The AI Decision Models for PE-Backed Tech CEOs to Predict Strategic Growth with Kam Star

In this episode of Future Finance, host Paul Barnhurst chats with Kam Star, a strategic product and technology leader with over 25 years of experience in AI, finance, and digital innovation. Kam shares his unique perspective on the evolution of artificial intelligence from early expert systems to today's powerful neural networks and how AI is shaping decision-making in finance and private equity. The discussion discusses the prediction, gamification, motivation, and the future of AI-assisted strategic decisions in business.

Kam Star is a visionary in AI and technology, with a rich background blending computer science, psychology, and architecture. He has led significant business growth initiatives, including scaling Blue Prism’s product portfolio and founding ventures with revenues exceeding £100 million. Kam holds a PhD in Computer Science and Psychology, alongside advanced studies in AI from MIT, Sloan, and LSE. Passionate about emerging tech, Kam currently focuses on predictive analytics to support better decision-making in private equity-backed companies.

In this episode, you will discover:

• The history and evolution of AI from rule-based expert systems to modern neural networks.

• How gamification and motivation impact team performance and customer engagement.

• The role of predictive analytics in supporting strategic business decisions.

• Challenges and opportunities in democratizing advanced AI tools for mid-market companies.

• Kam’s vision of AI’s future impact on decision-making, creativity, and personalization. 

Kam’s expertise and thoughtful insights make this episode a must-listen for anyone interested in the intersection of AI and finance. His stories about the evolution of technology, combined with practical advice on predictive analytics and gamification, offer valuable takeaways. Whether you’re a finance professional or a tech enthusiast, Kam’s vision will inspire you to think differently about how AI can transform decision-making. 

Follow Kam:
LinkedIn - https://www.linkedin.com/in/kamstar/?originalSubdomain=uk

Website - kamstar.net

Join hosts Glenn and Paul as they unravel the complexities of AI in finance:

Follow Glenn:
LinkedIn: https://www.linkedin.com/in/gbhopperiii

Follow Paul:
LinkedIn -  https://www.linkedin.com/in/thefpandaguy

Follow QFlow.AI:
Website -  https://bit.ly/4fYK9vY

Future Finance is sponsored by QFlow.ai, the strategic finance platform solving the toughest part of planning and analysis: B2B revenue. Align sales, marketing, and finance, speed up decision-making, and lock in accountability with QFlow.ai.

Stay tuned for a deeper understanding of how AI is shaping the future of finance and what it means for businesses and individuals alike.

In Today’s Episode:
[02:57] - Welcome to The Episode
[05:48] - Childhood programming story

[10:16] - Evolution of AI and expert systems

[13:26] - Gamification’s role in engagement

[19:07] - Predictive analytics focus

[22:25] - Cause and effect in decision making

[30:46] - Generative AI and creativity

[38:22] - Fun rapid-fire: Secret room

[42:48] - Ambitious bucket list goal

[43:32] - Closing thoughts and thanks

Full Show Transcript:

[00:02:57] Host 1: Paul Barnhurst: Welcome to the Future Finance Show. I have my co-host Glenn, here with me, and today we're going to be interviewing Kam Star. We're going to have a chat with him. So welcome to the show, Kam.

[00:02:07] Guest: Kam Star: Hey, thanks for having me.

[00:02:09] Host 1: Paul Barnhurst: Yeah, really excited to have you. So let me give a little bit of his background before we jump into the questions. Doctor Kam Star is a strategic product and technology leader specializing in driving exponential growth and value creation for private equity-backed companies with over 25 years of experience. Kam has a proven track record of scaling businesses, including growing Blue Prism's product portfolio from 50 million to over 250 million led to its $1.6 billion acquisition. As the founder of PlayGen, Kam has firsthand experience building and scaling ventures exceeding 100 million pounds. He also founded Digital Shoreditch, growing the festival to over 40,000 participants, establishing it as a major event in London. Kam holds a PhD in Computer Science and Psychology, a master's in architecture and Advanced Studies in Artificial Intelligence from MIT Sloan and the London School of Economics. Passionate about emerging technologies, Kam thrives on creating products and strategies that fundamentally transform how we work and live. Quite the background, Kam, I have to say. Impressive.

[00:03:25] Guest: Kam Star: Thank you, I try. I'm very curious about the world. So any opportunities to get any letters before or after my name or try something new, I'll be there.

[00:03:35] Host 1: Paul Barnhurst: I love it.

[00:03:36] Host 2: Glenn Hopper: I have to add here to the prophetic nature of combining computer science and psychology. Obviously now I mean with AI and the way we interact, that's perfect. I mean, I guess your studies looking at both of those was that your reason for picking? Look, Paul, I'm just jumping in straight to a question, but.

[00:03:54] Host 1: Paul Barnhurst: I knew you would. You get excited with these things. So go for it, Glenn.

