[Music] You're listening to the Ellen Gray podcast. I'm Horatian Idem McCarthy, a manager in the institutional clients team at Ellen Gray, and your host for this episode. Artificial intelligence is rapidly changing the way information is created, analysed and shared. Today, investors and asset managers have access to more data, more computing power and more sophisticated tools than ever before. In theory, this should make markets more efficient. Yet, history suggests that periods of technological disruption are often accompanied by excitement, speculation, and bold predictions about the future. From railways and telecommunications to the intellect, investors have repeatedly faced the challenge of distinguishing between transformative technologies, elevated expectations, and genuine long-term investment opportunities. At the same time, while tech changes, human behaviour often clings to the same patterns. Markets continue to be influenced by optimism and pessimism, compelling narratives, and the tendency to follow the crowd. So, if artificial intelligence gives investors access to better information and more powerful tools, should we expect markets to become smarter, all the same behaviours that have shaped market cycles for generations continue to create investment opportunities. Joining me today to explore these questions is Simon Skinner, director and head of our offshore partner, Orbusis London, based global investment team. Simon, welcome to the podcast. Thanks very much, great to be here. Now, Simon is always good to catch up with our friends from the other side of the pond, and I've always been curious about this. Amongst your many qualifications, you are also a lawyer. How did you find your way into the world of investments? So, I studied law at university, which I really enjoyed as an intellectual challenge. I found it fascinating, complex, but also really understanding building new principles of logic through that process. And I got swept up in the recruitment program for the big law firms in London, promising me exciting deals, interesting work, the chance to really make a difference for some big projects that were going on. When I arrived at the law firm, I found that I'd probably been oversold. I found that I was a very small cog in an incredibly large machine, and actually I didn't find the work that interesting, where I saw the interesting part of the process happening was before deals came to the law firm. So, it was a decision making that really interested me, not executing the decisions. And so, I felt like I wanted to move into a position or a place of work, where I was the one getting to make those decisions, and then being held accountable for the outcomes. And so, I knew the direction of travel. It was just a bit challenging to go from being a lawyer to being an investor. But fortunately, I came across a quite quirky, but very differentiated firm in the west end of London called Orbus, who were willing to hire non-investors into investment analyst roles. And so, I started a conversation and not quite understanding what I was getting myself into, joined in 2008, which was in the middle of the global financial crisis, which just turned out to be a wonderful time to observe markets and psychology and human behaviour, and really start to cut my cloth as an investor. Nice. I think it's apt that you're talking about those kinds of themes coming through, because today we're going to be talking about quite a big topic. And that is artificial intelligence and the tools that we have at our disposal. Now, artificial intelligence seems to feature almost every investment conversation at the moment. Why do you think it is captured investors' attention so powerfully? So, I think humans are very drawn to narratives, especially narratives which promise something meaningfully different and better than the past. And I think this time around, this technology change that we're in the middle of, it's not something which is remote to investors, like a railway, for example. It's something which investors are using every day. So now we have a super computer in our pocket. We have access to incredible, large language models on our desktops. Investors are using these models, and therefore it's easier to dream about the upside from the whole world using these models. And so, I think people get caught up in this narrative, and they start to see that this could change huge amounts of what human beings do on a day-to-day basis. And that allows for a future which looks very different to the one that we live in today. And it's quite exciting. And so, people see that the share prices have moved a lot. They dream big about the future and share prices could move further. And that kind of greed emotion starts to kick in. People want to get that upside. And I think that's where we are today. What fascinates me is you're talking about narratives, you're talking about greed and emotions that cause investors to want to act irrationally. And I think it's fascinating that you mentioned we have access to more information. And of course, bitter tools, connectivity than any other point in history. Yet opportunities still exist, which is great for country-renstock pickers like ourselves. But why haven't markets become perfectly efficient? So, I don't think this is going to be the first time that the access to information or the analytical tooling that we have has improved. If I was to take you back to the 1950s and 60s, when Warren Buffett first started investing. He was pulling out information with a huge time lag from a paper directory. And you could understand why somebody applying a very diligent value-based approach in that era might do well. But actually, if you roll forward and look through all the technological advancements we've had over the last 50 or 60 years. So, you know, electronic data and then the internet, mainframe computing, spreadsheets. None of these have changed the nature of markets. And I don't think AI is going to change the nature of markets either. Market is ultimately driven by human decision makers. And one of the most powerful dynamics we see in markets is herding or crowding. And so, when something starts to work, people build a narrative and they like to explain why price is going up or going