Blake Hall WARNS: "We're Losing Our Grip On Reality" | The AI CEO Podcast

Speaker A: We're stopping two North Koreans, um, a week from getting hired into companies in critical industries.

Speaker B: When we look at what's happening with AI we look at all the deep fakes, we look at the fact that you can generate faces today. Does it ever make you pause and think we're about to lose our grip on what's real?

Speaker A: Well, I think what is real has never been more important as the world is increasingly artificial. You know, like a mom gets a call where the voice emulation sounds like identical to their kid. Right. Um, they think they're kidnapped if they don't send the wire right now. You know, meanwhile, of course, nothing's happened.

Speaker B: What's going to happen with the companies that are just focused on incremental AI right now?

Speaker A: I'm terrified. You know, I feel like a lot of companies are going to be destroyed. If you're not infusing AI into everything that you do, I think you're, you're probably on your way to death.

Speaker B: Can you explain a little bit more of how that one person is doing what he's doing and change of mindset using of tools like in your organization or otherwise?

Speaker A: Yeah, sure, I think, um, from leading elite military missions to securing the identities of over 161 million people, our next guest knows exactly what it takes to protect high risk systems. He's an award winning entrepreneur, an inventor with 11 patents, and the co founder and CEO of ID Me, please welcome Harvard MBA and former Army Ranger, Blake Hall.

Speaker B: I've spent two decades guiding CEOs and founders to scale and reimagine what's possible. Now the biggest shift of our lifetime is here. AI I'm Seem Alexander, founder of Disruptive AI, co chair of DC Startup and Tech Week, and the host of the AI CEO Podcast. If you're ready to learn how AI is transforming industries, creating new growth and rewriting the rules of business, this is the podcast for you. Because the truth is, the leaders who adapt now won't just survive, they'll define the future. Let's tap in. Welcome to today's episode of the AI CEO Podcast. My name is Seema Alexander and I am your host. And today we have a very special guest for season three. His name, uh, I'm sure people know it. It's Blake Hall. He's a co founder and CEO of ID Me. Uh, him and I actually met initially on the DC Startup and Tech Week circuit only, um, a few months ago personally. But I've heard your legacy in this area for so long, being part of the startup ecosystem and the AI ecosystem. And what the topic is for today is all around identity and AI and business and humanity. Right. And I could not think of a better person to have that conversation with. Uh, and for my audience, they know that this season is really about bringing the top 5% of leaders together who really understand what's happening, um, both from a business side, but a reality of globally. You know, these are real things, real trends, uh, and, uh, real opportunities also for entrepreneurship, for innovation, uh, and also for protection, um, for all of us. So, Blake, thank you so much for being on the podcast.

Speaker A: Thanks, Sima. Great to be here.

Speaker B: Yeah. So I have a hard hitting question for you. Let's do it from the beginning. Okay. So when we look at what's happening with AI, we look at all the deep fakes, we look at the fact that you can generate faces today. You look at all the video documentation. Does it ever make you pause and think we're about to lose our grip on what's real?

Speaker A: Well, I think, uh, what is real has never been more important as the world is increasingly artificial and we are in a dog fight every day with nation state actors and criminal organizations that are trying to use those technologies to defraud the federal government to commit identity theft. We had a clip that aired on 60 Minutes 70 several months ago where, uh, a member of a Chinese criminal organization was targeting a random individual in Pennsylvania using deep fakes to assume, uh, his face to try to commit tax refund fraud. We're stopping two North Koreans a week from getting hired into companies in critical industries. We're going to publish a paper on that, uh, with Vanderbilt University's Institute of National Security. So it's not even just the sort of general concern that people have with deepfakes and knowing what's real versus what's fake. We are very much right in the thick of combat in cybersecurity against our enemies who want to actively harm us. And so, um, we take that very seriously. Uh, my. Why is protecting and keeping people safe and giving them the tools to keep themselves safe? Uh, but it has never been harder, uh, to do that than it is than it is right now.

Speaker B: Okay, so when we think about most people, when they think about deep fakes, when they're thinking about clones and all this things, they're thinking about the influencer side, the things that they see digitally. And what you're talking about is so much deeper and so much real and behind the scenes that people don't understand what's happening. Right. And those are the real implications. Actually, I had, uh, A couple friends who work a pretty high level of security in government. And they would always be like, seema stop putting pictures of your kids online. Know, and they're like there's a lot happening that they can't say that's right. You know, and, you know, and I think there's two pronged to this. I think there is the reality of business and privacy and, and sort of uh, you know, being able to patent things and all the things that a business person thinks about. And then there's the back end stuff. So when you think about your company ID me in the age of the AI era, um, does identity still belong to an individual?

Speaker A: Of course it does, yeah. And I think identity is going to be just this inherent part of the new fabric of the agentic web. But uh, there's a lot that needs to get sorted out. Like what you were just talking about. If our most powerful institutions have difficulty defending themselves where we are fighting every day, then how much more like a mom after dropping her kid off at school gets a call a few hours later where the voice emulation sounds like identical to their kid. Um, they think they're kidnapped. If they don't send the wire right now, the kid will get harmed. And meanwhile, of course nothing's happened. Um, so there's this dual mandate to both protect our institutions and also to give people the ability to know when you get that phone call, um, who's on the other side of the line. And um, humans adapt when you have that threat. I view it as a great opportunity because there is a ton of fraud and identity theft lurking underneath the surface. It's now become a crisis.

Speaker B: Mhm.

Speaker A: And I think that crisis is going to force digital uh, identity into the forefront where we have to have a strategy that keeps people safe, keeps our country safe. Um, and once that's there, uh, it's going to make life different in the same way that before Visa. I'm sure people paying with cash probably couldn't imagine a world where you could just give a card anywhere you go and have payments done right away in a very simple manner. Um, that's the opportunity that we have as an organization to influence my kids life and their kids life for generations. And so it's very cool to be at this moment in history where there's uh, great purpose to what we do and there's also great opportunity to make life better than it is today by taking friction out of it.

Speaker B: Yeah, I want to like, because I. You talked about a lot of different things here, but the children part I want to just mention one thing because I got some advice and just to take your, get your take on it, that they say always have a safe word at home. And I don't even know how to say it out loud because it was sort of told me. They're like, go in the woods in a random place with your family and have a safe word. And so God forbid you get one of those phone calls, you ask the question.

Speaker A: Yep.

Speaker B: And so if that person on the other end cannot answer it, you know it's fake. And do we need more of these strategies now that all of this stuff is not. It's so real. And this is the weakest the technology's ever been.

Speaker A: Like, uh, that's right. I think, um, you definitely want contingency plans to keep yourself safe. That is a very important part, uh, of protecting yourself. What we want to do as an organization though, is when you get that phone call, we want you to have context for that call is originating from. And if you can tie phone numbers back to people and organizations, and if ultimately we know that that's a phone number and maybe the call's originating from a device that's overseas, uh, then that context would really be helpful because what we're really talking about is context to know is this the real situation that I'm dealing with or is it a false one that's been created by bad guys? And so in the short term, you should definitely have things like safe words. All my kids have credit cards actually, because, um, the credit bureaus, the first time they see a Social Security number, they join it to whatever name and date of birth went on the credit application. And so what bad guys will do is they'll file, um, credit history, like credit applications with a child's Social Security number. That Social Security number then gets bound to a made up name and date of birth. They actually create this online Persona. They commit identity theft and fraud. And when the kid turns 18, they find out their Social Security number has been tied to all sorts of fraud. So I went ahead and registered, uh, their Social Security number to their real name and date of birth first in order to preempt that. And those are sort of tagged.

Speaker B: At what age can you do that?

Speaker A: Oh my God. They were like, uh, let's see, seven, five and two or something. I don't know, maybe I got through a loophole in the bank or whatever.

Speaker B: Good for you. That's amazing.

Speaker A: Yeah, yeah. Um, but it's really important. But I think part of what we're doing is so much more Scalable than what maybe the most proficient people would do to take care of their family. Because most people aren't going to have passphrases and things like that to keep themselves safe. And that's what we work on every day, is to figure out how can we create a system that keeps people safe and really reduces the amount of work so that, you know, your, your elderly parents or whoever you know needs to be protected. They don't have to, like, call you to be safe from scams. You can actually have a system that is like a sentry that keeps him, uh, safe even when you're not around.

Speaker B: Yeah. So let's talk about sort of where your mission started.

Speaker A: Sure.

