Brett Winton Podcast

In today’s episode, we speak with Brett Winton, the chief futurist for Ark Invest. Brett joined ARK in February 2014 and has worked alongside Cathie for almost 15 years since their time at Alliance Bernstein. As Chief Futurist, Brett drives ARK’s long-term forecasts across convergent technologies, economies, and asset classes, helping ARK dimension the impact of disruptive innovation as it transforms public equities, private equities, crypto assets, fixed income, and the global economy. Brett also serves on the ARK Venture Investment Committee. Brett joined ARK as Director of Research, guiding and managing the proprietary research of ARK’s investment team. Prior to ARK, Brett served as a Vice President and Senior Analyst on the Research on Strategic Change team at Alliance Bernstein. Brett earned his Bachelor of Science in Mechanical Engineering at the Massachusetts Institute of Technology (MIT).

In the episode, we covered everything from why disruptive technology now through the size of the opportunity for some of the portfolio companies like Tesla, Coinbase and DraftKings.

‍

Please see the transcript of the show below! 

‍

[00:00:00] 

[00:00:00] Ryan Loehr: Welcome to the exchange podcast by EWL. As advisors to some of the most successful families in the country, Craig Emanuel, Tim Whybourne and I, Ryan Loehr, draw upon some of the best minds in the country. We believe that by exchanging ideas, we can deliver better advice and better outcomes for the families we worked for.

[00:00:29] Ryan Loehr: Now, we're inviting you on this journey. In this podcast, we interview some of the country's best investment managers, business advisors, bankers, and founders to share their valuable insights. And our hope is that with better information comes better decisions, helping you to achieve more financially. 

[00:00:50] Tim Whybourne: Good morning, everyone, and welcome to another episode of The Exchange by Emmanuel Whybourne and Loehr.

[00:00:55] Tim Whybourne: My name is Tim Whybourne, and I am your host today. Today we have a slightly different format in a [00:01:00] webinar we recorded with Brett Winton from ARK Invest in mid December 2023. Brett's joined ARC in February 2014 and has worked alongside Cathy for almost 15 years since their time at Alliance Bernstein. As Chief Futurist, Brett drives ARK's long term forecasts across convergent technologies, economies, and asset classes, helping ARK dimension the impact of disruptive innovation as it transforms public equities, private equities, crypto assets, fixed income, and the global economy.

[00:01:26] Tim Whybourne: Brett also serves on the ARK Venture Investment Committee. Brett joined ARK as Director of Research, guiding and managing the proprietary research of ARK's investment team. Prior to ARK, Brett served as Vice President and Senior Analyst on the Research on Strategic Change team at Alliance Bernstein.

[00:01:41] Tim Whybourne: Brett earned his Bachelor of Science in Mechanical Engineering at the Massachusetts Institute of Technology. In this episode, we covered everything from why disruptive technology now, through to the size of the opportunity for some of the portfolio companies like Tesla, Coinbase, and DraftKings. Before we start today, we must disclose that Emmanuel Wybon Lewer and its directors [00:02:00] may hold positions in ARK and ARK related funds.

[00:02:03] Tim Whybourne: So without further ado, let's have a listen. 

[00:02:06] Brett Winton: For why we started arc as I think never been more vital and clear than it is now, from my perspective you know, we started arc to focus purely on disruptive innovation. We focus on 5 innovation platforms. AI public blockchains, robotics, what we call multi sequencing.

[00:02:26] Brett Winton: It's being able to read and operate on all of the molecules in your body. And energy storage. And those are 5, you know, big, major buckets that we think are hitting critical stages of inflection. All over the course of this, you know, business cycle, and probably the business cycle to come and a way to think about this is we think future historians will look back on each of those platforms individually and say.

[00:02:53] Brett Winton: Oh, my gosh, I can't believe that all of those were happening at the same time. Each 1 of those is the equivalent of [00:03:00] electrification or the internal combustion engine. Each of these platforms have have common characteristics that are carried across what are called general purpose technologies. They all have steep cost declines.

[00:03:10] Brett Winton: They all cut across vectors and they're themselves platforms of innovation. So, meaning things can be built on top of and and so if you go back and look at kind of. How often these general purpose technology platforms arise and how meaningful they are. It's kind of like you get if you're lucky, you know, 1, every 20 years and you have to go back 100 years or more than 100 years to see this amount of kind of technological fulfillment.

[00:03:38] Brett Winton: Then you had internal combustion engine, electrification of the telephone, all entering the economic marketplace at the same time. Now we have 5 and actually. We think they're, they're, they're even in some cases more meaningful than those previous like canonical examples of this is a major innovation platform.

[00:03:56] Brett Winton: And so, the, what's really interesting about capital [00:04:00] markets to me right now is 1, I'm actually more convinced that these are meaningful than I've ever been. The AI in particular is moving faster than any technology that we've studied as a steeper cost decline. It also has this kind of like.

[00:04:13] Brett Winton: Catalytic benefit to all the other technologies that it's pulling all of them forward with it advances in AI impact your ability to do Robo taxi obviously impact your ability to kind of like, get, you know, next generation drugs to market efficiently. Clearly feeds into your ability to, to, to develop robots that are going to work in the world.

[00:04:34] Brett Winton: And also makes kind of like the imposed scarcity that public blockchains offer a more valuable and differentiated kind of measurement characteristic of economic activity. And so it's kind of like the acceleration in the eye is accelerating everything. Simultaneously, the capital market seemed to think that innovation is with the exception of a couple companies, which I'll touch on is, it's basically like a [00:05:00] story that is past that that we're past the moment that kind of this was all coven inspired fever dream and you can see it in the last couple of years of performance where you know, it's been extraordinarily kind of painful and humbling for us.

[00:05:12] Brett Winton: And I'm sure for our client to have stuck with us, and we really appreciate that to see how. Kind of rate sensitivities assets are in the marketplace. And, and so kind of the, for a long time, the portfolio holdings that we've held have been treated almost as if they're all 1 asset, which is a very long duration asset that has a lot of rate sensitivity.

[00:05:32] Brett Winton: And so when you had an unprecedented rise in interest rates over, over a couple of years the assets as a whole, the innovation asset class is underperformed. And so, from my perspective, the you know, if, if anything, kind of our, our bias right now is actually to, to make our forecasts on the top down, like, value of these technologies more aggressive.

[00:05:55] Brett Winton: Technology continues to surprise everybody. Meanwhile, the capital [00:06:00] markets are you know, tending to treat the technologies as if they are less valuable than they've ever been. If you take our top down work across the technologies and you say, well, how valuable are they going to be? And we published this last year, but we think across.

[00:06:14] Brett Winton: All the innovation platforms that we focus on, and there's a set of underlying technologies that we explicitly model under those technologies that collectively they're going to comprise more than 200 Trillion dollars in enterprise value by 2030. and so. To give you kind of a sense of reference, there's at the end of 2022, there's a 100 trillion in market cap globally.

