Hello, and welcome to the Download. I'm your host, Dave Richardson, and it is tech time. Now, I want to be clear because we've got Marcello Montanari with us here, who is an expert in technology stocks, and we haven't had him on for a while. And a couple of people send me notes saying: you're avoiding Marcello because we've gone through a bear market in tech stocks. You're trying to time things and only have him on when things are hot. And I just want to say flat out that it's just been some scheduling issues. For those of you who listen to this podcast all the time, you know that I travel all over the place, and sometimes it's hard to connect. Of course, the investment professionals we have on this podcast are also extremely busy people. So sometimes it's just hard to line up. I'm actually more of a contrarian than someone who's following the herd. If you listen to the podcast, you know that. So, Marcello, I apologize for not being able to get you on sooner to get in front of this fantastic rally that we've seen in technology. And since we're talking about technology, we'll talk about artificial intelligence. Thanks for coming back.

Well, it's great to be back. I wasn't thinking you were avoiding me or anything, and I'm not hiding. Everyone knows where to find me.

That's right. As we were talking about before we came on, you have to appreciate that when you're investing in a growth sector, if you're making good picks— and we'll talk about a little bit about how you do that; and go back and listen to some of Marcello's previous appearances, you'll understand the philosophy, how he thinks about identifying these companies—, but you're in it for the long haul with this. And you should expect to generate above average returns over the long haul. But along with that additional risk and that growth orientation of these companies, there's going to be some dry periods along the way. And 2022 was certainly that. No, Marcello?

Yeah, that's an understatement. The tech market had been very strong since 2016, and it peaked out in 2021. And then we've come through a cascading correction, which started in the spring of 2021 with the concept tech stocks. And then later in that summer of 21, we had a period where the highly valued stocks, companies with great prospects, but still very highly valued, they started coming off; companies like the Shopifys and the Cloudflares of the world. And then they finally got to the generals, the big players at the end of 21. So for 2022, we've been basically through a bear market in technology, and that's reversed. If you look at some of the one-year numbers, like the Global Tech Fund in Canadian dollar terms right now, as of the other day, was up 38%, on one year. So a lot of that has been recaptured in the last six months. But at the moment, it appears we've come out of the recession and there's a school of thought that everybody's been waiting for this recession that's supposed to start, and each month it seems to get pushed back another 18 months. But there is a school of thought that says, given the upheaval from COVID, maybe what we're experiencing is this rotating recession where one sector after another is completely desynchronized. And tech was one of the first ones in because at the same time it's interesting to see that industrials have been killing it. Anyway, there's an element of bouncing back from perhaps what was a correction that occurred earlier than the rest of the market. That's one element.

And there's a great example of why we love having you on because you're the only guest who throws out words like «desynchronized», and we love that kind of stuff. Marcello, how do you explain what was really the driver of the downturn in tech? Is it as simple as these are growth stocks, some of the earnings are pushed out down the road, interest rates rise, and that's going to knock it down. Or is it deeper than that?

You can never truly tell. But certainly, interest rates. Because these are long duration assets. And anyone who knows anything about bonds knows that the longer the duration of the bond, the longer to its maturity date, the more sensitive it is to interest rates. And growth stocks, by definition, most of the value is basically contained within the present value of the growth opportunities. So you've got the current business plus the present value of the growth opportunities. And because they're growth companies, that tends to be very long dated. And so when the interest rates go up, that portion gets impacted more than others. So there's that. And then with interest rates rising, the fear became, oh, the Fed is behind the curve and they're jacking up rates too fast, they're going to throw the economy into a recession. Like I mentioned earlier, this recession that never seems to show up. But 99 strategists out of 100 have been calling for this recession for the last year. And I mean, that's going to have an impact on managements, on investors. That compounded the issue. And if you are a manager of a large company and you're making capital asset decisions and staffing decisions, all this negativity in the marketplace is going to give you pause and you're going to say, okay, maybe I shouldn't be hiring the extra five developers I needed. Maybe I'm going to put that product on the shelf or something. So it all just feeds off of itself. And just to add to that, we come back full circle. A big section of tech involves the entire advertising business with e-commerce and payments. And that was the big section or a big portion of the tech market. And when Apple basically started implementing all its privacy issues, it basically just put a wrench in the spokes. What happened was e-commerce was impacted because the ad platforms that direct business their way, the ability to attribute ads to purchases, disappeared. So then the return on ad spending disappeared. And so, e-commerce companies are saying: well, I'm not going to spend if I don't know if it's leading to a sale. Therefore, they stop. We're now at a point where this is all starting to reverse itself, which is a positive. But anyway, there's all sorts of things at play here and it all happened at the same time. Sorry, I dragged on a little bit too much.