Speaker B: Right. And you know, you went to Vanderbilt University. I know you were an rotc and you served our country, and I really want to thank you humbly for this. Right. And you were on the front lines and, you know, that passion then brought you to ide me. That was actually a different focus initially with the veterans. So maybe if you could share a little bit of context on, you know, you being in combat and what that, you know, because I can hear it. It's, it's really interesting sitting across from me because you're, like, piercing in my eyes. But, uh, there's always something about people in military. They're just so disciplined and like, it just. I love it because I'm like, you're so present. Right. And I think you learn a lot of that through the experiences that you had to, you know, grow up in, in a lot of ways and build that leadership arm in ways that a lot of us didn't have to. So maybe if you could just share perspective on sort of some of your experiences there and then how it led into ID dub me. But prior to ID dub me. Sure.

Speaker A: Um, I grew up around amazing men. I'm a third generation soldier. My grandfather, uh, fought in combat for 66 months, uh, from World War II, Korea, two tours of Vietnam, uh, was a war hero, uh, in St. Etienne, where he crawled out in front of three German machine gun nests to save his battalion from being overrun by a Nazi counter assault or, uh, counterattack, rather. Um, and my dad served for 30 years as a colonel. So when I went to Vanderbilt, it was PeaceTime, it was 2000. Just a way to pay for school. Uh, but I felt like I needed to serve the country just because of what I'd grown up around. Um, 9, 11 happened, changed my life, changed a lot of people's lives, changed my life. And that's when I knew I wanted to be. I was going to go to war, uh, when I commissioned as an officer and I figured if you're going to go to war, you might as well play with the varsity. So I knew I wanted to branch infantry, I knew I wanted to be an army Ranger, um, did that, had a rifle platoon for five months. Got picked up for recon, uh, pretty early, um, had a boss who changed my life. And uh, as we went into uh, combat In Iraq in 2006 in Mosul, as we were staged in Kuwait, some three letter agencies came in and said, forget everything that you guys have learned about precision engagement, uh, and reconnaissance. You're going to run kill capture missions, 20 minute string, day or night. When, uh, you get an alert, you've got to be out of the wire going after whoever we were targeting. So mostly Al Qaeda vehicle bomb networks, but suicide bomb coordinators, financiers, uh, you name it, some uh, of the cuds proxies that the Iranians had going. Um, and like the second or third day, these NSA guys, uh, that were attached to my platoon that specialized in signal intelligence, they were like, hey, sir, if you don't understand, uh, graph databases, weakest tower theory, you know, some of these are the things we're not going to be very good at our jobs. And I said, well, um, fears I need to learn about graph databases and weakest tower theory and everything. And uh, that was my introduction to digital identity.

Speaker B: Uh, can I ask something real quick?

Speaker A: Of course.

Speaker B: Before we go into all things I.D.

Speaker A: yeah, of course.

Speaker B: If AI was a thing, then sure. When you were in combat, what would it have changed?

Speaker A: Oh my goodness. Uh, well, what's amazing is how quickly AI can process data. So when we would capture a high value target, we would do something called sensitive site exploitation. And so all the digital media and everything that we would pull off of a target site. And then when you be like, all right, this was like an Al Qaeda m Emir, a prince in their jargon. When are we going to get the distilled intelligence from that hit? They were like, well there's like a nine month backlog and we only have so many Arab linguists who are like, God, seriously, we're going to get the intelligence after we're already back back home. Um, now you could just download all that intelligence. You could have AI process all of it, do semantic, you know, the translation is built in. Um, and you could do a follow on raid immediately coming out of that first one. Uh, so that will completely change war fighting.

Speaker B: Yeah, I mean, and that's what's changing every industry, every workflow, everything.

Speaker A: Absolutely right.

Speaker B: Because of the way intelligence can flow today, for sure. So. And, um, we're going to come back to that as we talk about your own organization, some of the reorganization of. Of intelligence that you're having here. But then let's, let's get into it. Right. So you, uh, you finished serving, right. And did you all of a sudden have an idea of like, I want to be an entrepreneur? There's a lot of vets that become entrepreneurs. You know, I don't know if it's just to figure out the next, because entrepreneurship is very purposeful, like, you know, and when you find that right balance, it's a beautiful thing. So. But it's also a lot of pivoting in the mix to get there. So tell me, tell us how that happened. And then, like, what was the initial mission of what you were building and where you are now?

Speaker A: Purpose is the key word. I went to Harvard Business School because back to my boss, he wouldn't write me a letter of recommendation unless I applied to Harvard Business, so. Changed my life in another way.

Speaker B: Yeah.

Speaker A: And, um, when I was at hbs, I just was looking for the same sense of purpose I felt when I was wearing the uniform. I felt like I'd lost a limb in a part of myself. Um, I did a summer at McKinsey. I love the people at McKinsey, but I just remember there was a Wednesday night I was looking at this excel spreadsheets, like 7 o'. Clock. And I looked over at one of the other guys and was like 18 months ago, I was leading a small team hunting a vehicle bomb network that had killed thousands of people. Um, I was like, I don't think I can do this. Um, so I don't know if it's kind of war puts lightning in your bones in a way, with the adrenaline and getting, uh, accustomed to maybe a little adrenaline. Sure. And I think, uh, everyone who's high performing is an addict in some way. And so it's just about making your addiction a one that helps people. I think for me. And so it was a combination of purpose, and I'd grown very accustomed to tolerating extreme risk. And I figured if I could lead a team to great success with folks shooting at me, how hard could it be to start a business? Now, that cockiness has since been tempered

Speaker B: with deeply humbling moments, if I'm being honest.

Speaker A: Uh, that's right, Absolutely. As you and I both well know, like, sometimes you need the ignorance of youth to take the jump because if you actually knew how hard it was, you just would never do it.

Speaker B: Yeah, 100%. So then you went from defending sort of this physical world to now looking at it and building sort of digital borders. Right?

Speaker A: Yeah.

Speaker B: So. So, and then ID me didn't start off as ID me.

Speaker A: Nope.

Speaker B: Yeah. So can you share more on the. Cause it was a focus on vets initially.

Speaker A: Yeah, initially, like, we wanted to create more. More of, like a marketplace that was built around, um, veterans. And I looked at how Facebook started, you know, just for students. It was a closed community. Um, and, uh, and there were a few things we needed to get distribution from USA and Military.com in my mind, in order to get to network effects with supply and demand side. But along the way, I noticed that we were spending most of our time actually working on authentication and identity, that if the only folks who are allowed to sign up were members of the military, all of our time was actually spent on the signup and the identity confirmation, uh, part of the flow. So USA and Military.com, it turns out, moved very slow. It would have been like two to three years to run a pilot, and that just wasn't fast enough. So I pivoted more to what living, uh, social was, uh, doing here locally in dc, which was. Well, what if I did, like, daily deals and targeted brands that already had special programs and discounts for the military? That way I wouldn't have to worry about the supply side. It's already there. I just need to get the demand aside from the military families. And I think I can do that if I can just kind of corral all these different benefits.

Speaker B: And what was the problem you were trying to solve at that first initial point?

Speaker A: Well, so military folks move quite a lot, and Craigslist had a huge trust problem. Problem. Um, all sorts of, uh, violence and crime and, like, there's all sorts of, you know, stories about this that it's still a problem. And, um, and so in a lot of these military communities, you. They wouldn't even have a Craigslist that was built for their city. And they move at, like, you know, two, two and a half times the rate of the typical American family. And so I thought that could be like, a great. If we're playing Risk, this is like my Australia to get going.

Speaker B: Yeah.

Speaker A: Um, and this is kind of how Sam Walton starts, you know, Walmart, uh, before it becomes this, like, amazing logistical, um, company that can move prices much lower than other, uh, companies can because of their economies of scale. He would target markets that were big Enough for one, but not for two. It's classic game theory. As we got going, uh, the feedback from the brands initially was, you guys don't have any users, so why would I want Marketplace?

Speaker B: It's two ways. Remember that?

Speaker A: It's still two ways, even if you're just trying to solve the demand side, if you don't have any users. But the feedback was two things. Um, it was. Your ability to confirm identity in real time is really interesting. And if you built that as a product that we could integrate into our own workflows and our own applications, we would buy it. And as I talked to more marketers and I was like, why are all these programs offline? Like, it's the 1950s. If you want to buy tickets for the Nats game that were discounted or get an E Learning voucher from Microsoft, you had to show up in person. And the consistent response was, we don't have a way to confirm identity or military service online. Uh, so we can't keep it safe. And that's why we confined distribution to brick and mortar channels. And I said, and if you had a way to keep it safe? And the consistent response was, we would bring it online.

Speaker B: So this is the whole time where the Internet, like identity theft happened. After the Internet post Internet, you solved for that with this problem. But it was also interesting because you're like, that's not where we started. And this became an accidental product in a lot of ways or much a more intentional product. And that's what happens.

Speaker A: Well, in some ways, I guess it's a more niche form of identity theft. It's stolen valor. It's this thing that, uh, on any, uh, dating site at any given time, or all the dating sites, rather, there's like 6 million Navy Seals, but in real life, there's only like 60,000.