down. And that tends to generate momentum. And then more people believe the same thing because it's happening. And so, as long as there is a human decision maker involved somewhere in financial markets, I believe they will be overwhelmingly driven by human emotion. And a human decision making rather than a kind of logical rational assessment of positives and negatives of any given situation. So, the crux of this is that behaviour is really the primary source of inefficiency. So, as an viscer, what does that mean for how we should approach the market? So, I think one of the important things to realize is that evolution hasn't served as well in terms of our decision making ability, especially under pressure. And so, 10,000 years ago, it didn't pay to be a contrarian and to be the one caveman who didn't run away from the Saber-tooth tyler. That's where that herding instinct comes from today. And it's something we had to actively fight when we're making decisions in financial markets. So, I think just being mindful that we all have these biases and that you need to build systems around yourself to try and protect against them. But also, you should be mindful that other people have these biases. And when you see people at all acting in concerts with very narrow view of one particular outcome in the future, you should be on guard. And I think if you think about, you know, when everyone's writing about the same topic with the same view of the future, that to me is a warning signal. I want to really interrogate these views to make sure that people are taking a balance view and not actually just following what other somebody else has already said. Because I've heard you mention echo chambers before. And that's exactly what's happening. You have the same narratives and you amplifying the same conclusions being drawn about certain investments. So, looking ahead five or 10 years, where do you think AI will have the biggest impact on the investment industry? And where do you think it'll be overhyped? The area where I have the biggest impact is it will increase awareness for more people of more situations. So, if you think about what AI is really good, it's really good at processing basic information and summarizing. And so, one of the constraints that investors have had in the past is just actually being able to consume enough information to be aware of all the different situations that are happening globally at any one time. And so, I think it will be increasingly common for people to have more awareness of more situations. However, these tools are not original thinkers. They are basing their views on what has been written or thought about or spoken about somewhere else already. And so, while you have more people more aware of what's happening in the world and more different places, they're all going to think the same thing. They're awareness is going to be constrained to one version of what's happening in the world and one version of what may happen in the future because these tools are all feeding off each other. So, a stat that I think is somewhat shocking is that there are something like 4,000 news or information sources globally which are just AI-driven. So, no human oversight. And if you think about where these 4,000 new sources are getting their data from it's each other. And so, you effectively have what feels like a huge diversity of sources, but really, they're all saying the same thing. And so, if you roll forward into the future, you're going to find very convergent narratives, you're going to find one version of what's happening on and you're going to have not much human supervision or independent thought going into those views. So, I mean, it's fascinating that you say there's 4,000 without human oversight. All come into the same conclusions and as is humans just picking it up and saying that must be the truth. People are excited about AI and what it presents as a tool and it is no doubt going to change the way we operate as a society just far outside even the investment world. And I think that we often view it as a once in a generation technological shift. But history is full of these transformative innovations from the railway boom to the internet. You've looked at it in a lot more detail. Which past technological revolutions do you think offer the most useful parallels for understanding AI today? I think there's probably two dimensions. So, at the moment,
AI is understood as mostly impacting knowledge work. It's taking information and it's taking the ability to process and analyze and it's putting more power into people's hands, which is ultimately removing the need for so much human intervention in some of these processes. You can see how many industries, companies, sectors, which are mainly focused on the flow of information somehow, so lots of administration bureaucracy, this kind of thing, could be really affected by AI. I think that hasn't happened yet but watch that should include his government. So actually public sector, you know, if you think about the amount of government which is basically the flow of information and you think about how many people are employed by governments globally, now as they spend, that to me seems like something we should transform over a number of years. And bringing that efficiency into the system. You'd hope so. I mean, I'm not very optimistic that governments will do this well, but eventually you think that there should be scope to improve that side of society. It is a delicate balance though. On the one hand, you want to bring more efficiency into the system and take out a sum of the bloat, but on the other hand, they are human jobs and employment that you need to consider. How do you think of that when you have to balance the two? Because I know a lot of the time we think of AI is this big, scary, you know, cull of human jobs and what that would mean for future economies and the ability for us to take care of ourselves. I think at any point of technological change, which unlocks efficiency, so you go back to the industrial revolution. In the moment, it's very easy to understand the jobs that will be disrupted, it's very hard to understand the jobs that could be created. And so I think it's quite natural that we look here and we say, well, this would