This week, a Chinese company named DeepSeek shocked the markets by claiming it had made huge advancements in Artificial Intelligence at a fraction of the cost and effort of the big Tech giants in the United States. Is it too good to be true? Further, what would it mean for the average worker if the AI industry were able to expand even more rapidly than currently imagined? Will there be any jobs left? Or will our computers, tablets, smartphones and other devices put us all out of work?
Listen to the full episode, here.
John Norris (00:30):
Well, hello again everybody. This is John Norris at Trading Perspectives. As always, we have our good friend, Sam Clement. Sam, say hello.
Sam Clement (00:36):
Hey John. How are you doing?
John Norris (00:37):
Sam, I’m doing fantastically. I hope you’re doing well.
Sam Clement (00:36):
I’m doing great.
John Norris (00:37):
I’m glad that you’re doing great and I hope that you’re always doing great, but there’s going to be a big threat to you moving forward, and that’s known as AI.
Sam Clement (00:49):
I guess that seems to be the current opinion.
John Norris (00:54):
Over the last several years. For a while, the biggest question I would get would be regarding cryptocurrencies, and now it seems like everyone is just laser focused on what AI’s impact is going to be on the economy. But before we get to that, you know what about this DeepSeek comes out this week on Monday and really just upsets the entire world markets. And for those of you who just really have not been following the media at all, DeepSeek is an artificial intelligence company that’s an offshoot of a venture fund out of Hangzhou China. And for you, if you’re not familiar with Hangzhou, well you are now. It’s a metropolitan area in China of about 12 million people, which would make it larger than Chicago. It is also one of the leading technology hubs in China.
(01:41):
So it’s not all that surprising that Deepseek is based there. And the thing is, within the artificial intelligence community, people know who Deepseek is, but the average American didn’t. But really for all intents and purposes, it upset the apple cart this week when it announced it had made all sorts of advancements for a fraction of the cost and without all the expensive hardware that firms like Nvidia and others have been producing. And if true, if only a part of it is true, it could fundamentally change how AI is developed moving forward. So Sam, when I hear something that is too good to be true, I generally don’t believe it because if it’s too good to be true, typically it is. Tell me what you think about DeepSeek. Is it too good to be true?
Sam Clement (02:30):
Well, especially when it comes out of a pretty opaque country in terms of business and there’s an embargo on the top; Nvidia chips that are supposed to be used for these models. And so they’re saying they didn’t use those.
John Norris (02:45):
They use some old ones that they had like, stockpiled or something.
Sam Clement (02:47):
Well, the smaller ones that are still out, they’re the lesser powerful ones. One hundreds I think. But nonetheless… so you’re getting this info, whether it’s true or not, and the Chat GPT people are saying, yeah, they used our proprietary algorithms or what have you. And I’m already starting to venture into a space that I’m not an expert on, but I think the point is we do that with some regularity. That’s just a normal Thursday, but so there’s this opaqueness that you don’t really know what’s true or what’s not, but it seems like, I mean, it’s clearly was a significant step. There was some efficiencies, and I think it may be this next iteration of using other models to become more efficient and what that means or could mean clearly freaked the market out on Monday. It was relatively focused towards those areas and sectors that would be impacted by this and maybe lower cost or lower margins, what have you. And then the market just kind of rebounded Tuesday and felt largely back to normal.
John Norris (03:56):
Well, in a lot of ways, some of the bigger AI players like in Nvidia is still not back to where it was. And really the problem with what DeepSeek announced isn’t a problem in terms of advancements in AI technology. It’s just how it’s developed. Right now, it’s been very hardware heavy and requiring a lot of very expensive equipment from Nvidia and others. And what DeepSeek is saying is we can use a lot of software and a lot of stuff that’s not all that as expensive. So that presents a problem because we in the markets have invested so heavily in the, for lack of a better word, “hardware suppliers” but when someone’s coming to coming out saying, I can basically do what you’re talking about here, but I don’t need all this expensive equipment, people start to get worried whether or not they’ve spent their money wisely.
Sam Clement (04:47):
And I get that. But again, this is not an area that I could speak to every aspect of…
John Norris (04:55):
Which is the reason why you’re doing a podcast about it as opposed to actually developing it.
