A (0:00)

Welcome to Money for the Rest of Us. This is a personal finance show on money. How it works, how to invest it, and how to live without worrying about it. I'm your host, David Stein. Today is episode 555. It's titled Five Practices that have Shaped My Career. Tomorrow I'm speaking to a graduate finance class at the University of Arizona. It's something that I do about once a year. The professor invited me to share what I've Learned across a 30 year career and as an investment advisor, an author, an entrepreneur. And as I thought about it, I've narrowed it down to five practices. The first I learned really my first day at university in high school, the teachers would write all the notes on the board and I would just copy them down. That first class, the professor just started speaking and I what do I write down? But I even asked the girl sitting next to me like, well, at least I watched what she was doing, that she was writing some of the stuff down. And so I started taking notes. That's a practice that I continued obviously throughout my undergraduate, my graduate, but I have continued to do that throughout my professional life. I started by writing down in hardcover Moleskine notebooks. And I would fill one up, I'd write the next one down. But problem with Moleskine is yeah, you can kind of go back and review, but you can't really search that. And so when technology came along where you could record online in somewhat of a wiki format, that's what I did. Starting with social text 2005, I used backpack by Basecamp, now known as formerly 37signals. That was in 2007. 2008 I started using Evernote and used primarily Evernote and Basecamp for our business as well as backpack up until 2019 when I started started using Roam Research for most of my notes with the exception of things that have to do strictly with organizing our business. And we continue to use Basecamp. I also have kept a written journal since the seventh grade. Now I generally just write a summary post by hand once a year. So in thinking about these, this note taking, I've not used Evernote in a long time, but I have all these notes and I figured out today how to export them as HTML and then I use Claude Cowork that basically formatted the HTML into a markdown file that I could import into Roam Research. So hey, there's something AI could do format an HTML file. So then I was reviewing the notes, which is why it's such a helpful practice occasionally go back and I found a note from 2009 on panarchy, I was in the process, I was reading a book by that topic, lance Gunderson and C.S. holling. I was writing our would have been our quarterly market commentary strategy report at FEG Advisors. And I was introduced to this concept of panarchy and how it describes how complex systems change over time. Everything's connected, but you have different layers. You have microbes that, that live and die very quickly. You have trees that live longer. I was recently visiting where I grew up in Cincinnati, and it felt familiar my my neighborhood and drove by the house I used to live at because some of the trees were still there. They've been there now 50 years or more, probably close to a hundred years because they were big even when I was small. So they change much more slowly. And so you have these higher up levels in the panarchy that are persistently there. So in that 2009 note I quoted from the book on Panarchy and it says, because the template formed by the panarchy is so remarkably conservative and persistent, animals can develop rules for actions that take advantage of that persistence while retaining enough flexibility to adjust to variability and the unexpected. And they called them rules of thumb or schema. And it minimizes the information we need, but it guides us in our decision making. Because these upper layers of the panarchy, where there's more persistence, where the changes don't happen that much, there's some predictability there, not complete predictability, but enough to have rules of thumb in order to continue to survive and really to thrive. And some examples they give is people learning to drive. First it's difficult, but then if you sort of drive the same way to work, a lot of it just happens intuitively. These are rules of thumb. And it's because of the persistence of the roads, a higher layer of the panarchy. I mean, there's aspects to roads that move very quickly. The individual pieces of gravel can shift if you're driving on a gravel road, but the road itself, those are hard to shift, those are hard to change the pathway. They developed over decades, if not in this case over hundreds of years. And so we can rely on them now. Things happen as we drive unexpected and we have to adjust. But we can use rules of thumb. They talk about insects and birds. A lot of their rules become genetically encoded instinctive behaviors. A beautiful example that they gave was the term is milpa M I L P A and that is the maize culture in Mexico. And I was, I became very familiar with that when I lived in Mexico, down in the Yucatan. And as we've gone back, Milpa is essentially the process. They call it a script, really an internalized plan that people in the Yucatan and other areas in Mexico and Central America use to guide their behavior, their activities in terms of growing corn in a way that they. They shift where they grow it over time, and they do it in a way that renews the. The forest ecosystem. So they won't grow, continue to plant in the same spot. They'll move it around. If you travel out in the Yucatan and some of the outside