[00:03:58] Host 2: Glenn Hopper: Was I mean, was your thinking at the time that when we moved towards AI, that your psychology is going to be an important component of it, or how did the combination of those two come about?

[00:04:09] Guest: Kam Star: So my life philosophy is not a dualist. I'm a monist, so I don't really see a difference between a beautiful painting or a beautiful mathematical equation. I'm a monotheist. It's all one thing. And what's really important is it's all one thing. Whether it's a physical, you know, something physical or something emotional. We do have to experience it. And experiencing things is a big part and component of the reality that we experience. So I've always loved computers, but at the same time I love understanding how people think, what makes them tick, why they make decisions. And so that's also perhaps the  reason I did architecture. I just couldn't make a choice between art or the sciences. So I had to go into something where I could have the physics and the chemistry of the wall, and then the. Color theory and form and shape and history and just it's something that's been with me since I was very little. I think I grew up back in those days. We had physical books for a lot of knowledge, so we had encyclopedias, a couple of full of encyclopedias at a at a height where my sort of seven eight year old hand could reach them, and I sort of one day just started reading it from one end to the other. And then when I finished, I did it again. So I really got spoiled by being able to just pick one thing. So computer science and psychology. Absolutely. history and culture and music and art with quantum physics. Yes, please. It's very interesting.

[00:05:43] Host 1: Paul Barnhurst: We got a learner here, I love it. There are a lot of different backgrounds there. That's great. And speaking of that, yeah, I was reading the story about how you developed your first product and sold it to a construction company when you were 12. Tell our audience that story. I love that story. Give us a little bit of background.

[00:05:58] Guest: Kam Star: So I have some of the listeners who would be old enough to remember. We had a theater called Sinclair ZX spectrum. This is in the very early 80s, and I got one for Christmas when I was 11. And back in those days you used to get magazines that had basics, that was the programming language, and you could play and make things and type things out. In any case, I was making little games and playing with my ZX spectrum. I went to my dad's company, he was part of a large construction company, and there were rows of engineers sitting at their tables with big sheets of paper. With these what are called slide rules, these rules that you can move that have particular calculations on them. And they were just sitting there moving these things back and forth and writing numbers down and moving them again. It just looked very, very odd to me. So I got very curious, what were these guys doing? I asked my dad. He said, well, why don't you find out? So I sat with one of the engineers and he very kindly showed me what he was doing, which was calculating the steel depth of these steel beams in this really large construction site for a Siemens factory. Absolutely enormous things, had hundreds of thousands of these steel beams, and they were all slightly different.

[00:07:12] Guest: Kam Star: So they all needed to be calculated. He showed me how he does it, and as he was showing me, I was sort of storing in my head his process. Oh, okay. So if then, then this, if that and then calculate these things. So I took that away. I think I made some notes or I think I went back and made some notes. In any case, I decided to make a computer program that would just ask you, okay, what's the span and what's the load? And then it would tell you, based on Young's modulus and the set of tables, that it could be just transformed into a set of if-then statements. What the steel beam, the I-beam steel needed to be. I wouldn't present it this back. well, first I showed my dad and he was like, hey, this does it work? And I'm like, I think so. So,I got invited to go in with, we had to take a TV in and a tape deck in the computer, set it up, and, one of the, the chief engineers, he brought over this whole table, and we started to kind of go through the first 3 or 4 or 5 or 6 that we tried.

[00:08:14] Guest: Kam Star: It was all actually correct because, well, guess what? It's just a set of predetermined calculations. I think the crunch came and I think it was maybe 5 or 6in. We put in one of the one of the beams and the software gave a different answer. My, my, my little program gave a different answer. And I was convinced, well, I must have made a mistake. Something's gone wrong. and I was quite disappointed, but actually, probably credit to the chief engineer, he decided to run, rerun the calculations by hand on that specific beam. And what turned out was that the original calculation that the engineers had done was wrong and someone had made a mistake, and he immediately was like, ah, we're going to buy this, you know, this is going to save our bacon. You know, I can't have this thing. So that was a really good accident for me. And,  you know, I never really looked back in terms of building stuff, but, you know, it's kind of really just started by spending the time with the user, understand what they're doing, kind of get into it, and then create something that solves their problem in a better way.

[00:09:23] Host 1: Paul Barnhurst: I love that story. Glenn, what were you doing at 12?

[00:09:26] Host 2: Glenn Hopper: So at 12, I was actually, I'll tell you, on my Commodore Vic 20 because we couldn't afford the Commodore 64.  I had the data set and I was trying to program my version of AI, which was centered around a Dungeons and Dragons game. So. But nobody was buying my game. I never got very far in it. I was using it.

