Second-Level Thinking: Why Dividend Stocks Win in the Age of AI

Everyone's talking about what AI is going to disrupt. The question most investors aren't asking: What happens after that disruption, and who actually wins? The obvious answer and the right answer are rarely the same thing. 

In this episode, Greg introduces a framework he first encountered through Howard Marks: first-level vs. second-level thinking. First-level thinking reacts to what's in front of you. Second-level thinking follows the chain of consequences and the ripple effects most people ignore. In an era where AI can reshape an industry in months, the gap between those two ways of thinking has never been more costly to ignore.

From there, Greg walks through real portfolio positions—Intel (INTC) and Accenture (ACN)—to show how second-level thinking plays out in practice. He also runs through a handful of names—Union Pacific (UNP), UPS (UPS), GE Vernova (GEV), Chevron (CVX), Lockheed Martin (LMT), General Dynamics (GD), Johnson & Johnson (JNJ), and Merck (MRK)—to illustrate which kinds of businesses AI threatens, which ones it quietly strengthens, and why some of the most "boring" dividend stocks may be the most defensible investments of the next decade. The core argument: brands, software, and moats built on perception are vulnerable. Logistics, infrastructure, and physical production are not, and AI may actually make them stronger.

Topics Covered:

[00:41] Introduction & why AI matters for dividend investors 

[04:47] First-level vs. second-level thinking — the Howard Marks framework 

[08:43] AI is accelerating disruption — and may be technology's own worst enemy 

[11:21] Are strong brands and moats as durable as we thought? 

[13:50] Why physical infrastructure may be the best AI defense 

[15:19] Intel ($INTC) — patience, conviction, and the US chip story 

[18:16] Accenture ($ACN) — the market's fear may be first-level thinking 

[22:21] Union Pacific ($UNP), UPS ($UPS) — logistics AI can't replace 

[24:27] Rapid-fire second-level takes: GEV, CVX, LMT, GD, JNJ, MRK 

[28:08] Final takeaway: the game has changed, sustainable dividend growth requires a new lens 

________ 

Disclaimer: Past performance does not guarantee future results. This episode is for educational purposes only and is not investment advice.

________ 

RESOURCES:

Schedule a meeting with us: Financial Planning & Portfolio Management 

Getting into the weeds: DCM Investment Reports & Models

If you enjoy the show, we'd greatly appreciate it if you subscribe and leave a review

Follow us on:
Instagram | Facebook | LinkedIn | X

Transcript

 [00:00:11] Greg Denewiler:

This is Greg Denewiler, and you are listening to another episode of The Dividend Mailbox, a monthly podcast about dividend growth. Our goal is to stuff your mailbox full of dividend checks. When they grow over time, a funny thing happens: you create wealth. 

Welcome to Episode 59 of The Dividend Mailbox. Now, we don't know the future, but of course, you've become well aware that AI is probably gonna take your job out, and you're gonna be unemployed like everybody else. But if you think about it from the standpoint of what we're gonna talk about today, you may avoid a lot of trouble, and from an investor standpoint, you could actually benefit from it. 

So today, we're gonna look at AI. It does impact dividend growth investing, and that's gonna lead us into a concept I first heard about several years ago from Howard Marks. It is the concept of first-level and second-level thinking. Personally, I think it will have the biggest impact in total return going forward. 

If you're not thinking about what we're gonna talk about today, you are likely asking for trouble. 

Just as a quick note before we get into this, dividend growth investing sounds simple, but doing it well over long periods of time takes discipline and patience. These episodes give you pieces of how we think about it, but if you're trying to build wealth, it helps to have a clear framework you can come back to when things get uncomfortable, which is why we wrote Dividend Growth: The Quiet Engine of Wealth. 

And ultimately, the real goal of it is to help you tune out the noise and make better long-term investment decisions. So if that interests you and you'd like a free copy, you can find it in the show notes or at growmydollar.com/dividend-growth-book. With that, let's get into the episode. 