Speaker B: I didn't know.

Speaker A: It's the catfishing and everything that goes along with it.

Speaker B: It's funny.

Speaker A: And so the enemy has always been fraud. That if there was no fraud, then all the things could be online and you wouldn't need to manage hundreds of logins. You could just assert who you were and get access if everyone told the truth. But stolen valor and things like that, um, I dealt with. So the marketers then said, it's not just military, it's all these other segments. It's students, it's teachers, it's first responders. And that's when I was like, oh, my God. It's like PayPal for everything. It's PayPal for identity is really the opportunity um, and then I'd run into like, you know, executives at different technology companies, and they're like, wait, does this mean I won't be able to use my student id? You know, when I. When I go out and shop? And I was like, you're. You're like 30 and you haven't been in school for eight years. And they're like, yeah, but it's an ego boost for me because when I

Speaker B: show my student id, you think I'm still that young. Are you serious? I thought they were going to say, at least I get my discount.

Speaker A: No, it was all about just the vanity of not being, like, for their age.

Speaker B: That is so funny.

Speaker A: Yeah, so funny.

Speaker B: Okay, so then you move from retail into more. Into federal governments and other things. What was the pivot there? And then what's the sole mission right now by dw? So let's get into that.

Speaker A: Yeah. Um, so. So the focus was still on military families. And it's to borrow a line from Visa, you know, where do military families want to be? And of course, they want to log into Veterans Affairs. And so that was always kind of

Speaker B: like access to their benefits and all the things that they deserve, right? Yep.

Speaker A: And that's, that's how VIS started, actually, out of bank of America, started in Fresno, California. And so they didn't focus on everyone everywhere, they focused on everyone in Fresno. And then where do Californians want to be? Probably in California. And then branched out and it took three decades before Visa gets to everywhere you want to be in the US and then mastercard merges with Eurocard to form a global network. So you want to be really deliberate about if you're trying to put identity in the hands of people and not corporations, which is really like, philosophically what we're doing, functionally, what we're doing is we make it easier to log in and prove identity. Emotionally, it's super stressful to manage hundreds of logins. And it can be very disturbing when you know you're entitled to something and you can't prove who you are.

Speaker B: Mhm.

Speaker A: And philosophically, the way the world works is it's controlled by data brokers and corporations that divorce you from your own data. And so philosophically, what we were trying to do is we're just building a digital twin to a physical wallet to put SEMA back in control. We don't sell data, we sell trust. You control your data.

Speaker B: So digital twin is a word that people are finally hearing in a more broader sense, not everybody, but it's in the world of AI right. So you were using AI before all the hype of it, after 2022. And I say hype. AI has been around 75 years. Right? The concept of it to where it is today, sure. You, uh, guys were using advanced machine learning and computer, all the things beforehand. So maybe if you could just, uh. So where you end up becoming is focusing on becoming the single login, uh, for identity, the PayPal for identity.

Speaker A: That's right.

Speaker B: And I loved hearing what you're saying, Blake, because that's not what you started. Right. And a lot of people, when they start, because, uh, I'm part of the tech ecosystem and the emerging tech ecosystem, and I always hear, well, we're going to be the Uber of this. And from day one, right. And it's like you haven't even built for that traction, and you had to learn to get to even being able to position yourself in that way. It's like what I've heard, like, from

Speaker A: what you just shared, my superpower is learning. And I just treat building a business like the scientific method. Um, if you have hypotheses and then you test them, you'll figure out what's valid and what's not, and then you adapt. And you do that enough and eventually you'll be able to separate the symptom from the real problem. I was just trying to help the community that I came from, so I was addressing that. But as I kept tracing symptoms back down to the problem, that's the real discovery. When you talk to marketers, they're like, it's not just military. We have all these bots that attack us. We have fraud and chargeback issues. We can't segment truck drivers at gas stations, interior decorators at Pottery Barn, podiatrists at Brooks, and you're like, oh, my God. So the learning, the very cool part, Robert Frost says this quote, no surprise for the writer, no surprise for the reader. And I love that in the process of exploring some topic and going deep, you learn truths that you would not have known as a newbie to this space. Um, and when you come from outside the industry and you don't have all this sort of preconceived wisdom, you can reject all of it and say, actually, like, no. Like, you're thinking about identity from the point of view of like a bank, or you're thinking about from the point of view of the government. That's not how people think about it. People think about it as theirs because it is theirs. And so then to be an advocate for that and to like, learn and to figure out how different organizations and different industries think about identity in a way that's important to them. It's been very, very cool. It is just constant learning.

Speaker B: Well, that consultant mindset, even though you only have one year, McKinsey hasn't. I'm just kidding.

Speaker A: Yeah. I would say between that and the military was the same way. This concept of after action reviews after every mission, and we ran over 450 of them. And we would say, what went well, what didn't go well. We want to sustain the things that went well. The things that didn't go well. We were very honest about talking about that every time. And over the course of, like, all those reps, we eventually became incredibly proficient at our task.

Speaker B: Yeah. So let me ask you. You again, left the military, started a company, had a mission, then it became tech infrastructure. But you're not a technologist. Neither am I, by the way.

Speaker A: Not by background.

Speaker B: Company, too. So, for me, my question goes to. We're in this, the biggest technological shift of our lifetime, and there's so many CEOs and business leaders that are looking at where we are with AI right now. Um, and I keep comparing it. I know you mentioned this is bigger than the Industrial Revolution. I mentioned this is 10x bigger than the Internet. Uh, most people don't see that. They're seeing this as features and capabilities, not as a. I don't even call it a transformational shift. I call it a rewiring.

Speaker A: Sure.

Speaker B: Right. Uh, and so you were able to, even with the earlier stages, advanced stages of AI back then, be able to put these Lego pieces together to build something that's never been built before. Right. And now we're at a place where we have all these Lego pieces and people just don't understand how to put them together to innovate and to do all the things. So my first question to you is, how did you do that? Like, you know, what was the genius behind figuring those pieces out when you weren't a technologist? And then we'll do a part B of how should CEOs be thinking about it today?

Speaker A: There's so many layers to this, but, um, maybe the heart of leadership is credibility. And So I led 45 men in a platoon when I was 23 years old. And what I obsessed about every day before taking over that platoon was, how can I make myself worthy to lead men who are often 10 years older than me and have been doing it longer in a way that they would look at me as their leader? And so I figured out, you know, Ranger tab, airborne. I got my expert infantry men's badge with them, which actually has a 10% go rate. It's much lower than Ranger school in terms of the pass rate. And so I had to earn it on the merits. And when I would talk to them, I'd say, look, I might not be the best shooter, I might not be the fastest, I might not be the best in combatives, but I'll be damned if I'm not top five in all three.

Speaker B: Mhm.

Speaker A: And that's how you know I'm the leader of the pack, because I am better than you across the board. And that's my belief in like, if you want to lead the pack, you have to really earn it. And I think as a CEO, there are four primary disciplines. Brian Armstrong at Coinbase wrote a great post about this. Like what the archetypes of great CEOs, there's go to market and sales and marketing CEOs like Marc Benioff, there's operational CEOs like Travis Clinick, there's product and design CEOs like Brian Chesky, and then there are um, technical CEOs of course, like Mark Zuckerberg and Brian Armstrong himself. And so you can come from any one of those backgrounds and be extraordinarily successful as long as you're self aware and you know how to build a team around you that's strong where you're weak. At the same time, if you want to become a great CEO in a generational business, you have to become proficient, if not excellent in all of those disciplines as time goes on. And so this is where like you don't stop learning when you graduate from school, you know, and, and I think even more so when, when you're running a business that matters and you found something that can change people's lives, you really need to learn, um, as fast as you can. And so I think right now with AI, uh, the issue is that you have skills that you need to develop. The capacity of the tools is so far beyond human, uh, skills where engineers like Andrej Karpathy, who's led AI at Tesla, is like, I've never felt more behind in my life. So if he's never felt more behind in his life, how much more the rest of us. And I think if you understand what matters the most for your business, what skills do I need to develop that will help us solve the most critical challenges at this stage? And then you are relentless about focusing on whatever that thing is. Um, you can move your business forward and at that point it doesn't matter what your background is. In fact, in the military, the, uh, whole point of a general officer is that you lose your functional designation, you lose the notion of infantry or armor, uh, because your job is now to manage the entire organization. I think the only ones who are actually worthy of that are the ones who develop skill across all the fundamental domains and the emerging technologies that are needed.