be so many jobs that would be created, but ultimately, that's cost that can go back into something else in society, whether it be better research and development, whether it be better outcomes in education, etc. So there's usually something which comes in to take the place of which or whatever has been displaced. I remember the story of an elevator operator who was a very respected job. And then you know, you could push the button yourself and a lot of people were so alarmed at the fact that these jobs would be wiped away, but we didn't seem to see much of a shift and those people were placed into other jobs, as you said. And you go back to kind of the agricultural revolution, so the mechanization of agriculture. Absolutely. You know, the majority of people in society used to work in agriculture. I'm pretty sure most people are happy not to do that anymore. So I'm not sure where incremental roles come from, but I'm pretty sure that society will pivot and innovate and find places for people. So now you've actually pulled on a thread that I think is quite interesting. It's us talking about how Orbus is approaching AI. And I think there's a lot of discussion in industry about AI replacing specifically investment professionals. Now how is Orbus actually approaching the technology? So the first is what I call table stakes. Okay, so this is the application of AI in a way. I think other investors will be able to replicate over time. And it's really about driving efficiency in your investment process. And so I now use AI effectively as a junior investment analyst that works alongside me all the time. And so I can ask it, you know, somewhat basic queries, get good quality answers and iterate and start to build understanding before I have to go and trouble an actual human with any of this stuff. And I think that allows us to triage through more ideas more quickly. And I expect all my teammates are doing this. So this would be using Claude and chat GPT and alpha sense and these kinds of tools. And really what that's doing is it's allowing us to consider more ideas early in the investment process to broaden our funnel. And then hopefully to triage to better quality ideas over time. But I think that will be something that most investment managers should do over time. That's the natural progression of things in the same way that, you know, you don't do paper based calculations anymore. It's just quicker to use this richie. But there's a way of implementing AI within Orbus, I think is quite powerful and is more proprietary. And so that relates to our proprietary data. So Orbus has huge expense really, maintained a library of all the research we've ever produced, of all the investment recommendations we've ever produced. So going back almost four decades. Exactly. And we've stored all of this in our proprietary systems for that period. So it has been useful in terms of training, in terms of building and understanding of the past. But the models we have access to now are allowing us to extract much richer insights from that historical data. And that extends both to the research we've produced, but also the behavioral analysts. And so now we can give analysts access to rich kind of pattern predictions through our historical data to say, look, in a similar situation, how do we behave, what worked, what didn't work well, etc. What are the lessons from history. But it's also allowing our portfolio managers to get a much better understanding of how our analysts have behaved in history and where the really strong alpha signals are in our investment engine today. You actually have a team that focuses just on the behavior of the analysts and the portfolio managers over time. That's correct. So we have decision analytics, which have been running for several years in London. They are helping us to understand our prior behaviors and particularly to understand the biases in our behaviors. And so that's been a very heavy analytical lift until recently. But AI is really helping them to go more quickly and into more detail in those behavioral biases. But now we're able to use AI to say, given what we know about the behavior of our analysts, what's the best portfolio we can construct today. And so I think about it was we have our investment engine, which is being tuned up with use of AI becoming more efficient. But AI really is building as a much more effective transmission or gearbox. And taking the best quality alpha signals we can from our investment engine and conversion them into client portfolios. And so that's taking into account, you know, the individual analysts, their strengths and weaknesses, the type of market environment that we're in. I'm really looking to optimize that portfolio construction layer. So before I get to the gearbox, I want to go back to your decision analytics team. What is it like? Is it like a therapy session that you have with the decision analytics and members of the investment team? How prescriptive are the recommendations or the biases that they've managed to tease out over time with regards to future decisions? So I think therapy is probably the right term. Investing is a very personal endeavor because it's about your own decision making. And there they can be, you know, it's very clear cut data. There's no fog on it, you either outperforming or you're not. And each stock is either a success or it's not. You know, the market makes makes clear whether you're good at in the analysis in that moment. So the the first process the decision analysis team does is they will examine someone's entire track record. And they will look for all the patterns, all the strengths, all the weaknesses and they will present your report on this, which is quite sobering reading because they highlight all the many flaws that each of us have as human beings. And then once you've kind of digested that and accepted it, which can take some time, we start to talk about practical ways that you could improve and structures and processes you could implement, which would help you improve on some of those biases. We then meet quarterly to discuss what's happening in our recommendations and whether we've been successful in starting to mitigate some of the biases we've historically shown. But this requires a huge degree of