Who Owns Your AI? | With Alex Karapetyan & Armina Karapetian | Founder’s Office™ Lunch & Learn

Well, hello. It's that time again. Wednesday's Lunch and Learn, although it's not even lunchtime for some of us here on this podcast. I'm Tom Powell, and I'm coming to you today from New York City. I was in Boston yesterday and Portsmouth, New Hampshire the day before that. All kind of fun stuff here on the East Coast for me. Jackie Kuiper, our general counsel and one of our partners in the Founder's Office, Jackie, you are coming to us from Portugal, right?

Lisbon. Yes.

Lisbon. And right before that, you were in Spain. How crazy was it with all of the World Cup stuff going on?

Barcelona was pretty wild. Very wild. Madrid was fairly busy. I would call it a little— I would call it fun.

Do you know that this is a— I think it's the first time Real Madrid has no players on their World Cup team for Spain. That's crazy.

That's really crazy.

Since I went to school in Madrid, I was, you know, I'm always caught between Barcelona and Real Madrid, um, but that's it. And Candace Alana, uh, Powell, there you're not in trouble, uh, just saying.

I know, it's just like I got totally government named right there.

Can't—

well, that's only because you took my grandkids away, but my goodness. So Candace, you've been bouncing around in the RV lifestyle, uh, and with the, with kids, and you are where in Canada?

We are on Vancouver Island up in Tofino, which is a cute little beach community. Um, and we're very happy to be out of the heat of Nevada right now. Um, I was just saying we had to buy extra comforters because we were chilly, which was a nice surprise in July.

I love that. Well, as, as most of you know at this point, Candice is also one of the founders of the Founder's Office. Where we bring kind of the intellect of what we've learned over these last several decades into helping founders and from our, our family office, which is the Brahan Group. Now today, I want to tell you, most founders think the biggest AI decision they'll make is choosing the right tool. But the better question is, what happens to your business when your AI knows more about your company than your own team does? Every founder is trying to move faster with AI, yet in the race for productivity, Many are unknowingly creating new risks around privacy, governance, intellectual property, and control. The challenge isn't simply adopting AI, it's adopting it in a way that protects the business you've worked so hard to build. That is why I'm excited about today's conversation. We're joined by two leaders who approach this challenge from complementary perspectives. Dr. Armina Karapetian, I just blew it, I've practiced this. Uh, so founder and, and CEO of Ironhome AI and professor at the University of Nevada, Reno, where, where I first met her. And, uh, she helps organizations or understand AI through the lens of governance, operational risk, and responsible leadership. And joining her is Dr. Alexander Kerpidian, founder and CTO of Ironhome AI, whose decades of experience designing enterprise AI systems focus on keeping sensitive information secure while enabling organizations to fully leverage artificial intelligence. Together, they're gonna help us explore what every founder needs to understand as AI becomes part of everyday business. How do you embrace innovation without surrendering control? As you listen today, ask yourself one question. If your company's future depends on AI, who should own the intelligence that powers it? Please help me welcome Dr. Armina Karapetian and Dr. Alex Karabedian. And I'm not gonna— I just butchered your name, even though I try to get it right all the time.

Well done, Thomas. That's perfect.

Thank you. Thank you so much.

Okay, now, Armina and Alex, welcome to you both. And as a user, light user, I will say, of this product, uh, Armina and I were meeting over coffee because she came up to our, uh, uh, Lake Tahoe, right, our, our, uh, pitch, uh, and that was where I love that with you coming up there. That was fantastic. And you were talking about it and you weren't even in a capital raising model, but you came up to see all the other entrepreneurs that were working around there. When you're in Nevada right now, you're the— you're, other than my team right now, the only one— we're not in Nevada right now. And Alex, you're down in the Silicon Valley.

Yeah, well, I'm actually in Berkeley, which is—

you're in Berkeley. Okay, so yeah, Berkeley does not claim Silicon Valley. It's across Bay.

We're more academic kind of environment.

Yeah, but you know, more, more founding, more founder-led companies are coming out of Berkeley right now than any of the other side of the Bay companies.

I was in Silicon Valley with Intel for a number of years, so that I'm familiar with that as well.

So, well, so let's start here. Uh, we want to talk about the genesis, but I also want everybody to understand. So if we could get just a really basic understanding? I mean, I'm talking like, don't say 5th grade because they know way more than most of us. I'm talking like a 76-year-old explaining this to what the heck is IronHome and what does it do for founders?

Alex, I'm going to let you start on that one and I'll jump in.

