Predictive AI Investing

Ryan Pannell: 

Thanks for having me, Elana. It's great to be here. 

Elana Margulies-Snyderman: 

Absolutely, Ryan. So, to kick off the conversation, tell us a little about the firm and how you got to where you are today. 

Ryan Pannell: 

Well, I mean, you gave a good introduction there. Kaiju is a global ecosystem of predictive AI companies focused on autonomous trading systems in capital markets, which we call synthetic portfolio managers. And each synthetic portfolio manager is embedded with a discrete trading ideology designed to achieve a specific goal, and they can work collaboratively or independently as need be. In terms of how we got here, my background is in applied cryptography and theoretical physics, and I come from a quantitative and systematic trading background professionally. So, it was a natural progression, I think, into predictive AI. I had used early predictive AI technologies in cryptography, and at least once the technologies became more accessible, and by that, I mean more affordable to work in, we started to add them exclusively to our portfolio management, which was back in 2018, 2019 so dinosaurs' age in our terms. 

Elana Margulies-Snyderman: 

Ryan, given your focus on predictive AI investing, I would highly welcome the opportunity to hear your high-level outlook for the space. 

Ryan Pannell: 

Well, I think when you're talking about AI as an investment or as an investing technology is, consider it in three different buckets. You either have the underlying technology which you're bullish or bearish on, that'd be investing in a company focused either directly or tangentially on AI, something like NVIDIA, Microsoft, Meta, et cetera, Google, et cetera. If you're looking at thematic exposure to artificial intelligence, there's a handful of ETFs, which basically bundle those companies together in different ratios. And then finally, if you are a diehard evangelist, which I guess is the category that we'd fall into when it comes at least to predictive AI and capital markets trading, there's using AI to make the investment management decisions. And so really it falls to the investor to decide on where their comfort level is. Are they sort of lukewarm on AI? They think it's going to be a dominant technology, but they're not sure where, they'd like to spread risk, they think one or two companies are doing it really, really well, or they trust the AI to make the investment decisions for them. 

Elana Margulies-Snyderman: 

And Ryan, what specifically, what are some of the greatest opportunities you see in your space and why? 

Ryan Pannell: 

Well, I think advances in, certainly data have been huge for us. AI depends on high quality data, and you're talking about predictive AI in capital markets. We have extraordinarily high-quality data from a number of different vendors, and it's not subject to errors or omissions that might result in hallucinations like generative AI is. When you have price time and quantity data and you have it from ICE or NASDAQ or Morningstar, CBOE, or all of the above, they tend to agree. We don't find a lot of errors in there. There's still data sanitization that needs to take place, but that's a huge boon for us, and that allows us to, with a high degree of certainty, find recurring patterns that you wouldn't be able to find in less well-organized data. 

Elana Margulies-Snyderman: 

Ryan, on the other hand, what are some of the greatest challenges you face in your space and why? 

Ryan Pannell: 

I think there are two. One obviously is skepticism. The AI explosion came right on the heels of the crypto implosion. So here you have yet another new technology that's a little difficult to understand. It seems nebulous. And then you have a bunch of experts saying, "Yeah, I know you don't understand, but don't worry about it. We've got it." And you remember very well what happened the last time a collective of experts, so-called experts said that. And so, there's a trust-building component right here at the beginning. You have to demonstrate value. You can't just tell people that they should trust. You have to show them that the technology can achieve its goals and why they should feel safe with it. And that really ties into number two. You're fighting up a hill against 60 years of pop culture. I mean, we've had AI as the bad guy in our movies and books for a very long time. It's never the good guy. It's always the bad guy. And even though we know that this is science fiction and not reality, it's in there somewhere. So, people handing over their wallet to a predictive AI system is a big leap. Although I would say that predictive AI systems like self-driving are bridging that gap. I mean, even if you experiment lightly with it and you don't have a nap in your Tesla or your Genesis, but you're lightly holding the wheel, it does a very good job today. And so maybe it's not so far away. Still, it's a challenge. We have to overcome that skepticism and build trust. 

Elana Margulies-Snyderman: 

Ryan, to shift gears a bit, ESG has been top of mind for the investment industry and investors, and I wanted to welcome your thoughts on if predictive AI is well-suited for ESG investment decisions. 

Ryan Pannell: 

I think actually predictive AI is a terrible solution for ESG investment decisions because AI period, whether it's predictive or generative, is terrible at context. It's terrible at nuance. It's really good at finding recurring patterns. But ESG requires multi-layered deep evaluation of not just the company, the product, the space, the sector, but on a global macro level, where will this fit? How feasible is it, and can there be enough uptake? I think actually where AI fits in ESG is mostly as a component to be considered. It's an enormous energy hog. It uses a ridiculous amount of water. It takes up space, it generates heat. I think you're going to see ESG press on the AI space, especially the hardware space and force it, I hope, to become a little more energy efficient, a little more responsible as a technology. Our solution can't just be more horsepower, more cycles, more power period. It's got to be responsible. It's got to be about refinement. So, I look at ESG as something that helps us be better rather than something that we can be a player in determining what investments should be made. 

Elana Margulies-Snyderman: 

Ryan, we've covered a lot of ground today and wanted to see if there are any final thoughts you'd like to share with us. 

Ryan Pannell: 

I'd hope that your listeners would keep an open mind when it comes to AI in investment management and ask really to consider the dual tracks of AI. Generative AI, I would not trust to recommend a movie to me, let alone pick a stock portfolio. It's not what it's designed for, and it's not good at that. It is good at collecting and sifting through large bodies of information and assisting analysts make a human evaluative decision. But predictive AI, you're talking about systems that watch the stock market at the tick level, every trade in the world at the same time, and can perform billions of discrete examinations on a nanosecond timescale. That's something we can't do. We cannot see those patterns, and those patterns do tend to repeat. So obviously it's a little self-serving for me to say given our focus. But I do think that predictive AI has a very strong future in capital markets trading and direct investment management, whereas generative AI is a more assistive technology. 

Elana Margulies-Snyderman: 

Ryan, I wanted to thank you so much for sharing your perspective with our listeners. 

Ryan Pannell: 

No problem. Thank you very much for having me on. I really enjoyed it. 

Elana Margulies-Snyderman: 

And thank you for listening to the EisnerAmper podcast series. Visit eisneramper.com for more information on this and a list of other topics. And join us for our next EisnerAmper podcast, when we get down to business. 

^{Transcribed by Rev.com}