Loading summary
Transcript114 lines
- [00:00]
A
Do you remember that pretty famous story about ChatGPT telling a lawyer that a bunch of fake court cases were real? So if you haven't heard about this, ChatGPT basically hallucinated a bunch of fake court cases and the lawyer presented it at court and well, that did not end well. Given the AI tends to hallucinate, why would anyone trust AI to analyze their investments? Well, that's a rhetorical question, and my guest today has some answers. Because my guest today has figured out how to turn to ChatGPT and other types of AI from a highly imperfect machine that hallucinates into a legitimate junior financial analyst that only uses SEC filings. Today's conversation is going to center around how to use AI responsibly when you analyze your investments. Today's guest is Brian Feroldi. He is a financial educator and author. He's written over 3,000 articles for the Motley Fool. He is the author of the book why does the Stock Market Go up? He runs the YouTube channel long term Mindset, which has about a quarter million followers. And he has more than 600,000 followers on Twitter where he teaches investors about how to analyze businesses. So he talks about how to read income statements, balance sheets, cash flow statements. And today he joins us to talk about how to use AI as your junior analyst. Welcome to the Afford Anything podcast, the show that knows you can afford anything, not everything. This show covers five pillars. Financial psychology, increasing your income, investing, real estate, and entrepreneurship. It's double I fire. And today's episode is all about that first letter I investing. Most of you, including myself, are index fund investors, which is amazing. So if you're thinking to yourself, hey, if I'm an index fund investor, why should I care about analyzing individual stocks? Well, a couple of reasons. Number one, even if you're 95% in index funds, many of us have a little bit of fun money in individual stocks, like some Apple shares that you bought back in 2015. So that's one reason. The other is that learning how to actually evaluate whether or not a company is worth your investment, that that is a skill set that has tremendous value. Because if you know how to evaluate businesses, then you as a business owner, as an entrepreneur, as a person who works at a company, you understand how businesses run. You understand what makes businesses profitable, and that makes you more valuable to your company, either as an employee or to the company that you might one day start and run. Learning business valuation learning company analysis is a highly adaptable and transferable skill set. And today Brian talks about how AI can be your partner in that. So with that said, here he is, Brian Feroldi. Brian, welcome.
- [03:05]
B
Thanks for having me. Paul. Awesome to be here.
- [03:07]
A
Thanks for being here. Can you tell us about the pros and cons of using AI to choose stocks?
- [03:12]
B
I am a fundamental investor. I have been researching stocks the traditional hard way for two decades now. And to me that means reading SEC filings, reading earnings reports, listening to conference calls, reading analyst reports, all that kind of thing when, when AI kind of exploded onto the scene. I guess we're in year three now. Initially I was a little put off by it because it seems so new, but over the last year I've really tried to adopt AI in my everyday life and, and what AI is so good at, it's taking huge amounts of data, huge amount of quantities of data and compressing it and explaining it to you in an easy to understand way. I think if the listeners here have used AI, and I'm guessing 100% of them have at this point, they're kind of familiar with the pros and cons of AI themselves. On the pros, it's really good about getting information fast and digesting it. On the cons, if you've used it for any amount of time, you, you know that it tells you incorrect information and it has an optimism bias built into it. So from a pro perspective, I think that AI can be an invaluable tool for fundamental focus investors like me to help them analyze companies and analyze stocks. But you have to put in strong guardrails in place because of that bias that's built into ChatGPT to be so positive and to, and to essentially make up stuff. So as long as you use ChatGPT the right way or AI the right way, I think can be a wonderful tool.
- [04:42]
A
All right, I've got some follow up questions both about hallucination and about the optimism bias. But before we get there, I'd first like to establish you are largely an index fund investor. Yes.
- [04:51]
B
Yes, I have a good chunk of my net worth. All of my retirement funds are in index funds. The bulk of my net worth, about 70% of my net worth is in individual stocks.
- [04:59]
A
Oh, really? 70%? Why is it that you're comfortable with such a major proportion of your portfolio.
- [05:05]
B
In individual stocks from an overall asset allocation perspective? All the studies that I've shown and all the data, the long term data is clear. Equities is the place you want to be in the long term. So I just believe that in my core, the reason that I have gone, the individual stock route is because I'm the type of person that likes to dive into the details. And when I first Learned of investing 20 plus years ago, I thought that picking stocks was the way to go. And I just, for whatever reason, the idea of analyzing companies and becoming a stock picker really fascinated me and really took hold in me. And I recognize that that is not for everybody. I'm the type of person that would read SEC filings for fun. Like to me, that would be more fun to analyze a company that I'd never heard of than to watch a Marvel movie. So I'm not up to date on the latest Netflix shows or anything like that, but I just really like the details of investing. And the reason it's become such a large portion of my portfolio isn't because I set out to do that. It's because my investments have done so well over the last 20 years that the individual stocks that I have honed as a group have come to dominate my net worth.
- [06:09]
A
Oh, okay. So in terms of a cost basis, how did it originally start from a cost basis point of view?
- [06:15]
B
Yep. So I tried to do it as tax efficiently as possible. So I've always tried to max out all of my tax advantaged accounts, my 401k, my Roth IRA, et cetera, for both me and my wife. Anybody that has a 401k, which is a majority of the way that most people invest into the markets knows that one of the downsides to a 401k is you're very restricted on what you can do. You are very restricted to the options that the company that you are working for give you. So I tried to invest in as low cost away as I could through those vehicles. So I put them into index funds whenever that was an option for me. And just as a broad category decision to make my life as easy as possible, I made the decision all of my retirement funds are going to be in index funds. Done. Easy, simple, no thinking, dollar cost averaging. Wonderful way to invest for the long term. But with the active portion of my portfolio with the capital that I wanted to put into the market that was beyond the limits of 401 and Roth IRA that I chose to be more tactical with and to go the individual stock route. That was my broad capital allocation decision making.
- [07:19]
A
Let's come back to then. AI we know that AI hallucinates. Sometimes it can hallucinate over even the most basic of details. Given that reality, is it really prudent to turn to AI for any investment decision generally, much less individual stock picking?
- [07:39]
B
Yeah, your first instinct is totally correct. You should be skeptical. After all, this is your money on the line, right? This is your hard earned money that you're putting at risk. I don't like to put my capital at risk unless I'm absolutely convinced in the fundamentals of the investment that I'm making. Given that AI hallucinates. This is one reason why I was super skeptical for years about using AI to help me make these decisions. Simply because I did not want to be making investment decisions based on faulty information. Right. That'd be like building a skyscraper on a foundation built on quicksand. That would not be a good, a good strategy. However, as I've studied AI and I've talked to some people that are way better at AI than I am, I think that there are some simple rules that you can adopt that will dramatically increase the quality of the information that you get back from AI. So much so that if you follow the structure that we're going to talk about, I now have full confidence in the information that I'm getting back from AI.
- [08:37]
A
Okay, let's talk about some of those. How is it that you put those guardrails in place?
- [08:40]
B
Yep. Rule number one to me is to insist that whenever ChatGPT or any AI gives you information back that you force it to use original sources that you trust. So as an investor, one original source that I trust is the sec, the securities and Exchange Commissions. So one of the rules that I put into the prompts that I built is you are only allowed to source information from the securities and Exchange Commission, or you're only allowed to source information from sources that I personally trust, very few sources. So for me, that's the SEC, that's company filings, and then there's a few third party sites such as, like Morningstar.com where I'm analyzing a company's competitive advantage. But if you just first and foremost insist on the sources that the prompt use, that alone will dramatically increase the quality of information that you get back.