vulnerability. You know, you really are walking in there, bearing everything in terms of, you know, your concerns, your fears, your anxieties. And so we've been very careful to structure it in a way that people feel comfortable talking very openly and making sure that this is a process of coaching and growth rather than any kind of sense of an assessment with a very, very strong meritocracy in August. And so the people going through this process are genuinely world class investors. They're at the top of their game. And yet we're taking them in there and we're explaining how bad they are. And so, you know, people have to detach from the egos and calmly, rationally assess where they can get better and think about the steps to do that. And I think that's quite unusual. You take people who have been so successful for so long and are willing to embrace a growth challenge like this. But I can say from personal experience and from the observations of people closely working alongside me in London, the benefits are real. And I think it's, I think it's very exciting that we can be taking people who've already been very good and further improving them over time. And I think it's usually culturally positive as well. Yeah, it's just 1% enhancement that can have very meaningful long-term outcomes. It's a fascinating way, which we're using AI, I think maybe a little bit differently tarpies. So I think it's a good example of how the collection of our data over time, we haven't appreciated in the past half-how for it could be. And so the fact that we have all this data, all these paper portfolios, all these recommendations going back nearly four decades gives us a unique ability to better understand ourselves and improve in the future. And I think if you're going to set out on this journey today, you probably have to collect data for 10 years before you saw any benefit. And so I think if you haven't done this already, you're in a really tough situation. You're going to invest for 10 years so that you maybe get a benefit at that point, but probably you have to wait a bit longer as well. Yeah. You know, it's a tough not to crack. But as you say, even 100 basis points additional return can be transformative for our clients. And so, we're chasing down every last game we can get. It's wonderful to hear. So what I'm hearing is so far in the investment process, you've got your AI tool acting as your junior investment analyst at your side, shifting through information, you've got decision analytics looking at previous data sets and trying to pull out the biases. So how do you balance the efficiency gains that you getting from AI with the need for independent thinking and original research? So I think one of the reasons that markets are not going to become more efficient because of AI is because you have to want to ask the right questions. And that's down to the human at the start of the process. I think there are very few investors who will sit down with their AI model and say, finding a situation which currently feels incredibly uncertain. But where if certain factors go in our favor, the payoff could be very favorable in five years time. Because that person has to know they're going to have to endure, you know, volatility, uncertainty, potentially feeling very wrong in the short term, which is just very
unattractive to humans. And so that's a bit of an extreme example, but if you are, if you're set up to look for situations where you know that other investors could be making a mistake, I think you're in a very strong situation. But to do that, you have to structure your whole firm to withstand the volatility and the noise that can come in the short time to do this. So that goes from the ownership of the firm to the incentives in the investment team to the clients that you've attracted. And you have to have all of those 100% aligned before you can embark on a journey that looks like that. And I think it's quite rare in our industry to find people who have done it. And so I feel confident that Orbis has built something which is well positioned for this environment. To ultimately the decision making that goes into making the big decisions for the portfolios comes down to the human. It's human judgment rather than led by the AI. That's exactly right. It's definitely human judgment. What we're trying to do is extract the best quality information we can for our investment engine to help those judgments be the most accurate they can be. So I hear that a lot of investors are are relying quite heavily on AI models and they've become incredible. I mean, I think back to when the early AI tools couldn't correctly identify the number of ours in the word strawberry to what they're able to do now. It really is a huge leap of our technological perspective. But what risks do you see with this over-reliance on AI generated analysis, particularly for investors? I think the biggest risk I see at the moment is false certainty, false precision. So I've been using these the AI models for a couple of years. It's very rare that I get any degree of uncertainty in an answer. If you ask a question, you generally get presented with information with no error bounds marked around it. And it's very rare that the model says I'm not sure. And so I think the the danger is false precision. People have a query or have a thesis and they check it with AI and it gives them the answer and they feel confident that well, this is an incredibly powerful model. It's reviewed all the information available in this area. It's come up with the answer. This must be right. The world is incredibly uncertain and the range of outcomes are far wider. Most humans can appreciate. And we're already overconfident in many degrees. And so I think it's going to be an extension of that overconfidence, which will cause people to take risks they shouldn't take. Probably to manage the portfolios in terms of sizing inappropriately. And we know that ultimately the human trend to herd or crowd is probably going to lead people into some trouble at some point in terms of risk management as well. So what aspects of AI adoption do you think investors are over-istimating in the next few years? And then conversely, what long-term implications of AI might the market be under-istimating today? I think