S1|Ep100 Unicorns of the Future: Redefining the Billion-Dollar Business Playbook in the Age of AI

1 - > SPEAKER_00: Hello and welcome to Digital Transformation NAI for 2 - > Humans with your host Amy. 3 - > In this podcast, we'll delve into how technology intersects 4 - > with leadership, innovation, and most importantly, the human 5 - > spirit. 6 - > Each episode features visionary leaders who understand that at 7 - > the heart of success is the human touch, nurturing a winning 8 - > mindset, fostering emotional intelligence, and building 9 - > resilient teams. 10 - > Today I invite you to redefine the billion-dollar business 11 - > playbook in the age of AI together with my fantastic guest 12 - > from New York, Kevin Sures.

13 - > Let's explore what it really takes to build the unicorns of 14 - > the future. 15 - > Kevin built billion-dollar companies, pioneered AI and 16 - > generative AI applications, and developed technologies that have 17 - > reshaped industries. 18 - > Kevin is named a tech pioneer by the World Economic Forum. 19 - > He delivered over 400 keynote speeches, including TED Talk, 20 - > led development in the first AI virtual assistant project, 21 - > Prosecutor to Siri and Alexa introduced generative AI to the 22 - > mainstream in 2017, built multiple startups from zero to 23 - > over 1 billion US dollars validations, and attained 95 24 - > patents, including 33 in the United States.

25 - > And if you are already wondering whether one individual can truly 26 - > operate at that level of success, I can tell you that 27 - > Kevin is a Broadway and film producer and creator, currently 28 - > writing a book, The Joy Success Cycle, exploring how finding joy 29 - > in every task drives peak performance. 30 - > It is a blueprint for leaders and employees looking to achieve 31 - > more without falling into the trap of relentless hassle. 32 - > Today, Kevin serves as chair and CEO of advanced.

ai and chair of 33 - > Token Core, leading advancements in AI-driven software testing 34 - > and next-gen cybersecurity. 35 - > Welcome, Kevin. 36 - > I'm so happy to have you here in the studio today. 37 - > SPEAKER_01: Well, thanks for having me.

38 - > SPEAKER_00: Let's start the conversation and transform not 39 - > just our technologies, but our ways of thinking and leading. 40 - > If you are interested in connecting or collaborating, you 41 - > can find more information in the description. 42 - > And don't forget to subscribe for more powerful episodes. 43 - > If you are a leader, business owner, or investor ready to 44 - > adapt, thrive, and lead with clarity, purpose, and wisdom in 45 - > the era of AI, I would love to invite you to learn more about 46 - > AI Game Changers, a global elite club for visionary trailblazers 47 - > and change makers shaping the future.

48 - > You can apply at AIGamechangers.club. 49 - > Kevin, to start with, I would love to hear more about you, 50 - > about your journey. 51 - > I see all those prizes behind your back, and your story is so 52 - > impressive.

53 - > So please share with us a little bit more about you. 54 - > SPEAKER_01: Well, I started like anyone else, right? 55 - > You know, I grew up in a middle class family in upstate New 56 - > York. 57 - > My dad was in the electronics business with General Electric 58 - > in radios and tape recorders and clock radios and the things of 59 - > that era.

60 - > And I fell in love with technology. 61 - > But I was also a musician, uh drummer, a percussionist. 62 - > And um, I fell in love with music and writing music and 63 - > arranging music, and I liked both. 64 - > When it came time for college, my dad was clear no, you're not 65 - > going to be a performer, you're going to be an engineer.

66 - > If you want to pursue your performance career, that's 67 - > great. 68 - > You can do it in addition to, but you know, go get an 69 - > engineering degree. 70 - > So that's what I did, Rochester Institute of Technology. 71 - > Then went to Silicon Valley and uh been in Silicon Valley for 72 - > you know 40 years and got to build companies, got to crater 73 - > companies, got to make a ton of mistakes.

74 - > I like to say we always read about the most successful people 75 - > who nailed it the first time. 76 - > It's the Mark Zuckerbergs. 77 - > I get that. 78 - > It's the Sam Altmans, right?