- [09:30]
A
Would another way of doing that be to find certain sources? So, for example, to download the SEC filing PDF, it, upload it and say, do not use any information from the web, only use information that comes from this PDF that I have uploaded here.
- [09:47]
B
Absolutely. Anything that you can do to restrict the information that AI is pulling from to a source that you trust, again will dramatically increase it. In fact, we're going to go through this prompt later. But one of the things that I insist on in the prompts is whenever you present with a number, you also have to simultaneously Give me the link to the information that you got that number from. So when I'm reading through the prompt and when I'm actually referencing a number that's pulling up, I I can click over to that original source and if I want to double check where the number came from right in the prompt itself. So if I'm like, okay, it's saying this company did 50 million in revenue. If I click over the SEC filing and I scroll down, I can verify that the SEC filing gave me that number. When you see that, then it's like, okay, I trust this information.
- [10:31]
A
Do you go through and methodically check each one or do you randomly spot check?
- [10:34]
B
So it depends on the information that I'm looking for and the quality of the information that I'm going for. So, so I primarily use ChatGPT as my AI and I use this mostly for high level overview of stock. So if I'm analyzing a company for the first time or I'm analyzing an earnings report on a company that I own, it's wonderful for turning over rocks, if you will, for like, is this a good company? Is this a filter to use? Because it can do that so, so quickly. If I was going to actually buy it, there's deeper analysis that I would want to do and with different sources. One source that I use is called Fiscal and that's a website that pulls directly from SEC filings and helps me chart and graph all kinds of numbers and stuff like that. So I think AI is a wonderful source for a first pass into getting to know a company. But if I'm going to go deeper on it, there's other sources that I use.
- [11:20]
A
All right, tell me more about the prompts. So you've talked so far about restricting the source material that AI uses. Beyond that, what else do you do within the prompts to reduce the risk of hallucination?
- [11:33]
B
Another thing that I do is I give the prompt exact step by step instructions that I wanted to do. So, for example, let's say I was researching Apple's stock. I wouldn't Type in the ChatGPT is Apple A buy? Right. Like that is such a broad statement. And that is that gives ChatGPT or any AI way too much freedom to interpret that. I think a good mental model to get in your mind is you have to think of AI like a junior analyst or a junior intern that is eager to do whatever you tell it to the instant that you tell it to. If you were hiring an intern to come in and you said something like, go research this guest and that was the only information you give them. They would go off and comply and come back with information. And it might be in a format that is completely incoherent to you. If you gave that same intern the instructions for, I want you to start by going to this person's LinkedIn bio. Then I want you to write down where they went to school. And then I want you to go up and pull up an interview they just did. And I want you to write down the questions that they had. And you gave them step by step instructions for how to do the thing. Obviously, the information that you're going to get back would be so much better because it's in a format that you want and it's following the exact instructions that you are giving it. ChatGPT and AI works the same way. So in the prompts that I built, I say, step one, use these sources. Step two, go to this part of the 10K and look up this information. Step three, answer this question. Step four, answer this question. Step five, after all this is done, give it to me back in this format and in this order. So by insisting on it follows a step by step function that I created. The information that I get back is exactly in the order that I want it to be in.
- [13:17]
A
You know, the comparison to a junior analyst makes a lot of sense. Or a junior an intern. Because so much of the time, you know, the challenge of hiring and training is being able. You know, ideally you hire and train because you're busy and you need to offload work, but it gets worse before it gets better because you have to do so much work up front in terms of preparing a clear set of instructions for any juniors.
- [13:44]
B
Yeah. If you've ever hired an assistant like that, I like to think of it as like 10, 80, 10. And as the broad framework, 10% of the work is just creating the instructions for what you want to do. That is on you to do that. You the employer or you the boss, you have to do that first 10% to say, here's what we want to do and here's the instructions to do it. The 80% is the actual going through and executing on the instructions. Right? That's what AI is designed to do. Right. If you give it a good framework, it will go out and do that hard work for you and come back with information. Your job is then to take that information, do the final 10%, which is interpretation, the results that you're seeing. So AI can be wonderful with handling the middle 80% the same way a virtual assistant would be good with handling the middle 80%, but you have to give them good instructions, and then you have to check their work that's still on you.
- [14:29]
A
Right. And then part of that final 10% would also be iterating on the first 10. Right. Based on, you know, giving feedback and iterating on that first 10% based on how well they've done in that middle 80.
- [14:41]
B
Yeah, absolutely. And I know that you've hired people before, and so you know the quality of the information that you give them or the standard operating procedures that you tell them to follow, the more clear you can be with that, the less errors there are, the less communication there needs to be back and forth, and the happier you are with the output of the work.
- [14:56]
A
Right, Exactly. All right, let's continue going through some of the prompts. Let's see the process for how you shape these instructions.
- [15:03]
B
So one key prompting technique, and this works for anything, not just for researching companies. AI works really well when you start out by assigning that AI to have a role. So, for example, if you're going to be an investor, or one thing that you can tell the AI to do is if you're a value investor, say, act as Warren Buffett, or if you're a growth investor, you can say act as David Gardner, recent guest on your podcast. Or if you're going to be a short seller or diving deep into the financials, you can say, act as a forensic accountant or act as a financial analyst. Just by giving it that instruction upfront to put it in the framework, suddenly, the information that it will pull from in its database, it is going to be restricted to that role that it has. So again, if I say to the AI, act as Warren Buffett, it's going to instantly know what the word moat means. It's going to know to emphasize things like competitive advantage and return on equity and a strong balance sheet and long duration of compounding. And it's going to reference all of the data that's out there on Warren Buffett and his investing style. So just by giving it that simple prompt upfront to assigning it a specific role, that will restrict the data that it can use, and it will dramatically increase the quality of the output that you get.
- [16:13]
A
Now, what I see here is an inherent conflict in asking it to be Warren Buffett while simultaneously not necessarily wanting it to pull from any source online, including garbage sources. So what's going through my mind is maybe uploading a bio of Warren Buffett as part of that PDF and saying, act as Warren Buffett based only on the information that's in this PDF, do not use any outside sources besides this PDF. Would that be a way to do it, or would you trust it to have at it to interpret who Warren Buffett is and what his philosophy entails?
- [16:47]
B
Yep. If you wanted to take the next step and kind of prompt it with a role and a really specific role and put that in as the very first prompt that you do, that can be one way to do it. From what I've found, just putting into the prompt itself, just giving it a role upfront and saying, act as Warren Buffett or act as a financial analyst, to me, that does the job. But if you want to go the next step and give it a more detailed profile so that it's using that. That more power to you.
- [17:09]
A
Hmm. Okay, so assign it a character role. What else?
- [17:13]
B
So step two is, like we said before, ground everything with trusted sources. So come up with a list of trusted sources for it to pull data from. I don't think you could go wrong with just saying the securities and Exchange Commission or company filings or company earnings reports. That will dramatically cut down on hallucinations. And you can also, if you want to go the extra level, put an extra step there for it to verify the information that it's getting. I've done things before like, say, something like, pull up the most recent 10Q. 10Q is a quarterly report that comes out that companies issue three times per year when they're reporting earnings. And sometimes it pulls up a 10Q, but it's not the most recent 10Q. So just by saying things like, the first thing is to start with, what year is it? Or what day is it? And then say, pull up the most recent one. Sometimes ChatGPT can get a little bit tripped up unless you give it literally that specific of instructions. But assign it a role and give it the sources to pull from. That's a great foundation.