villages or whatever, you'll see areas along the road that there's a bunch of weeds growing, but you can see the old corn crops. And so this script is a series of activities, but there's also room for experimentation and to adjust, and they become very localized. But it's an example of an encoded script, rules of thumb, that we use in a particular activity. And these five things I'm sharing in this episode are really that type of script, rules of thumb. It's a pattern that I have followed, and it gives me persistence, but also flexibility to see what comes up. And so that first one is to take notes, to continually review what I've said. I've done this for decades now. There could be notes on books I'm reading, I'll save newspaper articles or other periodicals. And then I can increasingly use AI to search them if I choose. Although, honestly, I tried using Claude, I connected it to Roam Research, and it's a little slow, just easier to use the embedded search functionality within Roam or whatever if you're using Evernote or whatever platform. And the point is, it doesn't matter the platform, as long as it has an ability to export, which they do, and occasionally to back it up, you can have a corpus of notes over decades that you've taken, which is what I have. Before we continue, let me pause and share some words from one of this week's sponsors. Delete me. DeleteMe makes it easy, quick, and safe to remove your personal data online. At a time when surveillance and data breaches are common enough to make everyone vulnerable, it's easier than ever to find personal information about people online. Your address, phone number, family member names are all hanging out on the Internet. And that can have actual consequences in the real world, makes us all vulnerable. And that's why I continue to use Delete Me in my personal life. I'm someone with an active online present, so privacy is important to me and I really appreciate that. DeleteMe is out there scanning the Internet looking for my personal information, requesting that it be removed. And then they send me a quarterly report showing what has been removed from the web. The New York Times wirecutter has named Deleteme their top pick for data removal services. So take control of your data and keep your private life private by signing up for Delete Me now at a special discount for our listeners. Get 20% off your delete me plan when you go to JoinDeleteMe.com data David20 and use promo code David20 at checkout. The only way to get 20% off is to go to JoinDeleteMe.com David20 and enter code David20 at checkout. That's JoinDeleteMe.com David20 code David20 for those notes. I'm getting continual input, which is the second thing. Get continual input. Feed your mind. We do that by traveling, we do it by reading. I think one of the most important things and I learned this in college, one of my first finance classes. We had to subscribe to the Wall Street Journal and read it and we actually had a lesson on how to read the Wall Street Journal. This was back when it was a newspaper paper. It had set sections and he went through this is, this has this section. This is how that works. This is, this works. This is how you read a stock table. Now, I consume all my periodicals on my iPad, including the Wall Street Journal, but to me, the best periodical. If you're in finance or you're interested in investing, I prefer to get a broader, more international exposure. So the Financial Times is my top choice for a ongoing periodical, especially if your emphasis is finance. Second would be the Economist. But I mean the Wall Street Journal is still solid and I subscribe to a number of periodicals but find at least one read it regularly, take notes and save articles, clip articles, just save them somewhere. And the act of doing so helps us remember we can also subscribe to various services areas that curate information for us that, that are grasping a lot of things from a variety of sources and curating them. They sharing their view. I've used Paul Kadrosky, the former Wall street analyst, but he's been on his own for decades now and we're friends. And he does a very good job of just summarizing what he's seeing. What's interesting. Used to do it on Twitter, now he does it. He does a little, it's like 10 bucks a month and I'll just see. He's and he's incredibly smart and it's been very helpful, especially figuring out what's going on with AI. Gregor McDonald is another one that I've used in the energy space. A lot of the input in college, I would just wander the library randomly. I'd pick up various academic journals and look at the articles. I definitely would look at the ones in the Journal of Finance journal Portfolio Management. And I was just feeding my mind into getting ideas and I would just. I would walk all over the library and just pick up books randomly that looked interesting. So we need that input. We take our notes. And from that, something I've done consistently is experiment, try things, particularly on what I call the leading edge of the present. With this concept of panarchy. Gunderson and Holling point out that there tends to be greater variability along the edges. It could be along the edges of an ecosystem, perhaps somewhere that's been disturbed, where new invasive species can start to take hold. It could be on the edge where you're starting to see extinctions. And so we're not trying to predict the future, we're trying to get to the edge of the present, where there's a lot of change, a lot of activity and experiment there. We want to see where, where