Sam Clement (04:59):
But my first takeaway from this is we’ve seen, and we’ve already seen it from some of these big tech companies, reiterating how much they’re going to be spending on those and tens and tens of billions of dollars from these companies.
John Norris (05:11):
Well, the US government’s talking about hundreds of billions.
Sam Clement (05:13):
Yes. And I’m saying for individual companies, I mean $30, 40, 50, 60 billion of CapEx all towards this a
John Norris (05:22):
Even Meta said that this week.
Sam Clement (05:22):
I just don’t believe that we have reached a point where those companies are all wrong and this opaque information we are getting from a Chinese company is correct. I mean, Nvidia was even optimistic in talking about the benefits of what DeepSeek is doing and being able to learn from previous models and becoming more efficient. And we saw the Microsoft CEO talking about Jevon’s Paradox, which is as things get more efficient, prices go down and demand goes up. And I think the rebound in the markets Tuesday kind of reiterated that. We saw the large, more transparent US tech companies and Nvidia who’s supplying all of this to everybody come out and say, this isn’t a bad thing. There’s some efficiencies. Maybe they took some of what we were already doing and spent billions of dollars on, and maybe they’re not being fully transparent and maybe they’re using more powerful chips than they say they are, which we’ve had a few people say… that this is net really just a good thing.
(06:25):
And that seems to be where NVIDIA’s not back to where it was. And there’s still probably some skittishness and there’s continued embargo talk with the Trump administration as far as that goes. So all in all, I think the market has kind of digested it as well as it probably could have going from this worst case scenario, margins are going to collapse. Nobody needs Nvidia chips anymore. And then you see all the big companies come out and say, from what we know, spit out the bones. There’s a lot of good that comes from this. We’re going to continue to do what we’re doing because I mean they’re best and the brightest and that’s what they see the future as.
John Norris (07:00):
Alright, right. Well thank you all for listening! Sam just kind of almost had a drop the mic moment.
But I’m going to tell you something, a little analogy that I used with someone here regarding what DeepSeek had to say, and it’s not perfect, but I think it’s okay for the average person at an average function talking about this type of stuff. Certainly, I’m not a technological expert when it comes to designing AI and what I’m about to say, my friend Sam, will probably be listening and probably be screaming at the podcast telling me to shut up and then he’ll send me a text telling me I need to shut up. But the analogy that I used was, imagine we’re starting off in, imagine people are building race cars and that’s the only technology we have for automobiles, race cars. We’re talking about Formula One or IndyCar.
(07:51):
We’re talking about IndyCar type stuff. And that’s what all the teams are buying. Just like, my gosh, we’re going to spend tons of money on these expensive cars. And then there’s some guy, I mean, yeah, we all know who he is. He’s a mechanic, comes out and says, Hey fellas, I’ve been able to build a decent car over here for a heck of a lot less, and it’ll go around the track fine, it’ll do great. However, and it’s a lot cheaper. However, the people that are building the Indy cars, they’re going, Hey, I hear you and I agree with you and it’s pretty cool. I want to find out what you’re doing, but I’ve got a race I’ve got to win here.
Sam Clement (08:27):
Also in that example, I would say that second mechanic engineer person learned a lot from the baseline of what’s happened with their cars. And they’re not saying, I need to learn how to create a car, the information and the templates are there, and it’s doing a better version of it or learning from that. And so again, I think the reaction we’ve seen once everyone took a breather for a few hours in the market has been optimistic. I mean, I just don’t believe that this costs $5 million, like they said. There’s just no way and can’t, we’ve seen other AI CEOs who are,
John Norris (09:07):
What is that? Like one 20th, what Otani is going to get paid next year? For world-changing technology?
Sam Clement (09:13):
They can’t buy you a commercial real estate building in Birmingham, but they sold the AT&T tower for 28 million. Yeah, I have a hard time believing that.
John Norris (09:26):
World-changing. I mean civilization changing.
Sam Clement (09:32):
For 5.6 million bucks.