Speaker B: I love that. So one of the things that you mentioned and I just want to share, it's like I get the developer at Tesla, he was like, I feel behind, right? Like I get it because he's looking at this probably a lot more critically and a lot more focused on whatever the next big thing he's working on. What I believe is that, um, I share this with you. Only 5% of business leaders across globally really understand what I actually is and what are the pieces underneath. And I call them Lego pieces. And I. Well, my belief is that if you put them in business layman terms, you can start to even fundamentally what machine learning is, what natural language processing is, what reasoning and intent recognition. It's like to be able to create intelligent automations, to be able to create things that are intelligent operating systems in your organization. Having those fundamentals will help shift your potential opportunities and other things at a business level. Because this is a business problem, this is not an IT problem.

Speaker A: That's right.

Speaker B: If you hand it off to an IT person, you're going to have other problems going forward and you will be left behind. Before we get back to the episode, I want to say something very important. We're in a three year window where businesses will either rewire or get left behind. And that is what I call a yes and moment. Yes, AI is going to disrupt how things work. You hear it episode after episode on this podcast. And the leaders who really understand it are going to be able to create massive advantages for their businesses, not just incremental change. But that's exactly why we created AI BusinessCon. It's not another AI conference. It's for non technical CEOs and business leaders who want to rethink how their businesses operate, reimagine what's possible and actually rewire how they compete. Now, if you want to be part of that 5% of leaders who understand this shift, who take advantage of this shift to take action, go to BusinessCon AI today and sign up. I'm really looking forward to seeing you there. Uh, what my point is, I think it's different depending on what type of CEO and what type of industry that you're in. And I say it because I feel like again, the Tesla example, uh, is like what is he working on? I'm just curious, what's the next big thing?

Speaker A: Is it. I don't know what Karpathy. I like follow a lot of his just like thoughts on, on AI and everything else. I don't, I don't know what he's like specifically focused on right now.

Speaker B: Yeah, yeah. I mean and no. So I just think there's some foundational fluency that if, if business leaders start to grasp that some of the stuff will start to make more sense. Sense, right. And how to actually leverage the technology is what?

Speaker A: Well I think the key thing is that um M. Technology doesn't exist for technology's sake. Technology exists to solve business problems in a particular domain. Whether that domain is healthcare or law. Um, transportation, you name is very rare to find an engineer who's capable of writing really good production code, who understands the domain. And what that means is that there's an incredible tax mhm on communications within most companies where you will have a business expert who knows it, like an accountant who could walk you through accrual and all these different concepts, or a lawyer who's really disciplined and understands litigation in a particular area and trial law and how that works. When they explain what they want in software, if the engineer doesn't know the actual domain of where the software solves problems for users, the way that we'll model it is this abstraction that wipes away concepts that end up being really critical. What is very cool about um, this current era is that you can have teams of one. So everyone now is a business problem solver. If you have a good design system, if you're in a company at scale, the role of designer, product manager, front end engineer, back end engineer, they're collapsing into one person. You might have a buffer with like a team at the infrastructure layer that can provide some resiliency. And as long as that person actually knows the domain, the business problems to be solved, that loss of fidelity, um, through all the telephone game handoffs is gone. You can move way faster and you actually end up with software that solves business problems the way that a business expert would want it designed. And that is really incredible. Um, we talk about maybe the people who are most positioned in terms of skill to take advantage of that. But that's very exciting.

Speaker B: Well, I think you're ahead of the game than most CEOs are. Right. So even having that as an example, being able to have one person to have all those different Skill sets. Now because of AI, can uh, you explain a little bit more of how that one person is doing what he's doing and change of mindset using of tools like in your organization or otherwise?

Speaker A: Yeah, sure, I think, um, so I think the folks who are most positioned to succeed are senior engineers, um, who know how to like harness AI and cloud code and everything else who are business problem solvers and they are quite rare. But those are the ones who, who are going to be able to take advantage of these tools first because they know the business problem they're solving, they're already very good at it. And now if they have like a design system that helps them create in a way that's aligned with the brand of the company that they're working within, they don't need anybody else to go develop software, um, that the challenge has really been about problem solving. Code is a task that has largely been eliminated. Now they still need to review it and things like that. The second cohort are going to be non technical business problem solvers who gain enough AI literacy and skill to be able to manage these tools. And to that would be me. There you go.

Speaker B: Category too.

Speaker A: And those are going to be the, I think everyone else who just isn't interested in solving business problems. They're not going to be that useful and maybe they were never that useful in the first place. Um, so once you have that you can use so many different tools. You can use Figma, uh, make for prototypes. Gemini 3 is incredible for image generation. Claude code, of course you've got cursor is like an interface that can draw upon multiple models that have different strengths in different areas. I was reading, uh, a lot of folks are using or one person actually that I follow was using uh, Claude code as a junior developer and then using Codex as a senior to review the code that Claude was producing. Um, this is now a single person who's using different tools and models and agents, um, where they can do it all themselves so they don't need to worry about when humans have meetings. You don't know when you're walking out of the room when everyone just heard. We would do back briefs in the military and sometimes it was shocking. It was like were we in the same room when we briefed this mission, not musical chairs.

Speaker B: What's the thing when you're like telling somebody something in their ear and it keeps going and then ends up being the wrong thing?

Speaker A: Yes, that's, it's these random like pointers and memory that you know and now like you can actually See it. So what people are calling like AI hallucinations and like, have you ever managed humans before? It's the same thing where like you say the same words and they're interpreted like 30 different ways. If there's 30 different people.

Speaker B: Yeah.

Speaker A: The ability to actually see and be like, no, no, no, that's not what I meant. Do this. And then to have feedback loops is actually quite incredible.

Speaker B: Yeah. No, I think everybody believes that the agents or anything, AI has to be 100% correct. And if it's not, it's hallucinating. And the problem is even data centers are 99.9% uptime. Like, you know, like there's always that point, 1%. And it obviously it's about context and training. Right. Because if you're not training and we have a lot of projects, we're like, man, the way the client is saying it. Ah, sometimes it's put in acronyms or sometimes it's done in this way. You have to train even the natural language to help support what the actual, what it means, um, how things are translating. Once you do that and there's feedback loops and it gets better and better and smarter. As long as there's intelligence behind it.

Speaker A: Yep.

Speaker B: I mean, it changes everything.

Speaker A: Yeah, yeah, it's cool.

Speaker B: So let's get back into, I think from a consumer or business side of things. I mean, agents, agentic systems, agents are going to be changing everything. Right. And um, I think agents are one, data is another. So actually there's something you had sent me which was a study by Duke University talking about the data brokers.

Speaker A: Right.

Speaker B: And, and this sort of like there's 10 data brokers that own a lot of personalized data of us. Like, you know, um, individual consumers from when you might go to a credit bureau to whatever. Right. Like, and in a sense they use that data for targeted advertisements and other things. And you can get a lot more detailed in terms of that. What I know in the age that we're living in is a lot of people and emerging tech startups are scraping data from every source that's possible. Some even with the ones that have terms and conditions that you shouldn't. But the reality is they're building data sets. AI. One of the biggest moats in the future is going to be unique data sets in your company to be able to create personalization and precision within experiences with their customers and potentially future business revenue streams.

Speaker A: Right.

Speaker B: But now going back to your whole model of like, data belongs to me. Like, but now data is belonging to everybody else. And because we have, as human as consumers, put all of our actions online, they're public and there's no regulations right now that really say that you can and cannot do unless there's like terms and conditions on a LinkedIn for example. Um, this is happening real time, so you have to ted big data brokers, but I promise you there are a lot of micro data brokers are all building right now. So what does that look like? What does that mean?

Speaker A: Well, I think there's the data side of it and then there's the identity side of it. I think from the data side of it, um, if your business depended on sort of dense public data sets, maybe legal filings or 10Ks, that business is gone.

Speaker B: Mhm.