there's some momentum, some, some trends and use our curiosity. And as I go back over the decades, when I first started as a credit analyst, there was no World Wide Web. This would have been the early 90s, but there was Usenet. There were these sort of these online bulletin boards and I got access to them at my firm at ncr and I started just exploring that. And then when the first browser came out, I wanted to get a version and they said you couldn't, you couldn't access it. I pretty close to the tech guy, yeah, no, I will not install a Netscape browser on your computer. But we eventually got to it. And then when I joined FEG in the mid-90s, then the Internet, World Wide Web's going. We started using email and I started buying URLs, like, I didn't know what I was going to do with them. But they're a form of optionality. You buy a URL, maybe I'll use it. And some of them I do. I now have 43 URLs that I own. And some of them we bought like acid camp. I bought that probably 13 years ago when we finally started using it. And I bought the URL names for all of our kids and they've used them periodically over the years. I built my first website to sell our house in 2004 when it became easier to do it. And eventually WordPress and some of these other things. I launched my first blog in 2005 and a lot of these were side projects. I launched a site called Real Time Reviews that was scraping well. The idea was to scrape reviews, sort of candid reviews of hotels and restaurants and aggregate them and put them on a website that didn't work out. I launched my first personal finance blog in 2008, shut it down for compliance reasons. I uploaded my first video that I recorded in 2008, launched YouTube in 2012, the podcast in 2014. A lot of these. I had no idea if there was going to work out. We did a mobile app in 2016 and eventually we launched software as a service. We'd never developed software before. I could do the same thing. When it comes to crypto, there was all experimentation, seeing what was happening on the leading edge of the present at feg, that's what I was always doing. And sometimes I got a bad rap because I was always trying stuff, or my attention span was kind of short. But it was that experimentation that got us to launch our outsourced CIO product. It was called Managed Portfolios. At the time, we used ETFs, but now that's the bulk of the firm revenue because that's where the industry eventually went. Now, we were early, I was experimenting, but it is that experimentation that has led to opportunities and basically has built my net worth, because some of the things worked out, many didn't. I am not a successful YouTuber, but I've been doing YouTube off and on for 14 years. We'll get to understand the probabilities in a minute, because there's a reason YouTube hasn't worked out. I'll get to that. But think about what's going on today with AI. I was at the. I must have been February 2023, I was speaking at a CFA dinner at. I think it was at University of Arizona, and somebody asked about AI, and I remember sort of dismissing it at the time as just basically a word predictor I didn't really understand. Now, this would have been just a few months after the first public version of Chat GPT, and I. I just didn't get it. But I was trying it out. I finally got it six months later. But I've continued to experiment with it, Both first with ChatGPT, now with Claude, now other large language models. I'm using Claude, Cowork. But it isn't just to use them, it's to understand what they're good at and what they're not good at, what their weaknesses are. The fact that most large language models are stateless. They don't know what day it is, they don't know what time it is, they don't keep persistent memory and it's that way because they're less likely to hallucinate. It's cheaper to run the models that way. Sometimes it's helpful to just talk to an LLM and have them give a fresh opinion without any context. But in my case, most of the time it's not. And it is something that the models struggle with. Sometimes they'll have memory files or skill files, but they won't necessarily always bring them in. And so there's a lot of underlying complexity there. Black box and the models are still struggling in in some regards. Before we continue, let me pause and share some words from this week's sponsors. Earlier this year, my son Brett and I on two different occasions we went to Palm Springs and we had a vintage clothing pop up. We were retailers and we used Square to take credit card payments. We we bought the Square device. It connected to our phone. It was incredibly simple. This set up and use Square is a business toolkit that helps you sell, manage and grow without the chaos. Whether you're just getting started or already running something great. Square gives you the tools to take payments, track sales, manage your team and keep everything organized all in one place. And I can tell you their iPhone app was super simple and we had no issues at all taking credit cards. And I was worried that this is going to work. It worked great. I'm not super techy and I everything was worked wonderfully. So Square helps you run your business with confidence, clarity and less chaos. And now it's easier than ever to get started. And why wait? Right now you can get up to $200 off square hardware at square.com Go David that's square sq u a r e dot com Go David. Run your business smarter with Square. Get started Today I saw an example this week there was a piece published in the Harvard Business Review based on some work from, I