Full Transcript
Noah: All right, so I'm here with Paul Cecil. Paul, tell me about yourself. Where are you from? What do you do? All that good stuff.
Paul: My name's Paul Cecil. I'm the head of strategy, VP of Strategy for Realpha Tech Corp. We're listed on the Nasdaq under the ticker, AIRE. My background, I'll give you the 60-second overview. So my background's in finance. I graduated from Ohio State, undergrad in finance, and then a couple years back had the opportunity to travel to India and complete an executive education program at IIMB also in private equity and capital markets raising. So I've been with Realpha since about September 2021, so I was employee number 10 and at the time we were a completely different company and so just a very small team. I've been with the company all the way since basically since the inception through our NASDAQ listing. I've done, I think, eight capital raises now and then six company acquisitions and I'm looking forward to seeing what the future holds for us at Realpha.
Noah: Very cool. And what does your day-to-day look like at Realpha?
Paul: It's a little bit different every day, I would say, but my role overall is to guide the company towards our vision and make sure that in the long run we have the right resources at the right time in the right place and the right people to basically execute on our strategy. So it's really, it's guiding the company towards the vision. So my day-to-day is anything from helping with capital raises to designing different strategies, interacting with the board of directors. I run most of our investor relations as well. So explaining our strategy to the broader investor group and that sort of thing.
Noah: Very cool. So the big topic today is AI. Talk to me about just your thoughts on AI in general, and then we can dive into how you use it for different workflows. Which ones you like to use, all that sort of stuff.
Paul: We're obviously very bullish on AI. Our ticker, AIRE, so AI for real estate, and so we're definitely an AI-forward company. We've been on the forefront of AI for a long time. So even before ChatGPT came out, we had designed our first kind of AI that its goal was to analyze different properties to see if they would be good short-term rentals and rank them on a score of zero to 100. So that was the original AI in real estate that we had built. And then when ChatGPT came out, obviously the world changed overnight and our CEO came in and said, forget everything, the world has changed. You need to unlearn everything and relearn with AI, right? So from that moment on is when I've redesigned all of my workflows at this point, and with AI in mind and the capabilities of that because, I think that the people and the companies that don't embrace this are going to fall behind very quickly.
Noah: And are you seeing with companies you work with or with investors you work with, how much adoption do you see across your space and meetings that you're having?
Paul: The tech space obviously is very AI-forward. Everything's about AI right now. Real estate, not so much. There's been some people were dabbling, but real estate traditionally lags behind in technology already. And they still use things like fax and data tapes and that sort of nonsense. And yeah, it sounds like a freaking VCR tape in my opinion, but yeah. Traditionally real estate has lagged behind in AI adoption, technology adoption in general. And so I think we're ahead of the curve on that. And, at least, at the very least on par with our technology peers and everything and the speed at which we're moving, we acquired an AI development company in Nepal actually. So they worked for us from the very beginning on doing all of our product and tech development. And so they became too important and we just went ahead and acquired them so that we're very dedicated to that. But, to keep it to the industry in general, I think real estate is slow, tech is fast. We're bridging those gaps.
Noah: Which AI systems do you use the most, would you say, on a day-to-day basis?
Paul: OpenAI for sure right now. 'Cause that deep research function with ChatGPT has gotten so bloody good. Just because it's so fast. It's so good and it's, from what I've seen, it's the one that remembers the context the most. So it knows us so well at this point where whatever question we ask, it knows the context, it knows the strategy, it knows the constraints, call it. And it's the best at providing recommendations. And, we regularly or my team regularly uses it to create PhD level reports that would traditionally take six months to put together, but now can be done in within 20 minutes.
Noah: Wow. What type of reports are those?
Paul: Call it case studies on different companies, competitor analysis, let's see. Like the other day I was curious about what, like Palantir has done a very good job at their investor relations strategy. So I was very curious as to what Palantir and specifically Alex Karp was doing to attract so much attention from retail investors. And so I had them do a full-on case study of Alex Karp and it took them 30 minutes to put it together with deep research, but it came back with a 40-page report with a hundred different sources and examples. Like, you know, who can do that fast? Nobody, right?
Noah: Yeah. That's fantastic. If you had to rank OpenAI, Anthropic, Gemini, and those guys from one to five. I know you said you used OpenAI a lot and ChatGPT. Where would you put them just in terms of the race in general for the space? And when we talk about race, it's not necessarily one company is going to win. Obviously they have their different specialties and area of expertise where they could win or they could excel, but just in terms of where they are right now, and if you wanna throw Nvidia in there too as well, where would you rank those companies?