Speaker A: Um, anything that's public, given the ability for AI to process it, translate it, synthesize insights, that's over. Proprietary data though, has never been more important. Um, if you truly have a proprietary data set and like network effects, AI is actually going to compound those advantages in the context of identity. What we'd already gone through before was um, the whole sort of process of trying to use data and knowledge of socials and questions, that was already over because of all the data breaches. Your Social Security number doesn't change, your name changes very slowly. Same thing with address. Once the toothpaste is out of the tube, you can't put it back in. Right now your data is for sale, I'm sure on the dark web for less than a quarter. Um, you can buy driver's licenses for folks, valid images for 15 or 20 bucks. There's a whole supply and demand economy that you can actually look at in terms of prices. Uh, NBC News has reported the Chinese military has the personal information of every single American adult. Uh, in fact the uh, Equifax hack, if you remember in like 2017, that was four Chinese military officers, they were indicted by DOJ, I think in 2019 or 2020. Um, so the game is now moved into different forms of media that are biometric in possession. Uh, is this actually your face or is that a deep fake? Um, is this actually your driver's license or is that an AI generated id? And um, in identity there's two techniques to confirm what's real. The first one is called verification, which is an empirical test in the world to say is that actually, ah, your face or is that a deep fake? Um, and then the second thing is validation. So after we say like, okay, uh, we think that this face is real, then validation would be like, well, whose face is this? Where could we match it back. Could you match my face to ID Me's website and go, oh, that's the CEO of ID Me. So you need both of those techniques to work. But the battleground is now around, um, possession of phone numbers and tenure to things. This is just like terrorists. Terrorists have very short, uh, tenure to things. They're swapping SIM cards and they're just doing enough to get over this trust hump. Um, same thing with the way identity thieves behave when they try to inject a phone number they control against personal information they've stolen. It's around, um, detecting deepfake ID cards and around faces. And then the validation step would be, even if they can maybe beat you on the ID check, if you can query back to the dmv. And the DMV goes, no, I never issued a driver's license with that combination before. So you could think about it, like, if you're walking through Times Square and, um, you saw, like, a handbag from a famous luxury brand, an expert could kind of look at it and say, like, oh, the stitching is not right. This is fake. Or they could say, let me get the, like, serial number and let go back, you know, to Prada or Gucci or whoever. And they'd be like, no, no, we never issued a bag with that serial number. So those are the kind of same techniques that are applied when you're trying to disambiguate what's fake from what's real. And, man, there is just an arms race on the, the techniques for.

Speaker B: Yeah, I was just at the 4A's marketing conference. So it's like a big conference for, uh, top agencies around the country, potentially around the world. And, uh, someone from Google was there specifically talking about content provenance, right? The CP, CPPA or CP2 or something. And that, you know, all the tagging on the back end of photos that they're going to start to put in or already started, right? For AI generated anything. The thing is, I mean, I literally did my headshots on AI. My cousin sent me hers, and I was like, that looks just like you, girl. Uh, but better.

Speaker A: There you go.

Speaker B: But, like, just like you, right? Where I'm like, you know, I literally did a post on my Instagram and I put 10 pictures. Like, guys, tell me which one's real, which one's not. And I got like, most people did not get the ones that are fake, right? They're like, you know, that looks just like you. But I was like, um, actually fake. Thank you. You know, but it's that Real, right? Like, I had somebody, um, you know, now there's 11 labs and all the avatars and all the things. And we know actually micro influencers are going to become avatars. That's the new. It's a big business model, if you follow Gary Vee and like, that whole model. It's like all the influencers of today are going to become AI avatars, and brands are going to be pumping money into these avatars. Um, and they're going to be very real. They're already real. You know, there's so. It's so. It's a crazy moment, Blake.

Speaker A: Like, that's why identity is going to be so foundational to everything. Like, it's. It's signing, it's authorizing agents, uh, with almost like limited powers of attorney to operate on your behalf. And what our content, as you mentioned, everything is going to need a signature to say who made this and, like, who authorized it. So if it's the government, if it's healthcare, if you want to get your health care records, like, you know, how do you know it was sema? And like, this is where at the root, we're certified against the federal standards to bind your identity to our wallet in a way that it's effectively treated like a digital driver's license. So we could say, well, here's the provenance, the credential, here's our signature. And then systems admins can authorize agents to act on your behalf because we have what we call the trust chain as an audit trail. It's almost like a certificate authority, where the root of the certificate authority is you, in this case, saying, nope, we authenticated Seema. She authorized, she wants to change your name, and she doesn't want to wait in line at all these different places. So the agent's just going to go ahead and do that for her. Uh, that only works if you have identity at the root. Um, I think for content, you'll need organizational signatures as well. If it comes from an organization like a media, um, outlet, uh, and from people. And if everything is signed, just like artists would sign paintings in the corner. Now you can separate trusted content from content that maybe isn't signed. And that alone would be a pretty significant signal of risk where algorithms could wait how they want to push things differently.

Speaker B: The thing is, there's no regulation. So any emerging tech company, like I talk to you about these headshot companies, they're not putting any digital, you know, any meta in the back end of any of this stuff. They're just not right. Because one they're not told they have to, to. They probably don't know. Like, you know, they're just doing what they. So I think even in that when they're consortiums for the big brands and they will follow some sort of protocol just to be good citizens potentially. Right. Um, there is the dark web as you mentioned. I just watched Startup by the way on Netflix and I'm like into the dark web now. But you know, what I'm saying is there's still gonna be gaps. Like, you know, and it's a really hard uh, unless it's a regulation, you have to do it. Like, and whenever that happens in this country, I don't know, I always say, hey, go follow the EU rules. Like if we want to do it, you know, let's follow something that is already in, um, if anything, because I

Speaker A: would disagree with that pretty strongly.

Speaker B: Okay, good.

Speaker A: I think there's a reason I want

Speaker B: you to disagree with it.

Speaker A: There's a reason I'm just such a, I'm a tech optimist and I believe in free markets that all the issues that you're seeing, like would you rather have, you know, a great technologist, ah, and innovator solve that through a new product or business model that doesn't exist yet, or would you rather have some government bureaucrat who's like never had a job in their life tell you what the rules are? Form follows function. And so what people just need to do is chill out. It's an issue. But I guarantee you there are entrepreneurs who are amazing who are coming up with solutions to these things because they're meaningful and that's what teams respond to. And if government regulates too soon, you kill it on the vine. And there's a reason that Europe is stagnating while America is soaring. And it's because freedom. If you just let it go, let it rip and uh, tolerate the short term turbulence, we will figure it out.

Speaker B: I love it. Okay, let's have this conversation because this is important. I agree. I'm uh, an emerging tech optimist. I, the reason I'm shifting everything right now to go heavy on AI and business is because I'm like, this is the most opportunistic moment but also the time that you have to, you have to take precaution because this technology is different. It's different and this is sort of that moment in time. And I talk about this a lot around like innovation has implications and not every, like, you know, when we looked at the social media timeframe and the tech bros and These rooms and you watch Social Dilemma and they're sitting there like their job as a product manager was just keeping people on social channels, right? So the dings, the notifications, the uh, dopamine hits, all the things, right? And now there's the guy who talks about doom scrolling today and he's like, I created that feature 15 years ago. If I knew it created all this anxiety and depression, all the suicide, all the things that happened, I would have never created it. Right? So my point being is there is a responsibility and innovation. I agree. Like, I don't agree that we need to be fully regulated, but things, the point was like, things like where I was like, if you really are saying that we need some sort of identification and pictures to be like, this is a deep fake and it's not part of our policy. The emerging tech folks are not going to do it. They just don't, they don't know. They're not paying attention. So that's what there's like, ah, are there the standards or are they not like, you know, and like that's what I mean.

Speaker A: Well, no, no, I think that's fair. There are definitely things where, um, there are perverse incentives that should be addressed. And I think if you get down to the incentives and you fix the incentives, that's the best way. And certainly there is room in that sense for targeted regulations. I would say we encounter an enormous amount of scam victims, um, that we are detecting the scam, protecting them from themselves, in this case, uh, an extraordinary amount, most of it from overseas. The way these victims get harvested is on social media websites and there's been some reporting about that. Um, and then the social media websites with lobbying and everything are able to like sweep it under the rug. But it's a huge issue. And when I look at things like that and I'm like, wait, you actually have an incentive to not stop it because the fraud rings are paying advertising fees and the folks who are hurt are individual Americans and the American government. That is terrible. And that is a breakdown in sort of the free market and incentives that should be addressed. And I think as long as it's targeted to things like that, um, what usually happens is very disturbing with like whenever I talk to folks on the Hill is that they are not experts and they want like sweeping like one size fits all policy. And I'm like, that's bullshit. That's the opposite of everything I learned as a military officer, which is define the problem specifically, make sure the solution solves the problem. And then, you know, and so right now, like a lot of things, even now that are being proposed as policy is like, put it all on the device and nobody can track you. Well, like 80% of the fraud we stop are scams. What if there's an elderly woman who like loses her life savings in her 401k plan and like in the client side model, she needs help. Me.

Speaker B: Yeah.

Speaker A: And everyone's like, sorry, privacy, we can't help you. That's terrible.

Speaker B: Mhm.

Speaker A: There's a reason that banks can see your financial transactions. So that consumers are not on the hook and liable for compromised credit cards and like $5,000 worth of charges. And yet in that model, that's exactly what would happen. And so when you're talking about like adult websites and kids and social media. Keep it on the phone.

Speaker B: Yeah.

Speaker A: When you're talking about financial risk, you need to bring it server side. You need to protect people so they're not liable and on the hook for tens of thousands of dollars when bad things happen. And there's just not that level of nuance because most people who are in politics don't know how to code. They don't even know how the Internet works.