So I'm gonna start with what brought about this idea and got us thinking about it in a significant way. A month ago, we decided to invest in a new phone system. It was really sort of by matter of necessity, not choice. We were using a traditional landline system because I had the opinion that landlines are just better-quality voice. 

And then, at least in the past, they've always been much more reliable. But as we went through the process of going to the new system, one of the big eye-openers was when we're having a conversation, we have the note-taking capability, and within about five seconds of hanging up from a call, we get an email notification. 

It's really kind of incredible. It's not perfect, but it summarizes the call, the key points. There's a place to forward it, text or an email, so that Lizzie, our trusty associate, can paste it into our CRM system, and basically, we have a great log now going forward, and that alone was significant. The quality issue became a much smaller piece. 

And then, talking to a friend the other day who used to work for AT&T, he made a comment: “You know, you're probably overplaying the quality thing anyway because most systems are now on a voice-over-internet platform. You probably were worried about something that nobody else even thinks about.” 

And my phone story is really a simple explanation. First-level thinking is, you know, phone quality. Second-level thinking is all the capability that it can do, and it actually makes us more productive, helps us create more value longer term. 

I originally heard second-level thinking from a book Howard Marks wrote several years ago. He was just talking about you gotta go beyond what's happening in the present and look at longer-term impacts and how they affect business. It's easy to think about, but it's hard to do. 

The reality is very few people, especially in this world where everything is right in front of you and social media and everything is trying to get you to click on something, that is all basically first-level thinking. Staying in first-level thinking potentially can really harm you as a long-term investor. 

So I think in the world of dividend growth investing, this is really going to become key. Second-level thinking goes beyond the superficial analysis. Unlike first-level thinking, which focuses on immediate results and obvious solutions, second-level thinking examines the ripple effects and unintended consequences of decisions, asking questions like, “And then what?” 

It anticipates outcomes over time. First-level thinking is simplistic, it's fast and common. It reacts to immediate information without considering long-term effects. For example, seeing a stock rise and assuming it will continue to go up is first-level thinking. Second-level thinking is complex, deliberate, and often counterintuitive. 

It considers interactions, time, and the broader consequences of decisions. For instance, instead of just eating a chocolate bar when hungry, a second-level thinker considers the long-term health effects and chooses a healthier option. Now, of course, that's a lousy example because we own Hershey's and we love chocolate. 

We would add in second-level thinking that it makes you feel better, which then leads to a longer life. But we'll just let that one go. 

You know, Warren Buffett has talked about this concept, whether he identifies it as such. He's really focusing on long-term results instead of luck, which is why you have to be a second-level thinker to even start to think that you're gonna own something for potentially decades. 

So is it easy? No. To practice it, one thing is ask, “And then what?” Evaluate outcomes in the short, medium, and long term. Analyze beyond consensus. Compare your perspective with common assumptions to identify overlooked opportunities or risk, and then connect separate pieces of data to form a coherent understanding of the bigger picture. 

To some degree, this is value investing from way back, where value investors historically were trying to look at the unconventional conclusions that were out there, figure out, okay, you know, what's the crowd missing and why do I see value there? But what is happening now, AI is speeding this up. It can change things dramatically in... 

You could even say maybe we're down to days. If you've been following the market very close at all, what you have seen is almost anything that has AI attached to it, the stock has reacted. And in some cases, some of these stocks are going up unbelievably fast. There's a real divergence between the haves and the have-nots. 

You can be in tech, but if you're in software, the market is starting to think you're history because AI can write software. You know, you don't need a software company anymore. But they're still in love with Nvidia. Their chips are driving the heart of AI. And I don't pretend to know the answer to this, but it seems like AI could potentially be its own worst enemy. 

Just take Nvidia, for example. AI can potentially design a chip. So where does the big value add? Why should it be a $5 trillion company if three years from now, five years from now, a decade from now, AI could be designing its own chips? Now, do I know that's what's gonna happen? No. Is that the market consensus right now? No. But this is where the whole second-level thinking really starts to kick in. 