- [18:08]
A
Okay. Assign it a role, give it the sources to pull from. Be incredibly specific and clear in your questions and your instructions. That's step one and step two.
- [18:17]
B
Yep. And then step three is to give it a stepwise structure to go through. So a step one, a step two, a step three, and break it down in specific sequence that you want it to do as if you were explaining it again to an intern, someone that had never done this before. So for me, that would be things like step one, acquire the data using these steps. Step two, verify that the data sources are accurate. Step three, now that we have that established, go through these specific questions in order and step four, report them back to me in this formatting. So 1, 2, 3, 4. The clearer you can be with the instructions you want it to follow, the better the information you will get back.
- [18:57]
A
How do you know? So I'm thinking about some of the people who are listening to this who are primarily index fund investors because we have many people who listen to this are, myself included, are predominantly index fund investors. Many of us are not super acquainted with stock analysis to begin with, even absent AI. So if that is your situation, how do you know enough to know the steps that you want and the formatting that you want?
- [19:23]
B
Yeah, great question. My analysis shows that there's about 60 million Americans that own individual stocks. So even if you're the type of person that has primarily put all of your assets into index funds, which I think is a fabulous strategy and should be the default strategy for 99% of people, even in my own friend group, people that do that, they often also own Nvidia or they also own Apple or they even own like our friend Brad Barrett died in the wold index investor, yet he also owns Berkshire Hathaway and he also owns Markel. So even him, who has been a huge proponent of index funds, still owns individual stocks. Now, as long as it's a small part of your portfolio and you're not risking your retirement on it, that's a perfectly fine thing to do. Some people call that fun money, some people call that play money, or they just want to learn about it. But if you're the type of person that's going to be investing in individual stocks, which many people do through stock based compensation, through the company that they work for, I think it behooves you to always start by educating yourself. Learn about the fundamentals of the company, learn the basics of reading financial statements, learn how to tell what makes for a good company. You don't have to go as deep as say Warren Buffett does, or as deep as I do, or as deep as professional investors. But if you can just give yourself a grounding education in the basics of business. How does the company make money? What are its key products or service? What are its growth prospects? What the heck is valuation? What risks do I think of? The more you can understand those basics, the more confidence you can have that owning that individual stock is a good choice.
- [20:58]
A
You know, and it's true many people have some proportion, you know, a fun money fund where they can buy individual stocks. But I think a lot of people are vibes investors. They heard about something at the proverbial water Cooler and on a lark, threw some money at it and it either worked out or it didn't. And if it worked out, then people can develop overconfidence and that can later, of course come back to bite them. So I think many people get, you know, without the grounding in how to conduct fundamental analysis on a stock, I think a lot of people can, can start as Vibes investors. Either it fails or it succeeds, which is sometimes worse success, which leads to an even bigger failure ultimately. And then ultimately people just sort of get burned.
- [21:47]
B
Yep. And to me, if you're the type of person that bought a stock because a friend told you about it and you did zero research on it, that's not investing.
- [21:55]
A
Right.
- [21:56]
B
That is just pure gambling. And if you look at the stock market, most people associate the stock market with gambling because they don't understand the absolute basics about what stocks are, what they represent, why their value changes over time, how the market moves in cycles, and what is it that causes stocks to appreciate or depreciate over time. To me, those are fundamental things. So I don't consider just knowing the name of a ticker, buying it and then hoping it goes up to be investing. To me, that is just pure gambling. Now there's nothing wrong with that if that's what you want to do. As long as you know that what you're actually doing is you're gambling, you're not investing. To me, investing is buying shares in a company that you have researched, approaching that company with a ownership mindset, with a long term mindset, tracking the results of that company to make sure that the thesis, the reasons you bought it are intact in the first place. And owning it, as long as those reasons are still in place. To me, that is investing, right?
- [22:51]
A
So that then goes back to the construction of the prompt. Do you have sample prompts that people can use? Step three is create steps. And I think a lot of the people who are listening are like, I don't know enough to know what steps to create.
- [23:03]
B
I have a prompt. If you go to longterm Mindset Co, which is my website, Backslash AI, I do have a free prompt that people can Download. It's about 2,000 words long. So it's a pretty meaty prompt. And what it'll do is, and we can certainly go through this, it's a prompt that follows all of the techniques that we've laid out. And what it's designed to do is to go into any business that you name that's publicly traded and it will answer seven questions. That I think are foundational questions about the business, such as, what does the company do? What are its main products and services? What countries does it operate in? How frequently do customers purchase that product? What happens in a, in a recession, and can the company raise prices? Questions like that, which are things that I want to know as just foundational information before I would do any analysis on top of that. So, yeah, if your listeners want to get a free copy of that prompt, just go to link. I'm hopefully in the description or something.
- [23:54]
A
Perfect. With regard to those foundational questions, one thing that strikes me is, okay, let's say I know what I ask those foundational questions and I learn what countries that company operates in. How do I contextualize that information? So let's just use a hypothetical company and we'll say it operates in North America, Malaysia, Singapore, Indonesia, and its basis of operations are those two regions. What do I do with that information? Is that good? Is that bad? I don't know.
- [24:23]
B
Well, now you know more about the company than you did before. For example, if you were to do this analysis on a company like Domino's, I think most people, when they hear the word Domino's, would assume U.S. pizza Co. And they sell pizza. If you ran this prompt, I think you would be surprised to learn that the majority of Domino's revenue comes from sourcing pizza. So selling pizza supplies to their franchisees in the United States, and they have a very strong presence in international markets. So just by doing this prompt, you would learn more about Domino's, the company, than you might assume because you're a consumer of that company's product. In fact, many companies that are in The United States, McDonald's, Starbucks, Nike, in many cases, those companies get more revenue outside the United States than they do inside the United States. And once you understand that if the US Market is not doing well and you're worried that a recession is coming, you can be like, okay, well, I know that this company actually gets the majority of its revenue from China or from Japan or from an international market. And even if our country is doing well, they that doesn't mean sales in those other countries aren't doing well. And the company's results will be maybe better or at least buffered from that fact. Or you have to think about, well, if this company gets 80% of its sales from Brazil, then all of a sudden you have to worry about exchange rates, right? This company's results are going to be heavily influenced by the relationship between the US Dollar and I think it's Brazilian real Real. Okay. That's important information that you should know if you're going to be investing in a company. So it's not to say that you should act in one way or another, but I think it's just absolutely foundational that you know where companies get their revenue from. That. That to me is essential information to know.
- [26:02]
A
You mentioned if you're worried about a recession in the US in an increasingly globally interconnected world, a recession in the US Would have spillover effects that could lead to recessions globally or downturns globally at a minimum. I mean, we saw, I think the ultimate example of that would have been 2008, which was truly a global. I mean, and it's when people call it the global financial crisis. I think part of the reason it has that name attached to it was because it was the most global crisis of its kind. And it was sort of the crisis that really cemented how very globalized all of our economies are.