No, and just before we move on to the rebound, so you're saying what's happening in that payment space and e-commerce, that's reversing. So what do you expect to see out of that? And why is it now reversing? Are those companies going to be fine now or are they impaired long term?

Yeah, actually another thing I forgot to mention was that we had all the supply chain issues at the same time. If a merchant can't get product to put on their shelf, they're not going to advertise it and they're not going to be able to sell something that they don't have. So all these things came together at the same time. That's why earlier I mentioned this desynchronized situation caused by COVID. Apple implemented its privacy issues. It was called IDFA. It was basically the ability to link an identifier with a purchase so you could link an advertising that was viewed with an actual purchase, and they broke that link. And Apple said, well, we're going to replace that with these tools that are more privacy friendly. And those tools were terrible. Nobody had any visibility on what was going on. What happened was that the big guys using artificial intelligence, by the way, or machine learning, started saying, okay, we got to figure ways around this. So they started putting all their brain power on solving this issue. At the same time, Apple realized that they did immense damage to the entire ecosystem, and they said, okay, we got to fix these tools. So they've basically come out with better tools, which launched last November, and then there's been a new iteration I believe as well. So those tools are actually better. And what we're hearing from advertisers is that they're starting to get signal back from Apple. On top of that, Facebook and Google have done a good job improving all of their algorithms at the same time. So basically, the attribution has come back and with attribution, they can target. And once you can target the e-commerce, sales start to improve. And you've seen it in the results. Shopify had good sales results; Amazon has had good sales results. In fact, I think yesterday there was a third-party data gatherer that said Amazon is probably going to exceed expectations on their retail side this year. So that entire thing to this downward spiral and now it's starting to work the other way around. And like I said, they're all symbiotic. The advertising is working, so therefore e-commerce sales go better and then therefore they can turn around and increase their velocity and then it helps the payments providers as well. So that's one element that seems to be working now. If we go into a terrible recession here, all bets are off. But again, that recession hasn't shown up yet and it seems like everybody's working.

Exactly. You know what, this was outside the direction of where I planned this discussion to go, but that was just fascinating. Again, I never heard anyone explain it that well. Obviously, I watch markets, and I know a lot of the listeners watch markets regularly, but I've never heard it explained that well in terms of what happened and caused that whole segment of the tech sector to struggle so much.

Yeah, because you need to really get into the plumbing to understand this stuff. It's very esoteric.

Yeah. And you get the plumbing, which again is why you do what you do and why it's such a hard job to do what you do. Because you've got to know in depth about all these things, how they're interconnected to know which companies will win and when, and most importantly, win out over the long haul. So Marcello, the next big thing that's on everyone's mind, and you tied it into how artificial intelligence and machine learning helped resolve this particular issue that we just discussed. But it's also been somewhat of the catalyst, particularly for the group of what are the strongest stocks so far this year. But it seems to be a lift right across the board in technology and it's just brought some momentum and some interest back to the sector. So, artificial intelligence, what are your thoughts on that and where we go from here?