people are quite optimistic short-term. So I'm not sure that I've got many good answers to where people are over-estimating the impact. I think this is clearly, if you look at the revenue that's required to justify the investments being made at the moment, people. It's far ahead. Capix. Capital expenditure. So there needs to be a strong follow-through of adoption. So it's justified these levels of investment. And so that to me tells me that most people are quite positive on the amount of AI that will be used over the next few years. If I look out longer term, I think there's two areas where I'm curious and could probably see upside or I don't see as talked about as much. The first would come back to be the use of AI in government. And I think, there could be some. We talk about demographic issues, we talk about budget issues. I think if AI could be used in a constructive way in the public sector, it could solve a lot of these issues. And that could be transformative for some countries. So that's quite interesting to think through. The other area which I'm. People are clearly talking about, but I wonder if they really understand the long-term implications would be robotics. And so, you know, to scope for a lot of manufacturing industries to look very different in the future if we can implant AI into the interface of physical manufacturing. And so I'm not really thinking about humanoid robots. I think they're probably a distraction. But when people imagine robotics in a car factory, they just imagine a human shaped robot taking over the same processes. In reality, you'd build the whole factory completely differently if it was fully automated. But I think that could happen. And so I think that's pretty interesting to think through as well in terms of the cost of production of some items, perhaps the cost of moving and transport of certain items. And so I think it could have some quite big impacts on supply chains, perhaps like national and international relations as well as it becomes easier and cheaper to move labor-intensive processes away from historic bases. It's an interesting one, especially given how noisy the macro is at the moment when you think about the fragmentation and the disruption in global supply chains and this propensity or rather shift for countries to want to unsure or racial manufacturing. I think that may be quite far off into the future, but I definitely understand where the gains would be in terms of bringing those efficiencies into those systems. It seems totally reasonable to me that if it's strategically important to your government and there is a cost efficiencies to do it, then it should happen pretty rapidly. I'm going to shift gears a little bit. Your enlistment philosophy at all this often involves looking where others are not. Where are you currently seeing the greatest disconnect between market attention and underlying value? I should be clear ahead of time. We haven't thematically invested in AI. That's not a way of approaching the investing problem. But what we try to do is to calibrate to what we can see is actually happening in the world. And then thinking through the second and third order consequences. Some of the things that we've been successful in spotting was that memory pricing would be positively impacted because of the demand that's driven from AI chips. While now it feels very obvious that the likes of Samsung and SK high-necks would be massive winners from AI. We were researching this 18 months ago when it was very far from the case that it would be obvious. But we looked through what was happening, the technology, the supply chains, and thought that these companies would have earnings in a very sharply positively for a sustained period. And so we've been invested there. And that's not because we were particularly bullish on any kind of AI theme. It's just that we could see that the CapEx was committed and was going to be spent. And we can see these companies would be beneficiaries in the forms of time. But we're also very aware that when the market's attention is focused in one area, you can have big opportunities in other areas. And so things which don't look anything like AI today, we think could be big beneficiaries over time. So for example, we like businesses where a very strong management team who tend to be very aligned to a big ownership stake and who have a great track record of implementing technology in their own businesses. And these businesses are going to be the beneficiaries of the massive investment that goes into AI because they can implement it in their businesses and realize efficiencies probably way ahead of their peers. And so you have a step change in potentially margins and returns. So the kinds of businesses we're thinking about would be something like Core Pay. Or which is payments business or QXO which is building products business or interactive brokers which is a online, broken trading platform. Very strong management teams, great ability to execute on in terms of technology. And we think very, very fairly value today. Even if there's no improvement and so there's a lot of upside in those names. The other area where we found great opportunities is in something like healthcare. So no excitement from AI in many areas of healthcare. Very uncorrelated to the border macro cycle. Went through a big cycle through COVID. And so we've been through and kind of I'll use this as a bit of a pun but surgically looking for opportunities and found some great names like Starris and Bruka which we think have very strong businesses have organic improvements which they can drive. And I think the outlook is very positive for those businesses not reflected in the share prices today. So you've mentioned the semiconductor manufacturers in the portfolio. You mentioned it's Kehinex, it's K2, you've got Taiwan semiconductor manufacturing company, what TSMC. There are other levers in the portfolio that have also kind of played in that AI ecosystem that you have in the portfolio that you were quite early to the party. And maybe the energy theme is an interesting one that's kind of bubbled up. It's an interesting contrast. If we look at the weight of technology shares in the US market, I don't think it's ever been higher. And if you were to roll back 12 