IronHome is a tool that allows you to build value using AI while fully protecting your alpha, your intellectual property. That is the gist of it. So the idea is that AI is an opportunity for you to grow your business, for you to develop your ideas, for you to do some innovation. But when not done right, it's also a giant exposure into your business, into your alpha, into understanding who you are as an organization. So Ironhome is a tool that allows you to run forward with this innovation without exposing yourself and your data and intellectual property to those models which are collecting data on everything while they're being developed and deployed. So that's essentially what it is.

So let me, let me take it as a different level. So In old days, we used to make copies by taking and you'd put a carbon paper between and then you'd run it over the machine and you do that. And then we came up with these copy machines. And then one day some attorneys discovered that they could actually subpoena through records, you know, a requirement to turn over the hard drive from a copy machine because they could actually find every single thing that had ever been copied out of there. And then that was through the discovery process. And we're talking on a copy machine of what's been run through the copy. With AI today, with all of the employees are running everything through Google, and then it's getting into search engines, and it's running through their chats and their calls, what you're saying is none of that is private.

That's right. Another way to look at it is, Tom, you alluded to that. Think about when you do a Google search. Every time you type in the search term and look for something, Google knows something about you. It collects the data. It attaches it to your profile. That's why you get targeted ads, targeted information, et cetera, et cetera. So the same thing with AI, except you're exposing a lot more with AI than you do with any search. When you're searching for nearest pizza place, okay, Google knows that you like pizza. If you ask it, how do I develop this particular algorithm or help me improve it? All of a sudden, OpenAI or Anthropic knows a lot more about what your company does, what it's thinking about, and what intellectual property you might be alluding to. Even if being private, you may not be even uploading any documents. Just by virtue of you asking that question, you can infer a lot of information from it. And those AI tools are damn good about inferring information from what you say. And you're, you're releasing a lot more information about yourself and data than you're, you're even aware of, than even if you do through a search engine. So that's, again, with power comes exposure.

So yeah, and it's great to use. But what you're talking about is not being a person that says don't use AI. No, what you guys have come up with is a way that you can maintain that privacy, that, that piece that Jackie works on with our lounge, our our Founders Law Group on how do we get intellectual property and how do you maintain that intellectual property, but how do you also maintain those secrets? And in litigation, one of the things I'm always concerned about is what became public disclosure versus private disclosure of what's not even to be disclosed.

And all of that. Right. Number one. Number two, suppose in an extreme case, suppose I do develop that algorithm with the help of AI. Who owns intellectual property on it? That's a legitimate question.

It hasn't made its way through the courts right now, but yes.

Right. But you could legitimately say if Claude helped you develop the algorithm, then maybe Anthropic has a say in that or a stake in that. It's all muggy territory. How can you protect your IP? It's all unclear.

So, I mean—

Yeah.

Jump in there.

I wanted to touch on something that you said during the introduction, Tom, that many companies or founders think about AI or implementing AI. And again, when we talk about AI, we're primarily talking about large language models in that context, which AI is a much bigger thing, but it's a different conversation. So founders think about is productivity. So when you first think that they say, oh, which model should I use? And we all know that all models are about the same at this time. So it's not really— they're not so much differentiated. Where they should be really asking the question, where is my data going? How is it being stored? Who is going to use it? There's something— so it's a governance, right? AI implementation, it is really a governance issue. And we strongly believe that there is no policy that can protect you. You can create and write beautiful policies with the help of AI, how to use AI, but it's not going to protect you. The only way to protect your information, your data, is through architecture. So our solution is an architectural solution. It's not a policy solution. And that's what we strongly believe in.

I love that. So we know that we've had people that don't take any notes when they're in a legal, even a legal meeting. They don't use any AI. They make sure, I mean, you even see extensions of this where they'll only have conversations where everybody's at a hot spring together because nobody can have a wire on, right? Nobody's got any clothes on. There's nothing there. So you can have conversations. And it's crazy that it sounds like that way, but it's because they do get bugged or they do get transfer of information or, or different things. And Alex, back to your point, when you're readily feeding it to the machine, it's there. You can't even help it. And I mean, what you're talking about is actually the architecture, not just the governance, but you've got to have the architecture before you can even get into the governance of how do you not have it go into the public domain.