[00:06:38] Brett Winton: So, so it's not hyperbolic to say that we think that two thirds of market value is going to be, you know, these disruptive innovation platforms and the tools that enable them and agile companies that are aggressively deploying them. And, and so from a kind of strategic allocation perspective, you know, the main thing I'm passionate.

[00:06:57] Brett Winton: About talking to people, the reason I talked to them [00:07:00] is because I think people are unintentionally short innovation and it actually is the thing that is going to drive long term wealth creation over the course of this business cycle. What do I mean by unintentionally short? Well, they've been scared off because of what's happened over the past couple of years.

[00:07:18] Brett Winton: So they anything they had, they got scared out of. And they probably have positions in their core portfolios that are at threat of disruption if we happen to be right about how these technologies are going to develop. And so the there's the, there's wrong way risk in that. If I happen to be right on how valuable innovation get, then that rail company you have in your portfolio, that traditional branch based bank that you have in your portfolio suddenly gets into, you know, really financial hot water and maybe becomes a smoldering crater in your traditional portfolio.

[00:07:55] Brett Winton: So if an EV autonomous truck is going to be cost competitive with freight [00:08:00] rail, which we think it is then probably your rail company, your corp. Portfolio becomes potentially a bankruptcy, or the traditional automotive manufacturer, who is right now pulling back on their ev spent becomes a bankruptcy.

[00:08:13] Brett Winton: So you better own the platform that appreciates a lot in value as it's disrupting something that you hold in your core, you know, set of positions so that you don't just to like right size your risk and kind of like. Measured over the course of the business cycle, I think the more than the majority of value is going to reside in these innovation assets.

[00:08:35] Brett Winton: So, from a kind of like, closing off your unintentional innovation short, and then kind of trying to position yourself to be, you know, to accrue value over the business cycle. I think it's a strategic allocation that people need to have. What's interesting now is there's also a tactical opportunity because it's clear that the interest rate cycle, at least from our perspective is, you know, beginning to turn and [00:09:00] just the, the performance of our style of positions of the last.

[00:09:04] Brett Winton: Month, month and a half is clearly indicative that there's a you know, a lot of kind of like, potential return just on kind of like the interest rate cycle unraveling a little bit, the Fed cutting rates, there'd be more liquidity in the system. And so you have an opportunity for a tactical allocation to innovation in an area where you need a very strategically aggressive exposure in our view of the course of the business cycle.

[00:09:29] Brett Winton: Now is the time. And so then, and in particularly, because the, the, the, the evidence that innovation is going to be meaningful and meaningfully disrupt and enhance our lives is stronger than it's ever been. So just to focus, like, a 2nd on AI and what that means for all of these technologies, but artificial intelligence has.

[00:09:50] Brett Winton: Of all of the technology, we study the steepest cost decline. We think that the cost to train an AI model is following 3 X per year. So [00:10:00] to people might be familiar with Moore's law, which is kind of a technology cost decline. Moore's law implies that every 2 years, the cost of a computer falls in half.

[00:10:09] Brett Winton: Now, how does that actually play out? It's not that I get a computer for half the price. 2 years from now, I get, I pay the same price for my laptop, but it's 2x more performant 2 years from now. Well, with AI, that's happening every 6 months, at least that. And we think it's going to happen through 2030, every 6 months.

[00:10:26] Brett Winton: And so we, we did the work deriving that cost decline. Publish that forecast a couple of years ago, when we published it, frankly, I was a little nervous. I was like, this is seems aggressive, but it's given the work that we've done. It was consistent with our cost decline methodology. And we thought it was, you know, it's what the forecast said.

[00:10:47] Brett Winton: So that's what we published. Since publishing it, we've actually been too conservative. So just even over the course of this year open AI upgraded GPT for the GPT 4 and a half turbo. [00:11:00] In doing so they cut prices by roughly 3 times and it wasn't a full year. It was over 240 days and the models got like 4 times faster at producing.

[00:11:10] Brett Winton: Kind of words in sentence completion, so adjusted for performance. It's something like a 10 X decline in less than a year. And so kind of the, the that's both useful on its face. Like, hey, now I can use a large language model to get my answers more quickly and less expensively. And it's also indicative of how performant these models are getting in all kinds of different areas.

[00:11:36] Brett Winton: So, rather than, you know, paying a 3rd, the cost, you can maybe string together a bunch of queries to allow the models to to produce more complex tasks for you as a knowledge worker inside an organization. And, or you can take that model and because it's less latent you can put it in a self driving car, or you can put it in a robot and [00:12:00] you're not subject to the fact that it's trying to think of the next word or the next turn to take and the car drives off the road.

[00:12:06] Brett Winton: And so kind of with this improvement trajectory. All of the other technologies that we work on are getting pulled on side more quickly, particularly in earlier stage technologies, like multi mix and robotics, where kind of like some of the humanoid robot capabilities that we anticipate over the next.

[00:12:27] Brett Winton: A couple of business cycles are now looking more and more likely that they're going to fall into this decade, given how fast things are moving. So that's like a high level overview. What do you do? You all think I'm crazy? Or what are your thoughts? 

[00:12:41] Tim Whybourne: Not at all. I've always been a. Yeah, big believer in the thematic.

[00:12:47] Tim Whybourne: I think it makes a lot of sense moving forward, but I think the technology is easy to get your head around and the market for how quickly the technology is moving, I think, obviously, in EVs, batteries have got a lot better with autonomous vehicles, the [00:13:00] technology has got a lot better and safer wIth genomics, I think that they are really solving or deleting cancers from, from from existence.

[00:13:09] Tim Whybourne: But we, I think what is hard to get our head around is just like the correlation between the price movement and the tech and the evolution of the technology. And the instance of, of autonomous vehicles. So obviously we have autonomous vehicles and I believe they're currently safer than a, a driven car.

[00:13:25] Tim Whybourne: The adoption is not there because we. It's not legislated. So when do you, in the, in the case of autonomous vehicles, as an example, like, when do you think that will translate through to, to return on share price? 

[00:13:38] Brett Winton: Sure. So let me disaggregate 2 things out of that question. 1, yes, autonomous vehicles exist.

[00:13:45] Brett Winton: I've ridden in a Waymo mobile in San Francisco. It's amazing. And and, but they don't exist at scale and the existing solutions that. Are working are actually not at least as far as we can tell [00:14:00] done in a way that allows them to scale. So, actually, the real puzzle of waymo, which is a Google subsidiary or an alphabet subsidiary is they can deliver autonomous rides, but they aren't aggressively.

[00:14:14] Brett Winton: At least given the size of Google's balance sheet building out a whole fleet of autonomous car in every single city so that these things are everywhere. And it's just the way people get around. Part of that is because the way mobiles. They, we don't know exactly, but, but we can triangulate and they cost somewhere around 150, 000 dollars to 200, 000 dollars per vehicle to deploy because they have.