months, if you look at the weight of the energy shares in the US market, I don't think it's ever been lower. And so that, I mean, itself is just pretty interesting. If you think about every data center, it's massively power-intensive. And the US largely is a fossil fuel driven power market. So in the same way that we looked at memory 18 months ago and said, you know, how much demand will there be for these chips? What do we think is going to happen to the prospects for these businesses? When we look at the power consumption that's required for data centers, we can't quite square the circle on how they're going to get built and powered at the same time. And so we think it's quite likely that particularly US natural gas is drawn on quite heavily. And so we've been working through the opportunities in that space to say, look, if these gas producers are able to secure data send contracts and move away from the market-based pricing which they've endured historically, we think shares are very good value. And so a great example is EQT, which is one of the largest gas players in the US. I think it's very well positioned such that as gas demand grows for these data centers which we think are going to get built over the next five to ten years, there has to be more gas used to power these data centers. But we don't think that's taken account of yet because it's not so showing in the gas price which tends to be at spot. And so if we're right, then these shares should do really nicely. Actually, if we're wrong, we think, you know, we're very unlikely to have any impairment on our capital, there's nothing at the price for it today. So maybe if we come back in 12 months time, you'll be saying, you know, these energy shares have gone really high because people now appreciate the story and, you know,
that would be great. But I think if nothing changes then we're still going to be absolutely fine with oppositions. And I think a nice tailwind is coming through from the fact that we've currently got a disruption in global energy supplies. Yeah, this is really interesting. So the market seems very relaxed about the situation in the Middle East. You know, we had a period where oil spiked up, spot oil spiked up, but very quickly came down again and people kind of looked through the disruption and thought things we get back to normal pretty quickly. I've no idea how that plays out and how long it takes, but it feels like it's certainly a risk. And so we have got some, I guess, a hedge if you like, in the portfolio from that situation continuing to get worse. It's so interesting. Earlier on you were talking about uncertainty and then you were talking about how we all start accepting the same narrative. You know the echo chamber and it was so interesting I was looking at polymarket, which is the largest prediction market in the world. And when Trump had started talks about a ceasefire, they had almost a guaranteed odds of the ceasefire remaining. And it's so interesting because after two weeks of negotiations, we now know that the truth is in tatters, the straight is closed, and we once again in the situation where oil is starting to spike as people realize that we may not be able to get oil supply. So it's a very interesting one because you've got tons of information and the surety, a certainty of certain things. And it doesn't actually materialize. Yeah, I think it's a good example of humans underestimating tail risks. People assume that things are going to happen in a fairly narrow range of outcomes. And you know, I guess if you just reflect on your own experience, every time you get surprised, that means you've been overconfident in the past. And you know, we can look back on the last five years and we'd say that was a wild ride. Yeah, we've been through COVID, we've been through tariff, we've been through a new war in the Middle East, we've got Russia at war in on the edge of Europe. And if you know, five years ago, he said, right down the most extreme thing that's going to happen, I don't think we would have got any of that on the paper. And so, you know, there definitely will be more surprises coming and we need to be prepared for that. As a contrarian stockpaker, what characteristics do you ultimately look in this noisy environment when searching for opportunities? So what I find very powerful to do is be a student of history. And so I like to look as much history as I can for any given situation. And so where I have a sense that people are not paying attention or are very negative in a certain area, I'll then try and bring myself as much context as I can or situations that are similar to that. And try to think through what's the range of outcomes that I could expect from the same starting point. Then I have to think through is something different this time. Is this going to be another version of history or is something structurally changed so that this time is different? And so if you were looking at electronics retailers in 2005, 2006, is Amazon started growing? You know, there could be a recession in the US and you know, best buy for example, I'm sure got very cheap in a recession. But something had structurally changed and that Amazon was going to grow and eventually kill your business basically. But it tends to be more the case that history repeats than it doesn't. And so if you can get fine situations where people are, you know, not paying attention or very negative, and you can build a relevant context to suggest that you think that this is another one of those. So something that should happen before which you can start to understand, then I think you can get interested, but you still have to do the work. You still have to go really deep on the company. You have to value the business in a number of six scenarios and get confidence that you have real downside protection. And you have to understand that you are constantly triaging capital amongst those situations where you have conviction. And so you need to find that asymmetric range of outcomes, which you have real confidence in. So in the portfolios as they stand, where do you see our downside protection sitting? So I think we've got to be realistic. Stop market to high. You know, if you were to value the global stop market on a on a number of different factors, I