Speaker B: They're not on stage too. That's what I'm saying. There's no level of fluency and they're making bigger decisions. So I will take back the whole, just be like eu. And that was a really good dynamic. I love that conversation. That makes sense. But I think it's the aligned incentives is big and who gets to who because then they create these big coalitions and they're supposed to create them. And I just, I never kind of follows through how it's supposed to be.

Speaker A: Totally. I think if you could kill the Citizens United Supreme Court decision and if you actually had experts who understood technology that were crafting policy, we might have a shot. But that is not the case. And so until that's the case, I think, uh, you want to keep the government out of it because we have to beat China and our adversaries here.

Speaker B: Yeah. There's a big arms race in the back end. A big AI race happening. Space race. A lot of things that people don't see in the back end.

Speaker A: That's right.

Speaker B: So I get that. So let's talk about agents for a second. And I go back to it. The reality is agents are going to be doing a lot of things on your behalf. You started that conversation earlier. It's going to become where agents and agents interact. And what I mean by that is, I know, like, even in, uh, chatgpt list some large language models, payment processing is going to occur. You're going to be able to purchase things, or in some ways you are. Right. And so there's all these transactions happening on your behalf. And to your point, identity. Now, it's not just about me as a human being or me as an individual. Now I have an agent transacting. So is there now an identity factor on those agents? Like, what are you guys thinking about? How does that going to work?

Speaker A: Yeah, there's got to be identity because the agent has to have authorization and a set of permissions to conduct activity on your behalf. So it's going to be very important from a security, uh, point of view of like, what should this agent have access to both within my organization and someone else's organization? Um, we protect many of the business programs for federal agencies like Social Security, Business Services online, IRS for Tax Professionals, for Business Services online, nih, and Health and Human Services for medical providers. I think something like a third of all the prescriptions in the country, like ID ME authenticates those doctors when they're prescribing narcotics and controlled substances. So, so there's, there's so much that goes into it. Where identity is going is going to be really critical. But the other part that is now available with AI is that you're not just automating workflows with agents. Um, and taking maybe, I think the cost per hour now is like $10 an hour if you had an agent like working flat out. And that's only.

Speaker B: You did a calculation. What, a compute calculator?

Speaker A: Yeah, yeah, I followed exactly.

Speaker B: It's so interesting.

Speaker A: It depends, it depends on what you're doing, but generally. And that's only going to drop over time.

Speaker B: Yeah.

Speaker A: Um, but the other thing that's happening is that you can understand why the agent made a decision right now in companies, there's what you know. Institutional knowledge is the knowledge about a company that's proprietary, that's locked in someone's head, and if they leave, you lose that institutional knowledge. What's now possible that provides compounding effects is you can see, like, what variables did that agent consider when it made a decision that resulted in a particular state? So as an example, like, if you're doing something with pricing and the VP of sales made an exception, there's something that happened, like some competitive dynamic, something about the nature of that customer that made them grant an exception in the old, like, way. You would just see that they had made the decision. You wouldn't know why. Now you Actually have the trace of like why that decision was made. And you can have a feedback loop to say was that the right call or is that the wrong call? And so now that you can actually get at data that was previously only in humans heads and have that refined and made, I mean that's just incredible because you can see why something happened, not just what happened.

Speaker B: Yeah. And that's why we're focused on building an intelligent operating system to be able to capture that for your own companies. Right. Because that knowledge transfer goes away very quickly. And so uh, it's a different world that we're living in. Again this conversation is very advanced than what people mostly know. And that's kind of where I see the biggest gap is fluency. Because this is where um, the beauty of being part of things like DC Startup and Tech Week and being part of uh, tech screenings for accelerators and stuff. I see where all the AI native companies are going right now. And what is your perspective? Actually I'd love your perspective on this is what's going to happen with the companies that are just focused on incremental AI right now.

Speaker A: Well, I can only speak for myself, but I'm terrified. I feel like we're doing okay in terms of how we're moving, but not fast enough given how much is at stake. Uh, I think there's going to be a great divide and a lot of companies are going to be destroyed. Um, this is like maybe Clayton Christensen, disruptive innovation. Some uh, are more insulated than others. But if you're not infusing AI into everything that you do, I think you're probably on your way to death. Mhm.

Speaker B: Yeah, there ah, there were two quotes. So when Mark Cuban. And um, it's not, it's Bill Gates. He had said the, when the Internet was going on, it was like in the next five years if people aren't focused on using the Internet, they're going to die at the vine. And then Mark Cuban said that in 2022 or 2023 of March and he was like, it's the same thing with AI, like this is it, right? And so like you start to see all these signals but not everybody's paying attention to when technology is truly crossing the chasm. And then there are all these open questions on still. I think people have now from three years ago when I started preaching about this stuff because I was exposed to one of the earliest agentic development environments before it was like a big mainstream thing and the company I was advising was consulting Google's agentic area. It was that early. And we got highlighted at Google IO. And that's when I knew when we were looking at the back end of this agentic, uh, system, I was like, this shit is. It's insane. It's insane, right? I was like, take things that are going to take six months and do them in minutes. How is that even possible? It's going to change the game.

Speaker A: With anything that's exponential. If you're waiting to be sort of in the middle of the bell curve, by the time you see it and know it's too late, it's already on top of you. And so you have to see the exponential early if you want to be able to capture it. You have to have conviction that this is. And that's why CEOs get paid and leaders get paid, is to see it before it's on top of you and to react. And with the magnitude of what's possible now, I mean, I'm already paranoid about everything. Because you just are. I think if you built something from scratch to where it is because you know how it's put together and how fragile, like, ultimately all this can be. Um, so, yeah, I know that this is massive and that if we don't go all in and have people who believe what I believe, that the company, uh, would be in trouble. And so a lot of it is just having the conviction to make that call, give people the tools and the ability, um, to come along for the ride. And then if folks just don't believe or they don't want to come, they can select a different tribe. And the market's a great teacher. And if I'm wrong, maybe I'm wrong, but I don't think so.

Speaker B: So tell me, what was your aha, oh, shit moment?

Speaker A: Uh, you know, I don't know that there was one. I would say what happened in November when Gemini 3 came out in Opus 4.5. I think that's a seminal moment in technology and computer science history. Um, that's where there was just a palpable shift in not only what was possible, but also the way that all the engineers I was talking to were talking about it, where it was like a reverse, um, of now it's like it's actually doing most of the generation. I'm just doing a little bit of editing, and it's no longer the handholding and, like, I don't know if this is really worth it.

Speaker B: Right.

Speaker A: And then the next models, which have just now released in the last week or two with codecs and 4.6, they're only getting better. Uh, so something happened in November, uh, where I already was pushing it and everything else, but this is the first time just across the board. I would say the majority of everyone who is on the technical staff. It was more coming up than going down.

Speaker B: Yeah. What about this past couple of weeks with Clawbot? Like, uh, I mean, yeah, uh, that's a whole different ballgame.

Speaker A: Sure.

Speaker B: Right. And it was one of the biggest shifts and uh, I feel like since ChatGPT and like, because I'm in the know of what is about to happen and where it's going, I'm like, this is a little closer to AGI. It's a little closer to this. This is what people are going to start to see AI as a system. And the creation of these operating systems becomes more and more real. What is. And you know, obviously it has its, uh, flaws. And what I always say is like, you know, if you're not a developer, don't use it. There's a lot of things and you need to use it on a separate system. It cannot be on your computer. But the fact that OpenAI snatched the founder already, sure. You know, he was further along than most people understand where this technology is headed. So what is your perspective on that technology?

Speaker A: Yeah, I think, well, the first wave is inference on all the pre training and then the reinforcement learning and really get inference. And what's happening now is memory, uh, and context is improving the context window. And if you think about memory itself, how does it work? Um, it's pretty cool. Think about your own relationships. There's the raw transcripts of all the conversations you've had, you know, with your kids and your friends and like everything else.

Speaker B: Right.

Speaker A: But usually how memory works is we like structure facts that are distilled from those conversations. So, you know, it might be like, oh, Seema likes Cabernet, you know, How'd you know that?

Speaker B: You did research. I know you did research.

Speaker A: Lucky guest, you know, or like cortados in the morning or like this person's like vegan and. And so spying on me.

Speaker B: I swear.

Speaker A: Yeah, so there's like a superstructure where you've got all the conversation, but then you're just kind of storing facts that kind of build a profile. Um, and I think that's really neat. And so the advances in terms of how memory works and how memory provides structured context without burning up, uh, m. Too much compute is really interesting.

Speaker B: Huge. I mean, it's, it's. I, I think context and content is everything. Memory just, it's going to make Everything so much stronger. Right. Like, like I'm, I'm blown away. But again, I keep thinking about, like, what part. Actually, this is a good question for you too. With the Internet cycle, I think we used to be at the DOS phase. I think we've crossed it. So where do you think we are? Comparing it to the Internet today?