- [26:39]
B
Yeah, absolutely. And you're right that if the US Isn't doing well, that typically does not spell good things for other countries around the world. And the inverse could also be true that the US Economy might be doing great, but China might be going through tough economic times. Or Brazil or Argentina or the UK or Europe. Right. Different countries have different economies, and while they are largely integrated with each other, they can move in different directions. So again, one of the things that I want to know when I'm researching a company that's built into this prompt is what happens to this business in a recession or what happens to this company's revenue during tough times? There are some companies that their revenue and profits drop precipitously when the economy goes downward. Auto sales, for example. If the US Is in a recession, people delay buying a new automobile. It's also very highly sensitive to interest rates. So automakers like Tesla, like Ford, like gm, you can expect that those companies revenue are going to fall dramatically if the economy, if the US Is doing very, very poorly. Other companies like Walmart or Dollar General, for example, those companies actually do better. Their businesses improve when times are tough in the US because consumers are trying to stretch their dollar further, trade down, and they shop at places like Costco or Walmart or Dollar General, so their sales actually improve. So again, that's kind of foundational information that I want to know what has happened to this company's revenue in past periods of economic stress so that I know if I'm investing what.
- [28:09]
C
Kevin Harlan here. This Black Friday prime is taking over with an incredible day of live sports Black Friday football is at the center of it all with the Chicago Bears taking on the Philadelphia Eagles at 3 Eastern. Immediately after that, it's an exciting Emirates NBA cup doubleheader. The Bucs will take on the Knicks at 7:00pm, then the Mavericks will be in LA against the Lakers at 10. It's the final night of knockout rounds as teams vie for a spot in the quarterfinals and the whole day starts on the links with the Capital One skins game as four elite PGA golfers face off with $4 million on the line. Black Friday game day tees off with the Capital One skins game at 9am Eastern. Then it's Black Friday football with the Bears and Eagles at three, and it all culminates with the doubleheader of NBA basketball, the Emirates cup action featuring the Bucks at Knicks at 7 and the Mavs at the Lakers at 10. And it's all only on Prime.
- [29:17]
A
It's time for Black Friday. Dell Technologies biggest sale of the year. Enjoy the lowest prices of the year on select PCs like the Dell 16 plus featuring Intel Core Ultra processors with built in advanced features. It's the PC that helps you do more, faster plus earn Dell rewards and enjoy many other benefits like free shipping price, match guarantee and expert support. They also have huge deals on accessories that pair perfectly with your Dell PC and make perfect gifts for everyone on your list. Shop now@dell.com deals Kevin Harlan here this.
- [29:52]
C
Black Friday prime is taking over with an incredible day of live sports. Black Friday football is at the center of it all with the Chicago Bears taking on the Philadelphia Eagles at 3 Eastern. Immediately after that. It's an exciting Emirates NBA cup doubleheader. The Bucs will take on the Knicks at 7:00pm, then the Mavericks will be in LA against the Lakers at 10. It's the final night of knockout rounds as teams vie for a spot in the quarterfinals. And the whole day starts on the Lynx with the Capital One skins game as four elite PGA golfers face off with $4 million on the line. Black Friday game day tees off with a Capital One skins game at 9am Eastern. Then it's Black Friday football with the Bears and Eagles at three. And it all culminates with the double header of NBA basketball, the Emirates cup action featuring the Bucks at Knicks at 7 and the Mavs at the Lakers at 10. And it's all only on Prime.
- [31:04]
A
All right, we prompt the AI and that includes these seven foundational questions that really give us a backbone of understanding what this company does, where it operates, how it makes money, how it operates in different Economic cycles. What next? What's step four?
- [31:19]
B
Well, so step four is run the prompt and analyze on the company and read and start to interpret the results. And this makes for horrible audio, so I apologize to everyone that's listening, but perhaps we could run this prompt on a company that you know or you're interested in and we can kind of talk through the results.
- [31:34]
A
Perfect. Okay, let's do it.
- [31:35]
B
Great. So I have. For people that are watching on YouTube, here is the prompt I'm going to be reading through. So at the top it says the identity is acting as a financial expert by doing business model analysis from SEC filings. I gave it a clear mission, which is to prompt the user for the company, a name. I gave it clear execution triggers for looking up the SEC filings in an order. I gave it seven questions that I wanted to go through and answer those questions specifically. And then I gave it the instructions for here's how I want this information to be presented back to me and I insist that there's sources on all the information and more. So I'm just going to push enter, run that into a chatgpt. Is there a company that you know fairly well or have, Want, want to analyze?
- [32:19]
A
There's a company that I actually know almost nothing about that I'm curious about and it's Cava.
- [32:25]
B
Cava.
- [32:26]
A
They are a fast, casual Mediterranean food eatery. So imagine like a Chipotle of Mediterranean food.
- [32:35]
B
Yep. Great way of saying that. If anybody's familiar with Zoe's Kitchen, Zoe's Kitchen was a company. Cava actually bought Zoe's Kitchen out and rebranded all their names to Cava. So I know a fair bit about Cava. Many people are calling it, quote unquote, the next Chipotle. If you don't understand that reference. Chipotle has been a sensational long term investment since it was spun out of McDonald's in 2005. I think something like that. But yeah, let's do Cava.
- [32:57]
A
Yeah, I don't know anything about it. I've actually never even eaten there.
- [32:59]
B
Oh, highly recommend it. I think you'll love it.
- [33:01]
A
Okay. I have long wondered if it's going to be seeing how quickly these locations pop up. I've long wondered, is this the next Chipotle?
- [33:08]
B
And it's been a very hot stock since it came public. I ran the prompt and ChatGPT is saying it understands the prompt and it says what company name or ticker? I like to include both. Would you like me to analyze? So all I did is type in Kava Here and I pushed enter. So ChatGPT is going to do this thing and I'll read from the top and we'll see how it does. So it says using Kava Group. So the name of the company is KavasGroup 10K as of December 29, 2024. And it's 10Q from Q1 and Q2 of 2025. So right at the top here, the information that's going to be using are from SEC filings that are the correct year. So that's a good start. So question number one that I have this prompt answer is what does this company do? So let's read the results. Cava is a US based fast casual restaurant chain focused on Mediterranean inspired cuisines, bowls, pita wraps, salads and related grocery store products. It operates its own restaurants over the Cava brand and it also sells Mediterranean dips, spreads and dressing into grocery channels. Now as part of this prompt, Paula, I can see that there's a link to an SEC filing right next to it. So if I can be like, well did it just make that up? If I click over to that filing, it actually pulls up the most recent annual report and this is where it got the information from. And if I want to click down to the business, here's a link to the what the business is looking at. So this is where it came from and I can verify that. Yes, that is indeed accurate information. So that's again how I have confidence in this. Right next to that information it says here's the SEC filing I got this info from.
- [34:46]
A
And if people want to visually follow along, we're showing all of this on our YouTube channel. So YouTube.com afford anything for anybody who wants to visually follow along with this.
- [34:57]
B
All right, so question number two I like to ask is how does this company make money? We know that it does something with restaurants. How does it actually make money? Those two things might be completely different from each other. So ChatGPT says revenue comes primarily through its restaurant operations. The grocery retail product line is mentioned but not separately broken out. And it's public summary I could locate and according to last year restaurant revenue, it says I cannot give you an exact figure because it isn't broken out in the summary. But I found but the company it noted it opened up 58 new restaurants. All right. Digital revenue was 38% of total revenue. And the grocery retail business was referenced in an article but not in the 10k summary. So this is actually, it's saying I couldn't find this information directly in the SEC filing. So at least we know it's not making up the results. But we do have the annual report here that we can pull up. Oh, it gave me a link directly to Kava Group's first quarter results as of 2025. So if we read through here, we can see that it reported 328 million in revenue in the first quarter and we can at least get some information on there. So kudos to ChatGPT for not making stuff up.