Paul: OpenAI I would for sure put at the top. Just for our use case and just with the amount of money they're raising and everything I think, I'd throw in one of these big guys is gonna come out at least in the top three I think, because, they have the sustainability and the resources to be able to push through all of the burn I would call it, and integrate with our ecosystem. And, Microsoft is in bed with OpenAI, so I'd probably put them in the same boat as OpenAI, but out of Google versus Meta, call it, right now. That's a hard one. I'm, I would say probably Google will come out on top of that, but I'm not entirely sure. I don't particularly like Gemini right now in its current state. But they've come out with some cool, because Gemini doesn't do the context very well, it doesn't remember things very well and provide it, it gives you quick responses and things like that, and it's good for summarizing emails and that sort of thing. But I, I just have not had good luck with Gemini. And we were testing it all the way back when Bard was a thing rather than Gemini. But they have some cool stuff coming out, like NotebookLM and that sort of thing. So I'd probably put, I'd probably put Google as a second and then maybe Llama as a third. Nvidia is gonna be a winner from a different perspective. I think they win no matter what. 'Cause they're the hardware layer. They're not really the, it's a different category. Call it, they're gonna win no matter what. So OpenAI, Google, Llama, and then I would actually throw Perplexity in the mix. I think Perplexity has gotten very good in its current state. Claude/Anthropic, I hardly use at all. So I'd probably put them last.
Noah: Yeah, that's fair. Given your role and what you do. I spoke to a developer last week who put Claude near the top, but also he is a developer and he uses Claude for code on a daily basis. So it's different for different roles. If we were to go back to, say, 2020 or 2022 and put your efficiency level at say a hundred percent, say you were a hundred percent efficient at your job when you started at Realpha, where would you put that efficiency level now? In terms of being able to use AI and how that's affected your work? Would it be, say, five times better, you're 500% more efficient? What would you put that number at if you could quantify it?
Paul: At least 50 times more. Wow. At this point, there is no way that I'd be able to handle essentially what is three teams worth of work and still have a social life without AI. Because even just the, because I'm doing the research and data analysis piece, right? That's covered most of the AI, my team could do that. We can, we can dive into this a little bit later, but, I've developed my own personal board of directors. That's been one of my cooler projects with AI and that has taken my own personal productivity and just fulfillment and in learning and things like that, to just a level that I couldn't comprehend before.
Noah: That's super interesting. Yeah. Let's come back to that 'cause I wanna hear more about that. What would be an issue that you've had with AI or what are some kind of drawbacks or problems that you've run into using AI and how have you been able to correct those problems or go around them at least?
Paul: I think the main one is AI has to be used smartly, not as a crutch. It, you can't outsource your thinking entirely. The way that you prompt it and tell it to remember things is gonna determine the quality of output you get. And, if you just throw some shit in there, it's gonna come back with shit. It's gonna use a lot of em dashes and emojis and stuff like that, which is, some of the stuff they give you is not always thought through or appropriate for a work environment. And so you have to, you have to know how to train the AI a little bit and say, "Hey, I don't like this. Don't put this in future answers. I don't like this kind of formatting. We're in a work setting, don't use any emojis," that sort of thing. And so we've had to do some training on that. And, just generally being careful what you feed into the AI especially. On a, if there's anything on a personal account, right? Like it'd be, it'd have to be time use of that, but mainly the first one I would say.
Noah: Interesting. Yeah. What the kids are calling the ChatGPT dash, the old em dash for sure. If you were talking to a senior in high school looking to get into finance, going to school, and then someone who just graduated college who doesn't have a job yet and telling them what, just giving them advice on kind of the scope of the economy, work, just what they should be doing and how they should be using AI to get where they want to go. What would you tell them?
Paul: There's a lot there. Since this is an AI discussion, I'll stick to AI. I think that a lot of the, especially if they're coming straight outta school, I think a lot of the wisdom from schools is severely outdated and they're not gonna be able to keep up with the pace of learning and teaching that is happening in AI right now. Like traditionally, this was already a problem where you and I have talked about this many times, but now the problem is exacerbated by a hundred percent or, not a hundred percent, a hundred times. And because what AI has essentially done, especially for entry-level jobs, it is, it has leveled the playing field for entry-level information. So everybody has access to the same entry-level information. And so onboarding processes are done faster. You're gonna be expected to learn at a higher rate, but even getting your foot in the door you have to have some, you have to have some unique thing about you, right? About your personality or about the way you use AI or some unique process that you have or some unique story that you have. 'Cause if you're just relying on a degree at this point, you're pretty screwed. I think, but that has opened up a lot of opportunities for experimentation on things because you know that information on literally anything is readily accessible. So show that you have interesting things about you and attempts and projects on your own. Because things can be put together a hundred times faster than before. If you're trying to get into development or, yeah, if you're trying to get into development, develop something right with AI and show us how fast you can do it. Or if you're trying to get into finance, right? Design an algorithm with ChatGPT and try it out on, throw some hundred bucks into it or something and see what it does. I think pretty soon here in finance in particular, Excel may not be the dominant anymore, right? It may be some other AI tool that comes along, although I haven't seen anything that's really impressed me as of yet. Except maybe was it poly... Polygon? No. I have to go back to you on that one, but there was one that you have to reset your thinking. School effectively has, think of school as, it hasn't taught you anything at this point, but it gets your foot in the door. Or it doesn't get your foot in the door. But it's a baseline of knowledge, right? You're out on your own. You've lived a bit. Now you have to come up with some way to get your foot in the door with companies.