[00:14:39] Brett Winton: Tens of sensors, they have, you know, multi thousand dollar LIDAR units. There's maybe 5 of them on the car. There's I think 29 cameras. They have ultrasonics, they have radar and they have some tens of thousands of dollars of computer equipment in the trunk to process all the information coming through those centers and [00:15:00] kind of like, yield the kind of driving plan for the car.

[00:15:03] Brett Winton: And so. Even at that price, if Google thought that the system could scale efficiently, you could actually end up with a price per mile. That was very attractive to consumers. If you rolled them out. If if you could roll them out kind of anywhere, or if you could fact rate a city with these vehicles the reality as far as we can tell is that that it's not.

[00:15:25] Brett Winton: That easy for way to take cars in San Francisco and say, okay, now we're going to have them drive around Charlotte or I don't know, pick a Brisbane or something. And and, you know, and there's potentially some like conservatism from within alphabets, corporate culture. That's preventing them from scaling is aggressively imagine.

[00:15:43] Brett Winton: You're the person who is like the middle manager inside. Waymo you know, your marginal interest in, like, pushing the frontier here is dulled a little bit by the fact that you have a great total comp package from Google and you get, you take Fridays off if you want. And and if you deliver [00:16:00] it across the finish line, you don't necessarily get any of the upsides and.

[00:16:03] Brett Winton: Search is such a cash flow machine that it doesn't like, cleanly translate into kind of economic incentive for you to really be aggressive. In fact, the worst thing that could happen is is. Even though it's the right thing to do, you try to aggressively expand the business and something happens that seems like it's like a reputational risk to Google.

[00:16:22] Brett Winton: And so you get fired because, you know, living the good life at Google is fine. You don't actually need to do anything to live the good life. And I think that's actually a structural problem that has that that I'll probably get into a little bit later. But the, you know, Tesla also has a full self driving solution.

[00:16:38] Brett Winton: You cannot have it drive around and take people around. For rides right now, not because there's a regulatory thing that stops them from doing it or not directly. It's really because the system's not good enough to do that yet. I use it all the time. It's an amazing luxury product for someone who has, you know, kids in the back seat that you have to, like, try to give snacks to while you're driving on the 4 or [00:17:00] 5 in L.

[00:17:00] Brett Winton: A. And it, and it definitely stress reduces driving. But it still requires interventions and help and monitoring. And so the combination of me, plus the car driving itself is actually statistically quite a bit safer. And we think almost 20 times safer than the average U. S. driver driving on city streets.

[00:17:20] Brett Winton: But not not just the system itself yet. So, the, the question is like, when does this translate into, into share price? Well, think about Tesla's business model in this way. When Tesla sells a model three to you, they're selling you a 50, 000 car, or at least in the U S it's 50, 000. I don't know what you can buy it in Australia, but and the net profit to them is maybe.

[00:17:44] Brett Winton: 5, 000 I'm using very round crude numbers here, but they roughly correspond with reality. Okay. And roughly 5, 000, but it's a 1 time sale. It's not a great business from a, like, high level perspective to have to build these really [00:18:00] giant factories and then produce cars out of them to a set of end users that have volatile demand patterns.

[00:18:06] Brett Winton: So, kind of like. Okay. The, the number of vehicles sold in a year falls by 30%, and then your factory is underutilized. And you know, this is the reason that automotive companies go bankrupt over time. What happens if they can deliver a software update to their vehicles that allows it to operate as a robo taxi by itself?

[00:18:25] Brett Winton: What does that mean to Tesla the business? And what does it mean to the owner of the vehicle? Well, they, we think that when robo taxis launch at scale they will at 1st, be able to price competitive with Uber. Here's a, here's something that provides you an Uber ride. But you don't have to chat with the Uber driver, which is great.

[00:18:45] Brett Winton: And it's probably it's safer than the Uber driver. So 1st, they should be like 2 and a half. Dollars a mile, but then as they penetrate the market, these robo taxis will have to come down in price to draw more people into the market. We think there's a trillion dollars or a [00:19:00] trillion addressable miles at a dollar per mile.

[00:19:02] Brett Winton: So let's take that as a kind of intermediate penetration point for Tesla. Okay. So the, they deliver a software update to that model 3 that they've sold to you and it can do 100, 000 miles a year if aggressively managed. So. Okay. 100, 000 miles at 1 a mile is 100, 000 in revenue. Tesla will take a platform fee just like Uber does.

[00:19:24] Brett Winton: Tesla will take probably a better platform fee because it's delivering all these. Autonomous features and there's no real competitor, but just say it's 50%. So 50 percent platform feed. That's 50, 000 net revenue to Tesla. And if they're getting software margins off that, which they should be able to then it's 25, 000 operating profit to Tesla off of that vehicle per every year that that vehicle exists.

[00:19:50] Brett Winton: So they go from selling you a vehicle once. 5, 000 in operating profit to 25, 000 off that vehicle for every year. It's in service [00:20:00] across all of their vehicles, because at least the way they're developing the software, they think they're going to be able to deploy it to kind of all of the vehicles in fleet.

[00:20:08] Brett Winton: And so, the whole business model transforms as it starts spewing out amazing cash flow, and it's very clear from the way the stock is priced. That's not. Underwritten in the stock, people are worried about next year's. Oh, gosh, they're going to have to cut prices. Their gross margin is going to drop from, like, record in the automotive industry into the high teams, which is so remarkably good for the automotive industry.

[00:20:33] Brett Winton: This is not a thing that most people don't believe that it's going to happen or think that it can't happen or just. Aren't willing to underwrite past next year's results, but you're not being honest with yourself if the advances in AI don't simultaneously, like, improve the odds, you think that Tesla is going to deliver this outcome.

[00:20:57] Brett Winton: And by the way, you who bought the, [00:21:00] the model 3, you know, that 50, 000 that Tesla takes, you also get net 50, 000 revenue and your cost to operate that vehicles around. 20 cents a mile, even accommodating, you know, you'd have to clean it. You'd probably have to do it as a full time job to really run it 100, 000 miles a year.

[00:21:17] Brett Winton: But then you, you'd clear 30, 000 in income off that car. So 1 Tesla could actually charge a higher platform fee than I'm presenting here. Or 2, that car is no longer worth the 50, 000 you paid for it. Suddenly it's an asset that generates 30, 000 a year. You know, it's worth like 250, 300, 000 or something, you know, so the, the, the, what will end up happening.

[00:21:40] Brett Winton: And 1 of the interesting solutions to it's like, what if they're regulatory barriers? Like, North Carolina doesn't allow it to happen. Well, then all of the model 3s that are North Carolina will be purchased by operators who are operating in Texas and be like, give me that. We need this for our robo taxi fleet over here.

[00:21:57] Brett Winton: Now, and so the, the the setup for [00:22:00] the business for, like, translation to business performance is like, financial metrics is really remarkable. Now, the problem with this whole scenario. There's a couple, but 1 is this is all stuff that if you're underwriting over 5 years, you can foresee, then you can say this belongs in a model.