- [36:03]
A
Right.
- [36:04]
B
That was not as helpful as I hope it was going to be with listing out what the revenue and sales are, but at least it gave us the sources that we could look it up ourselves. Okay, question number three, I want to know is who are the customers? And it says Kava's customers are primarily individual customers purchasing meals at its fast casual restaurants, walk in digital orders and delivery. And it also has grocery customers for spreads and dips. That's pretty straightforward.
- [36:30]
A
Very straightforward.
- [36:31]
B
That's pretty straightforward.
- [36:31]
A
Like any other fast casual chain, like most Southwest Grill, like Chipotle, like any place you go get lunch. Sweet greens.
- [36:38]
B
Yep, for sure. The next question where does it operate? Kava's operations are US centric. In Q1 of 2025 they operated in 26 states plus DC and geographic revenue breakdowns are not clearly provided in the summary. But at least we know it's basically a US based company and it's only in half the country as of right now. All right, question number five. This is when I get into some qualitative questions. So now that we know what the company is and what it does, I like to know some detailed questions about the company's business model. So one of the questions that I love to ask is how often do customers buy? As a general statement, I like to invest in businesses where customers first frequently make purchases from the company so they have a subscription based revenue or it's a consumable product. And I don't like to invest in companies where you buy one time from the company and then disappear for periods of years. So ChatGPT says many of Kava's restaurant transactions are repeat visits by consumers. The company reported same restaurant sales. It was up 10.8% in the most recent quarter, including guest traffic plus 7.5%. It's also expanding via new store openings which includes one time customers in that transaction. And in the retail grocery product line, the purchase frequency will depend on the grocery shopping habits. But yeah, so I would say that that's accurate if you're like most people, if you like a restaurant, you tend to not only buy from that company one time, but you do so every couple of weeks or so, Right? Next question I like to ask, can the company raise prices? If you've studied Warren Buffett, you know that his number one thing that he looks for in a business is pricing power, the ability to pass prices onto consumers. So if your prices increase, you can pass those along. That's called pricing power. Not every business can do this successfully. So I like to know, does ChatGPT think that this company can raise prices? And it says yes, there is evidence for Q1 of 2025, same restaurant sales growth was same 7.5% increase in traffic and 3.3% from menu price increases. So this is ChatGPT saying in the SEC filings, the company reported 3.3% increase in prices and traffic still went up. So that's a pretty strong sign that this company can increase prices if it needs to.
- [38:56]
A
Right. Can you give an example of companies that might struggle with pricing power? Companies that aren't able to pass their direct costs on to consumers even when those costs increase?
- [39:06]
B
Sure. Anything that's a pure commodity. So any product that you would buy and you do not care about the company behind that product, you only want the lowest cost price for that. Toothpicks. Yeah, toothpicks. Exactly. Or paper plates. I mean, you think about the grocery store anytime, you are not specifically brand loyal. So if you don't care what kind of cereal you eat, you just want, say, let's say you like Cheerios or whatever. But you're okay with the generic version of of Cheerios if they can't convince you to pay for the brand name Cheerios. But if you're happy to buy the generic version of Cheerios, that would be a company that does not have pricing power. Another way of thinking about this is some companies don't control the price that they set at all because it's dictated by the market. For example, if you're an oil company and you produce oil, you don't set the price of the oil, the market sets the price of the oil. Same for gasoline, same for gold or silver. Any commodity producer, they do not have control over their pricing. That's set by the market.
- [40:00]
A
Right. Actually, when you first mentioned pricing power, the first thing that came to mind were rental properties. There's a range in which you can price your rental property. You know, you could sort of be at the, depending on your level of finishes, you can be at the bottom to the top of a reasonable rent range, but you can't really go outside of the top of that range, unless you have a very, very specialty property. So for the most part, housing is a little bit of a commodity.
- [40:26]
B
Sure, yeah, it's definitely a commodity, but there are commodity components to it. But I would imagine I'm not a real estate investor like you are, but I would imagine location matters tremendously. The finishes matter tremendously, the quality of the building, the exterior siding, the landscaping, all that kind of stuff. I would imagine the nicer that is, the higher quality tenant you attract and therefore the more pricing you can extract from them.
- [40:47]
A
Right, exactly. And that's why I say there is a reasonable rent range for, let's say a three bedroom, two bath house in the 12345 zip code. Right. There's a reasonable range that you can be in depending on all of those factors. But oftentimes when you're looking for homes, if you're talking to a real estate agent who's not being a good guide and who's just trying to sell you something in order to close the deal and collect a commission, oftentimes you'll hear agents say things flippantly like, oh, you know, I don't know what the water bill is, but if it's high, just charge them more. That type of statement assumes a great degree of pricing power. Landlords simply are restricted in the pricing power that they have.
- [41:29]
B
Yeah, that's actually a great point. I know a lot of your audience is real estate investors. And if you're the type of person that's buying an individual property, you're not just looking at the Zillow listing and purchasing buy. Right. There's a huge number of details that you're going through before you choose to pull the trigger. What's the quality of the place? What's the location? What are the historic rents been? What's the tax? Right. How much leverage do I have to put on this for this to make sense? What kind of upgrades are those? Are all you knowing the details about that property before you choose to make an investment? To me, that's exactly how I think smart investors approach individual stock analysis. They don't just know the ticker and the price of one share and say buy, buy or sell. They get into the details and they want to answer questions like this. The more you can know about the investment, the better. Does the quality decision you make?
- [42:12]
A
Right.
- [42:12]
B
All right, so the last question I have here to answer is what happens to this company in a recession? So again, if tough times come, which they are guaranteed to, I like to invest in businesses that I think can grow Even during a recession, that's best case scenario. Or at the very least they won't be harmed that much during a recession. Those are the type of investments that I like to make. I do not like to invest in companies with highly cyclical revenue. So ChatGPT says. Well, as a fast casual restaurant, Cava is exposed to consumer discretionary spending. In Q2 of 2025, management noted that consumer traffic softened amid economic uncertainty. And although same restaurant sales in prior periods were strong, the Q2 results of the most recent result of 2.1% growth and the downward revision to full same store Sales guidance from 4 to 6% from 6 case indicates sensitivity to weaker consumer demand. Therefore, in a recession, traffic may drop which will hurt performance. I think that that analysis is highly likely to be accurate.
- [43:17]
A
Right. And would that same analysis apply to any fast casual chain?
- [43:21]
B
Yes, I totally think so. But then I also think about my own personal behavior. I do like to apply that to it. I'm a regular consumer of Chipotle. Like I like Chipotle's kind of food. And when I think about the various dining options that are out there for family, I have family of five. Chipotle, by comparison to going to a sit down restaurant is pretty affordable. So if tough times came, I might change my behavior to go to places like Chipotle or Cava more and go to sit down dining restaurants less. So you do have to think about those things too. But there's no doubt that if there was true tough economic times, many people would skip restaurants altogether and just eat more at home, which is the cheapest thing you can do.
- [44:02]
A
Right? Right. And what you're describing is substitution. In the case of a recession, there would be a certain class of consumers who will substitute a sit down experience with servers for fast casual. There will also be a different class of consumers who substitute fast casual for fast food or for more groceries at home.