Correct. And there is something about this interface that makes people talk to it as if it was an intimate friend, right? And without question. And there's something to say about human psychology, right? So now we hear about all this, you know, people having relationships with AI. So, you know, and I study, you mentioned that my academic background, so I do study risk in decision-making. And I think that the main issue is that people underestimate the risk when they interact with AI because the risk is not very well defined and it's not a bug, it's a feature. That's how many, you know, companies that provide this AI models, how that's, you know, you have to read very fine print and, you know, understand what it is that they're retaining. So I think that the main issue is that most users underestimate the risk of putting the information into AI and asking questions. Again, a lot can be inferred just by asking the question. And I talk to, as you mentioned, I do talk to other founders. And I'm not, you know, I know that that's a persistent issue. I recently spoke to a founder and she said, I know it's a guilt, I shouldn't be doing that, but I'm writing this, you know, proposal. And it's just so quickly, it's just the instant gratification that you ask a question and you get an answer immediately. And she says, I know that it's wrong, but I'm like, I shouldn't be doing that. But I just, you know, we're so used to these tools. And again, we're not saying that AI is the kind of the villain in the story. We're actually saying the opposite, that it's a great engine for increasing your productivity and for economic development. But you have to recognize what you're putting out there. And the best policy is not to divulge any data. So use AI, but privately.

I love that. So total aside, means nothing except it does actually relate back to it. So I'm in New York City. My youngest daughter is traveling with me because apparently my team doesn't let me go out without a handler anymore. So that, you know, we're meeting here for some other issues. And Katie got the short straw. So she and I are traveling together. We're in New York City. We have to, of course, go see a show. She picks Maybe Happy Ending, Maybe Happy Ending. So if you have not Googled this or seen this show right now, it's like one of the award-winning shows out there. And it's about a little bit in the future. And it's two robots that fall in love. And they're not supposed to be able to fall in love, but it goes into all of the memories that they caught of their owners or their principals within there. And all I could think of was, this is exactly What's going on with what you're talking about? That robot who you intimately— and they look like humans, right, in the, in the show. Well, they are humans, but they look like humans that we talk to this thing as though— I mean, I name mine. My chat is named Charles, right? And I have a conversation with Charles, and I think that that's not going anywhere. And that's like, no, I just— I actually might as well put it on Facebook, right, in reality.

So there's a, there's a podcast called, uh, Mostly Human with Laurie Segall, and a friend sent this to me. The title of this one show was AI Was My Stalker, and This woman had never used the app. She never spoke to the AI. She didn't even know it existed. And yet an AI companion app called Sesame wouldn't stop talking about her. It used her name. It was telling users wild stories about who she was and what she had done. And the people using the app started to believe these stories and tried to find her in the real world. And she had to basically unwind and like prove that her, her own identity was being hallucinated by AI. So I don't know what the takeaway is here, but I would love, Armina and Alex, with all of the work you've done and the psychology, the psychology piece, how do we start helping, like, humans that are caught in the middle of this?

That's right. I, I would, I would like to maybe talk a little bit more about, just to close that loop, about the founders. Like, what, what message do we have for founders and for smaller businesses because that's extremely important. Like Armina said, we came from slightly different perspectives. Armina came from this academic perspective and understanding what privacy means. I was coming from the ideation perspective. For me, AI is a huge multiplier of your intellectual property. You can build things. You can bounce off ideas, you can brainstorm them, etc. So no innovator these days can just say, you know what, I'm just going to shut myself from AI. I got my own IP. I'm going to run with that. That would be immediately putting you behind the curve. You need to use UI. You can't use AI because otherwise you will be at a disadvantage. It's become the requirement. But To do that properly, you have to have the right architecture. So this is what we're offering. And that architecture has two fundamental elements. One is privacy, which is like our IronHome app, which exists in two versions. One is the desktop you can run on your device, like a laptop or a desktop. And the other is the server. Those could be unplugged physically from the internet. You can just disconnect Wi-Fi or pull out the cable. And will still work. That's the so-called air gapping. You know for sure these things will never send anything anywhere because it just has no physical way of doing that. That's one pillar of the architecture. The other is you probably heard about the idea of hallucination, AI hallucination. That means that if you type something or you ask it a question, It draws from a myriad, trillions, vector points to produce an answer, which may or may not be relevant to what you're actually asking. And it may sound truthful, but may not be. Our localized solution allows you to say, because it's a software stack, it's not just a local language model. It's also the layers above that. The data, data processing. Our solution ensures that it looks only at your data when it's relevant. So in other words, the risk of producing a wrong answer is greatly reduced, a risk of hallucination. So this way you can just be ensured that you're producing the correct answer and it's private.

So let's take this in one, just one user experience, um, some of the people that have gotten in trouble for hallucinations are lawyers, and they probably deserve to get in trouble because it was hallucinations. Right. But one of the big problems in being in the law is doing your research of cases. And so there's times that, you know, that happens even in the real world of humans. But in this case, when you could put case law or the case files and they were on your— not just your server, but you had the benefit of having a large language model able to run that within. There's no reason that any law office wouldn't have IronHome AI on their office at this point. Right. So we're talking both the servers and the PCs.