Noah: That makes sense. And I was actually gonna ask you about that. I've seen kind of folks on Twitter talking about, "I just set my trading preferences using ChatGPT or Claude or this," and a lot of it is fluff but there could be some truth to it. What are your thoughts, just in general about AI trading?
Paul: It's already been happening for a long time. It's called quant finance, right? And that's what all the hedge funds do for the most part. That's why they're hiring math majors instead of, instead of just finance guys before, right? Like trading on the stock market in particular has very much become algorithm-driven. And trying to compete against it as a human is near impossible at this point. Unless you have inside information, which you can't trade on, so that makes it tough. But finance has evolved spectacularly anyway, and even without AI. So there's all kinds of other job opportunities for that. But I think that a hedge fund with a sophisticated algorithm is probably light years ahead of an end user at this point because they'll use some like 700 different data points to do algorithms and, they have the deep mathematics and financial knowledge that is necessary to put together something like that. So I think for the time being, probably your kind of hedge funds still have advantages over the traders. You might be able to beat the market for a while, just trading on ChatGPT.
Noah: Interesting. I got two more questions for you. First, let's bring it back to your personal projects that you mentioned. What type of personal projects are you doing with AI? Talk to me about the board of directors thing, all that stuff.
Paul: Yep. That's been my, that's been my personal favorite is what I've, what I did was I started a project in ChatGPT. For the, for, I'm sure most people know this, but you can start a project folder and then based on the kind of threads you put in there, they feed off of each other and you can give project-specific instructions about how it responds in any of the threads there. What I did was I put a project together, called it the Personal Board of Directors. And for anything that I feel like I need guidance on, I've created a different coach for it. And this is, this could be as general as, "Hey, help me with, help me with relationships," or it could be as specific as, I have one for what was it? Bonsai trees. Like I made one that talks to me like a monk and it gives wisdom about, while I'm working on bonsai trees. So it could be very specific or very general. And what I've done now is I've trained each one to talk to me like certain people that I would like to have as a mentor or in a certain tone or to either provide action items based on what I'm saying or, and so I've got it. I've got it very trained. At this point. I have one, I have one for, call it strategy. I have one for just general questions about being a man. I have one about Bonsai. Like I said, I've got one for as a sales or pitch coach, speaking. I've got one for what else, like religion. And there's a few, there's a few other ones that I've put in that I may not interact with as much. Every once in a while I'll go back. My favorite one, I fed it. You can feed it specific people, you can feed it specific ideas, specific books. So my favorite one to talk to, the one I talk to a lot is just like general man advice, and, I fed it people like Jocko Willink and books like The 48 Laws of Power and that sort of thing, right? And military books. 'Cause I, so I've got it trained where it talks to me in military terms because I really like the military metaphors. So it'll give me a military time, it'll give me an I'll follow up you at oh 800, that kind of thing. And so it's been fun and I can basically get advice on literally anything within 30 seconds now. And so whenever I'm just dealing with something throughout the day, I pull out the coach or I'll put a journal in there or something and just, you can get an immense amount of knowledge out of that.
Noah: That's fascinating. That's really cool stuff. I have one more question. And that's about the future of AI. What do you see five years, 10 years down the line? You can talk in general or just if you have specific use cases that you're thinking of, where do you see this going in the short term and medium term future?
Paul: I want to first plug what we're doing. Because, you know what I'd like, what I'd like to build at Realpha is I'd like for somebody to be able to go through the entire home buying process and including the beginning part, the getting a mortgage, closing, everything afterwards, like moving furniture, everything like that. All and feel like they're talking to one person the entire time. And traditionally you have to talk to maybe 10 to 15 different people. You got your title agents and your insurance agents and your inspectors and your mortgage brokers, your real estate agents, that sort of thing, right? It's very disparate. I'd like to make it so that you can. It's just as easy as buying a car or even a cell phone. So you feel like you're talking to one person the entire time. And I think AI can be that concierge in the home buying process. But just in general, I would say that, within five to 10 years, it's really hard to tell. I think we'll see probably AI super intelligence by then. And they're gonna start putting it into humanoids. They're gonna start putting it into, like machines and hardware are gonna start to have AI like fridges and laundry machines and all kinds of stuff. I really do believe it's gonna become a point where Andrew Yang's UBI might become a reality. And I say that half jokingly, but. I think in the near term it's gonna wipe out so many jobs, but in the long term it's gonna help the world solve so many problems that we couldn't previously. And it's gonna help answer questions that we didn't even know we had and it's gonna help make honestly everything a lot easier. So
Noah: Interesting. So Paul, I appreciate your time and yeah, I look forward to talking to you again and keeping up the conversation with you. Absolutely.
Paul: Thanks for having me.