John Norris (09:34):
By a couple of guys in Hangzhou China. Guess what we did fellas.
Sam Clement (09:38):
Meanwhile, Amazon, Meta, Microsoft, Microsoft CEO’s saying, I’m good for my 80 billion on this. We did it for $5 million. So it just sounds so absurd to me. And so maybe they’re just stealing some information from Chat GPT. Really,
John Norris (10:01):
It wouldn’t be unlike those guys to do anything like that.
Sam Clement (10:03):
But they are not creating a new proprietary thing for $5 million that Microsoft, that Apple, that Amazon, the five biggest companies in the world are spending tens of billions dollars each on. End of story.
John Norris (10:18):
I would tend to agree with you on that without a doubt. Matter of fact, I do agree with you on that because no matter how many times people will say this time is different, whenever something is too good to be true, it almost always is. And this is simply too good to be true. Not to say that they haven’t made some cool advancements that people won’t adapt to that. However, it’s not necessarily the be all and end all. It’s not going to kill Nvidia. It’s not going to kill Meta. It’s not going to kill all these people. Now with that being said, while the focus has been on DeepSeek this week, a lot of people are very concerned about AI and frankly what it means for their jobs. That’s what people are worried about. When I make a public presentation, I always get ask questions about AI.
(10:58):
And what people really want to know is they ask, what does that mean for the workforce? And when I hear that, I immediately hear, what does that mean for me? And I got to tell you, Sam, for a guy my age on the other side of the hill of my career, if you will, I’m kind of excited about it. It could potentially make me far more efficient. I would say your age, people that will use it adeptly, it’s those folks in the middle I think that are terribly worried about it. And I told a group recently, I said, the thing with AI is it really exposes how fungible we all are, how easily we are replaced, whether or not it’s with another person with a new technology. And what AI is going to do is just dramatically increase our fungibility and just accelerate it.
(11:51):
And those people that don’t learn to use Chat GPT or whatever iteration out there is out there next are going to be crushed by it. You don’t have an option of saying, I’m not going to do that. Because I remember when people started using emails, I knew people who were like, no, I’m good with writing letters. Well, no you’re not. It’s one of those things you have to adapt to this technology. It’s going to be around, it’s going to fundamentally change our lives and how we conduct business. So you better get on board and if you don’t want to get on board and say, I’ll never learn how to use this stuff, you’re going to be crushed.
Sam Clement (12:28):
Yes. My pause and I guess more of a Luddite than someone from my own generation who’s using technology for everything.
John Norris (12:38):
You’ve already disappointed me with whatever you’re about to say.
Sam Clement (12:40):
I think comparing it to the internet is very fair in a lot of ways in the sense that everybody was right that said this was going to change our lives.
John Norris (12:50):
Yeah, it did.
Sam Clement (12:51):
A lot of people were wrong about the timeline of it, and a lot of people were wrong about the severity of it. And I think it was, I mean, we’ve seen it. Email’s a great example. It’s created, yes, there’s new things that you’ve been forced to implement, but it’s created efficiencies. I mean, sure some jobs have gone away with it…
John Norris (13:13):
It’s also been an enormous time suck.
Sam Clement (13:15):
But I’m sure some jobs have gone away, but I would say most jobs have just gotten more efficient with it. And I think yes, there’s probably some jobs that can be done fully by AI and probably will fully be done by AI. And I get that this is exponentially more complicated than the internet itself, but I think that tends to be where these things go. Yes, this has long-term, it’s going to change the world forever, but probably in more positive ways and more efficiencies and things. I think at the end of the day, most of these careers, people are still going to want an attorney representing them and looking over documents, even if they leverage AI. People are still going to want advisors to be handling their taxes, their investments, their financial planning. I mean, yes, AI is a way to use these technologies and hopefully create more efficiencies and can drive costs down, which drives fees down, what have you, and maybe give us all more time back in our day. So there’s these benefits to it, but I just don’t see it anytime soon getting to the point where, yeah, I’m going to have an AI chatbot representing me in court or managing my finances or managing or filing my taxes.