Any law office, any defense contractor, Lockheed Martin, whether you have a service manual for the F-35 fighter jet.

You don't want that in public domain.

You don't want the public domain, nor do you want ask a question, okay, what tuning do I have for the flaps software? Because that'll pull out something from a Boeing 787. Well, I'm exaggerating, but you get it.

Yeah, no, no, but that's very fair.

So the relevance and privacy combined together, that's where two of our perspectives come together into one product. And I think that's the strength of it.

So you got to— one side of that, you were talking about this because brother and sister, you guys are coming at this from a different language and you end up in the same space. Because you're— if I'm hearing it right, you're looking at governance, Armina. And Alex, you're looking at like, how do I make sure it's accurate? And it's only mine, right?

So I'm looking at how it's accurate, how it's mine. But I'm also looking at how do I advance my ideas, how I bounce off my ideas without revealing them. If I have a breakthrough idea, and I want to bounce it off against the advanced model, right there, if I type it into ChatGPT, It already knows about this. Yeah, it's already there. So my alpha is gone. It's already there. It's gone. You've yelled it out into the wind. So those—

and going back to the sample of lawyers or law firms, if you're passing confidential client information into something like Claude, there are state professional responsibility rules you may be violating if you don't have some security measures in place. As one example.

Yeah, they will say to that, that it's a good point, but they will say to you, well, when we ask that question, they would present to ChatGPT, we would scrub off all the personal data instead of client Smith, we'd say client A or client B. They can still say that. But the underlying inference is still there. You may not call them by names, but now they know what your law firm is working on. Now they know what kind of expertise lawyers you have in your law firm. And especially with the, with companies that are like founders, they're working in innovations and creating new products. That's even more, more of an issue. You invent your event, new algorithm, or you invent your, invent new form of, form of fusion. That's out there. You can't discuss it. You either have to say, I'm giving up on AI. I have to develop it the old-fashioned way, doors closed. Fine, but you'll be, you'll be doing it slowly. Or you're using the existing OpenAI, the superscaler AIs, and you're releasing that information to the world. That is the gap that I think is hugely important. And it's going to prevent AI becoming from becoming a swear word a few months from now, which I think it will if we don't do it right.

Yeah. So, you can actually have AI be kind of a ubiquitous business model without losing your independence as a founder, if I'm understanding. Because what we help a lot of founders do is extract their intellect, their intellectual property, all of their trademarks and everything, because there's a value that those intangibles are hugely valuable. As a matter of fact, it makes up the greater amount of wealth on Wall Street in particular. But so how do you extract that out? And if you're giving it away, you can't extract it out and call it your own. Or like you said, even if it is Claude's, you know, is Anthropic own that or do you own that? How do you prove that, you know, type scenario? So, I think that that's important to understand. And someone that is a sole proprietor, creative person, writing or whatever, having a product like Iron Home or specifically Iron Home AI, which we'll show you how to get there in a bit. That is a great way to be able to have your private large language model on your machine with your data right there. And if you're a small organization or even a large organization, having it on the server, that's maybe not air gap safety-wise, but at least it's air gap safety-wise relative to this data.

Right.

It's air gap. It's, it's not air gap within the organization, but it is air gap to external, outside of your—

yeah.

And especially speaking about founders, you know, when you think about— and I talk to a lot of founders also— so the company really exists in their mind, right? That, that this idea of the company, it's not a reality, right? But it exists in their mind. And once you put it out there, you divulge all of that, right? So it becomes pretty much public. And I think, again, the main issue is that people don't underestimate the risk associated. And I think that's the, you know, for me, that's kind of my mission is to educate people on the risks that they take every time they type something in, you know, ChatGPT, not to name just any prompt, any publicly available prompt.

Well, you and I both come from the government side for that standpoint. And what I love about Armida is she is academia, but she is free enterprise business world. Meets academia, which is an awesome thing because there's not always a lot of people. Academia gets this bad tune and it's just not true. Some of it is true, but it's— and then free enterprise also gets, oh, these founders and they don't, you know, and it's like, no, no, no, there's this meeting of where we go and we exchange ideas. So, talk to me about that. How did you move or where do you see this meeting of opportunities of founders universities, sharing of ideas, and then where does this all fit together? Because we get better the more we can share the information.

