it's more like an inevitability, and this is what a lot of technical analysts have pointed out and and historical references right to every bubble pops for some reason. And so we can go back after it's popped and argue about what the what the pins were. But the bottom line is these dynamics occur on their on their own. So I also want to emphasize that we're not really talking about one company or one tool. In other words, whether DeepSeek vanishes tomorrow or not, it doesn't matter. What matters is the approach of using software. That's not going to go away.

Speaker 1: Right. If I can just interrupt you for one second, because I'd like you to weave this into your answer. I'm kind of likening it to Roger Bannister breaking the four minute mile. It was something that people just thought was humanly impossible for centuries, millennia. And then he broke it. And then I think in the year after he broke it, it was broken like another 12 times or something like that because people now knew it was possible. And so to your point, China basically showed the world there's other ways to do this. And now that people realize there is a ton of innovation can bloom as a result that wouldn't have happened before because people just didn't know it was possible.

Speaker 2: Yeah, that's that's a great analogy. And and again, the point here is global capital poured into the Mag7 and the AI space for what reason? Well, it was considered a slam dunk. It was a guaranteed growth story, right, with no end in sight. So why not put your money where it's going to be safe and and and expanding, growing, generating profits? And so that if that if that magnet idea has been mooted, has been basically evaporated, then. Trying to cling on to these, that idea is going to be disastrous because it's just no longer the case. So when when we saw the last kind of tech bubble 25 years ago, then the Nasdaq fell about 80 percent over the next two and a half years. And I've often posted a chart of bubble symmetry, which is bubbles that that have a sharp ascent. They tend to decline at basically over the same time frame and scale. So the Nasdaq went up for two and a half years, like in a parabolic rise from a thousand to five thousand. And then it fell in about two and a half years back to a thousand. And it took 13 years to recover. And if you adjust for inflation, it took 16 years. So it's hard not to look at history and go, well, who benefited? You know, who adjusted more wisely? The people who sold at the beginning when they realized that the story had changed or the people who hung on and believed that, you know, this was a minor problem and that, you know, they have a lot of reasons to think it's going to come back? Well, obviously, the people who just realized the story had changed fundamentally and got out and sought some other place to put their capital benefited far more than people that kind of clung on to the story, seeking some reason to believe that it was still true. And so I think we're at that that kind of moment where denial is extremely appealing, you know, because it's sort of like, well, wait a minute, it can't be this. It can't have changed. You know, there's the fundamental story still got to be there. And if you look at the dotcom era, it's instructive because, of course, the Internet did grow, but it was no longer profitable at the rate and in the places that it had been previously growing rapidly and expanding margins. And so we're not saying that the AI space is dying. It's simply that it's going to be expanding in ways that are that are not generating gigantic margins.

Speaker 1: Yeah, and you've you've talked about that in years past where even before I came along, you know, there was the argument that technology is, you know, going to continue to sort of displace human labor. But don't worry, because we'll be able to tax these corporations and then give everybody UBI and it'll be great. Everybody will have lots of free time and whatever. And I remember you saying, well, look, that's not how it works. Basically, these profit margins get competed away over time. So there's just not going to be the profits there to tax the way that people are sort of magically thinking. And of course, you and I, I don't think that would be a good thing for society anyways, if we just told people, hey, don't go to work and just collect your check. But you're sort of nodding as I'm saying this. So there is no sort of perpetual free lunch here, right?

Speaker 2: Yeah. And of course, there's a lot of debates about where what's the government role, you know, and the general view is that government shouldn't pick the winners because that typically doesn't work out that well. And so the government's ideal role, shall we say, is to keep the playing field level and not let a monopoly take over the market and stifle innovation. So I think that there's some truth to that. And we can look at that and say, well, you know, if the current quasi monopolies blow up and are no longer that profitable, then that frees up, frees up capital and talent to go elsewhere. And perhaps it's into a more decentralized use of these tools as opposed to concentrating the power in a handful of giant corporations. So let me let me take that to go into my next issue, but real quick, just to put a bow

Speaker 1: around this. You know, there's a lot of discussion even going on in the live chat here of some people saying, hey, this is a buy the dip moment. Some people still saying, hey, look, this is actually a great thing for the AI ecosystem, even the big players. You know, when the dust settles, this is going to be nothing but good news for everybody. So, again, we don't know yet exactly how this is going to unfold. But listening to you, Charles, it sounds like you're saying, look, one, just given concerns about bubble valuations anyways, but also given the nature of the disruption of DeepSeek, sounds like you're thinking this is a time to be more defensive when it comes to the financial markets than offensive. Is that accurate to say? Yeah. And again, it's because I've lived through bubbles and so I and I felt the same