John Norris (14:33):
Well, some of that stuff I might disagree with you a little bit on this, but however, we do have precedent here. While it might not be exactly accelerated computing, we do have historical precedent on what happens when there’s a major technological revolution, which again fundamentally changes our lives and how we conduct business. Prior to the industrial revolution, most people were peasant or yeoman farmers in Eastern Europe, they were kulaks or serfs. I mean, that’s what people did. They tried to scratch a living out of the land. And then guys like Cyrus McCormick invented the grain thresher or the reaper. And so all these people that have been cutting down wheat and barley and rye with sides now have something drawn by an ox that can do the work of 12 men in a very short period of time, find people, find themselves out of work.
(15:25):
So you see fewer farms out there, fewer farmers. And the same thing has happened again and again. The same thing has even happened with the steel industry here in the United States. People say, we don’t make as much. We don’t need as much steel as we used to make. Cars aren’t made out of steel the way that they once were. Buildings aren’t made out of as much steel. We don’t need as much. We’ve already built out most of our infrastructure across the entire economy, in any event, I’ve done a little bit of math on it and technological advancements have displaced far more steel workers than the Chinese could ever dream of, all of this type of stuff. So whenever we have technological advancements out there, people do lose their jobs. However, we’re still doing a lot of those things just because John Deere and Cyrus McCormick came up with better ways of farming didn’t mean that we don’t have farmers any longer. It just means in the United States, we don’t have a bunch of subsistence farmers. We have huge farms which make a ton of money, that sort of type thing. And the same thing will be with AI. We’re not going to not have CPAs. We’re not going to not have attorneys. And I made a group, made a presentation to a CPA group here this week. I said, we’re going to have probably fewer financial advisors moving forward, certainly fewer chief investment officers, all of these things, there’ll be fewer of us, you, me, CPAs, attorneys, all that, be fewer, but those that remain are going to be that much better.
(16:53):
And they’re going to be better and they will adapt AI, they’ll use it so they can service their clients that much better. It’s that person that’s in the middle. And kind of going back to what we’ve talked about businesses or even societies, you compete one of two ways:
(17:08):
Either on price or on product and service, and everyone else is stuck in the middle. Now moving forward, the same thing’s going to happen with human labor due to AI. We’re going to have those people that can only compete on price and these to be people that stick frozen potatoes into oil and grease. Or they cut your nails or they mow your grass, those sort of type things. I don’t know how an AI application is going to cut my grass, that sort of type thing. I mean, so we’re still going to have those and they we’re going to have people that are just so good at what they do and they use technology to make themselves better, that they will be up here and they will get paid that much more money. And then it’s those people in the middle where we’re going to see those people stuck in the middle where there’s going to be an enormous amount of shakeout. That’s the bad news for the average worker. I did point out when I was talking to this group and said the same thing to other groups, imagine something that we take for granted right now. I mean, you can put out your smartphone and you’re probably listening to this on a smartphone. You can pull out your smartphone. And I got to tell you, that thing came out in 2007.
(18:17):
I mean, if my Meemaw was still alive, she died in 1991, that would be so much science fiction.
Sam Clement (18:22):
Yeah. Even I learned, they said, you need to learn to do this math because you won’t have a calculator in your pocket. (laughs)
John Norris (18:28):
When I was in college, I mean the computer lab at Wake Forest was 10 computers. I mean just that sort of type thing. So all of this stuff has happened in a relatively compressed period of time. What changes AI is going to bring to us in terms of services and goods, I can’t predict right now, but I’m going to try to – to figure out those things that are out there that are so much science fiction here today. Again, historical precedence that says just because we have a technological revolution that puts some people out of work in the short term doesn’t mean that there won’t be other goods and services that the technology produces that will employ them, and we’re going to continue to vault forward.
Sam Clement (19:12):
Yep. No, I completely agree. There’s going to be changes, some positive. I mean, that frictional employment, you’re exactly right. It’s not like unemployment rates just tick up and stay elevated just because a job is no longer needed. At the same rate, it creates that frictional employment where we start to find new careers and new opportunities. And I wonder how much going forward there will be an increased emphasis or value for whether it’s handmade stuff or people still doing it.
John Norris (19:45):
I think we’re already seeing it.
