#27 - Jacob Ziemann: AI Explained

29:43And so there's a whole host of low-hanging fruit pieces that their job is happening right now where Gen.ai is stepping in to supplement people. and there's actually measurable value getting created as we speak. On your second question in terms of what's going to take longer, there are jobs that kind of go back to what we were talking about before, things where AI struggles more. And so I think for a lot of people, that's the creative professions first and foremost. So when it comes to originality, emotional intelligence, being a musician, being someone who creates art, working in advertising, those are really difficult things for ai to do because again it's focused on data that's come in the past it's tough to be it's very difficult for it to be original and so those sort of roles and jobs to be done on the creative side i think are just really difficult uh to be uh to be automated at this point i think a lot of legal and regulatory situations i think that ai can help with some of the road tasks.

30:49It can help with search. It can help do some common documents. But those are spaces where you can't have error rates that are even 5%, 10%. The potential issues of making mistakes are just too high. And I don't think AI is at a place where the thinking and care and duty that a senior attorney puts into a process is going to be easily replicated by where AI sits right now. And then I think also just healthcare, I think again, the stakes are very high. I think AI can supplement a doctor or a decision maker. It can help them with some diagnostics. But when it comes to actually generating a plan, doing all the cost benefit analysis, doing the emotional work with a client, that's just going to take a really long time for generative AI to have a big impact.

31:44I guess you could think of it as having a really smart analyst that has more data than anybody and is really smart and keeps getting smarter and smarter and can give you great information, but you make the ultimate decision. Absolutely. You mentioned financial services. Obviously, you and I both work in that industry. I'm curious what you think about investing. Is that something that can be taken over by AI or does it have limitations that prevent it from being as good as humans? Yeah, I think it's a nuanced question. I think there are definitely ways Gen.AI can help humans in decision-making.

32:24It can help us process a lot of the work we're doing. It can do great summary work. It can help pull apart different data points that reside in different work we're doing at the ground level, talking to management teams, talking to experts. And so from what I observe day to day as a fundamental equity investor who's trying to make the best decisions possible, it can be really powerful in terms of automating the work that we're doing every day. But when it comes to actually going out into the real world and collecting that data, analyzing people, figuring out those things that are difficult for a machine to really understand, I don't think it can really replace that piece of decision-making.

33:09I think it can make us a lot better. We use it every single day. I have the major LLMs open in my computer in a tab all day. Our team is using these systems all the time. But I think ultimate decision-making when it comes to understanding what businesses you want to own for 5, 10, 15, 20 years, that's very difficult. I think in the world of quantitative finance, AI has had a presence there for a very long time. I think it will continue to get better and be a part of that. But I think that's a different approach. Maybe that's a much shorter term approach to how investing is done. And I think you and I both know there are many different ways to be successful as an investor.

33:55One of my prior guests was Martin Escobar, co-president at General Atlantic. And he talked about now they have another person or a computer on their investment committee. She's not voting yet, but she's AI powered and she provides great insight and she's gaining credibility amongst these great investors. And you can think of it as just one plus one equals three, another example of that? Absolutely. Yeah. No, I mean, we're processing all of our notes in our main repository, but we're also processing all of our investment memos. And we're definitely trying to find blind spots or things that we've missed on some of our companies.

34:36And then we're feeding in a lot of third-party information too. So all of the transcripts, all of the additional data that's more readily available and mixing that with the data notes and the real sort of information we're capturing from meeting with management teams directly. So I totally understand you could have not just another analyst, but you could have this improving decision maker who maybe will never get to our team's level, but can be a really powerful voice around the table. Right. And the way I've heard it described, which I think makes sense, and I'd like to hear your thoughts, is AI works really well when you have a lot of data.

35:15there's actually limited data in financial markets you know compared to billions of data points elsewhere there's a low signal to noise ratio that you have players that are impacting prices and those players evolve if you were the only one with ai technology you'd have a huge advantage but when everybody has it it's not as big of an advantage and and also when you study history and you just look at financial markets history that was one path that history took but could have easily taken thousands of other paths. And so the data can become overconfident that the future is going to resemble the past.

35:52I'm just curious about your thoughts about all of that and how it relates to AI taking over investing. I think that's exactly right. I think that if you think about the extremes here, and as you said, there's relatively a small amount of data in the financial world. The data I'm talking about on the ground, meeting with management teams, that's in a tiny amount of data, not just inside of our firm, but across the entire industry. On the other extreme, we have the big large language models from OpenAI, where you're getting to a place where 70 % of all printed words are inside of those models. And we're getting most video globally into those models.