Speaker A: Comparing what to the Internet?

Speaker B: Sorry, comparing where we. So I'm gonna ask a question again. So where, in terms of where AI is today compared to where we were with the Internet today, where do you think we are in that cycle?

Speaker A: Oh my God. I mean, I think, um, it's like the dawn of a new age. Honestly, that's only gonna get more profound over time. You know, this, this whole sort of like deterministic computing and the way it used to work versus now like neural networks and having machines that are able to reason and can be trained in all these different ways, and the data exhaust it gives off, it is to me way bigger than the Industrial revolution because as a platform, what you can build on top of it, just, uh, the problems that can be solved in every domain so much faster, is really incredible.

Speaker B: Okay. A, uh, shift and pivot. An interesting one. So Blake, you're friends with Elon Musk?

Speaker A: He's. Yeah, he's great. It's a crazy. Yeah, he's been really good to me.

Speaker B: Yeah. And, and I've done a lot of research. I've started reading his book and doing, you know, his, his mind is so maniacal. He focuses. So it's like mind blowing how he works. Right. And obviously with Grok, and he's been part of the AI cycle with OpenAI. And has he given you any insight? Has he been anybody who's been giving you any aha, uh, oh, shit moments? What does that look like?

Speaker A: I think, um, it's very inspiring, uh, just the way he thinks. And it's all about learning and uh, being an autodidact in terms of just falling in love with the problem and then going deep. And uh, he's definitely, I think, the best entrepreneur ever. Um, he's the smartest person I've ever talked to by far. It's not close to, uh, just a master of so many different domains. And uh, I think maybe what he gave to me was just reinforced what I already knew, but really motivated me to build technical skill in a way that I hadn't immersed, uh, myself in as fully. And the statement was, you can't lead what you don't know. And whoever leads a company needs to be one of the most skilled people at that thing, which is essentially back to what I told my platoon all those years ago. Um, and so for me I was like, fundamentally we're an engineering company, we're a data company. So uh, I need to be one of the most proficient at those disciplines and really immerse myself into database design and postgres and data models and things like that. Very quickly became a much better CEO for it. In fact, completely rebuilt product and engineering coming out the other side. I couldn't even try to speak for Elon on this stuff, but I would say his work for what he did to build that data center out in Memphis in like six or seven months or is unbelievable. From scratch, it's like fastest ever. Um, Jensen talks a lot about him too, and he's got the best seat at Nvidia. He's like, his teams just work so much faster and I think that ability to cut out waste in meetings and just get to the core of what is the actual problem, what's the most efficient, creative way to hit that problem and to go is incredible. But it all starts with mastering the heart of the domain. First principles, thinking.

Speaker B: And so, um, and he builds teams like nobody else. I mean the way their, their conviction he, I remember in this book he's like, I call people at 2, 3 in the morning, like it doesn't matter, like they have to be in back at work at 4 in the morning. Because we got that idea, we had that breakthrough. And not a lot of CEOs will do that to their people like, you know, and, but that's a type of hybrid of the, of the people like, you know, and so it's a culture he's built around him where that is just the way if we were to do these big, big things, you got to be on board. Do you see that? With what he's built?

Speaker A: Yeah, I think uh, you know, he uh, he's not shy about stating, you know, what they're up to. And uh, and his, his companies have deep purpose to them. You know, Tesla sustainable abundance and SpaceX extending consciousness. These are like really inspiring missions that people uh, can glom onto. I think one of the, what uh, I was curious about when I talked to him was how do you find the people to run these companies? You know, I know how hard it is to build one company to you know, multi billion dollar valuation to do like 10 of them.

Speaker B: Sorry, I was giving you flowers. Multi billion dollar evaluation.

Speaker A: Yeah, but he's got like 10, they're all worth, you know, hundreds of Billions. And uh, uh, he said, well, you know, I asked them about their career and then I have them describe a hard problem they solved. And he goes, and I asked for more detail and more detail. Uh, and he said no one can bullshit three layers deep. And so if you're skilled enough in the domain, you can sniff out the people who just were there when the problem was solved from the people who actually know how, like what the problem was in detail and how to solve it. Because those are the people who could explain like this is the problem. This might seem like an intuitive approach, but here's why the intuitive approach fails. And this is actually the way to hit that problem because you're missing like these other variables behind it. When you have somebody who talks in intimate detail, it's like anything else. It's craftsmanship. Mhm. You can just see it like an expert. They are talking at a level of fidelity and integrity that most people just could never hope to achieve. It's like NBA athletes when they're looking at folks footwork, they're going to be able to diagnose and distill things in colors and shades that none of the rest of us mortals see. And so when you can see that in somebody and the ability to recognize that across so many different disciplines I think is really remarkable, amazing.

Speaker B: And then you augment it with AI and holy shit, you got to be careful.

Speaker A: Still, there's still one of our engineers has a channel in Slack called Hal's Idiot Son, uh, where it's all of his back and forth with Claude. So.

Speaker B: Oh my God, so funny.

Speaker A: Yeah.

Speaker B: So we also live in a world of convergence. Right. And so one area that I, um, when I think of identity and I, um, I've done a lot of work in web 3 or just like really dug deep into it and I always felt like original use cases weren't really speaking to some of the problems that people were actually having. And eventually it's starting to catch up. But the reality is when you merge AI with blockchain and leisure, there is a sense of identity and it's real, it's digital and can't be taken away. Right. So are you guys, how are you thinking about blockchain in some of the solutioning that you're doing?

Speaker A: Yeah, I think. Well, you know, remember like blockchain is just, it's a, it's a database that can't store, you know, private data. Um, so it has limited use. Uh, now I think when you think about things like um, what are use cases where we're sort of uh, removing the possibility for censorship are very important. This is why Bitcoin is kind of the killer app. Because currency, it's okay if it's public. You want to prevent double spend and everything else. And if you want to get rid of around these governments that really have a problem with fiat currency, Bitcoin's an amazing solution to all those problems. I think when you start thinking about things like uh, the tokenization of finance, then web3 becomes really amazing with smart contracts and how shares and fractional shares get traded that exchanges don't need to shut down anymore. Um, and it's going to be controversial because so much of the blockchain community is sort of anti identity and is more almost this like decentralization everything.

Speaker B: Yeah.

Speaker A: But the reality is society is rooted in our social contract with the government and we have all these laws around know your customer and preventing the financing of terrorism. Uh, I don't see that lasting as sort of blockchain and Bitcoin and different financial tokens enter the mainstream because these industries are so regulated by the government itself. And so I think that um, we actually won this Citi Tech for integrity competition all the way back in 2018 or 2019. And all the blockchain entrepreneurs came up and said my solution doesn't work unless your solution works first. Um, and this entrepreneur, he was Argentinian and there was so much corruption where whoever is running the RFP would release the bids to their buddy and then that buddy would change the bid to be like a penny cheaper. And so the idea was to use blockchain to see uh, which bids had been opened after submission. But the only way that works is if the audit trail records who opened it. And if you see the contracting officer opened it and then two minutes later there was a revised bid that was submitted by this vendor. You would be like, maybe we should investigate the relationship between the contracting officer and that vendor. Because that doesn't seem like it's right. But if you don't have identity as part of the audit trail, blockchain doesn't work. And that's where the convergence of the two, uh, to your point, becomes really powerful. And so I think you won't see a lot of use cases tied to sensitive data because the nature of blockchain is to be public. But for finance man, I think the opportunities are really almost limitless in terms of. It's just weird that exchanges like Monday through Friday and that they close should be, should just be liquid and fluid anytime you want to trade.

Speaker B: Yeah. And that Word convergence is funny to me because we had Ted Leoncis, uh, as one of our keynotes at Disturbance Tech Week. And we were in the green room in the back end, in the back room. And he goes, so, Sima, what do you think the next 10 years are going to look like? And I was like, and our theme was Decade of Disruption. Like, Decade of disruption. And he goes, okay, well, uh, he's like, that's old. He goes, I used to say that AOL days come. No, no, he said, he said that's. That's not even. No, he's in the main street. What are you talking. I was like, all right. I came back and I said convergence. And why. I said it was. It was blockchain, it was, uh, it was obviously AI, but it's quantum, it's robotics. It's. Everything is happening in a lot of ways at once. And then. And I want to talk to you about quantum after. And he said, convergence is what I said during the Internet days. I was like, okay. I was like, what do you want me to say, Ted? And I remember him, like, after his talk, he literally came up to me and whispered. He goes, a Gentek. And I was like, I've been saying that for the last three years. Uh, but. And this goes back to the fact that I, you know, for him to say that to me when I've obviously been living this thing for the last several years was like, he's one of the 5% of business leaders that gets it. Most people don't.