[00:22:19] Brett Winton: In fact, I'd argue you were doing. You're not operating responsibly. If you're not underwriting some probability that this happens in a Tesla underwriting measured over a reasonable timeframe. And we underwrite all our positions in 5 years, but if you're, if there's like, say, there's a set of buyers in the capital markets and equity markets.

[00:22:41] Brett Winton: Some of those buyers are ones that are kind of like, willing to not do real work, but kind of like, take a risk on an idea. And then some other set or ones that aren't willing to take that kind of risk. And then those buyers and sellers are, they're controlled by the interest rate cycle as to whether or not they [00:23:00] enter or not.

[00:23:00] Brett Winton: Okay. Then the marginal price of Tesla is determined by people who are partly underwriting this when you're in a risk on period, and then a set of people who won't pay any attention to this when you're in a risk off period. And so that's part of the reason why innovation companies have. You know, more of this kind of rate and liquidity sensitivity is because kind of like the people who are setting marginal prices in the market are right now.

[00:23:27] Brett Winton: They're not underwriting that possibility. Now, I argue over the course of the business cycle. That's that's irresponsible actually. And and from a, like, you know, I want to, I want to maximize wealth for my clients over time. So I'm willing to look through that. And it's a reality of the way capital markets work.

[00:23:44] Brett Winton: That, that, you know, people, some people will refuse to account for anything in a company's financials until they see it reported in the 10 K. 

[00:23:52] Tim Whybourne: Yeah. So, so what probability does ARC put on that becoming a hundred percent certainty? 

[00:23:58] Brett Winton: Well, there, at [00:24:00] this point we think it's a when, not if there's a question of when.

[00:24:04] Brett Winton: So we have a distribution of potential timeframes in which they successfully launch full self-driving. And we think late 24 is a likely 1st commercialization at this point, they add by the end of this year, and they've launched it to employees only at this point. But by the end of this year, they claim, or they said that they'll release a new version of the software, which is neural nets from AI software from top to bottom though.

[00:24:33] Brett Winton: And by doing that, they're going to eliminate 300, 000 lines of code out of their software. And so it's like, what are those lines of code? Like, imagine that you're trying to, you're trying to think of how you would tell a car to navigate, like, a 2 lane roundabout. Right? And you have to think of all of the conditional statements of, oh, if there's a car in the other lane, it's not safe.

[00:24:55] Brett Winton: Or how do I determine what lane to get into to get to the, like, Okay. [00:25:00] Exit lane that I need to and so you can imagine that if you're developing that you're not going to really pre anticipate all of the conditions that you need to pre anticipate. And so then it'll operate on the road and it'll fail in certain spots.

[00:25:14] Brett Winton: Right and so then you have to find the spot that fails and then you have to have a smart software engineer be like, okay, let me think about how to insert another set of sub clauses to accommodate this particular corner case. So, by that's the state of. Their technology today for most for all of their customers, excluding their employees.

[00:25:33] Brett Winton: Once they're full AI and the problem with that is, like. Every unique problem you end up with requires its own unique solution, right? You have to like, then assign a smart software engineer to take the time to think about what's going on here. Let me figure it out. So once you go eliminate that code, what you're replacing it with is a high performance compute and and lots of data.

[00:25:56] Brett Winton: And so, kind of if you have a problem with 2 lane [00:26:00] roundabouts once you're full stack AI, the solution to that is just feed it lots of data from people successfully driving through 2 lane roundabouts. anD so every unique problem has the same solution, which is more data and more computing. And so it both means, like, the, the, their rate of progress net gets controlled by how much high performance compute they have, which is why they're building out their own chips and data centers.

[00:26:26] Brett Winton: And the, I think that their rate of progress will become more predictable. Even for them internally because before it's kind of like, oh, there's a problem here. If we fix that, we're going to be here, but actually there's an intermediate problem here and here and here, each of which require their own fix.

[00:26:42] Brett Winton: Whereas once you're full stack AI, it's kind of no matter what the problem is, it's like, hey, kind of we're solving these problems at a consistent rate based on how much additional compute and data we're throwing. At the problem so net, we think commercialization in 2024 is the likeliest timeframe [00:27:00] measured over the time that we underwrite like a 5 year underwriting.

[00:27:04] Brett Winton: If it happens in 25, that's actually, it's not like that meaningful difference from happening in 24. And, and the, if you look at our expected future value. Of the company, so 5 years forward, 70 percent of the attributed value is is for kind of, accrual to revo taxi, basically like that magical cash flow machine.

[00:27:28] Brett Winton: Even though we don't assign an abnormally high multiple to it, we do we do treat like, different business lines. Subject to their capital intensity and kind of their margin structure. If I'm a 4th seller of that cash flow to like, private equity at that time frame. And so it is higher multiple profits than their existing EV business.

[00:27:51] Brett Winton: You can underwrite the business, excluding Robo taxi entirely. In fact, we think fair value of the electric vehicle business alone is. [00:28:00] If you look at our open source model, it's around 700 a share or so. So imagine you pretend. Robotaxi is not going to happen. You still actually can be relatively aggressive with Tesla.

[00:28:14] Brett Winton: But the, the, the odds of it now are much higher. And, and just because of how quickly AI is moving. And like another way to think about is the way most doing it with like, call it 150, 000 worth of kit. On a car, Tesla's trying to do it with maybe 2 or 3000 dollars worth of kit on a car. That's why it's so hard.

[00:28:36] Brett Winton: But their advantages, they have literally millions of vehicles on the road, collecting data for them. Whereas has hundreds of vehicles. You know, and so kind of like the, the data advantage they have was not as much an advantage 5 years ago as it is today, because they didn't have the, the, the software to take advantage of all the data they're generating now, kind of like the software tools are [00:29:00] right out there.

[00:29:00] Tim Whybourne: Okay. And so that that's that's, that's 1 thematic. So on the genomic thematic is artificial intelligence making a medical difference to the acceleration of the problems they're solving? 

[00:29:13] Brett Winton: Yeah. So the way we think about we call it multi omics now, just because the number of tools and the variety of data you're able to collect has expanded and that's actually where AI is really important because just like call it an explainer of different types of, of data.

[00:29:29] Brett Winton: You get off the body. If you're doing gene sequencing, every 1 of yourself has basically a recipe for how your body. Yeah. Constructed and a gene sequencer tells you, hey, here's what that recipe is. But the reality of what's going on in a cell is it's not just reading that genome. It's actually reading different patches of it.

[00:29:47] Brett Winton: And the different patches depend on whether or not it's, you know, a liver cell or a lung cell or a cancer cell. And so you can do what's called RNA sequencing, which reads not just the genome, but reads the expression. [00:30:00] Of the genome, or then there's the epigenome, which is kind of like a layer on the genome, which determines kind of like how the RNA plugs in.

[00:30:09] Brett Winton: And then there's the proteome, which the RNA reads that recipe and produces proteins, which are actually what are like the molecules operating in the body. And so we have. Tools for measuring all of these things. The real challenge is like, well, I have 1 data set here and 1 data set here and 1 data set here and 1 data set here.