36:31And so those models are getting incredible at certain tasks because they have almost the history, not just of human knowledge, but the history of how almost every word is relate to every word on the internet and more expansively. And so I think you're absolutely right. The problems you can solve here when you have a quantity of data aggregated and trained against are going to be different problems than when you have a relatively smaller data set, and you're going to need different levels of human judgment input there. So in terms of AI, we've seen some pretty big bugs that need to be resolved.

37:09And we've seen examples of this brilliant machine that makes obvious mistakes that a human would rarely make. Would you talk through some of those? Yeah, absolutely. So, you know, I think the one we see pretty frequently is just the lack of common sense. So there's just a whole lot of, you know, world knowledge and common sense reasoning that we're learning our entire lives from our parents, but also our experience that we move through the world. And, you know, we see machines fail to do that pretty frequently. Going back to the autonomous car piece, autonomous cars can take us through dangerous neighborhoods or cause us to do things that aren't perfect for the social context or raise certain safety concerns.

37:51And so there's just all of this common sense that you use throughout your entire day, that you use in everyday life that isn't easy to pick up from that big backward-looking generalized data set. They're things that you've experienced. I think another Another big bug is just decision-making. So like we talked about it, we talked about earlier, the quality of the data really matters. And so if you're feeding in a whole bunch of data that's biased, what you're going to get on the other side is very biased. And that's going to have discriminatory lending practice issues. You're going to have potential issues in how you screen candidates for jobs.

38:28And so there's this big bug around bias that's driven by the ultimate data that goes into these systems. And then I think we've talked about this a number of times, but this, again, this ethical and moral reasoning piece, simple things like autonomous vehicles deciding what's the optimal decision when they're thinking about potential harm to this human being, this piece of property, etc. Many decisions that are very common sense, somewhat easy for human beings to answer. Many difficult decisions that you would need an ethicist to work on with maybe a priest or a rabbi. And so that piece, the ethical and moral reasoning piece, is a big bug that we're going to be working on for a long time.

39:13It's not easily solvable. How does the AI breakthrough that we're experiencing compare to other technological innovations in the past, like the railroad or the internet? Yeah, I think that's an incredible comparison. I think comparing it to the internet, that's maybe too close a comparison to what we're experiencing. I think my personal opinion, I think the opinion of a lot of AI experts, this is bigger than the internet. You do, again, have to go back to the railway and electricity. And so maybe I'll just focus on the railroad one because I think that it is different enough. Enough time has passed.

39:48And so the middle model we work with internally, and it's one that we're not geniuses. I'm sure we've taken it in over time from reading and from what we've learned in the world and other people have learned, is that any big tech innovation has to improve the lives of human beings on a really massive scale. It has to transform industries. It has to improve connectivity and transform industries while it's doing that. And it has to create new markets. And so in the case of railways, you had this enormous improvement in daily life. People could travel and see each other and see their relatives. The world literally just got smaller over time.

40:31And that improved not just their ability to see their families, but also the opportunities they had in jobs and education and access to markets. And so daily life is dramatically improved. I think in AI, we're seeing daily life improvements with agents inside the cars or personal transportation or Apple intelligence in your phone. But that's just the very beginning of the transformation of daily life we're going to see over the coming decades. And then from a transformation of industries, you know, improved connectivity, we've been having this conversation here about the innovation and change we're going to see in healthcare, in finance, in all of these critical industries that touch all of us every day and do make our lives better.

41:15The railroad obviously revolutionized transportation, you know, and enabled, you know, fast movement of people. There was every single industry was touched by what happened. And then on creation of new markets, a whole bunch of the industrialization of America and globally, moving coal and steel and commodity products in a way that made human life better. Railways did that. And I think we're stimulating all of our existing industries in a way that's going to create new markets, new knowledge markets, but also ultimately going to create entirely new markets and entirely new jobs to be done that we can't even understand at this point.

41:55In the same way that if you were sitting there at the advent of the internet, sorry, at the railroad, you couldn't have really seen those new markets, the new facilitation of trade that was probably coming our way. We've seen this before, but it's not exactly the same. No, there's, I think, up and down the stack from that high level piece, all the way down to the investing stack. There are commonalities, there are things we've seen, but there are also things that are distinctly different this time. I've heard the argument that AI progress will be constrained not by the technology, but by society's acceptance and ability to keep pace and its interest in protecting the people.