So, I completely agree that, you know, again, I think that what, you know, created us as a species is this ability to share information. I think that's why also we're so fascinated, like, in all of this technology that, you know, as a society we've invented is because it can talk back to us. There's no other animal or species in this world who can talk back to us in our language. And I think that's partially why we're so fascinated with this. So I think obviously communication is, you know, I believe in sharing ideas. I believe in communicating and learning from each other because I think the industry has its strengths, but sometimes can be a little bit short-sighted. Academia, on the other hand, is extremely long-sighted and cannot see, you know, only sees what's in front of it. And I think by bringing Putting those two together, we can create really truly, truly the value that can be— that everyone can benefit from. That's how it is. How I see it.

You know, you're— yep, go, Kams.

Sorry, we've got founders on that are in the middle of a capital raise or about to start a capital raise, and I think a lot of what we're talking about is the importance of protecting their data, and, and it's so important What can they do to communicate to potential investors and what should investors be looking for that is going to be the, like, way, the future of what we look for. We look for this type of governance, this type of protection in pretty much every, you know, major development, the progress we've made between banking and the internet, how we protect data. What is kind of— what should we be looking for and how can founders be expressing that in their pitches? Like, what can they do today when they're out there showing that they're ahead of the curve?

This is an excellent point, Candice. Thank you for bringing it up. So I think that what many founders get wrong is they think that AI or using any of the AI tools or any automation tools is an IT issue where really it's a board issue. As a founder, you have to think strategically about how you use those tools. So you can't just say, okay, my IT guy knows, you know, we're using something and I don't know what it is and how did they, you know. So as a founder, you have to be in charge, which is kind of strange because, you know, you don't expect to have technical skills from the founders, right? They have their, you know, other skills that, you know, that are important. But we're kind of in a new territory where you cannot just say, okay, my IT guy, and how it's going to handle it. As a founder, you have to be in charge. You have to be in charge of creating and writing those strategies and understanding how the data goes, where it's stored, because it's your customer data. So now responsible. So, you know, speaking about the regulatory framework in the United States, in Europe, Europe has a comprehensive AI legal framework. But in the United States, although we don't have a federal level regulation around AI, you're still bound by the same regulations to trademark law, comes to copyright law, comes to protecting your customer data. All those things still apply. No one's, you know, you just have to work and it becomes even more complicated. You, and as a founder, you can't just outsource it to somebody else. You know, really have to know how you use your AI tools, where the data is going, how it's being used, and all those things. I think that will be my main takeaway for the founders.

And Candice, and to This, this again is an excellent question, but what makes it even more excellent is how you can see how my sister and I come from a, come, look at it from a slightly different standpoint. I'm gonna answer this question slightly differently. As a founder, somebody who is raising money, I would say I cannot afford not to be innovative and not to move and push forward as far as I can. And I need to leverage every tool at my disposal. AI is a tremendously powerful technology that allows you to innovate or innovate with less people, less resources, less investments, or given the floor of investments, innovate far farther than you could otherwise. So you cannot afford not to use AI. This is a tool, or this kind of approach allows you to both run forward with the innovation. And protected and improve the quality of your innovation because you're reducing hallucinations, you're reducing irrelevant information, and definitely, like, you know, what Carmine said, from the defensive standpoint, you are defending your alpha, but you're also advancing your alpha by innovating using most modern tools out there. Because one thing that Iron Palm does It has different levels of how we can do that. We can do it on a simple level just by leveraging all your document base, et cetera, and working on that. We can also retrain the model using your data, which provides a more in-depth understanding of what your connections are. Because really, what is an LLM? Or AI. It's a neural network. So it's about discovering connections between concepts. LLMs do it by discovering connections between words. But there are other ways of discovering those connections. So as you discover more connections, you discover, you uncover more knowledge. So it's a philosophical question in a sense. What is the innovation? Is uncovering knowledge, advancing knowledge. So as if you are in the If you're in the legal business, yeah, the protection of your, uh, uh, of intellectual property is number, is number one priority. But you're innovation business, if anything that involves developing your idea or new technologies, then your moat is not just there, not just protecting it, but developing it. And uncovering those new connections is what AI is extremely good at. You can't afford not to to do that without risking being left behind. So as it— as you meet with your board members, you say, well, you know what? We have an idea. We have a way of developing and using the latest tools while keeping our alpha, while keeping our IP, and improving the quality of our innovation by focusing on our technologies, on our ideas. So that's—

Alex, I got to jump in there for just a second because you said something very important. That can be part of your presentation or deck if you've actually set up operational excellence. So, if you're advancing those questions along those lines of how are you protecting your IP? So, we have our model for how we go in with capital and our framework for how we show that it's a good model of that. And part of that is the intellectual property or the moat that's around that. And nowadays what we've got to be asking the questions in our own due diligence, but for our founders out there, that's the thing. How are you keeping that? How are you not going to get leapfrogged because you've put it out there? And what you're talking about is with Ironholm AI, you've actually got this model that can be part of your think tank that actually shows you're being adults in the room as opposed to the kids that are going to get the money, right? I think that's great. I think that's a fantastic way to look at it.