[00:09:48] Host 1: Paul Barnhurst: I just played the games. I had a Commodore 64, we had the Magnavox Odyssey, We had all of it. But unlike U2, you know, we're using it to learn. I was just like, shoot them up, let's have fun.

[00:10:01] Guest: Kam Star: Hey, that was fun to write.

[00:10:04] Host 1: Paul Barnhurst: It had its moments, I won't lie. All right. So I love that story. And, you know, I want to talk a little bit. You know, obviously, on our show, we talk a lot about AI and technology and where we're going. You published a master's thesis back in 97. You did it in 95, 96 about artificial intelligence in the built environment. Talk a little bit about that. What made you so interested back in that? You know, kind of period on AI and what was the. Tell us a little bit about the paper. Just a little background on it.

[00:10:33] Guest: Kam Star: Sure. Absolutely. I don't know if you remember Lawnmower Man. It was a movie that came out in the very early 90s. You know, it was all about virtual reality. And how are we going to live in this virtual space? I was, you know, I was doing architecture. I was studying architecture at the time. And, you know, I've always been quite techie, quite geeky, and it's just sort of I got really engaged with that. And, you know, there's an eye in that. And of course, you know, 2001 Space Odyssey, but we all know about that. But in any case, I was always really interested in systems. Back in those days, artificial intelligence was really seen as expert systems. And the promise of it was that we could use it. I mean, later on, I would just call it artificial stupidity because these are just heuristic rule based systems. There is no intelligence, there's no thinking if it's just a set of rules. But there were there were machines that and and really the idea was there will come a time soon where we can not only design much better buildings using these systems, using these expert systems, but we can use them, leverage them in the built environment to control everything from heating and cooling when things need to be made, how you know, the kind of homes for the future, that kind of a thought. And, you know, I think some of those things have come and gone. Expert systems are really kind of not the thing anymore. And I think what we mean today by artificial intelligence is probably much more recent phenomena around deep learning and these emergent behaviors that we're seeing out of very large neural networks. Well, to be frank, at the time, neural networks were really still in their infancy, you know. And what we consider to be artificial intelligence is probably as sophisticated as what, you know, a model 40 is compared to, the latest, hydrogen run car that you can buy. So, you know, things in common, but really fundamentally different.

[00:12:30] Host 1: Paul Barnhurst: Sure. I was going to say I might have to go watch Lawnmower Man. I briefly remember seeing ads for it. Did you watch that one, Glenn?

[00:12:36] Host 2: Glenn Hopper: Yeah. It was an adaptation of a Stephen King short story, I think two from Skeleton Crew. I can't remember which book it was. It was one of, . Yeah, but it was. Yeah, it was a good movie. It was. It's funny, everything you're talking about, I'm just picturing it all building to where we are with AI right now, talking about going from expert systems. And then I guess the perceptron and Hinton's, you know, his approach to it and all that. And then but looking at your PhD research and gamification, it's making me think of reinforcement learning and like,the AlphaGo kind of stuff. But in the research that you did, for gamification impact on team performance, I thought with your work in finance also, I mean, this would be for teams or even maybe for training, training models in it. But, um, in the context of financial services, how could game gamification strategies be employed to improve customer engagement, promote better financial decision making? I mean, how could we use gamification in areas like that?

[00:13:40] Guest: Kam Star: Very carefully.

[00:13:42] Guest: Kam Star: Yeah, yeah.

[00:13:43] Guest: Kam Star: You have to be very careful not to make things worse. So the games fundamentally are competitive, right? So most games are competitive in nature in some ways. And and but there's a lot of good things in games and in gaming. and in that, that whole space which you can leverage to engage folks. So, you know, and when we think about it, I think maybe there are two different things. The customer engagement, probably the simplest thing, is whenever you collect coins or tokens, when you go shopping or something like that. Right. It's a very, very simple thing. It's a kind of a reward system. It's a very basic reward system. It's actually not that exciting. But, you know, and it's hard to kind of keep it exciting. But the framing that goes around gamification like, oh, you're playing a game or there's some kind of progression there can be quite useful the idea of progression in that, you know, you're leading towards something. I mean, you know, your question around kind of financial decision making, one of the fundamentals of what makes games quite interesting apart from the competitive and social nature oftentimes is this idea of mastery, autonomy and purpose. So if you can somehow create a system which seems to have a purpose that would lead you towards mastery, towards something that gives you some sense of autonomy around the decisions that you're making, puts decisions in front of you that perhaps has a, you know, not just a celebratory aspect, but an aspect of how you are social signaling in, in your own ability to become better at something. Um, I think those can be made in a way to make things engaging.