42:36Do you agree with that? Absolutely. I think right now we're seeing the technology advance at such a rapid pace. I think as I walk around San Francisco, as I spend time with my friends and network, and I spend time inside these companies, it's just this incredible explosion. It's reinvigorating San Francisco, but I think it's just reinvigorating technology, broadly speaking. And so that explosion, how fast it's moving day-to-day, week-to-week, month-to-month, society is going to take longer and it's going to have to catch up and it may always be a little bit behind but if we're going to actually use these big advancements and not rein them in too fast society is going to have to accept these these technologies or find a way to accept them and so i think people are going to have to learn to you know trust these systems they're going to have to have great experiences that they feel are safe they're going to make sure that they feel these systems are safe, that they're ethical, that they're monitored by the right regulatory bodies.

43:41We're going to need to see governments step in to some extent and deal with difficult, thorny issues like data privacy, but also regulation stepping in and mitigating some of these really wild risks we talked about earlier. How do you regulate artificial general intelligence if we get there. And then I think as a society and just how do we think about economic distribution and potential unrest, this whole idea of how you protect jobs, but also allow new jobs to be created as a result of AI. So striking that delicate balance is something that as a society we'll have to figure out, the regulators will have to figure out.

44:26And so I completely agree. And then overlaid on top of all of that, which I think in Silicon Valley, we maybe talk more about, but I don't know if society talks quite as much about it, is just the global security and competition issues. And so if we're not moving fast, if we're not enabling this innovation in other countries where we are competitive, where there are national security interests are doing that, are we leaving ourselves vulnerable? And so we have all of these competing concerns and we have to strike the right balance between, you know, how fast the technology is moving, the benefits from that, you know, society accepting that, us regulating that, and then making sure that we're not constricting it so much that we're falling behind or putting ourselves at a disadvantage as a society.

45:13And I suppose one lesson history has taught us is you can't fight technological innovation, but you can try to manage its pace, but to a limited degree because of reasons like you described in terms of competition of others and just the natural progress that humans experience through time. No, I think that you're absolutely right about that. And in many cases, at the front end of every one of these big innovations, we see some commonalities, as we said. And one of the big ones is this, we will lose a bunch of jobs, the economy will be ruined, this technology innovation won't work for most people.

45:53and maybe it takes some time to sort that big change out. Maybe it takes time to sort out the winners and losers of industrialization or the internet or globalization. But eventually, new jobs get created to replace those old jobs. And we look back and say, wow, I'm so glad that technology happened because it actually did make all of our lives better. It made all of us more prosperous. There was this difficult period, but a whole bunch of new jobs were created. And ultimately, we have a much bigger pie that is beneficial to many more people as a result of this change in our society. Let's dig into the key components of an AI infrastructure and trying to keep it not as technical if possible.

46:41That would be ideal. Yeah, definitely. I think just right up front before we jump into the infrastructure piece, we've talked a little bit about large language models. Those are the big breakthrough. That's what we're seeing in ChatGPT. That's really the foundation of this big AI shift we're seeing right now. This whole infrastructure today is really built around enabling that big change, the large language models and all the benefits they deliver. And so as you're thinking about creating an entirely new technology stack, an entirely new infrastructure, it's all in service of making those foundational models better user experiences.

47:28And so when you think about a technology stack, you start most simply with the infrastructure at the bottom, which tends to be hardware, all of these middle technologies, which help enable the usage of that infrastructure. And then at the very top, you have applications. So this is where the difficult work actually gets done, where humans are actually interfacing with this core infrastructure, with these enabling technologies. And this is why we build infrastructure. We build infrastructure to do useful work with the application layer. And so for AI right now, at this foundational level, that's where most of the dollars are flowing today.

48:06That's where most of the revenue growth is. And we're putting that infrastructure in the ground for the very first time. And so that's why NVIDIA has just been on this incredible ride, not just as a stock, but obviously on revenues and its criticality to the development of the space. And so NVIDIA, for example, it's added almost$2 trillion of market cap a year a day. That's an entire Google. It's more than 10x free cash flow in a matter of quarters at incredible scale. And so we've never seen anything like that at any point in modern financial markets. And it's really because the foundational market, this infrastructure, it absolutely requires graphical processing units from NVIDIA.