Well, I agree with the premise that it's brought excitement back and in my view it's warranted. But we've been working on machine learning and artificial intelligence since probably the 70s, if not earlier. It's basically data sciences, and it's the same data sciences that we've been using since then. The thing that's different is the amount of compute power that we can throw at it now. So basically, these Nvidia H100 GPUs have immense processing power. And then when you combine that with cloud computing where you can basically put all of these things into a handful of data centers to put it to work and rent out by the hour for those that need it. You bring all of this capability into the hands of more and more players. So we were getting hints that that we've hit an inflection point probably in the summer of 2022, last year when Chat GPT-3 came out. And then 3.5. And then there's this great story about Bill Gates sitting down with the OpenAI guys and saying, hey, this is really cool, but why don't you call me back when it can pass an AP biology exam. And he thought they'd go away for three years, and they called him three months later and said, we're ready, Bill. Wow, it floored him. When some technology can floor Bill Gates, you need to listen. Then we end up with Chat GPT-4. And the big trigger point here was basically Microsoft unleashing this, like tying it to Bing, and basically unleashing it on 7 billion people and saying, here's this cool thing that you've never seen before. And this is the power of AI. Have a go at it. 7 billion people who don't all have access to the internet, but a lot of them do, all of a sudden everyone can play with this and test it. So that's definitely going to get interest up. And then you get this flywheel effect as more and more people use it. What artificial intelligence machine learning is, it's basically computer programs using data and then recursively using the output to improve its algorithms, in a circular manner. You launch this to 7 billion people, and you know they're starting to use it. And it's going to get better and it's going to accelerate like probably nothing we've ever seen. It hit a million users in five days. We've never seen that before. And technology over time has diffused faster and faster into the economy, starting with whatever, the steam engine, and the internet was quite fast, and then, you know, social media even faster. But this, AI as the consumers see it right now, we've never seen something diffuse this fast.

Yeah, and I'm sure many of the listeners have played around with this, or they've seen some examples of what it can do with other people who have played around with it. It's really quite remarkable, and it's the kind of stuff that comes right home. I know we were sitting in a meeting, and we had it generate a complaint letter and then answer the complaint letter. And both the complaint letter and the response had really human elements or an understanding of the techniques of how to be effective in your complaint and be effective in the response. And not just effective in the details of the response, but the empathy and the business acumen to make the right call around helping the customer out. It was really quite remarkable. Again, that's just one example and there's been millions if you watch and follow the news. But I think that's what makes it so compelling is it comes right to the front door, whether it's my daughter at 19 or my mother at 80, they can grasp it pretty simply and figure out a way to use it to make their life better.

Yeah, for sure. We can see how this is going to be used. Microsoft did this acquisition with a company called GitHub, which is basically a repository of software that basically software developers contribute their software and they pull stuff out. If a developer needs a calendar for an airline app or something, they don't have to recreate the calendar, they can just go in there and get one and basically just dump it into their program. Dumping it is oversimplifying it. So it's this gigantic repository that developers use and basically using large language models. On top of GitHub, you've got a situation where Microsoft calls it copilots where using natural language you can ask the copilot through GitHub to basically start compiling code for you. It's not going to do all the work, but it gets a big chunk of the work done. And we're already hearing that developers are seeing 20-25% efficiency improvements already by using this and you can see it all over the place. The other copilots that Microsoft was talking about is like a copilot for Excel or for your email program to help you write letters and stuff like that. So you can already start to see the use for it and the way Microsoft is framing it; you can also see that they'll be able to monetize this, which gets excitement behind it. The tech industry now has a shiny new thing to focus on. It uses immense amounts of compute power. So that's great for all the cloud computing guys and the chip guys and everything. As long as it can be monetized. Nobody really knows what the numbers are right now. So how does the market react to that? I don't know what the numbers are, but I know they're going higher. So what they do is, in the absence of real numbers, they basically start taking multiples higher. So what we've seen as of late is basically multiples have gone higher for most of the sector, especially the names that are involved with AI and can point to it and say that we're going to be beneficiaries of this. Either we're picks and shovels for this or we're going to deploy it in our own operations to get better. So the market has basically reacted by raising the multiples in the meantime. Have they overtorked and brought the multiples too high? It's hard to say. Probably in the long term, the answer is no, it hasn't. In the short term, you never know. Because now we're going to face a situation where nobody other than Nvidia— and I'm sure you saw the numbers out of Nvidia; they were monstrous, because they're right at the front of this. So you can't really do this until you get the Nvidia chips. So they're right at the front of this. And there'll be a couple of other guys near the front, but nobody else is going to be able to say, hey, we have AI revenues today. Hopefully the market isn't expecting to see AI revenues coming in the next couple of quarters. So once we start looking into 2024, I think we're going to start to see various pockets that AI revenues are starting to come in here and there. But like I said, you've already got situations where Meta or Facebook has been using AI to improve their algorithms, which should improve their ad sales. So it's happening that way. It's not totally direct, but you know that there's involvement there.