[00:15:27] Guest: Kam Star: Now, as with anything specifically gamification, a lot of the times, if it's just using a blunt instrument, it may work for a while and it may work in certain contexts. I don't think it's a it's a, it's a, it's a rule for in all contexts. What I was very interested in is given how competitive nearly all games are. But given that our social structure calls for deep cooperation, we don't compete with each other. Generally, on a day to day, what we do is we cooperate with each other. We do get into little groups and we compete against other groups, whether those are sports teams or companies or something else. But generally people are working together in a cooperative sense. Right? How do you frame something, and does it matter what kind of a person you are and what personality you have as to how you frame it, so that the contributions that you make towards something is is celebrated or that you know you are competing in a group against other groups, or you cooperate in a group and competing in other groups because that, that that turns out to be the most efficient way to get people motivated and that that solving the motivation issue, I think is oftentimes the biggest question,  you know, engagement is around motivation. If you are motivated, you will engage. So how do we build? How do we build motivation? And these are some of the tools. And I think so long as you apply them carefully. So as long as you keep them fresh and there is a, you know, there should be a sense of fun and celebration.

[00:17:03] Guest: Kam Star: It's hard to do sometimes, but as long as brands do that, that brings a sense of magic to what it means to engage with them. And everyone's like, oh my God, this is so much fun. This is interesting. This is a pleasure to use. Um, and people are sometimes surprised that it's, it's working. It's like, well, compare it to nothing. And I did a randomized controlled trial for my PhD where I set up an experiment where we had three groups. One set of folks were in a competitive setting, one set winner in a cooperative setting, and another set. They just had to do the job without any kind of framing. And it turns out any framing is better than no framing at all. So, you know, that's helpful. But it also turned out that if you're a highly extroverted individual, then you probably will prefer to be in a competitive setting. Whereas if you're a highly agreeable, conscientious person, then you would probably much prefer being a cooperative farmer. So there isn't one thing that is going to work better for everyone. and that personalized kind of approach. We haven't got to the point where all software is entirely personalized to our usage. I think that day might come just like our search results are very personalized. I think having interfaces that are personalized to us is that we don't have to necessarily craft. That could be something that would be on the horizon in the next five years or so, maybe sooner. So yeah, I think there is a lot in there. But approach with care, because you can make things worse.

[00:18:40] Host 2: Glenn Hopper: Yeah. And I just your, your research just all seems so prophetic because I'm thinking of, homo ludens, I guess is the book on, um, basically argues that, game playing is a fundamental part of being human. And, and I think about it, recently, Demis Hassabis was on, um, Reid Hoffman's podcast. I can't remember the name of the podcast, but they were talking about it.

[00:19:01] Host 1: Paul Barnhurst: Masters of Scale, I believe. Is Reid Hoffman. Right? Or Scale Matters or something like that?

[00:19:07] Host 2: Glenn Hopper: Yeah, I don't know. It came up in my playlist. Um, but it was, it was fascinating like that, that move 37 and AlphaGo, where it's just all the reinforcement learning and the gamification, they just you've just been you've been right on the on the line with kind of where the technology is going and where it's going now. And I'm, I'm curious with, with all that, I mean, and as you're seeing what's going on with these LLMS and, and how far the neural networks have gone, what are you seeing? You know, where do we go from here? Kind of. Where is your focus right now?

[00:19:37] Host 1: Paul Barnhurst: Ever feel like you go to market teams and finance speak different languages? This misalignment is a breeding ground for failure in pairing the predictive power of forecasts and delaying decisions that drive efficient growth. It's not for lack of trying, but getting all the data in one place doesn't mean you've gotten everyone on the same page. Meet QFlow.ai, the strategic finance platform purpose-built to solve the toughest part of planning and analysis of B2B revenue. Q flow quickly integrates key data from your go-to-market stack and accounting platform, then handles all the data prep and normalization. Under the hood, it automatically assembles your go-to-market stats, makes segmented scenario planning a breeze, and closes the planning loop. Create air-tight alignment, improve decision latency, and ensure accountability across the team.

[00:20:45] Guest: Kam Star: My main focus at the moment is around prediction and predicting the outcome of very difficult decisions, especially strategic decisions of enormous consequence, where there is nobody, no single individual, that can hold all of the different variables in their mind, in their head. You know, a company that needs to make strategic decisions about something that's happening in it. You know, they have high churn. What do you do? Do you hire more staff here? You know, maybe the CRO is saying, well, you just need more support, folks. The product person is saying we need to improve the quality of the product so that, you know, people don't need support or the customer success person is like, we need to do proactive customer success so that, you know, folks are onboarded and they don't check. Well who's right? Because they're probably all right in some way. How do you make that decision? And I think one of the interesting things that the, the scale and I can, can offer us is actually being able to build these prediction machines, not just to predict the next word, but to predict far more complex situation, or at least support people in being able to understand what the parameters are and be able to then explain what the thinking around a particular decision might be.