48:54It needs those semiconductors. And the demand from the big clouds for those semiconductors is essentially limitless right now. NVIDIA can't make enough to build infrastructure into the ground. And then riding on top of that infrastructure are these large language models from the companies you hear about every day, like OpenAI, where they're training on these massive data sets, data sets like I talked about earlier that look at 50 to 60 % of all the data, all the printed word that has ever been created. And training one of those models is incredibly expensive. It costs hundreds of millions of dollars.

49:30It will cost billions of dollars in the future. And once you've done that training, you can actually start to do useful work for people. You can actually make ChatGPT get better and better over time. And then that middle level, we have all these technologies that essentially help advance LLMs. So help technologists utilize the underlying technology, make LLMs better, build better models, essentially get the data and the access to the infrastructure ready for that application layer where humans are actually going to do useful work. And I'm happy to talk about some other companies that are in that middle layer, that application layer.

50:12There's businesses that are at massive scale today that are putting AI into their applications. There are native companies that are getting built with AI at the very beginning that are doing application work. And obviously, if you've worked with ChatGPT, that's the quintessential AI native application that's solving real problems for users today. Yeah, let's talk about some of the leading public and private companies in this space. What do they do? What makes them the leader? Yeah, absolutely. So I already talked about NVIDIA. They make semiconductors, specifically graphical processing units. These graphical processing units allow for parallel computing, and parallel computing is absolutely essential to the machine learning and the building of LLNs we've experienced so far.

51:04Obviously, you also have the big hyperscaler public clouds that have really revolutionized the internet the last 40 years, but have particularly revolutionized the cloud and are now critical to the adoption of AI. So that's Microsoft Azure, Amazon Web Services, and Google Cloud. They're obviously all enormous beneficiaries. Then as you move up the stack, you end up with a number of public and private companies. Some of the big private companies are really around how do you store these LLM models, how you collaborate on them, how do you store the weights that you see inside of these models. These are businesses like LangChain, Llama Index, Weights and Biases, Hugging Face.

51:50They're all solving different parts of that issue. And then you also have businesses that are trying to sort of monitor what's happening around these models and look inside them, hopefully, and figure out how the algorithms are actually working and how they can be tuned and improved. And then that application layer, there's a long tail of AI native companies that are trying to build applications from the ground up that are heavily reliant on AI. But then there are public companies today like ServiceNow and Microsoft and GitLab and a long, long list that are taking all of this AI technology that's being built by the public cloud companies and businesses like OpenAI and embedding that in their applications today to make those applications better.

52:43And so the quintessential example is ServiceNow, which is a business that helps in one area deliver customer service, deliver service from HR to employees or from customer service agents to consumers. and they've embedded all of these Gen.ai products throughout those use cases. And they're actually able to charge meaningfully more, up to 50 % more, because they're really getting to measurable returns from integrating this technology. We look at the news, OpenAI and many of the LLMs seem to get a lot of the headlines. What's your outlook for the LLMs? Yeah, that is something we wrestle with all the time.

53:27And you're absolutely right. Those companies are pushing the state of the art. They have the capital from VCs, but also from strategics. They have all the leading researchers. They have incredible technologists. And now they're building out commercial teams for the first time. And so those LLMs, those foundational models, there's probably three to five that really matter right now. Plus, you know, an open source model called Lama that's developed by Facebook. In the near term, I think we're very positive there. They're advancing, again, the state of the art very fast. They're training these LLMs at incredible speeds.

54:08And I think the distance between where they're at today and somebody who wanted to launch a new foundational model is enormously wide at this point. And so we're very positive on those companies in the near term. I think over a longer period of time, there are real risks there. So, you know, there's definitely situations that could arise where they're all training on the same data and they all are kind of getting to, you know, a similar sort of terminal level of effectiveness. And at that point, I do think it's more of a race to the bottom. And there's also potential issues where, you know, good enough may be okay for the vast majority of AI use cases.

54:53You may not need to be on the leading cutting edge model for every single AI use case. And then I also think that there are just issues around they're providing, you know, API access to sort of core generalized AI function. but really most of the value creation is happening at the application layer. And so if you think about the historic model for technology innovation, you usually put these infrastructure companies at the bottom at the very start. And so in the internet, you had the big investments in semiconductors, in Cisco, in all of the physical infrastructure. And then you had the application companies like Google that came at the very end.