- [44:21]
B
Yep, absolutely. And some restaurants are more insulated from this than others. For example, as we Talked about previously, McDonald's sales tend to increase during periods of economic stress. As people say, I can't afford Chipotle or Cava, but I can't afford a value meal at McDonald's. So there are companies like that that their sales do go up, even if they are restaurants and stuff like that. But again, if you're going to be an investor, I think it's really important that you think through the exercises of what happens to this business during a recession. The same way I would hope you would think through as a real estate investor. Well, what happens if unemployment increases in this area? Is my place good enough that I can still attract renters?
- [44:58]
A
Right.
- [44:59]
B
So that is the prompt and that is one of the first prompts that I do when analyzing any business. And in just a couple of seconds, if I've never heard of a company before, I can get a pretty good gist about what the company does, where it gets revenue from, what happens in a recession, and and I think it's a really good starting prompt for analyzing any company.
- [45:17]
A
Excellent. All right, so now that we know this about Caava, where do we go from here?
- [45:20]
B
Well, so we can decide this is the kind of business that interests me and I want to dig further into it or I'm not interested at all and I would just toss it aside and just go on to the next idea if I was interested in this business. I have a whole bunch of other prompts that I've built that helps me analyze the competitive position of the company, the long term growth trajectory of the company, the management of the company, the risks of the company, the valuation of the company, all of which are critical to the making an informed investing decision. Again, I have a series of prompts that I've built for myself beyond this, but I think that this is a great prompt to start with because it gives you an overview of the company.
- [46:01]
A
We've all set health goals, exercise more, eat better, lose weight. But without a plan they often fade. That's where Prolon comes in. Its five day fasting mimicking diet gives you a structured approach to stay on track and see real results. Prolon offers an injection free way to get help to lose weight, get help to burn fat while protecting lean muscle and maintain results in just five days. It can help activate fasting pathways to help support metabolic health and rejuvenate cells, all while letting you enjoy real food. Prolon is a plant based nutrition program featuring soups, snacks and beverages that are designed to help nourish the body while keeping it in a fasting state. If I were looking for something like this, this is what I'd use because Prolon has been shown to support metabolic health, fat loss and energy. And NextGen builds on the original Prolon with 100% organic soups and teas and ready to eat meals for a limited time. Prolon is offering listeners 15% off site wide plus a $40 bonus gift. When you subscribe to their five day program, just visit prolonlife.com Paula that's P R O L O N l I f e.com Paula to claim your 15% discount and your bonus gift. Prolonlife.com Paula these statements and products have not been evaluated by the Food and Drug Administration. These products are not intended to diagnose, treat, cure or prevent any disease or condition. Fifth Third Bank's commercial payments are fast and efficient, but they're not just fast and efficient. They're also powered by the latest in payments technology, built to evolve with your business. Fifth Third bank has the big bank muscle to handle payments for businesses of any size, but they also have the fintech hustle that got them named one of America's most innovative companies by Fortune magazine. That's what being a fifth third better is all about. It's about not being just one thing, but many things for our customers. Big bank muscle fintech hustle. That's your commercial payments A fifth third better Are you hosting people for the holidays? The holiday season's right around the corner and if you're hosting people, you're going to need bedding and sheets. You may need some cookware. Of course you're going to want holiday decor. While Wayfair's Black Friday sale is the perfect time to score huge deals on everything for your home, no matter what your style is, you can get 70% off during Wayfair's Black Friday sale. The Black Friday deals run all month long. Wayfair has everything you need for your home, every style, every budget. We're talking bedding, bath, big stuff like furniture to detail stuff like pillows. They have an absolutely massive selection of home items. Living room, outdoor areas, bedroom, you name it. I have lots of shelving from them, shelving in the entryway, shelving in the bathroom. I have a daybed from them that's very multifunctional. It has a built in bookshelf, it has drawers underneath. A great space saver. Get your home ready for the holidays. Don't miss out on early Black Friday deals. Head to Wayfair.com now to shop Wayfair's Black Friday deals for up to 70% off. That's W A Y-F-A-I-R.com sale ends December 7th. Tell me about the prompt. Specifically you mentioned you have a prompt for analyzing management. So much of management are people, it's characters. It's the talent and skill of the individuals involved. And that is incredibly difficult to assess as an employer. It's hard even when you're sitting face to face with somebody to assess their skill level. How do you do that from a distance?
- [49:43]
B
It can be tricky to do and you might think, wow, you really have to get into the soul of somebody to judge what kind of person that they are. But to me, it's a fairly straightforward process. When analyzing a management team, you might think that's really impossible to do or it's hard to do. As an investor, one of the things that I like to do when I'm analyzing a company from afar is just do a couple of quick checks. I think it was David Gardner that came up with the framework, but he called it the Oats framework. Oats O stands for ownership. How much of the company's stock does the manager or the CEO of the company own? As a general statement, I want owner operators. I want management teams. That network is going to be impacted way more than my network is going to be impacted if the stock goes down. So you can set up a prompt that checks for their executive compensation. How much stock do they own? How many millions of dollars of stock do they personally have invested in the company? As a broad statement, I want that number to be very high because if the stock price goes down, their net worth is going to be impacted way more than than mine will be. A stands for allocation. CEOs of companies are making capital allocation decisions. They're choosing to invest in R and D. They're choosing to open up new locations, they're choosing to make launch new products and services. And you want to check their historic track record for allocating capital. And does that lead to increased revenue growth, increased profitability? Do they have a history of launching new products and new services that drive the company forward, that grow the company over time? The T stands for tenure. This is how long have they been doing the job for? If they were ideally the founder of the company, they've been with that company since day one. That to me is best case scenario. And there are still plenty of public companies out there that the founder is still in charge. I love to see that. I don't like it when the CEO is a hired gun. They came from some other company and they were brought in and they're likely attracted to the high compensation or the prestige of the job. So the longer they've been at that company, in my view, the better. And the S stands for stewardship. So how well do they treat employees? How well do they treat suppliers? How well do they communicate and treat shareholders? You can actually measure those things by looking at the returns of the stock over time. You can go on places like glassdoor, indeed, LinkedIn and see what kind of reviews the employees give to the CEO. Another one that I like to do is just Execution track record. I want to invest in companies that have as a history of making proper acclimation, saying we're going to grow 10% or whatever, and then actually growing the company 10 or 10% or more. So there are some things that you can look at from an outside observer that can give you a real good sense of who is the person in charge and are they capable of taking the business forward.
- [52:31]
A
Mm. The last thing that you said, the proclamations retrospective, I am a huge proponent of that because we see so little of that. I did an episode earlier where I talked about how we see very little of that in the financial media where people will often make predictions, but they never do the retrospective to evaluate how their predictions went. And that's why they say, you know, you've called 20 out of the last two recessions. And so the proclamations retrospective I think is essential. How do you do that when it comes to a management team? But you want to curtail the sources that you're putting into AI because you don't want the AI to hallucinate. How do you make sure that the AI is mining valid data when it's doing a proclamations retrospective?