[00:30:26] Brett Winton: And the output is some kind of disease or thing going on in my body. How do I know where it's coming from? And what, what AI is really great at is, is actually translating from 1 data set to another. That's in some ways, you know, I don't know if people realize this, but the. The, the major advance in language model generation that we're seeing that open is using and and Google is using for all these things that are in the news.

[00:30:51] Brett Winton: The, the paper that introduced that advance was actually trying to solve the problem of translation from 1 language to another with. So. There's a, [00:31:00] it's a famous paper called attention is all you need. It was really like, this is a breakthrough in language translation. And it turns out that capability of like translating from one sequence of words to another in a, in a, in a robust way is also like useful for generating new words and, and all of the magic that we're seeing.

[00:31:18] Brett Winton: So practically how does it play out? Given kind of like all of this data that we're getting and, and, and how we can use AI. To kind of help us interpret it and even generate new data well, using kind of the same transformer architecture that I just described. That's what's used in these large language models.

[00:31:36] Brett Winton: They're also able to read the, the, the DNA, the genome in a person and predict what kind of protein it's going to spit out, which is something that couldn't be done before. And that really helps. Early in the drug development pipeline, when you are saying, hey, this person has something going wrong with them.

[00:31:54] Brett Winton: We know what's wrong with this gene. What is the structure of the protein that's coming out of that gene? [00:32:00] And then how can we plug something into it to stop that protein from hurting the body in some way? And so, there's actually a company in the portfolio that that is really driving this forward called recursion.

[00:32:11] Brett Winton: That that, they present that they can more than half the cost for early stage discovery. Of so pretty getting into the clinical trials, but the 40 percent of of, of R and D spend in pharma is early stage roughly. And so they can have the cost and have the time to go through that early stage discovery process by using kind of data in the AI systems that have come out.

[00:32:38] Brett Winton: So, but they're an interesting exposure from that perspective. And then at an aggregate level, kind of like the biotech and pharma industry has, with the exception of the covet year, which is dominated by, you know, all the. Vaccine sales had a decaying return on R and D rate over a couple of decades. And we think that kind of [00:33:00] starting at the early pipeline.

[00:33:01] Brett Winton: As people are deploying against it and able to reduce the cost to get to an interesting kind of drug that they're going to release and I and reduce the time you're going to begin to see a tip up in that return on and so you can actually make a case for, you know, the entire space, the entire biotech space being interesting on that basis, because it's not it's not priced particularly now, given what's happened in markets as if there's going to be like, return on those dollars.

[00:33:32] Brett Winton: That's that's, you know, any better than there has in the past. And if you also weight the portfolio towards earlier stage in the clinical pipeline and towards those companies that are more aggressively deploying these tools, we think you get actually a better kind of return on invested capital than even that.

[00:33:50] Brett Winton: So, that's how I is impacting the front end. There's probably a, like, a later advance. That's not being deployed today, which is imagine once I get into the clinic with [00:34:00] patients, and I need to find. 1, 000 patients for my clinical trial. That's actually a big part of the cost is like. Recruiting those patients, the time it takes to recruit those patients, and then the fact that you have to administer the drug to all 1, 000 of those patients across some period of time.

[00:34:18] Brett Winton: Half of those patients. You're not actually giving the drug to because they're the control group. So so you have to go through all the trouble of recruiting them and bringing them in and administering sugar pills to them. And so there's the possibility of being able to say, hey, instead of having this control group that we're not doing anything to accept, just measuring, you know, what happens to them.

[00:34:41] Brett Winton: If we don't give them any treatment. Do we have enough data from hospital systems from HR systems to synthetically create that control group? And so we no longer have to, you know, get 1000 patients to get the same. Understanding of how efficacious the drug is, we can do it with 500. Well, that would [00:35:00] also have costs.

[00:35:01] Brett Winton: It might not. And in fact, it probably wouldn't have the time, but it would reduce the cost of getting a drug to market in a really meaningful way. And so across both, you could potentially have that as they both get fully deployed, have the average cost to develop a a working drug. But the synthetic control group stuff is a little further out.

[00:35:23] Brett Winton: And then you know, these things proliferate into the to the aggregate landscape at the pace at which kind of companies take them up and deploy them. And so kind of more innovative and natural companies. We think we'll have differentiated performance given the, the power of these tools.

[00:35:39] Tim Whybourne: Okay. And I remember asking I think it was Kathy a while ago about the most exciting dynamic.

[00:35:46] Tim Whybourne: Thematic was and the answer was genomics at the time in terms of just market opportunity and how big it could potentially be when you look at the ARKK portfolio, so the flagship portfolio, is that, do you think that's still the case or do you think there's there's [00:36:00] a bigger opportunity as in terms of the general thematic?

[00:36:03] Brett Winton: I mean, I think it's hard, like genomics is definitely at the earlier stage and there is more. I think every advance there's there's more headroom in terms of advancement. Like, it's very clear that kind of the tools that we're deploying are in their earliest stages. And that that there's a lot of kind of opportunity to come.

[00:36:24] Brett Winton: Like, we've been talking about crisper gene editing for for years now and just today. Was the, or not today, but this month, the 1st you know, CRISPR therapy got approved in the U. S. and, and it cures sickle cell anemia or seems to cure sickle cell anemia in the U. S. there are 100, 000 patients with sickle cell anemia and the, the avoided cost of treating a patient with sickle cell anemia.

[00:36:50] Brett Winton: Like, if I, if I take somebody born with sickle cell anemia, and I see the access I have to. Pay on just managing that condition. Given the current state of play [00:37:00] prior to this cure coming out, it was a note, I think $1.4 million over the course of that patient's life. And they lived to an average age of 50.

[00:37:07] Brett Winton: And so, you know, CRISPR treatment is, you know, priced at 2.2 million, which sounds like a lot, but it's actually, that's just like a net cost savings for 1.4 of that, you know, and you're giving the person like a normal average. Lifespan, you're, you're effectively eliminating the multiple hospital trips.

[00:37:25] Brett Winton: They have to take a year, you know, all of the transfusions and you're shifting spend out of a category. That's very hard to cost reduce kind of like hospitalization and into to Madison. And so, we think that. This is the 1st in a whole set of new drug modalities that are going to be able to transform what are things that that people kind of tolerate and live with with drugs that are like, actually poorly targeted at them, including supported by a lot of kind of.

[00:37:56] Brett Winton: You know, visits to clinics and, and going into the hospital and [00:38:00] doctors with cures and effective cures for these conditions. So, and I, I think that it's the, the, the way I think about it is there's, there's really a synchronous wave of innovation that's happening or ways stacked upon waves, partly because of how AI is accelerating all of these things.

[00:38:19] Brett Winton: So, even. Yeah. Robotics, which, which, you know, we think is interesting the way we think about the robotics opportunities. There's reusable rockets, which are. You know, super fascinating and mostly dominated by space X. There's kind of like, adaptive robots that can work alongside humans. And then there's the, the 3D printing space.