55:41And the amount of value that gets created at the end of the application layer tends to dwarf the amount of value that gets created in the infrastructure layer. And so right now, we may be at the phase, and I think we're at this phase, where we're generating a ton of value as we put the base layer infrastructure in. And that's NVIDIA. It's these large language models like OpenAI. But ultimately, most of the recurring value that's going to be sustainable is going to reside at the application layer. And it's probably going to reside in many companies that haven't been built yet or are being built as we speak.

56:16There's been a perception among investors that larger incumbent tech companies are the primary beneficiaries of AI. How do you feel they are positioned today? Yeah, right now in the near term, that's an absolute fact. Going back to that infrastructure piece, you need to put this infrastructure in the ground if there's any hope for delivering great experiences for people and actually advancing AI over time. And that looks like an incredible amount of investment today. So it's billions of dollars spent on NVIDIA chips. It's tens of billions of dollars invested in these models. That's all flowing to the semiconductor companies.

57:00It's flowing towards businesses that are building out data centers, building out just the core physical infrastructure to deliver that computing. And it's flowing to the big clouds who are making those resources and those data centers accessible to AI-driven companies. And so right now, most of the value creation is happening at the largest companies because they're best equipped to go put that infrastructure in the ground. But if it plays out like the internet, that probably won't be the case five to 10 years from now. What do you think the AI funding environment looks like today? It's been very healthy, as you might imagine.

57:39And valuations have obviously been pretty difficult to swallow, depending on your investment approach and what stage you're investing in. I think it's important to sort of dig in the covers of what that environment looks like specifically. And so most of the capital, call it 60 % to 70%, has flowed towards those large language models. So headed towards open AI, towards anthropic and competitive open AI, towards all of that tooling around those foundational models. And then there's this long tail amount of capital that's been focused on some of those applications that are going to get built that are AI-specific.

58:18Some of these technologies that help the consumption of these core LLMs in different ways. But I think from a quantum of capital, it's still very focused, that infrastructure layer. It's very concentrated in a few companies. And you have this long tail of companies where valuations are all over the place, but definitely high and potentially rising. But it's very difficult to pick winners and losers at this point in that group. Yeah, you talked about high valuations. We see that in public markets, as there's so much optimism surrounding AI and companies that are touching AI. Are we in an AI bubble akin to the internet bubble?

59:01Do you expect a correction? How do you compare those two? You've definitely seen some multiple expansion. So certainly NVIDIA's multiple has expanded for just looking at that company as the quintessential explanation for this. But also, free cash flow in that company has more than 10x. And that's different than what we saw in the internet period. we saw multiple increases, 5, 10x, 15x, increases in fundamental revenues and a handful of businesses, but not nearly as fast as what we're seeing with NVIDIA where that's all happening in this compressed four to eight quarter timeframe. And so for these leading companies that are really driving stock market returns around AI, a lot of that, if you deconstruct it, really is driven by fundamental growth right now.

59:50On the other side of that, whenever we potentially go into your typical sort of, hey, we're really positive, but there's potential concerns in the near term, maybe the narrative shifts, and there is some deleveraging happening there, some of that growth may start to tail off and then valuations will decline. And so I don't know where we are along that wave of valuations relative to fundamental growth. I do know, unlike the 1990s, the leading companies right now are actually generating a lot of fundamental growth right now. It's not mostly valuation expansion that's driving fundamental returns.

1:00:32I guess the key question is, how long will the fundamental returns persist? You know, when you go through a phase where all of a sudden you get this explosion in growth, it's easy to extrapolate that over a long period of time. And there's a huge uncertainty around, are we underestimating the growth? Are we significantly overestimating the growth? It's hard to know. So I guess it becomes riskier at that point. Riskier to invest and riskier not to invest. Yeah, it's a perfect, that is such a good summation you just put out there, especially the very last point. You don't want to miss the value creation when you're at this steep part of the curve and it's very difficult to predict what NVIDIA's revenues could be in two quarters and four quarters.

1:01:20But at the same time, we know these things are never straight lines. While we may be super positive on where AI is in 10 or 20 years, just like we're super positive in 99 or 2000 around where the internet could be in 20 years, along that path, there's going to be periods where, you know, maybe some of these companies don't grow and you actually see a lot of, you know, deleveraging in a business like NVIDIA. Even if, you know, shutting your eyes over 10 years, you're going to end up with a business that's 10 or 20 times larger from a fundamental perspective. My last question is, if we look forward long-term for the potential for AI, Are you confident that it's going to be transformative?