[00:21:59] Guest: Kam Star: You're probably right. You know, ten years ago, I was talking about how we were going to have neuromorphic computers and neuromorphic chips in all of our, you know, our computers within a few years. It took ten years. Okay. Now it's in our phones and it's all on our laptops. I don't know if it's been used or not, but, you know, and I do like being on the very much on the frontier side of things, but I think in, in at least for me, for the next, this, this, this chapter, I mean, right now, what's really interesting to me is how do we apply some of the very latest techniques around causality, around being able to figure out what the consequences, you know, the cause and effect, you know, what are creating what if scenarios, particularly around decisions that are of consequence for large organizations where, you know, one person really doesn't have all the depth of knowledge, and even a group of people may end up in a group, think or worse still, might ask ChatGPT. You know, once upon a time, if we needed a new idea, Glenn, we would just probably go in a room. Maybe the three of us would go in a room with some post-it notes and we'd play some games.

[00:23:08] Guest: Kam Star: Let's do the five whys. Let's keep asking why, why, why until we get to something and after, you know, and maybe we'd use the six hats of thinking, and then we'd come out several hours later with maybe 2 or 3 ideas that we thought, you know, were quite good, sounded quite good. We could pursue one of these. And now I, you know, my 11 year old, can go onto ChatGPT and say, hey, ChatGPT, give me 1 million ideas of how to improve the EBITDA margin on this business and make it sound entirely convincing. And it's. Well,  But how do you then choose when the stories are so convincing, in the absence of data, in the absence of some analytical thinking?  and even the analytical thinking that was fine for, for, for the year, for, for the ages that have gone right. The world that was when you're facing into a new world of fundamentally very different possibilities, particularly around the ability to compute in an intelligent way. Right. Or at least seem to compute in an intelligent way. Right. What does that mean? Um, I think it probably brings back to focus this, this need to do, more analytics, you know, be more, be more data driven rather than just.

[00:24:28] Guest: Kam Star: Well, it sounded really good, because once upon a time, it sounded good because someone had experience and they were convincing. Now the machine can sound like it has experience. Whether or not it does, it's hard to know. So those kinds of things and, you know, we've seen things like digital twins or synthetic controls often several projects in the past around creating large agent based models where you create synthetic populations that then you can use some mathematical techniques and Markov techniques to be able to decide, you know, decipher what are the main building blocks, what are the causal points? You know, look, I'll go back to the business case. Look, sales went up when we did X. That's what people see. That's what you see on a dashboard. Well that's not that, that doesn't tell you why. And it doesn't say that, you know doing X caused the sales to increase by this y percent. And here is why. And I think that my aspect is one that's really interesting for me. Because when we know the why of something, we can really double down on it and make better decisions.

[00:25:39] Host 2: Glenn Hopper: Yeah. And, you know, there's so much focus now in AI on the generative part of it. But there is still that sort of classical, you know, just machine learning and taking this data and really analyzing data that way. But it is interesting how much better these models are getting. And then I keep going back to generative AI because it's what everybody's talking about right now, which could even be a gateway to, you know, more classical algorithms. It could be, you know, the interface that lets people interact with the models when they can't write Python or R or, you know, whatever they were doing, um, previously. So.

[00:26:14] Guest: Kam Star: Oh, 100%. You know, what you can do today with me by the way, I love my coding vibe or whatever it is you call it these days. I love all of this. I have them all. Someone asked me the other day, oh, which one do you have? I'm like, I think I have them all because they all have, you know, you know, each like cursed as these things with these models. And you know, when that thing and lovable will give you this and Gemini will give you that and DeepSeek will give you this, and Claude will give you that. And, you know, each of the models have their own kind of nuances and peculiarities, and it's always good to get a second opinion as far as I'm concerned. And the model goes right. ask one model to check the other one, you know, and but, um, what's really interesting is, what it allows you to do is it allows you to, you know, once upon a time. And I think it's Einstein that's been credited with this, with this quote, which is, you know, it's 1% inspiration and 99% perspiration. Well, guess what? This world just flipped upside down. It's 99% inspiration and 1% perspiration, because the time it takes for you to build the damn thing is almost. And it will be instant. It will be near instant. If I was, if I had enough money to buy myself like a couple of a100s or whatever they call the 200, you know that can do, I don't know, several hundred thousand million tokens per second. The moment you write the prompt for the app, it's finished.