Speaker A: Yeah.

Speaker B: Do you agree with that statement?

Speaker A: Yeah, I think the trend of humanity, um, because I'm economics and American history, by really focusing on the transition of an agrarian economy through the Industrial revolution. And so you're a nerd. I'm a nerd. Super nerd, of course, if it's not already obvious, huge nerd. Um, but the trend of economic history is fewer people doing more. Uh, before the industrial revolution, over 90% of jobs were in the agriculture sector, uh, because making food was really hard. And as soon as you could substitute machines and capital for that, uh, we need far fewer farmers and people to feed all of us. And so the nature of work just completely changed. You had society urbanized, uh, knowledge work, cities emerged. And now this sort of domain of software isn't only open to maybe 60 or 70,000 software developers. Anyone who cares about, um, solving problems can master those skills where the barrier to entry is far lower. And so I think there's going to be massive deflationary, uh, effects over time, especially as robotics hit So I think the cost of goods and prices are going to plummet, and I think there's going to be an era of abundance that we just have not seen before. Um, what the world looks like 100 years from now I think is about as impossible to predict as farmers who are like, wait, what's going on with this assembly line thing? And steam engine and combustion engine? Um, but, uh, I think it's going to be incredible in terms of productivity and that you only need one or two people to solve a hard problem that typically would have taken a large group of humans years to solve. Just from first principles, you can already know, um, productivity alpha and economics is the most reliable predictor of wealth because it's not like how much GDP is growing, it's GDP per capita. So the standard of living is going to go way up. Now it's the smart people that are getting super rich. It's not just the rich, it's the really smart people who are going to get rich and get very rich. How that is distributed, the bigger divide

Speaker B: is going to occur.

Speaker A: It's all on intelligence. It's never been more important to be smart, uh, and to be a business solver because they're going to learn the skills the fastest. Now, how we deal with that in terms of how we distribute money will be its own issue. And again, it's not inspiring to see the background of most folks in policy and government are not necessarily equipped for this in terms of economics or technology. But, um, it's a good problem to have. Whenever you have a much bigger pie, figuring out how to divide it, uh, hopefully shouldn't be a hard problem. It's when you have a small pie and figuring out how to divide it up is where things get really tense.

Speaker B: M. Do you believe we're going to have universal basic income?

Speaker A: I think it's de facto coming, yeah.

Speaker B: Do you, uh, trust AI?

Speaker A: Well, that's a big question. I think I would turn it around the other way. I would say regardless of what tool you're using, including AI, it is your responsibility to make sure that whatever your output is, you're accountable for it. And so, um, what I would tell my own team is like, I don't really care about what your belief system is according to AI or not, but if you put your name on it and want to ship it to production, you own it and you're accountable for it. And so individual responsibility is still incredibly important. You can't sort of, um, delegate your thinking to a machine, uh, and experts will know where to trust it and they'll also know where it's like no, no, no, you have that backwards. And they'll teach it.

Speaker B: Yeah. And so you are doing some reorganization in terms of roles. Another thing is leveraging AI in your own organization. Can you speak to, to how you're going about that to help support other CEOs? Think through it.

Speaker A: Well, I think back to fewer people doing more. The role of designer, product manager, engineer. It's all front end, back end. It's all collapsing into one business problem solver. And so the question then becomes like what systems do you need to put around folks to make sure that they are ready to build an AI first way? Um, we have a lot of regulations we have to look at like Fedramp and everything else. So it's helpful like Google, Gemini, um, and things like that are already inside of the Fedramp cloud, which is really helpful. Not all the new tools in our industry, uh, because the regulation are available for use. Um, and uh, you need design systems. Ah. So that as you're prototyping using Figma and things like that with figma, make that whatever you're prototyping is being produced in accordance with the brand. Um, yeah. And then Claude through Anthropic and everything else too like generating the code and, and giving people all the different training and tools they need in order to take advantage of that. But you can move right now into a prototype, have a conversation with engineering the same day and then move to generating code and pushing to production in just like a day or two. And I think even that division as the models get bigger, I do think you're going to have teams of one. So in terms of designing AI, first the question is like what's the right sort of scope to give to, you know, teams of one or two. Um, and Elon as usual is way in front of everybody. I think Grok, uh, for like Xai redid all their public APIs. It was one person. The entire mobile team is three. So it's just crazy. Amazing.

Speaker B: It is amazing. So do you have one enterprise large language model? It sounds like you use a lot of different.

Speaker A: We have a lot. They're good at different things. Yeah.

Speaker B: Okay. Which is not uh, normally again what people are doing is choosing one because that's the way it's always been done and not like, you know, I actually just put a post out of like I use these four large language models. They're not interconnected, they're used for different reasons. They're trained for different purposes. Right. Between Manus And Claude and like, you know, and others. So again it goes back to fluency. People don't understand it. So the way that IT folks are pushing their CEOs, we need one enterprise large language model wonder Microsoft. So let's, let's focus on that. And what do you think that approach is going to do to companies?

Speaker A: Yeah, well you're going to be limited because they're kind of like humans. Claude right now. And Anthropic, clearly the best at generating code. Um, Gemini is really good at image generation prototyping with figma. So uh, once you understand again, if you're thinking from first principles, this is the problem I have to solve. What's the best tool to solve that problem? You're not going to have one across the board. I mean they wish um, they've all made different decisions. Ah, ChatGPT is focused more on consumer, um, certainly they're trying to catch up with Codex, but that was a choice Anthropic focused all on business and specifically on development with Boris the other engineer who I think built Claude code as a side project.

Speaker B: Side project. I love that.

Speaker A: Pretty good side project worth hundreds of billions. But um, you know, so, so I think those decisions like matter and if uh, you know, and if you're not sort of uh, an unsophisticated observer but really understand what these tools are capable of doing, you would definitely not have a one size fits all approach.

Speaker B: Yeah, again, goes back to fluency. So I have one last question before we get into rapid fire. If you were an entrepreneur today, what would you start?

Speaker A: Um. Wow, that's a great question. Knowing what, you know, given how much I've leaned into the nerd side, uh, you know, and how much I miss my um, infantry days, I would, I would start something in the experiences, uh, section because I think that when smart people have incredible wealth, uh, what do rich people do? They like to travel and do like really kind of um, extravagant things. I think that's where a lot of the world is going to go to that as people become a lot more wealthy and uh, what's real and experiences become more important, I would start looking ahead to say, look at how rich people behave and how they travel. And I'd start businesses that more cater to how the world is going to change when a lot of wealth uh, is more democratized.

Speaker B: Okay, so rapid fire, you ready for it? All right. AI optimist or AI realist?

Speaker A: Optimist for sure.

Speaker B: Passwords dead or alive in 10 years, especially after quantum comes, uh, mostly debt. Biometrics Inevitable or risky?

Speaker A: Inevitable.

Speaker B: Regulation first or innovation first?

Speaker A: Innovation.

Speaker B: Human in a loop forever. No AI will create more fraud or stop more fraud.

Speaker A: Ooh. Well, I've got a stake in that one. It's got to be stopped. It's founder of ID Me. Like we're betting on the Home M team here. Oh, my God.

Speaker B: Is anonymity. Is it dead?

Speaker A: Yes.

Speaker B: Um, in one word, the future of identity is

Speaker A: trust.

Speaker B: When your kids grow up in an AI world, which they are and which mine are too, what do you hope being verified means for them?

Speaker A: I hope it takes friction out of their life. And everywhere I see people waiting in lines. I see people wasting time, and time is the most valuable currency. And so if they can live a life where they're safe from identity theft and fraud and they're able to get access to what they need, and they don't even realize, um, what it was like before with these long lines, I think that would be an amazing gift to them.

Speaker B: Wow. That was a perfect way to end our podcast episode together. Blake, what a pleasure to have you. Thank you for a very dynamic conversation. I love the back and forth, um, and perspectives, and this is the real conversation I want to continue to have so other business leaders really understand what's happening. And so thank you again. Of course.

Speaker A: Thank you.

Speaker B: And until next time, thank you for listening in to another episode of the AI CEO Podcast. If you, uh, absolutely loved this episode, which I know you did, please, like, subscribe and comment and we will see you next week. That's a wrap for this episode of, uh, the AI CEO Podcast. If you found today's conversation valuable, don't forget to rate, comment, and subscribe. It helps me reach more leaders like you and continue bringing in top AI first founders, industry experts, and visionary CEOs to share their insights. AI is evolving fast, and the CEOs who harness it today will lead tomorrow. Stay ahead of the curve, keep learning, and keep building. Thanks for tuning in and see you next time.