Sam Clement (19:45):
We’re already seeing that. And I think in the construction space even you’re having these prefabbed homes and the technology increases there. But then on the flip side, on high end homes, and I get these are different parts of buildings, but plaster walls are back in again, which is ridiculous. These hand done plaster walls that, I mean, they look amazing, and you see the talent that these humans have to do some of the things they do on high end homes. And so you’re getting kind of one or the other, which is the same thing we’ve talked about with restaurants with what you just mentioned, competing on price or really quality of goods or service one of the two. And so there’s going to be efficiencies that create cheaper goods on one side of things, and there’s going to be increased value for the hand made, the artist doing something that is unique across all things, whether that’s a landscape designer, whether that’s a home builder, what have you, continuing to split off into one of two categories.
John Norris (20:54):
Well, I got to tell you, you kind of spurred me on this. You take a look at some of the stuff that we have now, and I’m looking right behind Sam right now, and I’ve got a, we’ve got, what is that? What’d you say? About a four feet by four feet sort of canvas print.
Sam Clement (21:12):
It’s a canvas.
John Norris (21:12):
Yeah, it’s canvas, but it’s a print. It’s Trading Perspectives. The logo that you might see on your podcast outlet, I mean, 20 years ago, yeah, maybe I could have gotten that. I guess. I don’t know. It would’ve cost an arm and a leg. I mean, just an absolute fortune in order to get that. And it wouldn’t look as great as that does right now. However, when we got that Sam, I think it would cost us 150 bucks, something like that. I mean, it wasn’t that much. So we can have that type of thing, but I got to tell you, although that I would not have easily been able to get that thing done two or three decades ago, and now I can get it for a relatively cheap price, would you consider that a fine piece of art? No. Of course not. So you wouldn’t pay much more than 150 bucks for. However, for a really talented piece of art, people will be willing to spend thousands of dollars. And that’s kind of going what you’re saying. It’s the same thing that what AI is going to do here for that skilled craftsman, for that person that might even be able to use the advancements in technology to make themselves that much better. There will always be a market, in my opinion, for that type of person. And there will always be a market for the type of people just that will be willing to do stuff we don’t want to do at a relatively cheap wage.
Sam Clement (22:36):
I agree.
John Norris (22:36):
And so that’s where AI comes in, where it upsets a lot of people and they’re really scared about the future, what that means, it means to them, because Sam, let’s face it, the vast majority of us are fungible. So with that being said, taking it back full circle and going to Deep Seek, I think that’s what scared the hell out of the markets earlier this week, is if what Deep Seek said was true, they just made an entire sub-sector of the markets in the US economy fungible, and people have spent tens of billions, hundreds of billions of dollars because they didn’t believe it was fungible. The problem with believing what Deep Seek had to say is it simply was and is too good to be true, that it baffles the imagination. You’re going to completely alter/ upset the apple cart for an amount of money that frankly, Jalen Milroe probably made that last year. I find that incredibly hard to believe.
Sam Clement (23:40):
I do too.
John Norris (23:41):
Alright, well guys, thank you all so much for listening. As always, if you would like to drop us a line at or you can leave us a review on the podcast outlet of your choice. As always, if you’re interested in hearing more or reading more about what we’ve got to say or how we think, you can go to Oakworth.com, O-A-K-W-O-R-T h.com and take a look underneath the Thought Leadership tab and find access to all kinds of exciting information, including previous links to Trading Perspective podcasts, links to our newsletter / blog, which comes out every Friday called Common Cents, and also links to our quarterly piece called Macro & Market Perspectives. All the individual analysis that we have put out there that is now live and fresh and getting a lot of positive feedback on that.
(24:34):
And of course, it goes without saying that there’s going to be good stuff out there by our advisory services team and Mac Frasier and matter of fact, Sam, with that being said, I’m going to tell you maybe some AI will help me end this podcast at some point because I butchered the hell out of that part of my language there. So I’ve got to close by just asking you one simple question. I ask you the same question every week. Got anything else to say on this important topic?
Sam Clement:
That’s all I got.
John Norris:
That’s all I’ve got today too. Y’all take care.