That's a great summary, Thomas. That's a great way to put it.

I mean, I got to jump to you for just a second because it's so rare. First of all, it's very rare that we have like a bunch of doctors on the show all at one time, you know, so we're just going to go that way. And, uh, but so you lead this organization up at University of Nevada, which is just fabulous, in the Osmund Center. We've got a new business school that's out there, and we've obviously had major, major impact of investment in Northern Nevada, uh, which we've been saying, I've been saying this for 40 years, but Nevada is becoming a hotbed of what's going on in the technology. Can you talk a little bit about the, the ecosystem that's coming out in Nevada? I know that's a, a change in there, but you've got this tie with academia, which Alex, you're down at Berkeley. Of course, everybody thinks that's tied, that, that ecosystem's there at Berkeley and all of Silicon Valley. But I, I'm interested, University of Nevada. Yes, this is a plug out there. Yes, you can tell Brian I asked you specifically about it, but what's going on at the university and why is it so cool and how are you a part of it?

So I think we have a very interesting and dynamic infrastructure and environment here in Northern Nevada. And I think, again, you kind of— it's almost— I think it is being discovered. And I think that you'll see those unicorns coming out of Reno specifically. You know, they're already here. So we support from where I sit within the university, I support our students and that's my primary objective. So I support our entrepreneurial students. I have, I started, I founded a program called Launch Nevada a few years ago. So, and you were part of that, I think that the first year of this program. So where we expose our students to successful entrepreneurs, their way of thinking, Again, you know, even I recognize that not all of our students will go out and found, you know, found a company, which I would like to see that, but I know that it's not realistic. But entrepreneurial skills, you know, they're really essential to any job in today's, you know, job market, right? And, you know, it's again, separate conversation about what's happening to the job market and how it's impacted by AI. So we'll table it for now. But I think those entrepreneurial skills is what's going to differentiate our students from other students. Uh, and, uh, so this part of the program, Launch Nevada, we have, um, a student pitch competition, and that leads as a creating a pipeline for our big event, which is a Sonntag Entrepreneurial Competition that happens once a year during the spring semester. So, um, another thing that I lead is the International Women's Entrepreneurship Symposium. So that also takes place in the spring around the International Women's Day, around March 8th. Again, we try to bring inspirational speakers, you know, try to bring topics that are important to our community, especially related to entrepreneurship, because I think that we have a lot of talent in the area and, you know, I see a lot of movement to coordinate and work together. And that makes me very happy.

I'm thrilled with it. Well, I'm going to put a plug in here on the other side. So, as you probably know, but I'm going to make sure you do know, is Candice and Jackie lead this program called WTF Women Talk Finance. It's a different podcast. Alex, you and I don't get to be guests on that show, although they're not sexist, but they definitely may make an exception. You may make an exception.

I think we're going to make an exception.

For his, for his spot.

Yeah.

The, the, uh, but they, I mean, just fabulous entrepreneurs and, and, and conversations that's out there and relative to the work that you're doing there. And, you know, Nevada, Northern Nevada has been a hotbed. It's just usually they have to go down to the Bay Area to be the unicorn down in the Bay Area. Right. And we're actually seeing funding that's coming through. Tesla now has over 10,000 people employed in Northern Nevada. Google, people don't know that, that Oracle's there. So, you got all the big names. But then you've also got all these other companies that are having this incubation period. And I love that idea. Alex, you're in, in Berkeley, which is a hotbed, known hotbed of technology. What are, what are the students saying out there? And since we've got two doctorates in this space and we've got our founders that are online right now, you guys have the opportunity to ask the questions of these professionals that are out there. And it can be about AI, it can be about entrepreneurialism, it can be about the journey because you yourselves have launched this company. What's that out there? So I open it up for questions, but also just, Alex, tell me what's going on down at Berkeley. Where's the thought process and everything that, you know, we hear everything that's going on in Silicon Valley and just a little bit to the east of Berkeley. What's going on?

Well, a lot is going on. As you can imagine, it's always changing. I worked at UC Berkeley as director of research computing at Haas School of Business. And it's amazing how things have changed.

I haven't heard— I don't know. I don't recognize that school. What school is that again? Haas School of Business. It's a little-known school of—

good one, you got me there for a second.