- [53:14]
B
Yep, that's fairly straightforward when it comes to judging public companies. They report numbers every quarter. And the two most important numbers that Wall street looks at are revenue and earnings per share. Those are like key terms that most companies operate by. And you simply look at how does the revenue and earnings per share compare to prior periods. And if that management team gave guidance, which some companies do, which is when they say, in the next year we're going to grow 15% or in the next year we're going to report $2 in earnings per share. Well, analysts track. What were the actual numbers that were reported compared to the estimates that analyst teams were putting out? As a broad general statement, management teams that have a track record of consistently exceeding, exceeding Wall Street's estimates are rewarded in the market. And management teams that have a history of underperforming Wall Street's estimate have very poor stock performance. So that's actually a fairly easy thing to do and look up numerically to say, did they beat estimates or did they underperform estimates?
- [54:13]
A
Right. Okay, so you're using purely their formal guidance in their filings.
- [54:17]
B
Yep. If they do give guidance, you can use guidance or you can use analyst estimates. That information is. Is very easy to look up nowadays.
- [54:24]
A
Okay, now, outside of the management team, you talked also about assessing the valuation of the company. Can you describe some of the prompts that you've created in order to do that. Because when you're looking at valuation, you as an individual capital allocator, you know, you're trying to decide if these $10,000 should go to company, to Kava versus Starbucks versus some biotech firm. Given the wildly different risk profiles, opportunities, how as a capital allocator do you design prompts that can assess not just what the public valuations of the company are, but that can assess the best use of your dollars given your own set of risk tolerance goals, you know, individual considerations.
- [55:13]
B
So I'm going to take those as two separate questions. One question is how do you tell if the company is a good value, what the valuation? That to me is the first step. The second step you have to do is given everything I know about the company, the risks, the business, the growth potential and the valuation, what should I do with this information as a capital allocator? So one prompt that I have analyzes the stage of the business's life that it's in. I'm a huge believer that you should not use the same metrics to analyze all companies. One analogy I look at is think about businesses have distinct phases in their life. Would you judge a three year old, a three year old human by its SAT score? That would of course be ridiculous because the kid is too young for that kind of metric to be valuable. You have to think the same way when it comes to public companies. A company like Kava that is rapidly growing and rapidly expanding throughout the US I would not necessarily use the price to earnings ratio, which is a very common ratio that investors use to judge valuation to analyze a company like Kava, simply because Kava is in growth mode, it is rapidly opening up new stores and because of that, its financial statements are not fully optimized for profits. So the price to earnings ratio would be a poor metric likely to look at to analyze a company like Kava. If you're talking about a biotech company, for example, many biotech companies lose money for years or even decades because they're investing everything into research and development with the hopes of getting that drug to market. So it's very difficult to value those companies. And I would certainly not use multiple metrics like price to sales ratio, price to earnings to try and value them simply because they're in a different phase of their lifestyle than in another company. On the flip side, a mature company that is profit optimized, like Apple or like Meta or like Walmart, those companies, you can look at their price to earnings ratio and make a valid decision because they are fully Optimized for profit. So one of the things that I have the prompts do is they analyze which phase of the business growth cycle is the company is in. And then depending on which phase it's in, which valuation methods are most appropriate to determine whether that company is overvalued, fairly valued or undervalued. And then from there, let's say it comes back and it says, according to this valuation metric, the company is undervalued. That's obviously best case scenario if you're hoping to purchase it, then you take everything to account as a capital allocator and say, okay, this is a business that I like, this is growth potential that I like, this is a management team that I like, and it's trading at a low valuation. I'm willing to allocate some capital to it because I think that I'm going to get a higher return on the risk reward spectrum for making this investment versus another one.
- [57:54]
A
Right.
- [57:55]
B
Again, the same way that you would analyze any real estate portfolio. Right. You would consider, well, what are prices in say, Manhattan apartments versus and what's the rent to buy ratio in Manhattan versus, you know, the suburbs of Alabama? I would bet those markets are incredibly different from each other and the numbers make sense, sense in one and they wouldn't in another market. So the same way that you allocate capital when you're buying properties, same exact mindset to buying stocks.
- [58:20]
A
As you've developed out the series of prompts that you have. And people can download these prompts?
- [58:25]
B
Yes, the business analysis prompt that we discussed before. Yes, that one you can download.
- [58:28]
A
Yep. Okay. If people who are listening want to construct their own versions of these prompts, what advice would you give them?
- [58:34]
B
Take the free prompt we just mentioned and just study the structure of it and look at the execution triggers. Look at the name I assign upfront or the role that I'm assigning up from. Look at the stepwise function. And if you want, you can just copy and paste that and say, help me adapt this to analyze a company's competitive advantage. Help me adapt this to analyze the company's valuation. Help me adapt this to analyze a management team. Or if you're into prompt engineering and you want to develop them own, you certainly can use that. But I think the structure, the structure is a really important thing here for people to look at.
- [59:03]
A
The structure is important to look at, and it sounds to me like the curtailing of sources is critical. With structure and curtailing of sources, we reduce the risk of hallucinations, but we still don't fully account for optimism bias. And we know that that is baked into a lot of AI. And where that can be particularly difficult is if we ourselves are also as individuals prone to optimism bias. Then you've got two entities with optimism bias talking to one another, which can then become this unspoken unknown. Unknown. Right. How do you adjust for that?
- [59:39]
B
So I think as long as you give it the stepwise structure and you give it the reporting system, that the output that you get is in a format that you want, it's not inherently going to be doing this is a great idea or this is a terrible idea because what you're asking it to do is to go out, find information from valued sources and report it back to you in a format that is useful. I don't think you should outsource your analysis to AI. I think AI is wonderful for gathering information, for presenting it to you in a format that you think is correct. And then it's up to you, the investor, to do that last bit of analysis, to take that and say, is this a good company or is it a bad company? Is this a good buy or is it a bad buy? Should I put this on my watch list or should I forget about that company? So that's how I combated it. If you want to go the next step, when we're assigning a role, you could say, take the exact opposite approach. So act as a short seller, right? Do that same thing and say, what would you say about Kava? And that way you're telling the think negatively about this company, point out the reasons why you wouldn't invest in it or short selling. For those that don't know is when you are betting, you make money when the stock falls. So by prompting it, by giving it the role of think of a short seller, it's going to by definition, go out and look for everything that it could possibly go wrong.
- [60:49]
A
And so fundamentally, you're assigning it the devil's advocate role. And that as a step in the process, makes a lot of sense because you want to stress test all of your assumptions functionally. By assigning it the role of a short seller or a devil's advocate, you are baking into your process as part of that checklist. Let's stress test this by hearing every counterargument.
- [61:13]
B
And that's exactly what good investors do. They think of what is the bull case or what is the upside scenario if everything goes right, what is the bear case or what could happen if things go wrong. And if you're an investor, you know, it's Always a good idea to spread your bets and to diversify. I would never put all of my capital in into one stock, no matter how positive. One of the things that I insist on in the prompts is whenever you present with a number, you also have to simultaneously give me the link to the information that you got that number from. So when I'm reading through the prompt and when I'm actually referencing a number that's pulling up, I can click over to that original source and if I want to double check where the number came from right in the prompt itself. So if I'm like, okay, it's saying this company did 50 million in revenue. If I click over the SEC filing and I scroll down, I can verify that the SEC filing gave me that number. Right for one personality is not the right thing for another personality. There are 20 year olds out there that are scared to death of volatility. And the idea of their portfolio declining in value would just keep them up at night, right? Like investing through 2008, 2009 or 2020 was not fun. Like it was not. It's not a fun period to watch a huge part of your net worth decline in real time in front of you. And just some people aren't emotionally capable of seeing that and dealing with that. If you're the type of person that if your net worth fell 30% in a matter of weeks because of a recession, perhaps you should really cut back in your stock allocation and put more into less volatile assets such as cash or bonds. And that might not be the optimal thing from a long term growth of your portfolio, but it might be the right choice given your personality. I could use my mom for a second. My mom is a type of person that's scared to death of volatility. She just is scared to death of it. So she invested in her 401k for a couple of decades and always put it into CDs. Cringing is the right thing to do, right? But for her, the thing that gave her pleasure from that was in 2008, the value of her portfolio went up and she slept soundly at night knowing that her portfolio was safe. Now, I tried to point out to her, well, do you know what your value would be if you invested it in the stock market versus keeping it safe? And she said, I don't even want to think about it. That's not a fit for my personality. And to me, while that might sound like an abhorrent asset allocation strategy from me or you, that was the right asset allocation strategy for her. So Yeah, I haven't explored that deeply with AI, but I can be good about teasing that kind of thing out about yourself.