So it has a little bit of a feel back to the late 1990s where you put «dotcom» behind something and it drew attention. In investment dollars, you say AI and you just even watch corporate reports that they're throwing artificial intelligence, AI machine learning into language in their report. And the stock pops a little bit because there's that excitement around it. So how do you play that as an investor? How do you avoid the traps that were certainly there? Or is this just completely different Marcello?

In a way it's different because the guys who actually have a play here are existing. The reason the market is so narrow right now is because it's those seven big guys, the biggest market caps on the S&P 500. Those are the guys who've been identified as being the most exposed to the positive side of AI. That's where the interest has gone. And those are the stocks that have risen. And in this business, there's the bit that is art, that’s trying to understand the psychology of the market. And you could see that this was building and building and building. Did I expect Microsoft to unleash this on 7 billion people? No, but when it happened, it was like, okay. The psychology of the market went from something's going on to wow, this is big. On the market, everybody was negative, and tech was terrible and people were unloading or taking their Microsoft weights and Google weights down below market or completely selling it in the case of many hedge funds. And all of a sudden, overnight, everything changed. So now the psychology is, this AI thing is going to be big, I need exposure. You get this rushing effect going into it at the moment and the subject could change, obviously, and unfortunately I won't be able to tell you when that happens. But the psychology at the moment, even if you get a pullback, there's enough people that sit there that don't have the exposure going to say I know this is coming, I know it's big. I'm going to go in there and start buying some of these names. I think you're going to have this kind of support behind a lot of these names. And again, coming back to separating the wheat from the shaft, our view is, in one sense, it's kind of easy. You know Nvidia is going to be a winner here. You know Microsoft is going to be a winner. Google bought DeepMind in 2015. When Microsoft first came out with the Chat GPT-Bing, everybody thought, oh, Google is in big trouble, and they don't know what they're doing. But hang on a second, Google has been working on this since 2015. It'd be naive to think they don't know what they're doing. Did they get caught flat footed? Yeah, they did. So did everyone else though. And then once they had their big conference— I think it's called I/O—, they started showing you some of the things that they're using AI for and it was like, okay, these guys do get it. Overnight, Google went from, oh, you're a loser in this, to oh, actually you're a winner. Same thing with Adobe. Everybody thought, oh gosh, people are going to be able to create spectacular marketing packages out of thin air just using language and Adobe's toast. And then it turns out, well, there's all sorts of issues around this in terms of copyright and things like that. By the way, Adobe has this thing called Firefly where you can actually apply language to their program and get it working much better and easier and you open up the market for them to sell to more and more people. So all of a sudden, even Adobe went from a loser to a winner anyway, so it's just keeping an eye on things and making sure you do your work. And one of the last things to think about here is, we just think that any company that is sitting on massive amounts of data, in particular if it is domain specific, and that they'll be in a pretty good position if they do this properly. A company like Thompson Reuters, it's not the financial services company it used to be, it's basically a Westlaw. But it's one of the biggest repositories of legal information on earth. They should be able to create products that make lawyers super-efficient going forward. So you think that they'll probably be a winner. I'm not saying that's what's going to happen. They have to execute. But that's the type of company that could really benefit from this and the insurance industry could make hay over this, I'm convinced. Intact has been so well run in part because of their use of data and AI.

And that is important to recognize that a lot of these, what you'd classify as old school companies or old-time existing companies, they're able to take the technology and apply it and they have the resources to do that. And if they do it effectively, it's going to unlock opportunities in those businesses that it wasn't possible to unleash them before.

