Yahoo Finance

Today we're talking about how investors should think about investing in AI. I'd love to just start with your overall thesis here for picking companies.

Sure. Well, I come to this industry from software engineering where I was doing AI. Can you believe the luck of that, right? So today is your lucky day. Um, but regardless, I started in the industry in '99 where dot-coms were a craze. I did not fall for them because I had worked in software and I'm looking at them going.I don't really get it. This company doesn't have a product or this company is years away from delivering the product, so I have a very practical approach to looking at and evaluating technology. The first thing I ask is what problem is this technology solving?And then at what price can it be delivered? And if that price is higher than the product or than the problem that it's being solved, I take a pass on it and that might hurt like in the short term, but you're gonna be right in the long term. So that's what I'm doing is I'm evaluating all this new AI stuff saying what problem does it solve and what's the price to deliver it.

Yeah, no, that's really brilliant. So as we're thinking about the investing and the different buckets, they, they're kind of three main buckets of generative AI investing and you know I, I kind of harken back to what we had heard from Mark Zuckerberg many moons ago. I think it was about 2 years ago, essentially when he was laying out the vision for what their own investments would look like, and he kind of laid out the investment thesis.Well, in the buckets, applications, those are the products that people get to use like Chat GPT, Microsoft co-pilot. There's also learning models, the models, the so-called engine under Gen AI applications such as Chat GPT 4 or GPT-4 rather, which powers Jet GPT, and then there's the chip makers that all of this sits on top of and the data centers like Nvidia and Taiwan Semi.So break down how you think about investing in each of these buckets and perhaps we start off with applications where consumers are touching it.

Sure, well, I, I think it's important to note that most of the, um, uh, stuff that we're doing is a replacement for search engines, and there's a good reason for that. People who want to use a search engine want to get an answer to a question, not kind of figure out what the sentence should be to the computer and then scroll through.All of the links that come back as potential answers. So and it's at the face, chat GBT lets you ask a question and it gives you an answer. Now how much are you paying for that? You might be paying the 20 bucks a month or whatever to get premium service, but I'm thinking that is way, way under what it costs to deliver that.So in the long run, either the, the price of computing has to come way down or the price of accessing that has to come way up. The other thing is, if you're an investor in Google, you really have to um watch this because, you know, GPT and friends are eating Google's lunch. So that's one thing.Um, I think playing in a semiconductor layer is really smart because we don't really know what the end application is gonna end up at, what problem it's really solving because there's a lot of problems, but we know that it's gonna be delivered on a chip. So that's been an area that I've been playing for the last, um.Well, we'll say 6 years, not necessarily in reaction to AI, but there hasn't been a whole lot of compelling software companies to invest in that we're solving good problems in my paradigm where there's a clear problem, there's a clear solution, and it costs less than the problem. OK,

and so the companies that are best positioned, who are your top picks and why?

So, I'm gonna concentrate on the semiconductors, and this is kind of weird. I love Micron. Micron is a, a weird, it's not really a semiconductor company. They're more in the area of memory.Um, but they have been able to come up with fast memory that Nvidia is using on their chips, and I'm thinking whoever is Nvidia's competitor is going to be forced to use there. It's called DRAM. We're we're nerding out right now, but DRAM is really important for these chips that are doing the AI processing, so we like them.The second way to play this is software that designs chips, and we like synopsis and um they are buying a favorite local company of ours called ASIS which does um engineering simulation where you can, you know, test out, do, um.Online prototyping instead of physical prototyping and they do have a lot in the electronic design automation space as well, which is why Synopsis is buying them. But anyhow, so we like that company Synopsis and Micron, and they're not necessarily directly in the line of fire, but I think a lot of whoever wins in the AI chip race will might use their products.