think it would be something in like the top 10% richest valuations that we've seen in modern history. And there's a reasonable correlation between the starting valuation for the stop market and what you end up with 10 years later if you just buy the index. And so I'm not particularly optimistic about stop market returns from here. But if I look at what all this owns, especially our largest positions is very different. So we look at, you know, what our peers own in terms of large positions. And we have very little overlap. And we really think we've gone deep into areas which are contrarian. We found idiosyncratic situations where the drives are quite different to those which are driving the broad stop market. And we have confidence that either there's a management team in place who is going to add value through a deep cycle, all that the business is uncorrelated from what we believe the drivers of that cycle are likely to be. And so we don't feel like we're taking much of the same types of risks as the market. I think it's also interesting to reflect on the concept of diversification. So conventional wisdom is that you diversify your portfolio to minimize the risks. And you can, you know, you have a smoother ride less volatility if you own, let's say a broad stop market index. If you go and buy the global index or the US index today, you're overwhelmingly buying one thing, which is confidence in AI. You're a big tick. Yeah. You're a big tick growth. And if that, for any reason, falters or confidence in that falters, I think it could be an incredibly bumpy ride for investors. And I think, you know, much like if you were investing just in the internet at some point between 1998 and 2000, all your eggs were in one basket. You know, you had one thing which is going to drive the market. And yet there were some fantastic stocks. Very reasonably priced and completely unrelated areas of the market. That will under the radar. So you can go and buy a house builder or a brewer or a supermarket at great value prices. And five years later, if you came back, you'd done fantastically well while the stock market was down, you were up. And so I don't think we just have to think about ourselves as having downside protection. I actually think we want to be making attractive investments, which can work with regardless of the market or the macro cycle. So I've often heard you say, be the fox. What do you mean by that when it comes to thinking about the portfolios? So this came through thinking about the market dynamics I've discussed and the size of the herd. And so I think about the herd is the sheep. They tend to follow each other around. And if I think about what's happened inside markets over the last 20 years, I think the number of sheep in the herd has got much bigger. And so you've had the growth of passive investing, which has been, you know, very strong. You have the growth of momentum investing, also very strong. And you have a number of funds offering trend following strategies, which is different type of momentum. That as a portion of the stock market has become much bigger over the last 20 years. The number of people doing deep independent research and coming to an independent appraisal of security value has shrunk. And like these people are the foxes. So the foxes operate on their own. They're cunning, they're adaptable, they're opportunistic. And they stand apart from the herd. But there's not many of them. And I think in an environment like we're in, and given the market dynamics, and the starting point for where sentiment is, I think it's more and more important that investors focus on being the fox, making sure they're not part of the herd. So for when riskers who feel like they have may have missed out on the most popular names in the AI trade, and particularly in the ones like, you know, semiconductor manufacturing, we can they find other alternate opportunities that play in that AI ecosystem. So I think the first thing to say is that anytime you you feel like you've missed out, that's an emotional trigger. And it should put you on guard. You are chasing gains that you've seen other people already make. And you're unlikely to be basing a decision just on intrinsic value available in the shares. And so I know in myself, when I start to have any kind of those sensations, I need to put my pen down and get away from my desk for a bit, probably get a cold drink. But these emotions are real. I think for most people who are feeling that now, I think it's probably safer to take an approach which you could take in any one of the next 10 years. So find a style of investing or an investment manager who you believe can perform through the cycle, whatever the weather, whatever the macro, and make the same decision today that you can make next year and the year after the year after that, which is to slowly compound your capital in a way that you can sleep well at night at and not feel like you're going to be panicked. If it doesn't work in the next six months, that you're suddenly going to need to change your mind and pull your capital out of the market or whatever it is. So I think one of the lessons we can take from history is that it's much more important to have time in the market and to try and time the market, which is going to make the biggest difference over the long term. And so I always think about, make a decision which you can be comfortable with in a multi year view. And don't worry too much about what happens over the next six months because frankly that's impossible to know. And I think another question that I've often thought about is that I think many people will find interesting as a fundamental investor. When research like the Citrini Research report came out about a very bleak future and the impact of AI was a thought experiment of what would be in June 2028. And then we had this cell often softwaist docs that many people labeled SAS POCALYPS. What was the reaction of you and your team when you started to see this massive cell often softwaist docs? Curiosity. Why would things suddenly change like this? The evaluations of these businesses suddenly change? It's probably because you had people owning them who didn't have high conviction in what they were worth. That's what tells me when you have these sudden movements in a whole sector, particularly in big stocks. But if you understand the history a little bit, you probably get less interested in thinking, is this an era of opportunity? So two things I will draw on the first is software has been until probably 12 months ago has been seen as the single best sector globally over the last 15 years. It was the safe place to put your money. It was the highest quality business model around and it was seen as a winner in all weather.