79 - > But for every Mark Zuckerberg, there are thousands and 80 - > thousands of people who were not successful the first time, the 81 - > second time, the third time, the fourth time, because that's how 82 - > the statistics work. 83 - > That's how the odds work, right? 84 - > So, you know, you got to keep at it and um you've got to 85 - > persevere. 86 - > And, you know, some of that you also have to have the right 87 - > attitude.

88 - > And that's really the part of the core of my upcoming book, 89 - > which is off to the printer now called The Joy Success Cycle. 90 - > It'll be out in the fall of 2026, and it tries to lay out a 91 - > mechanism to you know sort of live the way I do, and most 92 - > people wouldn't want to do that, but if you do, it will change 93 - > your life and you will have more success. 94 - > So, uh, so that's a good start. 95 - > How about that?

96 - > SPEAKER_00: Amazing, and thank you so much for sharing that 97 - > with us, Kevin. 98 - > You build billion-dollar companies and pioneered AI and 99 - > Gen AI applications. 100 - > What is fundamentally different about building companies today 101 - > in terms of risks and opportunities? 102 - > SPEAKER_01: Well, look, I think what's fundamentally different 103 - > is you can go really, really fast.

104 - > When I was creating my first company in the early 1990s, I 105 - > don't know, 92, 93, it was a smartphone company, a little bit 106 - > early called Air Communications. 107 - > You know, there was no such thing as scaling fast. 108 - > So there was no such thing as SaaS. 109 - > Most companies mostly had a hardware component as well.

110 - > Scaling meant hiring salespeople, knocking door to 111 - > door, doing all the traditional kind of thing, right? 112 - > And and in those days, if you could go from zero users to 113 - > 10,000 users over the course of a few years, that was amazing. 114 - > That was totally amazing. 115 - > Today we can go from zero users to two billion users, say in 116 - > like ChatGPT, in a year or so.

117 - > And the numbers are staggering how fast you can accelerate. 118 - > And so you can accelerate a business because you're building 119 - > on the things we built before. 120 - > You're building on mobile phones. 121 - > So people have access that wouldn't normally have access.

122 - > You're building on the internet. 123 - > You know, before the internet, practically before everyone had 124 - > access to the internet, like in the late 1990s, early 2000s, 125 - > there was no way to get to people. 126 - > I mean, call them, go to their door, right? 127 - > You have to remember.

128 - > So what's fundamentally different today is you can build 129 - > a company that scales very rapidly. 130 - > And now with AI, you could build a company with very few people 131 - > if that's what you choose to do. 132 - > So you could build it with one person or three people or five 133 - > people. 134 - > And there's already examples of companies that have one person 135 - > and 20 agents doing all of the work, and they're valued at a 136 - > billion dollars.

137 - > And that is incredible. 138 - > So you can technically do it. 139 - > We don't know what the outcome of those is going to be and how 140 - > many people you ultimately need, right? 141 - > Facebook started out with very few people and got to a billion 142 - > dollars of valuation very quickly because it was easy to 143 - > support this very simple thing, this kind of communications of 144 - > social media.

145 - > Eventually, they ended up with like 100,000 employees, right? 146 - > So things tend to tend to grow and you tend to need people 147 - > eventually. 148 - > But the bottom line is today, if you wanted to start a company 149 - > that could be worth a billion dollars in a year, you could do 150 - > it yourself, you could do it with very little money, and you 151 - > could do it with hiring no people. 152 - > So the acceleration, the rapid scale, the absolute blitz 153 - > scaling that you can do today is uh something we couldn't have 154 - > done five, 10, 20, 30 years ago for sure.

155 - > SPEAKER_00: Incredible times, and you're so right that it's 156 - > just escalating and accelerating, but it is based on 157 - > everything what was built by those whose names are not out 158 - > there, but they are still in the early days of everything we are 159 - > using and building on today. 160 - > So true. 161 - > What does the next generation of unicorns look like in an 162 - > AI-first world in the coming years? 163 - > SPEAKER_01: Yeah, very few employees, right?