Yeah, exactly. And like I said at the beginning, there's not really anything new here. What's new is just the amount of compute power. A lot of these companies have been working on this for a long time. Like I said, Google since 2015, probably even earlier than that. That's when they bought DeepMind. So they were on this well before that. But a lot of the insurance companies— and I'm sure that Thompson Reuters has got great stuff in their Skunkworks programs— but we've been working towards this for a long time. So it's not a secret to any well-run company, I would say.

Yeah. What's the old music line on that? Overnight success after 20 years of performing kind of thing. So Marcello, I start to put my investment hat on here. I think for a lot of investors we would immediately think, okay, well, I've got to buy a stock that is specifically tied into this. But as I listen to your explanation and thought process through this, this seems more like something where you want to have exposure across the sector. But if you're in a diversified equity portfolio, you're already benefiting from it and you're probably going to get plenty of benefit out of it just because of the tech exposure and then the exposure in the businesses that will be winners because of this technology. You don't even really need to act, and you may not even realize that you've been benefiting from this. So when the neighbor comes over and says, well, I bought this hot stock in AI, and I'm making some money on it, you can go and say, you know what? My diversified portfolio is up overall over the last six months, and it's related to a little bit of the pop in AI, but I've got lots of exposure myself.

For sure. That being said, I think you'll have a situation where companies, like in the old saying, «get it». I think you can have a situation where a lot of companies that get it will outperform companies that don't get it within their sectors. You look at a company like John Deere, these guys get it. All their equipment is being digitally transformed. They're able to apply a one-hit fertilizer down to a specific plant and they're using GPS to do this. These guys get it. So if you're comparing them to some other equipment, like farm equipment manufacturer that doesn't get it, John Deere is going to outperform again. Having listened to Thompson Reuters last conference call, they appear to get it. Some of their competitors might not. So you want to be able to pick and choose and identify the guys who get it more than the others. And earlier I mentioned Intact because I remember sitting down with the CEO. What's his name? Charles Brindamour, I think. I hope I'm not mixing him up with another company, but I just remember the discussion that we had about how they were using data, and this is like over ten years ago. It just floored me, and I said, when you look at the Canadian P&C insurance business, Intact has far outpaced everybody else and maybe it's because of their use of data and the other guys just don't get it to the same degree. But at this point, everyone's going to get it because they're going to have to. So that'll probably be good for the picks and shovels guys. A lot of tech plays off of what's happening in cloud computing. This is another big data intensive hog that's going to help support that business as well. So people are already saying that whatever the slowdown in cloud computing is probably close to being over here.

Wow. So Marcello, we better track you down more regularly. This is just fascinating stuff. I came in again, I read a lot, I thought I had a pretty good understanding of everything that was going on in this space and some of the background behind it. A lot of the applications, you just blown my mind in terms of exploding just the possibilities of what this can do. So thank you so much for joining us today.

Actually, I wanted to put a disclaimer up front, which is anything I say here is probably going to be out of date in a month, because things have been moving so fast, it's been crazy. And I subscribe to a couple of newsletters and stuff where guys are just highlighting some of the applications that are being built off of Chat GPT and the large language models. Chat GPT doesn't do math very well unless it's a really simple problem. Anything complex, it doesn't know what to do with it. So there's a plugin that can send it off to a calculator. So they're getting around problems to make it better by having plugins and extensions and what they call APIs into other programs. So you see all of these innovations of building around this. Like I said, you wake up one day, go like, wow, I wasn't expecting that. So, like I said, a month from now we could say, hey, boy, this has veered off in a direction I wasn't expecting.

Yeah, maybe we'll save this for the next discussion. We just don't want to get the terminator. That's where we don't want this to end up. We wanted to go a long way and help things out. We just don't want it to go too far. And there's going to be a lot of challenges along the way to make sure that we've got the right controls in place so that doesn't happen. But again, Marcello, thank you very much for the update on tech and particularly this discussion on AI. And we'll get you on after you have a well-deserved summer holiday. Get you back on in the fall.

Sounds like a plan. Thanks a lot.

Thank you.