[00:27:44] Guest: Kam Star: It's done. Like you press enter. It's done. Right. So then it's about, well, what's the inspiration behind it? Where are you getting your ideas from? Are you just trying everything and seeing what works? Or are you really thinking about what it is and how you know how it feels to someone? And I think we've seen it in, in, in art, in digital art and generative art, insofar as it initially is a big rush and everyone's creating images. And after a while, those who are really thinking about the images that they're making, they're telling a story. They're really thinking through what is interesting. So it just becomes a tool to be able to communicate well. Those are the ones that are really valuable, right? And I think maybe we'll find the same way for software in that we'll just have this enormous proliferation of apps and software in the same way as YouTube right now. For every second, there's more than a lifetime's worth of video going online and being shared. And yet, even there, there are people who know how to tell a story, who are who people want to follow and watch. Maybe sometimes for hours. You know, this idea that we've got really short attention spans? Well, do we know, maybe. But, you know, people still create long videos and they still watch them in the millions because it's interesting. Right? And so I think we'll have the same thing. So yes, there will be a lot more dross. There'll be a lot more terrible videos, much more terrible writing, far more terrible apps. But the ones that really resonate with us for whatever reason, whether they solve the problem in a better way, whether the experience is one that we really covet and prefer, I think, you know, those will rise to the top.

[00:29:30] Guest: Kam Star: And yeah, I completely agree with you. While we're still we're Where today? Sitting here with a mouse and a keyboard. You know, there would potentially come a time where maybe that is really no longer efficient. I don't know. I don't know whether I could do everything that I can do just by telling the computer what I want to do, because I think sometimes it's easier to jab at something. But if it could preempt what I'm thinking and, you know, if I was, if I, if I'd like to, you know, if I had a crystal ball, I'd like to get to a point where my computer is. And I don't mean this in a creepy way. I mean, this in a, in a, in a privacy sensitive way. But it's watching me. It's watching me because it wants to, you know, actually do what I want so it can see where I, where I'm looking at. It can detect from my tone what I mean with, oh, can you just, just like, make that a little bit better because it knows what I mean, right? I can't wait for something like that. You know, a tool that is so hand in glove with my thinking and has learned my ways that I really can just flow and create without feeling encumbered by a tool that is kind of getting in the way and slowing me down.

[00:30:46] Host 2: Glenn Hopper: Yeah. And, you know, that's Sam Altman's vision of the kind of the infinite context window where and, you know, and open. I just turned on memory across all your chats and everything. So they expanded their own context window. So I think, you know, that's the way that something would be truly helpful. It knows it's like targeted ads on steroids. It's because it's. Yeah, but for every aspect of your life, you're.

[00:31:08] Host 1: Paul Barnhurst: Concerned about privacy and knowing everything about you. Now just wait.

[00:31:12] Host 2: Glenn Hopper: Yeah.

[00:31:14] Host 1: Paul Barnhurst: As long as it's a good way. I mean, there are scary things around that, but it can be incredibly helpful at the same time.

[00:31:21] Guest: Kam Star: And you know the problem? My problem is my younger daughter insists on using my OpenAI, ChatGPT for all kinds of making up stories and talking about capybaras and all kinds of stuff. So whenever I talk to them,

[00:31:34] Speaker6: My daughter, too?

[00:31:35] Guest: Kam Star: Right. Okay. So if you say I created an image of my life, um, based on what you know about me, there's, like, little toys. There's all kinds of strange things. It's like, oh, you know, based on what? What you know about me. Right.

[00:31:48] Speaker6: But you did assume.

[00:31:49] Host 1: Paul Barnhurst: You were a female. That's what it did for me when I asked.

[00:31:52] Speaker6: This is it.

[00:31:53] Guest: Kam Star: Right.

[00:31:53] Speaker6: So, you know, I think I think it's good. I think.

[00:31:56] Guest: Kam Star: It's fun. But, yeah, I think that with these machines, we are very, very predictable. Right. what we do, how we think, um, you know, as long as any of us, me included, would like to think, oh, no, I am the master of my own destiny. I am making new decisions today. Oh, no, I'm not. Oh, really? I'm not. I'm a highly predictable individual with the same set of. Ah, he's reading the New Scientist again. Yep. No, he's going to be reading. He's going to be on the slash science Reddit now, and he's going to be reading into that thing that it's just it's going to be the same thing. Right. And I think that's um, yeah. That being able to do your point, Paul, Pull around advertising to us. I think, you know, we saw that with WhatsApp and Facebook, listening to people's conversation and then shaping the adverts that are sent to you or the stories we hear about people, preempting someone's pregnancy without them even knowing it because of the changes in their habits. I think those are quite creepy, but I think probably we've seen nothing yet. Right? It's going to become so personalized, um, that we'll probably feel. And if it can be done in a way where we don't feel threatened, but it feels supportive, I think that's the real. Can we do that? I don't know, but that's what I think the next playground for something like a Google if it could indeed be Google. But you know, the kind of the advertising world would want that so that the adverts that we see are totally personalized to us, not just because of the, the, um, cookies that we have of the web page that we went to. But because of a much deeper understanding of the factors that motivate us to make decisions. And so I think that's I think that's quite interesting whether it could be misused. Yeah, I think it probably can.