Investors right now think technology is just the greatest thing ever, but in fact, maybe AI is going to be technology's worst enemy from an investing standpoint. The entire sector potentially is open to disruption, and ironically, some of these businesses that seem boring and very nondescript, they manufacture things. You know, just take PepsiCo or Coke. 

They're not exciting, but lo and behold, AI, in a lot of cases, helps those businesses, and I don't think you're gonna open up a bottle of AI and drink it here anytime soon. It's not too hard to imagine that some of these boring dividend growth stocks could actually outperform because they have less risk of disruption in them, and maybe technology ends up in a place that nobody expects. 

So you might be thinking, okay, well, what's your answer to all this? Well, you know, you can no longer just look at ratios, PEs, dividend yields, debt levels, profit margins. All of that really, to some degree, is first-level thinking. And one thing that I think you have to be careful with now is really think about, okay, how strong are these brands, and are they just based on consumer perceptions, or are they based on a specific product? 

Just think that AI can create a brand in a matter of minutes. It can come up with advertising. It can come up with a logo. What AI cannot do is create the impact that it takes to actually build a brand, but it can probably speed that up. You have to step back and say, “Okay, are these brands gonna hold up like they used to? 

Are these moats gonna be able to remain like they have been?” Or at the moment, and this is where I think it's second-level thinking, in some cases, the market's probably overreacted, where they view 'em as dinosaurs, but some of these brands, I think, are still gonna be worth some money. PepsiCo, Coca-Cola, Dr Pepper, it's gonna be pretty hard for another competitor to come out in a major way. 

All the restaurants that use either Pepsi or Coke or Dr Pepper, it's gonna be really hard to replace that. You know, will AI help PepsiCo in distribution? Yes. Will AI replace the drink Pepsi? That one gets a little harder to really conceptualize. McDonald's, I would say, is one example. Strong brand. 

Everybody knows exactly what you get when you go into a franchise. The product is very consistent. Well, how does AI really replace that? There are efficiencies that can definitely come out of the business, but they have a franchise network. They have something that consumers, to whatever degree you wanna call it, trust. 

That is something that AI can't replace overnight. So brands are one thing. Looking at, you know, actually making stuff, something like that concept is something you may wanna own going forward. 

In today's world, we personally are starting to think more towards companies that have infrastructure, supply chain, production. It's just harder for AI to disrupt them, and if anything, they could be big positives. It potentially lowers their costs, improves their distribution, but it doesn't replace the business. 

So we're gonna get into a few examples of how we look at that right now because, going forward, as you've listened to our podcast in the past, and it's really the very heart of it and why it exists, is how are you gonna get dividend growth going forward? 

Because the world is changing, and it's changing fast. 

So now let's get into a few examples of how we've seen it working for ourselves. And the first one, we've talked about this in the past, and it's Intel, and we sold it out of the model portfolio a few years ago because they cut the dividend, and we are a dividend growth story. 

But fortunately, we left it in virtually every other account because first-level thinking was they're struggling, they're not a competitive chip maker, they couldn't get their speed up with some of the competitors out there. But one of the big reasons why we decided, you know, we're gonna hang on to this thing is because they are basically a U.S.-based chip maker, or they're in very friendly political areas, where Taiwan Semiconductor, they are trying to rapidly move to more of a U.S.-based model, but that's not where they are now. 

So it took a few years, but it turned out that the U.S. component of Intel became much more valuable than the market placed on it just a few years ago. It took a lot of patience. It was dead money. The stock had really turned into a losing proposition for a while, but the core story was still there. They had a new CEO come in. 

It became a little bit of a bet, you know. Can this thing get turned around? Not only did it turn around, but the thing just rocketed up and made all of that patience pay off. 

One of the problems with second-level thinking is that you're almost always early, and sometimes if you're too early, then you're really wrong. 

We were down... Some of the stuff that we had purchased earlier, we were down 50% or more. It was taking several years for the story to really play out. You never know it's actually going to work until it does. If you're not patient and you're not willing to really get invested in why you own something, second-level thinking is not the place for you because it takes a lot of patience. 