And so that attracted a lot of capital. And what happens when you have a lot of capital trying to buy a few shares, well, they get very expensive. And so starting point for software businesses 12 months ago was a very elevated set of valuations. And so even if the sector halves, frankly, it doesn't look that good value compared to where it was in early 2000s or in 2010, for example. The second thing, I would say, is I talked about being a student of history and having to bring context to any given situation. And then try to figure out is, is what people are worried about structural or is it temporary? And in our eyes, that's a really, really tough question to answer for many software businesses right now. Given the rate of progress of AI, it really could be disrupting some of those business models. And so we've been quite cautious on looking in that space or certainly making investments in that space because it's not clear whether it's going to be a temporary effect and its sentiment or there's going to be a structural impact. It looks like there will be some businesses which get disrupted here and it's quite hard for us to see exactly where the line gets drawn between those which are and aren't disrupted. I guess one thing we can say is though, we have some businesses we think are very attractively priced who are within close to this area, but actually have something different about them. So maybe I could just talk to a business called Experian, which has recently been bought for our global strategy. One is one of three global credit agencies, particularly focused in the US. And so they have a data set which is decades old on most US consumers in their credit performance which is then used by banks and financial institutions to make lending decisions. They also sell tools and software which goes on top of the data set to help those businesses run more efficiently, make better quality decisions in a number of areas. They've been caught up in this kind of software sell off. But the anchor to that business is the data. And the data is impossible to replicate with AI or not and it's hugely regulated. So there's no way using AI to get my credit score? Nope, it's completely proprietary to Experian. And so sometimes you see that when the tide goes out, there's a couple of boats that get left behind that you think are really good bargains. And so that's the kind of situation where we're saying actually we think people are probably too pessimistic on this business because they've assumed it's just software. And actually there's something different to that business, but that requires getting bottom up stock by stock. And that's how we like to invest. So great businesses that actually would be resilient rather than be disrupted by artificial intelligence. And it may even benefit. So firstly, Experian has a huge amount of technical staff, many of whom are kind of much more efficient as a result of AI. And what we're finding is many of experienced customers are now drawing on that data more because it's easier to process with AI. And so the data is more in demand because of the AI that's available and used by its customers, which increases the value of the underlying data set. Fascinating. So if AI changes many of the tools that investors use, but not necessarily the behaviors that drive markets, what is the one lesson you hope listeners remember from this discussion? I think it's just to be aware of how AI is going to change the way that information is presented. And so it's probably going to narrow the range of views which get presented in the market and make it hard as a tip for people to form independent views towards going on in the world. And therefore it makes them more likely to be caught up in something which is actually a narrative rather than the truth. That brings us to the end of the episode. Thank you for joining me Simon. Thank you for listening. Some of the key takeaways for me were question the narratives aim to always be curious. And in the current environment where hurting is prevalent, be the fox. Warren Buffett famously said that the stock market is a device for transferring money from the impatient to the patient. AI may give investors access to better tools, faster analysis and more information, but it doesn't eliminate the need for patients, judgment and independent thinking. In many ways, the principles that have underpinned successful investing for decades remain unchanged. The tools evolve, the philosophy, and yours. If you would like to get in touch, please send an email to
[email protected]. You can subscribe to this podcast on your favourite podcasting platform to be notified of new episodes. Lastly, Alangray is an authorised FSP. To view our turns and conditions, please visit alangray.co.ca. Until next time, I'm Horatian Idem McCarthy and this podcast was produced by Volume Podcasts. Time. When you're a kid, it moves slower than a sloth trying to get through upward security. There is just so much of it. Just give forward a few years to your 20s. Time is still your friend. This has some balance. Go party and have fun. Then your 30s hit. And time suddenly steps on the accelerator like it's trying to catch a traffic light before it goes red. 40s. Now time is moving faster than your 10-year-old and sugar high. 50. You're desperately trying to save as much as you can, but time is moving so fast. And all you can think is, where did all the time go? That's why after 50 years of investing, we've learnt that time is the greatest gift of all. The good news is, if you invest early, time gives you money and then money gives you more time for some much deserved U time. Now isn't that a compelling enough reason to invest? Ellen Gray. Long-term investing. Ellen Gray is an authorized FSP.
PodcastTranscript AI