1:02:03And are there certain proof points that you're looking for in terms of its potential, particularly relative to what people expect? Yeah, of course. So I think over the long term, we're very positive. We're already seeing enough point use cases and adoption that are broad enough. You see how those could scale and expand from here. So we are very early, but we're seeing measurable adoption, measurable returns, and we're monitoring an enormous number of those proof points. And so I kind of think about those proof points in two categories. So there's laying the foundation, which we spent a lot of time just talking about, which is, are we putting the capital in the ground?

1:02:45Are we building the infrastructure? Are you putting out the data centers, deploying the chips, building the foundational models that will allow us to deliver incredible experiences that will allow for adoption. Everyone's really focused there. A lot of great proof points there. I think we're going to get there. I think we're going to build the infrastructure we need for the next 10 or 20 years in the same way we built the infrastructure we needed for the internet. Maybe we overbuilt in 1995 to 2001 with Cisco and all the fiber we put in the ground, but that fiber became the backbone of Verizon.

1:03:19And that was required because the requirements of the internet ended up being thousands of times larger than we thought. And so maybe we're at that point right now. But I look at how we're laying the foundation. We may be overlaying it for the near term, but we're probably underlaying it for the long term. And so we're measuring those proof points, those adoption, those laying the foundation proof points. And then on the adoption side, thinking about it on the consumer side, thinking about it on the company side, On the consumer side, we're looking at user growth. We're looking at how frequently people are using these products, the sort of work they're doing.

1:03:53We're looking at the applications and the user experiences and the user interfaces. Are there applications arising that make this feel like magic, where someone doesn't even understand or really care what's happening under the covers? They're just having magical experiences. And most of the time, they interact with these AI-powered features and applications. They're getting incredible results. They're not getting turned off. And so you can really measure that adoption and engagement and then look for those proof points of, are there these incredible applications that have great experiences on the front end?

1:04:28And then on the enterprise side, are real business problems getting solved that are measurable, that have real ROI attached to them? And across many industries and many use cases, we are seeing exactly that. We can look at usage data. We can look at some of the revenue opportunities that have been unlocked by Gen.Ai. We can look at some of the measurable increases in productivity for existing employees. We can look at some of the cost savings that have come about as a result of Gen.Ai. And so, you know, I think on the adoption side, we see, you know, both quantitative and qualitative ways to measure that consumer enterprise adoption.

1:05:07This was such an interesting conversation, Jacob. I appreciate you taking the time. I could do this all day. And we just scratched the surface of the qualitative piece down to the weeds, right? Yeah. And your passion comes through. and I wish I could sit in your seat for a week and be in front of all these great companies that you're constantly analyzing. And to me, just being on the front lines during this technological revolution is pretty fascinating. So thank you. It's never been a more exciting time personally to do this job and it's never been a more exciting time to be in this industry and actually to be back in San Francisco and just the Bay Area in general.

1:05:54where that you can feel that energy in a way that, you know, my peers who were investing in the 90s, it's the closest thing they can describe to that, that sort of spirit of innovation and excitement. That's awesome. Thank you, Jacob. Thanks for listening. We hope you enjoyed this episode. Please visit our website at insightfulinvestor.org to access past shows and learn more about our podcast. If you have questions, feel free to email us at info at insightfulinvestor.org. And if you enjoyed the discussion, please subscribe to this podcast to ensure you don't miss future episodes. And don't forget to forward today's conversation to others you think would enjoy listening.

1:06:37This podcast is provided for informational purposes only and should not be relied upon as legal, business, investment, or tax advice. All opinions expressed by podcast participants are solely their own opinions and do not necessarily reflect the opinions of Evoque Advisors, their affiliates, or companies featured. Due to industry regulations, participants on this podcast are instructed not to make specific trade recommendations, nor reference past or potential profits. And listeners are reminded that securities trading, commodity trading, and alternative investments are complex and carry a risk of substantial losses.

1:07:12As such, they are not suitable for all investors.

1:07:19Listeners should be aware that guests featured on The Insightful Investor may have current or past associations with Evoke advisors or the host, including as an investment manager of a private fund opportunity by Evoke, or access through an affiliated Evoke fund, or as a client. Participation as a guest on the podcast should not be perceived as an endorsement or testimonial with respect to Evoke Advisors, the podcast host, or their services. Similarly, the inclusion of a guest on the podcast does not imply that Evoke Advisors or the host endorses the guest or any company with which they may be affiliated or employed.

1:07:58Evoke has neither paid nor received compensation from guests for their participation.