To some degree, it takes a lot of guts. You know, Intel, did we get lucky? Well, you know, the U.S. component, which is why we held it, what was luck was how fast and to what extent it did start to work. Who knew the U.S. government was gonna take a position in Intel? They wanted to make sure that the company survived. 

The U.S. government normally doesn't invest in companies, and we can argue all day long whether that was a good thing or a bad thing, but the reality is it was because we wanna own something that's in the U.S., and it worked. 

So now we're gonna move into another, I think, a great illustration of first-level, second-level thinking, and you can maybe guess what it's gonna be on. 

It's a company that we've talked about a lot recently. It's one of our highest personally conviction trades. It's one of our larger positions now, and that is Accenture. You might say, “Why are you talking about something that you're down anywhere from 25% or more on?” 

Investors right now seem to be taking a first-level thinking mentality on Accenture. They are worried that AI is gonna replace consultants. Why pay Accenture to go out and do something that AI can do faster and cheaper? On just the basic foundation level, I would probably be out looking for another job if I just consulted on certain business practices or whatever. 

But Accenture has pivoted, and they've reinvented themselves on some level. They've gone from just really a contract consulting business to going full steam ahead into looking at how AI impacts an enterprise and how to integrate it in an entire organization. 

If you go to second-level thinking, even though divisions or certain parts of companies may do great at implementing AI, from an enterprise standpoint, how fast AI potentially moves, companies may find it virtually impossible to keep up, and that will and is going to become Accenture's focus. 

The U.S. government, I think, is just kind of the extreme example. So in the last year, year and a half, there's a big move to cut costs in government spending, and one of the first things and easy targets to go was Accenture's U.S. government contracts. It was about 10% of Accenture's business, so it had a pretty decent impact. 

But if you go out to second-level thinking, what is probably the most inefficient organization out there? The government is so large and so complex, they're gonna need someone like an Accenture, one of the only ones that's gonna be able to do it on a global enterprise basis, to go in and start to tie AI together. 

I would guess that AI could probably take out half the employment of the U.S. government right now. So you wanna cut government spending, AI is potentially gonna do it. Now, to what degree should it do it or how fast? That's a whole 'nother question. But it's just ripe for some major gains, and who's gonna step in and do that? 

And then think about this, okay, so AI is rapidly developing. It's changing extremely fast. 

You're going to probably need somebody like an Accenture to come in and to be able to continually stay up to date with how fast technology and AI is changing, and it's not too hard to imagine that most of corporate America is not gonna be able to develop that talent internally. 

Even though the stock is down, big enterprises are gonna need a global approach to how AI works beyond just simple solutions in different parts of a business. Could be wrong, but we are committed to holding Accenture. 

Well, we're gonna move from the creating world to the world of actual physical things. 

And my question to you is, you tell me, if you had to pick a company or an industry, okay, what basically has no threat from AI? What would it be? I really find it hard to believe that AI is gonna... all of a sudden there's gonna be a whole army of robots that are gonna start building railroads across the country. 

Basically, a railroad cannot be replaced. What AI does do is it makes the railroad much more efficient. It potentially allows maybe the trains to run closer together. The technology improves the cars, the whole distribution, how they're linked together, how they are loaded, their routes, and the railroads are already using it. 

But first-level thinking in Union Pacific is, well, they're a railroad, they're boring, the growth is relatively slow, and you're not gonna have major technological breakthroughs there. But as we've talked about before, they have actually beat the S&P 500 long term. Second-level thinking is, these things will continue to be great stories because the only thing that AI does is make it more profitable. 

It maybe makes it a longer-term story because it makes them even more competitive with trucking and some of the other sources of transportation out there. 

To take it one step forward, another stock we've talked about is UPS. This one gets a little bit more into the gray area, but I think the outcome is still somewhat the same. 