[00:38:38] Brett Winton: Well, like, the adaptive robots in particular, we're definitely kind of like that out years of our forecast, but the rate improvement of. Of humanoid robots right now it's really accelerated and partly because you have the software where they can, they can operate in the world. And so kind of, if even a couple of years ago, if you'd asked me if, [00:39:00] if like, there's going to be a meaningful market for a home help robot by the end of.

[00:39:05] Brett Winton: By 2030 you know, I'd be pretty skeptical, but now it does seem to be more likely that there will be something that you know, will hit a price point that consumers will accept and offer meaningful utility at that price point, such that it will be like a very expensive home appliance, but that is much, much more capable than your washing machine.

[00:39:26] Brett Winton: And so that. Yeah. 

[00:39:29] Tim Whybourne: It's pretty scary to think about.

[00:39:30] Tim Whybourne: Might make sense to go to a few specific companies. So I think one that we have here to talk about was Coinbase. So can you talk a bit about that opportunity and just around because in my mind, if this Bitcoin ETF gets approved, and I think ARK obviously has one waiting to be approved also what that would do to Coinbase, I'm in two minds whether it would increase or reduce demand for, for all of Coinbase. But what's your thesis? 

[00:39:55] Brett Winton: Yeah I think there's, you know, plenty of indications that a spot Bitcoin ETF is [00:40:00] likely to be approved in the U. S. So, you know, nobody knows for sure. And I think the, the best analogy is actually the gold ETF got approved in 2004. And that triggered a wave of inflows into gold from financial market participants directly.

[00:40:17] Brett Winton: But it also created kind of like a permission structure for people to consider gold and really other commodities as something that they should have as a part of their portfolio allocation. And so the, the, the ETF was a vehicle by which, you know. Retail and small market participants could get gold allocations, but then also, it means that all of wall street strategists had to come in and be like, oh, this is what the performance characteristics of gold are and why you should consider it as an allocation your portfolio.

[00:40:47] Brett Winton: And the, you know, to, to be right at the efficient frontier, a couple of percent makes sense and gold, et cetera, et cetera, et cetera. And as a result, the the price of gold basically 4x. [00:41:00] Over, you know, the course of 8 years after the not like, actually, I think when the ETF launch, because there had been some run up to anticipate the launch gold sold off slightly.

[00:41:09] Brett Winton: And and I would not be surprised if there's an ETF launch in the US if crypto assets broadly sold off upon the news that it came out because people, you know, I'm sure information is leaked. You know, people are. Anticipating this to a very high degree, and I think over a multi year timeframe, just like with gold, it likely it basically formalizes or the idea that this is an asset class that people should pay attention to.

[00:41:40] Brett Winton: And now there's a difference between gold and Bitcoin is when the price of gold goes up 4 times. You get a lot more people like digging in the ground to find more gold. But no matter what happens to the price of Bitcoin the amount of supply coming online daily is going to fall 4 fold. Over the course of, you know, 8 years, [00:42:00] I mean, really, it's going to fall in half next year.

[00:42:02] Brett Winton: It's going to fall in half again. 4 years after that, then again, 4 years after that and so, like, you can make adjustments to, like, if you assume that kind of like the, the demand response for Bitcoin is similar to the demand response. For gold triggered by kind of the gold ETF filing, even adjusting down for, you know, Bitcoin would be much more valuable today.

[00:42:24] Brett Winton: If they're the same demand for Bitcoin as there is for gold. That's why gold is a much higher market cap, but even adjusting down for current demand. So you just say on a relative basis, there's that much change. Then you would end up with an expectation for Bitcoin of around a million dollars a coin by 2030, given the difference in the supply characteristics.

[00:42:43] Brett Winton: Okay. Thanks. Of of the 2 assets. And so I don't think it's that's actually consistent with price targets that we've published on on Bitcoin over time. I think this, this is an orthogonal way to get to the similar type price targets. And to me, it seems. [00:43:00] Reasonable and consistent with how kind of financial market participants are going to respond to the fact that the SEC has.

[00:43:07] Brett Winton: You know, assuming they approve a spot ETF blessed the asset with a, okay. It's okay. You're allowed to own this thing. There are a lot of people thanks and others currently held out in the market because they're like, well, we're not going to go offsides to the sec, that would be a crazy thing to do, you know?

[00:43:24] Brett Winton: And so, Coinbase itself one, it is the for, for. Every filer for this ETF that is not self custody custody, Coinbase is the custodian. So it validates them kind of as a good actor in the space that that is blessed by the, I mean, despite its battles with the at least seen as the safe and responsible player and.

[00:43:49] Brett Winton: To some degree, the business model of Coinbase is, you know, if all crypto assets go up, Coinbase is going to do well. That's, you know, that will prompt a rise in the amount of trading [00:44:00] volume. They facilitate a bunch of trading, you know, their highest revenue generating activity is still right now like retail trading volume.

[00:44:07] Brett Winton: If a lot of people in the U S get really excited about crypto assets and use Coinbase to trade, that'll be extremely valuable for them. And they have a bunch of call options we think they're going to develop over time. They have staking as a service in the Ethereum ecosystem. They have kind of a layer 2 protocol called base that's built on top of Ethereum that allows people to launch decentralized apps.

[00:44:29] Brett Winton: They have kind of their own crypto wallet that allows you to, you know, safely hold on your own account Bitcoin and other crypto assets and web3 assets rather than relying on a central counterparty to do it. So there's a, as both the last supplier standing in a space that has been totally decimated by a bunch of I mean, combination of people who are fraudulent and then people who were unintentionally exposed to the people who are fraudulent, it's actually quite a sweet spot [00:45:00] and they've been.

[00:45:02] Brett Winton: Building all the way along Coinbase has had enough solidity in terms of its balance sheet and ability to invest in talent that that they haven't taken their foot off the gas. And so I'm kind of like, assuming that kind of the spot ETF filing. You know, prompts are continued, call it demand for crypto assets.

[00:45:21] Brett Winton: Broadly, I think Coinbase is extremely well positioned. You could think about it is like, imagine you owned like the NYSE and Bank of New York, which custody is all these assets and Goldman Sachs and Schwab and Robin Hood. We'll just throw it in there you know, all in 1 corporate entity.

[00:45:38] Brett Winton: Exposed to crypto assets. That's like their positioning. You know, and and it's, it's, you know, I, I don't know that they necessarily will. Like, if, if Bitcoin goes up 5x, we don't have Coinbase underwritten as if it naturally goes up 5x, there will be competition in the space. There will be peak impression.

[00:45:59] Brett Winton: But you could make [00:46:00] an argument that they have all of these kind of like ancillary opportunities they're building on top of their platform that could enable them to keep pace or, or pace ahead of, of, of where crypto assets broadly are, are doing. So I, there's no, I anticipating what you were thinking.