[00:33:45] Host 1: Paul Barnhurst: Right. The internet's been misused. Technology can be used both positively and negatively. That's never going to change. I mean, right, it's like hackers. You build something to keep them out. They find a way in. You build again, they find a way in. You got really smart people on both sides. And I don't see, at least in the short term, maybe in some distant future we might have some solution. But I don't see it anytime soon that, you know, you can have something just be used for good.

[00:34:10] Guest: Kam Star: Absolutely. I think that would be nice, a kind of a society of not just universal basic income, but universal generous income, where everybody gets to have all the things that they want is a generous income. You know, that kind of Star Wars. That would be a lovely thing. I've read far too much about human behavior, you know, struggles around power and dominance to fool myself into thinking that's going to be an easy one. but, you know, we'll try.

[00:34:39] Host 1: Paul Barnhurst: Yeah. I mean, it would be great if everybody had what they needed. And we live in a world where we're all content, but we all know human nature, and I don't think any of us plan on that anytime soon. So I want to ask a question first about your company that you recently built. So you're working on a company called Cosafa, and you talk a lot about this earlier about decisions and how you're using AI to help big organizations. So it feels like that's kind of what this company is doing, right. Data driven value creation for PE backed companies. You're really helping them make better decisions with predictive analytics. Is that kind of the purpose there with what you talked about earlier?

[00:35:16] Guest: Kam Star: Absolutely. That's that entirely you know, private equity is incredibly effective, you know, and some very, very smart folks who build, who buy and scale businesses in a way that that business was never going to be able to probably do by itself. But even the best firms have this fundamental challenge of certainty. You know, how do they know? How are they sure about a particular value? Creation is actually going to move the needle for that specific company in that tight timeline. And I think that's and that's where I'm building right now to create those causal discovery algorithms that help you and adapt it to the business context, so that you can do scenario modeling so that you can actually have a tool that starts with ingesting some of your information and some of these. And this is quite frictionless. And some of these algorithms are quite data sparse. You don't need to have a huge amount of information, right, in order to create hypotheses and then run simulations on top of and then be able to go a little bit deeper in terms of the causal inference. So I mentioned the digital twin, but a digital twin of the company. And seeing how it would perform with or without a particular intervention, what that might look like in order to move away from just a pure gut feel. But I think I need to stress here that this isn't about replacing all those many years of experience and expertise that folks who are in this space have.

[00:36:47] Guest: Kam Star: It's actually the vision is to support, support, better decision making, having more predictable, more efficient, data driven decisions around things that are not easy around product decisions, around particular trade offs that you might have. Um, and here I'm not talking purely about financial engineering. I think the financial engineering time space is very well understood. It's almost commoditized. You know, it's of course you're going to run a model, of course. Right. But I don't mean that. I actually mean anything, everything outside of that, that is to do with, with actively building or refining particular capabilities of a, of a business applying data to that, sort of data insights to that. And, you know, to that end, I'm pretty convinced that within a few years we're going to see, this causal inference, you know, why something happens to strategic validation across the board. It's not going to be a novel thing. You know, I know right now the top tier funds already do this, but they can afford to have armies of data people. What I'm really interested in is bringing this to democratizing this, actually bringing these solutions to, if you like, the lower mid market where you don't have half a million, 2 million to spend to figure some of this stuff out, you actually want to have but you want to have access to this kind of insight. And I think that's an interesting I find it really interesting because it's really challenging.

[00:38:20] Host 1: Paul Barnhurst: Love it. Thank you for sharing that. That's great. I think we're going to move on to a fun section we have. This is one we kind of surprise our guests with. So he's like, oh, what am I in for? So I'm going to let Glenn do the first question on this one, because I've been hogging it lately and then going first. So I'll let him explain it and then I'll ask my question after him.

[00:38:39] Host 2: Glenn Hopper: Yeah. So I actually wasn't set up for the, for doing my question, but I'll tell you what, we're what we do. So we used to do like a bake off every week where we would have ChatGPT versus Gemini versus Bing back in the day and all that. But, um, then we just kind of settled on and they're, they're so similar right now, all of them for all this. But we take your, your profile and your, um, whatever information bio we have on you, we load it in there and we say, come up with some fun questions about this. And so Paul does this a little bit differently. But then I then take it and feed it into which is what I'm doing right now. I normally do this while Paul is talking, but I, um, but I go into another, GPT and I say, I can only ask one of these questions typing this out live here?

[00:39:27] Host 1: Paul Barnhurst: He asked, which one should it ask?

[00:39:30] Host 2: Glenn Hopper: So I'm going to do that now.

[00:39:32] Host 1: Paul Barnhurst: And which tool do you use?

[00:39:34] Host 2: Glenn Hopper: So I'm using grok again. I've been playing around with this more. I paid for the upgrade. I'm trying like you were saying earlier, Kam, I feel like a split between Manus, I'm now on the paid version of Manus. I'm on. I don't know, I'm I'm I'm sort of just blissfully ignorant and willfully ignorant of how much I'm spending every month on these subscriptions.