First-level thinking is there could be autonomous trucks, drones. You know, there won't be UPS drivers anymore. Well, second-level thinking is the big part of their cost basis is routing these trucks and routing the packages that they're distributing. Somebody's gotta run the network. Somebody's gotta warehouse the stuff. 

Whether it's a drone, a truck, or whatever, somebody has to own that infrastructure, and there still has to be access to global logistics. So they're still moat businesses, but they are going to evolve. 

And here's just a few thoughts of first-level, second-level thinking. 

GE Vernova, originally just tied to utility infrastructure, and utilities are a boring, regulated business, basically slow growth. 

And then lo and behold, we've got AI data centers consuming lots of power, and GE Vernova is really sitting at the heart of grid infrastructure and how to power the AI economy going forward. 

Chevron. Will AI eventually take out oil? When you start to go into second-level thinking, it's gonna take energy and natural gas to power these AI data centers. You have to have redundancy with wind or solar. We need oil for plastics, for construction, and AI helps in trying to find oil. It potentially will probably lower the impact on the environment. 

Lockheed Martin and General Dynamics. First-level thinking is defense stocks are politically vulnerable. 

Government contracts can be cut or renegotiated, or in peacetime, we tend to cut spending because they're an easy target to save money on. But second-level thinking, at least in my mind on these defense stocks, AI is just making the world a more dangerous place, and now virtually everybody has access to phenomenal technology. 

Johnson & Johnson and Merck. Um, this one gets a little bit more complicated, but first-level thinking is AI potentially takes out some of the research and the value of these drug patents because it makes it easier to come up with new concepts. Startups could become a bigger threat, but second-level thinking, it just means that these larger companies are gonna be able to accelerate their drug discovery. 

They're the ones that are the experts on regulatory approval, that have distribution systems, have relationships with physicians. They potentially can get stuff into their pipeline faster or may even benefit from some of these smaller companies that come in and partner with these larger drug companies. 

So there are businesses out there that are actually only going to become better and potentially more profitable. Ironically, where some of the best performance has been over the last several years could become a challenge. And without really consciously doing this, we're looking at our model portfolio and a lot of the things that we own, a lot of times they're simple businesses. 

Some of those are probably the best defense that AI can benefit, but they're not gonna replace. 

So as we wrap this up, the key point to take away is AI is changing the game, and everybody knows that. It may seem obvious, but you just don't know the future. You have to be thinking about, okay, I can't just go out and buy a stock because it has a high yield or the valuation is cheap. 

The historical parameters where you purchased a stock in the past quite likely are not gonna hold into the future. It's gonna be more dynamic. It's gonna take a little more thought into, okay, how are we gonna get sustainable dividend growth going forward? What are the threats to it? How do these companies adapt? 

That's a huge piece, 'cause some of these companies just go out and make acquisitions trying to buy growth. Are they really addressing the long-term problem, or are they just looking for a short-term solution? Brands, technology, and moats as we've seen them in the past will likely see erosion in the new AI era. 

But logistics, infrastructure, you know, actual physical products may and will benefit from AI. 

From a standpoint of being a dividend growth investor who applies second-level thinking, start trying to think past what immediately seems evident. What's the current yield? What are all the numbers, PE? And start to think, “All right. How is my investment gonna continue to grow its income, and where's the threat? Maybe this is not the threat that I thought, or maybe I need to reevaluate here, and there could be a problem.” 

AI can both make dividend growth a great story going forward, or it can turn it into a disaster. Just make sure that you're aware the game has changed. 

If you enjoyed today's podcast, please leave us a review and subscribe. If you would like more information regarding dividend growth or our investment strategy, please visit growmydollar.com. There you will find previous episodes and also our monthly newsletter. If you have any questions or anything to add to today's episode, please email Ethan, E-T-H-A-N, @growmydollar.com. 

Past performance does not guarantee future results. Every investor should consider whether an investment strategy is right for them and all the risk involved. Stocks, including dividend stocks, are volatile and can lose money. Denewiler Capital Management may or may not have positions in the publicly traded companies mentioned herein.