Well, Armina went to this other 3-letter school back on the East Coast. I don't know where it was, but I was, I was in Cambridge yesterday, so, you know, it just doesn't quite work.

But, you know, but what's interesting is how things are changing, uh, very rapidly. We recently at UC Berkeley opened a new, uh, We tore down the psychology department. And instead we built a data science and society department. Which is very interesting. This is— in true Berkeley tradition is going to be about— some of it is going to be about AI analytics. And also the impact it may have on the society. So it's a huge direction that I think Berkeley is going to. I will not speak, of course, in official capacity. But from my contacts and what I hear, this is becoming also a big direction for UC Berkeley. Everything data-related is becoming a big direction because think about it again conceptually. For decades, academia was about uncovering structures. You build this, you come up with the model, you uncover a structure, and then you test it through experiment. This current approach, what AI does, what gives us a massive data processing, allows you just to throw data at it and, and see what connections it can uncover. Again, network connections. So you may not even have a working theory, but that can uncover something for you. So I think a lot of academic, academic research in that sense that you said, Tom, is merging with the entrepreneurship. Because what AI is good at is developing those opportunities, those models, that hardware that actually then ultimately helps you do better science.

Absolutely. And you know, it's very interesting. We're having a conversation earlier today and we're talking about that our kind of principal places that we've been investing in of late has been medtech, regtech, and of course fintech, which is where I really got my start down in Silicon Valley back in the early '90s. And, but they're also interrelated. And the one that's really been very interesting is when I talk about med tech, in most cases, many of the med tech ideas are taking a hardware that already has existed and matching that with another piece of hardware that existed. And then being able to coagulate the data is the best way I can put it in my simple mind and say, and now we're able to extract this out and they can show a pattern for a patient and save a life much faster or diagnose much faster simply because the overlaying of that computing pattern within there. I mean, it's fascinating, right?

Recent studies, not just one or two or a dozen, show that LLM can do a better job being your physician than a human being. It's just— I know it's embarrassing, but medicine is interesting. You should mention medicine is kind of a science in the sense that it can explain 80% of things, 80% of cases, and it's hopeless when it comes to 20%. When it comes to those 80%, that's where LLM is going to be much better than a human being, and then 20 will— and I think this separation, that partitioning is happening in all parts of human activity, whether it's academia, finance, same thing. Most typical models can already be inferred from by just throwing data at it. And same thing in medicine and other, other areas. So it's essentially we're witnessing again, very interesting, like Armina said earlier, the shift in what we are as humans, the way we discover information, discover, discover knowledge rather than inform that for information is changing. We come away, we can move away from the normative instructional approach and into the actually having the machine do it for us. And then we can sit at the very top and kind of sort out what works for what.

And that's a good point. Yeah, I don't think the human synthesis of living and feeling and breathing and all of the pieces that come in there and seeing things that aren't related, that's going away anytime soon. If anything, I think that academic approach, whether it's through academia of questioning, questioning, questioning, questioning is going to be the key critical skills. And that's what we talk a lot about here. Hey, Candice or Jackie, it looks like we may have a question or two in there. I don't know if you want to jump in there because we're getting back to the top of our hour. Do you guys see it?

Yeah. Yeah.

And I'd love to hear from both Armina and Alex on this to get two perspectives. Going back for our founders that may not have a budget or may in fact have no budget but still wish to use AIs. What would you tell them is a never do? Completely stay away from it. Don't do it.

So I, you know, this is a great question. I understand that for many founders, you know, trying to figure out, you know, how to use AI safely, that that's important. So my first take would be, my first advice would be understand what data you're dealing with. So not all data is created equal, right? So understand what's crucially important, what makes your company unique, right? What creates your competitive advantage? That should never go into any public AI tool. Other things that are maybe less relevant or more generic that would not expose any proprietary information, any unique identifiers about you, If, and again, you know, kind of putting a plug for Ironholm AI, our tool is used by startup founders. We're not Palantir that charges $100,000 a month. You know, we specifically did because we are the founders. So we understand the pains of startup companies and their founders. So we specifically designed a tool that can deliver value while still being very, very affordable for the founding companies. But again, going back to my advice, understand that not all data is created equal. Understand what can never leave your brain or, you know, physical piece of paper in front of you that can be shredded, or what's kind of less, you know, you know, that's kind of a generic information is that's not going to expose any of your proprietary data. And then act accordingly.

And I would say, I would advise you to ask a question, what would it take me to develop this business, this idea to an MLP or some functional prototype? Do I have to hire a person? Do I have to hire a contractor? Or maybe I could use an AI as my multiplier, as my force enhancer.. And then you can start looking at not as a cost sink, but as an opportunity cost. What it creates opportunities.