[00:46:16] Brett Winton: It's kind of like, Oh, if there's another channel by which people can buy Bitcoin, a spot ETF, what do I need Coinbase for? But the, the Venn diagram overlap between the clients that are going to buy an ETF. And the clients that like, there's a similar set of clients that might operate in both. But the reason the ATF is important is because in, you know, the US, there's all these financial advisors that they can't take their client money and put it into Coinbase.

[00:46:43] Brett Winton: They could tell their client, Hey, you should go to Coinbase, but then they lose revenue. So they never do that and their banks might fire them. So instead this operate, this gives them a channel to be like, Hey, your retirement money can go in here. Hey, you know, this like managed portfolio that I have for you.

[00:46:57] Brett Winton: I can do a couple percent into Bitcoin in a safe [00:47:00] and responsible way. And. And I can do so in a way where I'm not taking any personal or professional risk. And so that opens up a whole set of buyers into the market that previously were, were really kind of probably own some on Coinbase and their personal account, but aren't allowed to allow their clients.

[00:47:18] Brett Winton: To own many and so I think that's a distinction here and it probably enables coinbases institutional business and that there's all these pension funds and and kind of, endowments that. Really, it's been like, well, you know, the SEC is really against this stuff. Maybe we shouldn't even think about it.

[00:47:37] Brett Winton: And now, you know, they read all the strategist papers. It's like, well, it's dumb to not allocate to it. It's a different asset class, and it has different return and volatility characteristics. And, you know, over time, you know, there should be an allocation, but we're not going to buy the ETF. We're a giant institution.

[00:47:54] Brett Winton: So we'll go to Coinbase institutional and be like, hey. Bye. Bye. Can, how can you help us out here? And then they'll have, that'll help them [00:48:00] build their institutional business. 

[00:48:01] Tim Whybourne: Okay. That makes sense. I don't think we have that much time left, but in the short amount of time we have left, can we just touch on some of the newer additions to the portfolio that clients might not be as aware of in DraftKings and Unity Software?

[00:48:15] Brett Winton: Sure. So DraftKings is the, basically the way in which sports is monetising is changing. In our view you know, if you think about, like, the old school in the way in which all of all of like, media has monetizing is changing, but sports is particularly interesting because we think it's like, on the cusp of a monetization change.

[00:48:35] Brett Winton: That could be profound. You know, it used to be that sports was sports rights were enough and you could sign these, you know, multibillion dollar sports rights deals just on the basis of the kind of advertising. Attention hours, you would deliver to your partners. And now there's in the U. S.

[00:48:52] Brett Winton: Increasingly kind of regulatory leeway leeway so that you can place wagers on sports matches just through your [00:49:00] phone. And the as that happens, it actually allows the leagues to diversify the revenue stream by basically taking some of that. Wagering revenue and so then DraftKings, you know, and other, you know, sports are wagering providers serve as an important front end to how these leagues are going to monetize.

[00:49:20] Brett Winton: And the, the experience of of watching. Kind of anything is, is tending towards becoming more interactive. And so, you know, not just wagering on a match before the match happens, but wagering on individual plays or quarters or players is. Kind of like creates a more interactive experience, which all of these entertainment options need because they're all like facing the prospect of of all attention hours getting eaten by video games and tick tock unless they create a more engaging experience.

[00:49:53] Brett Winton: So then draft kings is, is essentially well positioned to facilitate all of that. The, the easiest way to [00:50:00] think about it is kind of like they're, you know. They are in a competitive, you know, relatively small set of competitive players in a category of growing consumer spend and and they collect basically a cut of all of that.

[00:50:18] Brett Winton: Wagering volume. And then as you like the, the real time betting, which, which in some markets, it's like 70 percent of the activities of being able to bet during the match is greatly facilitated by being able to, you know, take AI and take the data you have to create interesting and provocative things that people could bet on that are actually, you've gotten the odds right on them so that you don't get off sides yourself on your book.

[00:50:43] Brett Winton: As people are, are doing an increasing number of things during these events. 

[00:50:47] Tim Whybourne: A quick introduction to Unity Software. 

[00:50:50] Brett Winton: Sure. So unity provides a game engine for development, particularly for mobile platforms, but also for outside of game. So, [00:51:00] there's, you know. We think it's the, the idea of a game engine is, is a really kind of interesting platform idea as more and more time and attention moves into games and the ways in which games monetize.

[00:51:14] Brett Winton: Again, diversify and so, what a game engine does is it provides a substrate on which game developers can kind of, like, have the building blocks to build video games. Think about it. But having, like, a good virtual environment where you can build video games also, as it turns out. As the game engines get more performant, a good environment to even build virtual buildings.

[00:51:37] Brett Winton: If you're trying to show off an architectural design, you did or build virtual products. If you're trying to understand how users will interact with them. And so, kind of unity provides that. Game engine and then on top of the game engine are kind of mechanisms by which people can monetize. And so, you can imagine like video games is increasingly a business of you're not just kind of selling somebody a video game [00:52:00] up front, but you are monetizing their activity in some way, either by advertisements, or they're buying virtual items.

[00:52:06] Brett Winton: And so game engines become. Platforms that facilitate both that attention monetization, and then the virtual item sales and then unity is able to, like, facilitate that and then take a piece of platform feed off of all of that activity. 

[00:52:21] Brett Winton: I did want to cover one more topic because I think one thing that has happened in the market is people feel like, oh yes, innovation's happening. And then they like, I know the solution. I've read the newspaper.

[00:52:34] Brett Winton: Well, Alphabet, Google is clearly an AI company. So if I put money in Google, I'm fine. And then I've learned that Nvidia is creating all the AI chips. So, if I put money in NVIDIA, I'm fine, and then I've got my bases covered. And there, I'll cover one, then the other. Okay. So, actually, NVIDIA for within ARC, I think if you stack rank all our positions, it's around the 20th highest position across all of our portfolios.

[00:52:59] Brett Winton: Really [00:53:00] respected we think that more than a Trillion dollars are going to be spent on a hardware by 2030 and we think the majority of that is going to flow to Nvidia. The reason it's not a larger position in our portfolios is because it's given our underwriting methodology over 5 years. It's there are better places.

[00:53:18] Brett Winton: To put money than NVIDIA. It's, it's, it's really like you need that trillion dollars to come true and you need them to still have what are extraordinarily high margins for a chip manufacturer. And you need them to get the majority of that market. And there's like competitive threats. Tesla's developing its own chip.

[00:53:38] Brett Winton: Amazon meta Google already has its own chip. So, and AMD has, has a potential competitor in the marketplace. So, it's a way to interpret it is NVIDIA has to sell 110 billion of chip this year next into data centers. Somebody is buying those chips and thinks they're going to get software revenue off of it.

[00:53:56] Brett Winton: Like that's like, and so if you do the, the, the [00:54:00] math for them to successfully sell 110 billion. In chips, somebody has to believe that they're going to generate 700 to 800 billion dollars in software revenue on the back end. And you can translate that into 3 or 4 Trillion dollars in software market cap that has to be generated.