Speaker 2: emotions, you know, like we're still dealing with human wet wear 1.0 is what I call it, which is, you know, when we are really committed to something and attached to something, then we find reasons to support our attachment to it. And it's very hard for us to let go of it. And so that's that's the split that we're in. That's the Sputnik kind of analogy here, that if the world has indeed changed essentially overnight, then we're forced to adapt to it. Or we're going to suffer the consequences of of clinging onto something that's no longer valid. Right. And so and technology does this all the time to us. It's it's not exactly new. all the time to us. So it's not exactly new. And so I think the challenge for individual investors is to struggle with our own emotions, really, and to try to be as objective as possible about the nature of this change, right? And to your point, you often counsel people to consider diversification. So if somebody's account is heavily concentrated in these MAG 7 stocks, I mean, just kind of common sense suggests, well, maybe now's the moment to diversify, you know? And I think that I would kind of stress here that the Sputnik moment is you're riding the elevator up, and you think it's just rock solid. Profits are going to be higher, and the market cap's going to go higher, and so on and so forth. And then suddenly, the floor of the elevator just drops out, and you're staring down a 90-story shaft. And it's like, wait a minute, what happened? And I think we're kind of at that moment. And so it's going to take a while for us to reconcile our, get our minds wrapped around that. And so what's going to happen in that period where we're trying to grasp the scale of what's happened and what will happen, then there's going to be a lot of uncertainty. So we have to kind of accept or embrace that uncertainty, as painful or difficult as that is. Well, very well said.

Speaker 1: So Charles, I'm seeing, well, I'll just pull it up here. Moustaki says, yes, I love Charles Hugh Smith. How have you not had him on more? So thanks for putting that, because I actually invite Charles on a lot. He's just, he's a very busy man. And very much appreciate you coming on today on incredible short notice, Charles. And I know you have a meeting that's coming up soon. So we're all going to keep you for a couple more minutes. Real quick, folks, if you've enjoyed, you know, if you enjoy this format of bringing on a big thinker like Charles to react in real time to big world developments, like, you know, this whole deep-seek launch, let us know in the comment section. Actually, if you do, obviously, I want to do more of this. And specifically, if you love having Charles on the show, please let him know that in the live chat too. And maybe it'll convince him to come back on a little bit more often in the future. All right. So two last points, then we'll wrap it up. So we talked about the threat of AI as a job killer earlier. You know, you mentioned about, you know, people getting, large amount of people getting displaced by AI. And it's good if you can put them to additional productive use. But we were kind of going back and forth on this yesterday. I think one can make the argument that this technological transition is quite different than many other ones that we had in the past, where, you know, if you were in the horse and buggy whip industry, and you lost your job, well, you could just go walk across the street and apply for a job at the Ford Automotive Factory, right? The whole difference in this automation, robotics, AI world is to get rid of human labor. So when the human job goes away, there's not necessarily another human job that opens. Now it will create opportunity for aspiring entrepreneurs where one person now can do the job, you know, can create a company that can do the work of a 5,000 person company without the real people, right? It can leverage technology to do that. The question is, is what happens to the 5,000 people that were displaced, who in many cases might have a skill set that's low enough or easily replicable enough by AI or other types of automation or robotics, that there's just literally not that need for them, right? And we end up having this massive unemployment crisis where we've just got a

Speaker 2: ton of people who just literally are unemployable. Right. So let's differentiate the economy from the society. And so the general sense over the last couple of decades of neoliberal capitalism or the whatever you want to call it, is the economy dominates everything and finance dominates the economy. So we just let finance and technology run and then we sort of pick up the pieces of whatever's been creatively destroyed or disrupted. And actually, I think we've sort of lost sight of the fact that we also are a society. And so there's a social system, a social platform, a social set of values. And we're going to have to make some social decisions here that will influence the economy. In other words, you can't just let the economy become your society because things like unemployment destroy your economy by first destroying your society. Your social order unravels, the social contracts disrupted and so on. So people look at this as sort of, well, what's the government going to do or shouldn't do or should do? And the government is in some ways an expression of our social awareness, our social values, social changes. And so if we as a society decide that having paid work is a priority one way or the other, then we'll have to figure that out and it'll have economic impacts, right? How are we going to pay people? And what do we consider useful work? If the private sector can't fulfill that need for employment, then should the state step in and provide some kind of platform for employment? And of course, we saw this in the Great Depression with the public works programs. And so those are questions that we're going to have to answer as a society. And that's all I would say about that.