believe all the professors were from Harvard. And they did hundreds if not thousands of simulations testing out all the latest LLM models, GPT5, Claude, Gemini, Grok and others. And they were focusing on getting strategic business advice. They wanted advice and there was along different avenues that you can can use when you're deciding how to run a business, whether it should be centralized or decentralized. Should you focus on the short term versus long term? Should you compete ruthlessly or try to collaborate with others? Should you make huge radical moves or should you innovate incrementally? Should you differentiate your business or focus more on commodities? How much automation do. So there's these different tensions that you find. So they would ask questions to these models and surprisingly all the models gave the same advice on one side of the tensions. For example, they all said businesses should differentiate rather than focus on commodities. Well, there's a lot of successful commodity based businesses long term focus versus short term. It should be decentralized, overwhelming decentralized structure versus centralized incremental innovation. They should collaborate rather than compete ruthlessly. Now that's a problem if executives are getting advice and the models all say the same thing. We talked about this a few weeks ago, the homogenization of these models. Well why is that? Because they were trained on the public web for the most part and there are biases in the web, the biases at least in here case in terms of strategic advice. And as a result the models favor one framing of how to run a business. Now that's just one aspect of think about any other domain. Which is why it takes judgment and wisdom to use AI. That could be why a recent Financial Times poll found of the 4,000 US UK workers they surveyed, the best paid workers were more likely to use AI daily compared to workers that weren't paid as much. And it also was highly correlated to education. The more greater your education, the more likely you are to be using AI. The greater your your pay, the more likely to use AI. But if you're using AI, you need to know its weaknesses, know where it's just making stuff up. And if, if you're in a domain outside your area of expertise, that can be difficult to do. I saw one study and I don't have the link here, but they were, they were querying these large language models on health advice and half of the advice was wrong. So we have this amazing tool that has sucked up the Internet and we're still navigating how to use it. Now here's something that people are definitely using AI for coding. So one article that I'll link to from the New York Times, it gave the example one company that used to produce 25,000 lines of code per month, now using AI, it's producing 250,000 lines of code. Joanie Clippard, who's co founder and chief executive of Stackhawk, a security startup, said the sheer amount of code being delivered and the increase in vulnerabilities is something they can't keep up with. She points out that the ability to create new code, new enhancement, there's so many engineers, senior engineers that need to be there to review the code. But they're still producing more and more. That's putting stress on sales, marketing support. They're having to work faster and faster and faster just to keep up with the software engineers. Now is that volume going to be better? We don't know. Ideally, if there's a greater variety of things being produced, there'll be some breakthroughs. On the other hand, there's vulnerabilities in the code. And even the ability of AI with Claude Smithas model, which they haven't released publicly, apparently is incredibly adept at finding weaknesses in codes. And if that got into the hand of hackers, et cetera. But this is the leading edge of the present. That's where we should be spending some time either at work, at home, getting that input, experimenting. That's how I've gotten a lot of my ideas. Looking at what was happening on the leading edge and trying things out, seeing what worked. Often it doesn't, but sometimes it does. And then you double down on those areas. It's a form of optionality. You try things, you're looking for asymmetric payoffs, positive upside, protect on the downside. And that's why it's important, which is the fourth thing is to understand the probabilities. Recognize what is the median outcome of a particular activity. What is the average outcome? How likely is it to be successful? Now some things we can't know the probabilities. And that's the difference between risk and uncertainty. With uncertainty, we don't know the probabilities. Which means we need to do even more to protect ourselves. But with something like writing a book, working on my second book, my first book did about what the median published business book does a little better. I was disappointed. I still think it's a good book. You take an author like Morgan Housel. He was with Herriman House when he published the Psychology of Money. Their typical business book would have sold 5,000 copies, maybe 10,000. Morgan Housel's book sold over 4 million copies. That's an outlier. That's positive skewness. That brings up the average. You have one of your business authors selling 4 million. Suddenly the average goes up, but not the median. And so when we do something, if you write a book, and I highly recommend it, assume you'll sell a thousand to five thousand copies. You write the book because it helps solidify your thinking. It can give you authority. You don't do it because you're