[00:39:55] Host 1: Paul Barnhurst: I bet between the three of us, we get over 500 a month on these things.

 Well, no.

[00:39:58] Host 2: Glenn Hopper: I'll tell you alone. I'm paying for the $200 ChatGPT, so I've got us almost halfway there.

[00:40:04] Host 1: Paul Barnhurst: So yeah, I think you two are both going to have me beat by quite a bit. I think I'm around 60 a month, maybe a little more. You know, I have about three different tools I pay for right now.

[00:40:14] Host 2: Glenn Hopper: Okay. All right. I'm going to tell you, before I tell you the question, these and these responses are getting better. It says to get the most engaging and revealing response. I recommend asking. And, well, let me tell you what it says. This question stands out because it's imaginative, open ended, and invites came to share a mix of creativity, personality, and personal interests. And it goes on like that for a whole other paragraph. But I'll skip all that and I'll tell you what the question is. If you could build a secret room in your house, what would you put in it?

[00:40:44] Guest: Kam Star: If I could. Okay. I mean, I would make the holodeck.

 Sure. Yeah.

[00:40:50] Guest: Kam Star: Because, like, you know, why not? Instantly. I have every room imaginable in my room. But assuming we can't manipulate, you know, the quantum fabric of the universe. With that assumption in mind, if I could build a secret room. And this is going to sound very boring to most people who don't live in central London, but I would, I would put in and this is very personal to me. I would deck it with just the, I don't know, as many layers of, of soundproofing, and I would turn it into an anechoic chamber so that I could for maybe ten minutes a day, just go and sit in it in complete and utter peaceful, peaceful silence with no noise from the many millions of inhabitants who seem to just be outside my window most of the time. Um, that's what I would do. I'm sorry if it's a bit.

[00:41:46] Host 1: Paul Barnhurst: My daughter would love that room and peace.

[00:41:50] Host 2: Glenn Hopper: I was going to say, as someone who lives also in the heart of downtown in my city, not quite as many people as London, but, um, with the the ongoing, ongoing, God knows what kind of vehicles and people fights outside, just drunks yelling around outside. It's like, if I could just go and we, you know, get glass walls everywhere. So if I could just go shut myself in somewhere, I'm right there with you. So great answer. And I think Gronk Gronk picked a good question. You know.

[00:42:16] Host 1: Paul Barnhurst: All right. So I take a different approach. I'm not quite as sophisticated as Glenn is what it boils down to. So you have two options here. I use a number generator and you can pick it. You and I can randomly pick a number, or you can pick a number. And it's between 2 and 25 because Glenn used question one. So one's off the board. You could pick the number or the random number generator.

[00:42:38] Guest: Kam Star: Let's go for 25.

[00:42:40] Host 1: Paul Barnhurst: 25. You have no idea what question you asked. And you got it.

[00:42:45] Host 2: Glenn Hopper: The Alpha and the Omega.

[00:42:46] Speaker7: All right. I like this one.

[00:42:48] Host 1: Paul Barnhurst: What's one ambitious goal you still have on your bucket list? And then it adds this. It adds a little bit more that might surprise our listeners.

[00:42:57] Guest: Kam Star: Okay, I think my. Well, let's see. It may or may not surprise some of the listeners. My one ambitious goal is to list the business on the stock exchange. That's what I would like to do. I have had the pleasure of ringing the bell, but it was not my company. I would like to do it when it is my company, and for no other reason than it's just fun. No, there isn't anything other than that.

[00:43:22] Host 1: Paul Barnhurst: If that happens. When I should say not. If when that happens, we'll bring it back on the show to talk all about this.

[00:43:28] Guest: Kam Star: Please. Right.

[00:43:29] Host 2: Glenn Hopper: I thought.

[00:43:29] Speaker7: You were going to invite me.

[00:43:30] Host 2: Glenn Hopper: To a Company.

[00:43:31] Host 2: Glenn Hopper: Oh.

[00:43:32] Host 1: Paul Barnhurst: All righty. Well, thank you so much for joining us. It's been a real pleasure. I've enjoyed it. And, you know, I love all the learning you've done around AI and the knowledge you're able to share. So thanks for joining us.

[00:43:44] Guest: Kam Star: Wonderful. Glenn. Paul, thank you so much for having me. It's been a real pleasure talking to you, sir. You've made my Friday afternoon, so thank you.

[00:43:52] Speaker7: You're welcome.

[00:43:53] Host 2: Glenn Hopper: Likewise. Thank you.

[00:43:55] Host 1: Paul Barnhurst: Thanks for listening to the Future Finance Show. And thanks to our sponsor, QFlow.ai. If you enjoyed this episode, please leave a rating and review on your podcast platform of choice, and may your robot overlords be with you.