Speaker 1: And I guess my question to you is, do you think the deep seek development is accelerating, pulling up the date by which we as society are going to have to begin to face those decisions?

Speaker 2: It seems highly likely.

Speaker 1: All right. Last question and then we'll close it up. Does this potential introduction of cheap, ubiquitous, good AI accelerate us towards the day where Skynet awakens and all the robots decide they don't need humans on earth anymore? And one of the reasons why I ask this is really kind of up until about now, AI was sort of looked at as a sovereign pursuit. In America, our big tech companies are pursuing it, but there's a lot of government involvement and there's lots of people at these companies thinking about how do we prevent AI from eventually metastasizing into something that we don't want it to be. And there've been worries that there would be other countries that maybe didn't have as strong ethics controls as we do, which is true potentially. But now that it's kind of given to the world, are we increasing the risk that somebody somewhere can create an AI that decides it doesn't need humans or want humans and something bad happens? It's probably not gonna happen tomorrow, but could it happen in five, 10 years? I don't know.

Speaker 2: Well, that's a great topic, Adam. And clearly you could, for just obvious example, you could take a drone and you could load it with software that allowed it to autonomously kill some other humans without any human intervention, for example. That would be very doable with today's technology. And there's no way to constrain that from bad players pursuing that, for example. I think your larger question is, is AGI, general intelligence within reach? And of course, this is like, we can debate it. But I think there's a lot of limits on general intelligence that are poorly understood. And I would recommend the book, The Myth of Artificial Intelligence, if anybody's interested in exploring the conceptual limits on general intelligence and the kind of AI we have now. So that's a separate question. Are we about to enter the dystopia of robots take over? That's one question. The other question you asked is more relevant, I think, and more immediate, which is, can AI be used to do things that are destructive to human life and human standard of living? And then the answer is, unfortunately, yes.

Speaker 1: All right. And I'd like to continue pulling in that thread with you, but it's too big a topic that we don't have enough time. But I appreciate you giving the concise answer there. All right. Well, Charles, look, again, thank you so much for joining us here. Very quickly, I want to put up the link to your Substack here. So if folks weren't familiar with you before this interview, and they'd like to get familiar with your work, they can go to your Substack. Anything you want

Speaker 2: to tell them about it in particular? No, go ahead and subscribe for free. And

Speaker 1: if I don't annoy you too much, keep going. Yeah, not at all. I don't even need to really tell folks here how much I value your thinking. And they've just seen it for themselves here. But yeah, folks, if you like what you've heard from Charles today, you're going to love what you see on the Substack there. Now, folks, for those of you that are saying, hey, look, Adam and a lot of his guests that have been on the program have been warning about the market's vulnerability to some exogenous shock. And now maybe DeepSea could potentially be one of those shocks. What does that mean for my portfolio? What steps should I consider taking, especially if I'm regretting maybe not having been a little more diversified before this weekend? I highly recommend that you sit down with a good financial advisor who understands all the macro issues that we talk here on this channel. If you don't have one or like a second opinion from one who does, obviously, you can just go to thoughtfulmoney.com, fill out the short form there and have a free consultation with one of the financial advisory firms that Thoughtful Money endorses. And these are the firms you see with me on this channel week in and week out. I also just want to remind folks too, you can bet your bottom dollar we will be talking about the implications of this at Thoughtful Money's fall conference, which again is coming up on Saturday, March 15th. And if you want to get your ticket for that at the lowest early bird price, which we're still offering it at right now, go to thoughtfulmoney.com slash conference. All right, folks, please let Charles know how much you appreciated him here, both in the comments, but also by hitting the like button and then clicking the subscribe button below, as well as that little bell icon right next to it. Charles, my friend, again, I really appreciate you hopping on with zero notice to do this today.

Speaker 2: It's been a great discussion. Yeah, it's been my pleasure, Adam. Thank you very much for the

Speaker 1: invitation. All right. And everybody else, thanks so much for watching.