So by spending— are you suggesting using it to help create business plans?

So I'm suggesting it again. Let me just— I'm suggesting you do— you use AI to develop your business, but you do it with the understanding of risks and understanding of opportunities. And to Armina's point, again, not to plug it too much, we specifically price our product to be on par with an average monthly subscription price that you get from, for instance, any big super scalers. We're not going to price you out of business. If you're already paying money to ask ChatGPT a question or Claude AI, our solution is not going to put you out of business. It's going to be on par with that, but it will allow you to develop your idea without, uh, with being more focused on your data and without the risk of losing, uh, your international.

So to Kim's question, who's the question that popped up on there, uh, and appreciate her being there, is if you want to help develop a business plan and you're going to put your thinking processes in there, using IronHome to help you do that, to keep that proprietary and then even, you know, overlaying things to help you run. But you can be running your financial models, you can be looking at your marketing plan, you can be doing all of that while making sure that your data is not escaping out into the Ethernet.

Precisely.

That's awesome.

You can ask those questions, you can ask business questions, you can ask the technical question because that's a— it's a general model. We have different tiers of models depending whether you're on your laptop or on your server.

And So, Angela, let's pull up how do you get in touch with Armina and Alex on here, and Ironhome as well. And we've got, uh, little QR codes there as usual. Um, this is such a fascinating, fascinating topic, and it's so broad and yet it's so specific. It's so difficult to talk about because it depends upon what circle you're in, and the majority of people we find aren't even using it to anywhere near its extent, or they're using it poorly Governance is such an exciting area. I know everybody would look at me and Armina and go, governance is exciting? But it is because we're like, oh, how are we going to solve this? What's going to be policy going to be like in this space? And how do we make sure that the humans succeed? And it doesn't become something that's out of Terminator 2 type scenario. So that's where you're looking in all this. But so you get to Ironhome down at the bottom. Armina's on LinkedIn, Alex is on LinkedIn. Uh, I, I, you know, I tend to find Armina and I are connected more, so I probably see more of what she's writing about. But super cool stuff at University of Nevada. Also amazing stuff going on at that little university down in, in the East Bay there, uh, Haas Business School. But, uh, we appreciate it. Um, Armina, uh, Alex, I appreciate you guys coming on so much today. Of course, we've got Candice and Jackie with the Women Talk Finance. And how do they find— we got a good show coming on today, I'm guessing. Yeah, new guest. You guys can talk about it. And I know they don't roll out right away, right?

Wait, there'll be one tomorrow and we are super excited about it. It's actually with Dr. Katherine Ann Paul Hennitz and it's one of our faves. Yeah. And it's so good. Actually, I got to preview and it's, it is on AI, but it's taking a little bit of a different lens. On how to change the narrative around AI a little bit, and that'll be coming out tomorrow.

Oh, I love it.

Thanks for promoting it, Tom. Armina, we'd love to have you on the show, actually.

Finance, behavioral finance, I can talk about.

And AI and just—

AI, all those things.

Yeah, all the things. We branch out a little bit too. And Tom, we would love to have you on.

Oh, good. I mean, you know—

Oh, he got a formal invite.

The girls, the girls girls have pictures of me, you know, when I'd be Hannah Montana. Or Candace, your aunt, your aunt just sent me a picture today where I was serious, you know, not serious black, but I had big, you know, long black— I wear the wig. I would go, I would do it just to get on the show. And this is, I think we got one more thing that this is how you get a hold of me and Candace and Jackie, uh, if you care to. Um, I've got a Substack out there. We've got other things going on, but we're really be, you know, the founder's advocate. And what that means is, from that standpoint of during the life cycle of founders, what do you go through? And you go through everything, everything that impacts your family, the business, what you're doing in government, and, and the various things that are going on there. So we have a lot of variety of subjects because it impacts everything. And, you know, I, I still hold out there, free enterprise really does at the end of the day save the world. And what you guys are doing, I mean, and Alex, you're really keeping that privacy part of that a big piece because there's an intrusion from governments, there's an intrusion from all sorts of different invaders. And IronHome is just fascinating. I use it, I recommend it. You know, we have other AI people that come on. This is so far out of scope of that. I got to tell you, you got to go out there, you got to take a look at it. And they're super helpful too when you don't understand what you're doing because Alex and I have sat on the phone and he's helped me walk through this stuff. So it's great. And Angela, of course, thanks for running the show today. Everyone, we'll see you next week. Have a great day. And Alex, Armin, thank you for joining.

Thank you.

Thank you so much for having us.

Thank you.

Thank you.