- [63:37]
A
Let's say that you have the personality type like your mom, where your inclination is to put everything into CDs. You want to play it safe, but you want to feel differently. You want to prompt yourself to take on more risk. Can you design an AI prompt that will help you become more of the investor that you want to be? And I use in this example, you're too conservative and you want to take on more risk. But it could also work the other way around. You're a little too reckless and you want to be corralled in.
- [64:08]
B
Yeah, that's a great idea. And it sounds like you're giving me a good homework assignment here, which is almost to come up with a prompt that takes you through an investor assessment of form where it asks you a series of questions and you fill those out and that can help you to tell you what type of investor you are and perhaps some asset allocation strategies to think of. There's lots of ways to invest and what's mathematically correct by the academic definition may not be right for you given your personality. There are certainly young people out there that have overallocation to bonds because they don't want to deal with volatility. And there are certainly 80 and 90 year olds that are investing in a high risk, long term 100% stock simply because that's a bit fit for their personality. So that, that might be a great prompt to build or something that helps you to discover what type of investor you are.
- [64:49]
A
Okay. If you did want to build a prompt, people listening at home who want to build their own investor assessment prompt, do you have any guidelines around prompt construction? Best tips?
- [64:59]
B
Yep. So I would say give it a roll. If there's an investor that you would admire or is there a capital allocator that you admire. You can even just say act as a financial advisor, right? What, what kind of questions would you want to know from me? And just say act as a financial advisor and what kind of questions would you want from a new client for them to figure out their asset allocation? AI is wonderful as a tool for coming up with prompts to prompt AI, right? You just have to ask it the right questions and you have to be willing to iterate again and again and again. And sometimes building the prompts like the prompts that I've built have taken me hours upon hours upon hours of refinement to put in them to get them the way that you want. But I used AI the entire way to help me build those prompts. So yeah, give it a roll and then ask it to prompt you, the user, with what type of questions that you think it would ask. That's a great way to start.
- [65:46]
A
Thank you for these tips and thank you for spending this time with us. Where can people find you if they want to learn more?
- [65:50]
B
If you're on a social platform, just type my name in Brian Feroldi and I'm likely to be there. And if you want to see these prompts in action, I have several videos on my YouTube channel which is called Long Term Mindset, where I go through them in detail.
- [66:06]
A
Thank you. Brian. What are three key takeaways that we got from this conversation? Key takeaway number one AI can be your junior analyst for stock research, but ultimately the decision making does have to be up to you. Brian talks about how to use AI effectively for analyzing investments by treating it like a junior intern who needs extremely specific instructions. And the key to good use of AI is constraining AI, limiting it, for example, to only using very trusted sources like SEC filings, and then also giving it detailed step by step instructions. Because the thing is, you want to make sure that you're preventing hallucinations and if you don't give it those kinds of constraints, it might hallucinate.
- [66:55]
B
I think a good mental model to get in your mind is you have to think of AI like a junior analyst or a junior intern that is eager to do whatever you tell it to. The instant that you tell it to. If you were hiring an intern to come in and you said something like, go research this guest, and that was the only information you give them, they would go off and comply and come back with information. And it might be in a format that is completely incoherent to you.
- [67:23]
A
That is the first key takeaway. Key takeaway number two Force AI to show its work. Every single number that you see. The AI needs to show you exactly how it got to that. Make it provide a clickable link to the exact SEC filing for every single number that it gives you.
- [67:43]
B
One of the things that I insist on in the prompts is whenever you present with a number, you also have to simultaneously give me the link to the information that you got that number from. So when I'm reading through the prompt and when I'm actually referencing a number that's pulling up, I can click over to that original source and if I want to double check where the number came from right in the prompt itself. So if I'm like Okay, it's saying this company did 50 million in revenue. If I click over the SEC filing and I scroll down, I can verify that the SEC filing gave me that number.
- [68:14]
A
That is the second key takeaway. Finally, key takeaway number three. The quote, unquote, best investment might not be the optimal one. Sometimes the approach that is mathematically rational might not be right for your personality. So Brian shares a story about how his mom kept all of her money in CDs and missed out on a bunch of stock market gains. But she also slept soundly during the 2008 crisis, and that fits her personality. She is very afraid of volatility. She doesn't want to put her money at risk. And it's true. There's a famous quote that more money has been lost in anticipation of recessions than in recessions itself. And mathematically we might know that. We might be able to pull data that shows that. But that doesn't change how people feel. Feelings turn into behaviors that become suboptimal, and so you can't rationalize your way into taking on risks that don't suit you.
- [69:18]
B
My mom is the type of person that's scared to death of volatility. She just is scared to death of it. So she invested in her 401k for a couple of decades and always put it into CDs. Cringing is the right thing to do, right? But for her, the thing that gave her pleasure from that was in 2008, the value of her portfolio went up and she slept soundly at night knowing that her portfolio was safe. Now, I tried to point out to her, well, do you know what your value would be if you invested it in the stock market versus keeping it safe? And she said, I don't even want to think about it. That's not a fit for my personality.
- [69:51]
A
Weigh your own personality, your personality, your risk tolerance, your ability to sleep soundly at night. Weigh all of that when you're making your investment choices, because that stuff matters. Those are three key takeaways from this conversation with Brian Feroldi. Thank you so much for being part of the Afford Anything community. If you enjoyed today's episode, please do three things. First, share this with friends, family, neighbors, colleagues. Share it around the Thanksgiving table. Share it with the people in your life. Because that is the most important way that you spread the message of investing. Which is the first I in fiire. Second, please open your favorite podcast playing app and your second favorite and your third favorite, please open all podcast playing apps on your phone. Hit the follow button so you don't miss any of our amazing upcoming episodes. And while you're there, please leave us up to a five star review. Write a few words, tell us what you enjoy about us. That means so much and it also helps us book really amazing guests. So thank you so much in advance. I appreciate you doing it. Third please hang out with other members of our community. Affordanything.com community it's free and it's a great place to meet like minded people and talk about the things that that are on your mind as it relates to money. Are you thinking about retirement? About debt payoff? About investing in the stock market? About the use of AI? Talk about what's on your mind and you'll meet like minded people in that space. Affordanything.com community and 4th please subscribe to our newsletter affordanything.com newsletter it's free and we share ideas, concepts, updates that we don't share anywhere else. Totally free. Affordanything.com Newsletter thank you again for being an afforder. I'm Paula Pant. This is the Afford Anything podcast and I'll meet you in the next episode.