going to make a lot of money selling Books. Same with the podcast. The median podcast gets 125 list episodes. If you're in the top 20%, that top 20th percentile gets a thousand listen, which is great, but you're not going to build a business on that. We're a top 2% podcast. I could tell you that you're not going to build an ad generating podcasting business based on that. You're going to live off the ad. So you think about the top 1% podcast gets 29,000 downloads per episode. So they get about $700 per ad gross, maybe 500 net. You run three ads, that's $1,500. That's enough for one person to live modestly. But then if you have to pay production staff, et cetera, an entire team, and that's the top 1%. Meanwhile, the median accountant makes $82,000. So it's not the average isn't driven in. Accountancy isn't being driven up by outliers. It's much more predictable. Now that doesn't mean you don't launch a podcast or write a book, but you understand you're just not going to be highly compensated for it. I mentioned YouTube. We don't have a big YouTube channel. It, we've never broken out, we've not been. It's just never really taken off for some reason, lack of consistency. But the algorithm hasn't picked us. But we still publish there occasionally, so that's important. Now another aspect of when it comes to understanding the probability is in the investing markets. This is an investing podcast. In my first book, Money for the rest of us. 10 questions to master Successful Investing, I quoted the finance professor Andrew W. Lowe. He said the wisdom of the crowds depends on the errors of individual investors canceling each other out. And that's what really the wisdom of the crowd is. You, you have all these different views and some are overly bullish, some are overly bearish, some are willing to pay more. But you get this consens and you get errors canceling each other out on the upside downside to where the return, for example, the stock market is driven by the dividends, the earnings growth and changes in valuations over time. Now Low pointed out that investors can exhibit certain behavioral patterns that can be irrational in the same way and that can lead to bubbles. He likens it to a defective scale that has an upward bias. And so the crowd in aggregate, you know, if they're starting to behave irrationally, they can not be as wise. So I thought about this because of a piece a friend of mine Sent me that was focused on prediction markets like Poly Market, for example. In fact, they studied Poly Markets and they were trying to see does the wisdom of the crowd apply there? That you have the aggregate of diverse participants, the idiosyncratic errors cancel each other out and that's what drives the returns and the outcome. And they found, no, that wasn't the case, that prediction markets are remarkably accurate. But the source of the accuracy is a minority of informed traders. About 3% of all the accounts make the vast majority of the money and the vast majority of the accurate predictions. And they know this because Poly Market, it trades on the blockchain. So they're actually able to get all the different bets and to analyze it. So you have 3% that are incredibly informed. They make the majority of the profits. You have what they found were lucky winners. That's another 29%. They, they made insignificant profits in aggregate. And then the majority, 61.4%, they incur insignificant losses in aggregate. But 6% were unskilled losers who lost large dollar amounts. Predicting inaccurately. Now, there's an area where, okay, if I'm going to be on Polymarket or one of the other platforms, Kalshe I believe it is, I want to say, Kashi, the serial we need to know the probabilities that unless we're informed, we need to be informed, not in an illegal way. This week, the justice department indicted a US army soldier who made $400,000 trading on poly Market because he predicted that Nicolas Maduro in Venezuela would be captured by the end of January 2026. And he was, because he had essentially classified information. Now there's a lawsuit against him. But the point is there's another example of we need to understand the probabilities. So when we look back, what are the four things? Take notes, get input, keep reading, learning, traveling, building relationships with people, building out your network. This is all really under the umbrella of our capital. Our capital isn't just money. It expands our choices. We're experimenting on the leading edge of the present. Trying things, getting ideas, iterating, exercising options, committing, getting feedback, trying again. And the fifth thing is just get out and move, Walk, step away. And I remember doing this during the great financial crisis. I just had to turn everything off and get away. And now I walk regularly twice a day. Sometimes it's a hike, play tennis. But the idea is to don't get constant input, get it selectively at certain times. So then it allows your brain time to sort of your subconscious to process it to get these ideas. I get so many ideas when I'm out exercising because I'm not focused on it, they just surface and that that's important to take all that input and then make sure that we're we're stepping away and getting good ideas. Those are five things that have really helped me in my career. Let me know what what you have found helpful what is on your list, what's in your on your script, your rules of thumb that you have codified and you have found very helpful in your professional life or your life in general. This is episode 555. Thanks for listening.