[00:54:17] Brett Winton: If these chips are going to work, if it's really going to be a thing so we're positioned in the companies that we think are the software. Companies that have yet to, I can look around, you know, the capital markets. I can tell you, there is not 4 trillion in AI software market cap that exists there today.

[00:54:34] Brett Winton: And that excludes Google, which has its own chip. So, kind of the, we think the right positioning right now is actually in the AI software plays where companies that are generating have proprietary streams of data. They're agilely deploying AI in, in the markets, not giving them credit for it. At least yet so there's there's a set of conditions in which NVIDIA doesn't work because they don't sell 110Billion dollars in and data.

[00:54:59] Brett Winton: And then it's like, [00:55:00] look out below problem for NVIDIA. There's a set of conditions where they just meet expectations. And then their stock is flat, but just meeting expectations means. That there has to be some multiple trillions of dollars in software market cap somewhere that gets realized over the next couple of years.

[00:55:17] Brett Winton: And and so that's how we're positioned on that front. And then Google is, I think there is a wider potential dispersion, potential for turn decision dispersion for Google, then then perhaps there ever has been, or at least since early in its corporate history. They are. By all evidence and by my analysis in the thorns of the innovators, the AI is, is they, they have all the reason to win.

[00:55:43] Brett Winton: They have the best talent. They have the best data. They have a distribution. They have the front end to the entire internet. It's a big search bar, a prompt in which you type things. This is exactly what AI models are. You prompt in, you type in a prompt and you get an answer. And think about how Google [00:56:00] currently makes money off of search and always has.

[00:56:02] Brett Winton: It's you type something here and you get delivered over there and then they collect a toll for delivering you. Like, it's really a you prompt and then you go somewhere else. And what AI models do is, is they take everywhere else and they can press it and they bring it. Back to you, they never deliver the customer to another experience.

[00:56:22] Brett Winton: And so, like, if I'm trying to shop for a vacuum cleaner 2 years ago, how do I shop for a vacuum cleaner? Well, I type into Google. What's the best vacuum cleaner? Because I want the best. And then I go to Reddit. Because Reddit has the debates about vacuum, like, I ignore the 1st links. I'm not going to buy their stupid product that they advertise.

[00:56:41] Brett Winton: I want to know the best 1. so then I read the Reddit page and then I find out. Oh, I, I found the right packing cleaner. Then I go back to Google. I search for that vacuum cleaner. I buy that vacuum cleaner. That's probably 2 cost per click revenue events for Google in buying a vacuum cleaner. Right? So what happens now?

[00:56:57] Brett Winton: I asked chat. GPT. What's the [00:57:00] best vacuum cleaner and it has read not only the Reddit page, but it has read or at least synthesized all of the nerdy vacuum cleaner debate sites. And it gives me the answer. These are the vacuum cleaners that you should consider. And then I can say, well, actually, I have pets.

[00:57:13] Brett Winton: I need vacuum cleaners. It can pick up pet hair and it further refines its query for me. On that basis, and then at the end, you, I can see clear that there's no longer going to be a, and then I go to Google and buy the vacuum cleaner. No, it's going to happen right in stream in that AI model. And so all of the economic architecture that Google has built around cost per click is at risk.

[00:57:37] Brett Winton: Like, they have so much. Efficient optimize. This is how we sell ads. This is how we generate demand for ads. This is how we like optimize ads into search that are is maybe sunk costs. And so from Google's perspective, even though they know this is the future of computing, at least at the top, [00:58:00] even though it's obvious, this is the direction things are going.

[00:58:02] Brett Winton: There are probably thousands of people in that organization who are incentivized to slow down their rate of adoption and who throw sand into the gears. You say, ah, you know, yeah, sure. I'll do that. But have we considered. Whether or not these things are dangerous, or have we, you know, doing everything they can to slow down because they have a promotion coming if they deliver on their KPI, which is dependent upon optimizing the 3rd link in Google AdWords.

[00:58:32] Brett Winton: And so, though. It's obvious that Google should win in AI. It's obviously dangerous for them and for the people in that organization to do so. Typically, that's how big companies get disrupted. So I think that there is a, cause it kills their revenue line and delivering large language model results is much more costly than delivering.

[00:58:57] Brett Winton: A Google search, so it also hurts their [00:59:00] cost basis and the entire corporate architecture of alphabet is built on kind of like the seeming inexhaustible stream of cash flow that comes off of search. So, I would be cautious it's in it. So it's interesting to me that kind of like, the people have have seemed to solve their innovation problem by flooding into mega cap.

[00:59:22] Brett Winton: Because particularly in, in the Google scenario and like Apple, like Siri is a joke. Siri is like the equivalent of those old touch screens that didn't work. And then the iPhone came out and you're like, Oh, it's a touch screen. Touch screens are terrible, but no, it's just because those touch screens were terrible.

[00:59:37] Brett Winton: Siri is that terrible touch screen. So, both of those companies are showing a lot of evidence of corporate luxury. That I think should be concerned 

[00:59:47] Tim Whybourne: That is thought provoking. I was trying to use Siri in my car yesterday as I was driving and it's trying to get it to understand me. And... 

[00:59:58] Brett Winton: There's a band called TV on the [01:00:00] radio and I don't think it's possible to get Siri to actually play.

[01:00:03] Brett Winton: It gets so confused. It's like, it hears radio and TV and it's just, it just gives up. So... 

[01:00:11] Tim Whybourne: That's a good point. That is thought provoking. We yeah, we've got a few clients that are in love with Google and and we've never been presented with that particular case. Interesting.

[01:00:21] Brett Winton: Well, and there's, there's actually really strong evidence that they are there because they're not shipping.

[01:00:25] Brett Winton: Like as in every time open AI releases something, Google says, Oh yes, we've had that. We just haven't released it yet. And then, and then they, they just announced their Gemini models. And the announcement was so, heavily marketed. Relative to a release that's not even available for people to play with until sometime next year.

[01:00:46] Brett Winton: It really makes me suspicious about what the output was because the, the stuff that was available to play with, they claim was performant with GPT 3 and a half. But I, my 1st tried playing with them. I demonstrated like, at least in [01:01:00] math. It's definitely not. Nearly as good. 

[01:01:03] Tim Whybourne: We're not there yet. But no I really appreciate you, your time.

[01:01:08] Tim Whybourne: Please note that Emmanuel Wybon and Loew may hold positions in ARC. All statements made regarding companies or security are strictly beliefs and points of view held by ARC and are not endorsements by ARC of any company or security or recommendations by ARC to buy, sell or hold any security. Historical results are not an indication of future results.

[01:01:25] Tim Whybourne: Certain of the statements contained in this podcast may be statements of future expectations and other forward looking statements art's current views and assumptions, involve known and unknown risks and uncertainties that could cause actual results, performance, or events to differ materially from those expressed or implied in such statements.

​

‍