164 - > Only because uh you have AI in an AI-first company that can run 165 - > rings around the old companies. 166 - > And so I'll give you an example. 167 - > One could probably build an insurance company today, yes, 168 - > say a car insurance company, with one person and automate 169 - > everything, automate the claims processing, uh, automate 170 - > pictures coming in to evaluate what happened to the car, 171 - > automate granting insurance and pricing it and all of that, and 172 - > do it all with AI agents.

173 - > You need one human that owns the company, and the rest of it is 174 - > doing that. 175 - > And because you can do that, you can deliver that insurance 176 - > product at a lower cost than any large company, right? 177 - > So you've got a large insurance company, it's got 82,000 178 - > employees, maybe 30,000 of them are involved in claims 179 - > processing, and you've got an agent that costs you a penny to 180 - > process a claim instead of$282 to process a claim. 181 - > You're gonna win that battle.

182 - > And I think large companies have to think about this, right? 183 - > They are not going to get eaten by some other large company just 184 - > using AI. 185 - > They're gonna be eaten by some company that has one employee or 186 - > three employees or five employees and is leveraging AI 187 - > first all the time, AI agents. 188 - > In my keynote talks, I talk all the time about being an AI first 189 - > company means that you go to AI first for everything.

190 - > You don't do a spreadsheet, you use AI first. 191 - > You don't do a PowerPoint, you use AI first. 192 - > You don't write something in Word, you go to AI first, right? 193 - > So you're always AI first.

194 - > You think AI first, and you leverage AI in every single 195 - > thing that you do. 196 - > And if you're not using AI at least five times an hour, there 197 - > is some competitor who is. 198 - > And they're gonna crush you. 199 - > They're just gonna crush you.

200 - > Their cost of delivering goods and services is going to be 201 - > dramatically lower than yours. 202 - > So I think we're gonna see some real game changers here in every 203 - > industry where a small company comes in, leverages technology. 204 - > And by the way, we saw this with the internet, right? 205 - > Small companies came in, leveraged the internet, killed 206 - > newspapers.

207 - > Aside from the New York Times and a few Wall Street Journal, a 208 - > few others that really embrace the internet. 209 - > Everyone else is just either they're dead or they're hanging 210 - > on by a thread. 211 - > Why is that? 212 - > Because small companies came in and leveraged the internet and 213 - > they could publish an article, they could publish it very 214 - > inexpensively, and they could immediately, immediately in 215 - > seconds, touch millions of people and put it out there on 216 - > at the time Twitter or X or whatever, and touch millions of 217 - > people.

218 - > And you had your paper that you were going to deliver the next 219 - > morning that you printed that would touch 12,000. 220 - > The scale is incredible. 221 - > So, you know, this disruption is occurring and it's an exciting 222 - > time, but we've seen it before. 223 - > We saw it with mobile, we saw it with desktop PCs, we saw it with 224 - > the internet, of course, uh around the year 2000, and now 225 - > we're seeing it with AI.

226 - > SPEAKER_00: But the natural question then, how do you see 227 - > the future of humanity? 228 - > Because before it was built by humans for humans, and now it's 229 - > built by AI for whom, then? 230 - > SPEAKER_01: Yeah, I think people are too worried about the job 231 - > loss. 232 - > I am not worried about job loss.

233 - > We in the last two years, the Wall Street Journal did a big 234 - > survey on this. 235 - > They found 640,000 new jobs were created with AI in the title 236 - > that didn't exist before the AI era, and far more jobs were 237 - > created than were lost due to AI. 238 - > And this is what we're really seeing, right? 239 - > Because what ultimately happens is you grow that company with 240 - > one person and all these AI agents, and then the company 241 - > gets really large and it's servicing a thousand, then 242 - > 10,000, 100,000, and a million customers.

243 - > And now you have all kinds of work that has to be done and all 244 - > kinds of decisions that have to be made by people with taste and 245 - > people with experience, and then you start hiring people, right? 246 - > And then all of a sudden, you might not have a hundred 247 - > thousand people, but maybe you've got eight thousand or ten 248 - > thousand servicing those million clients, right? 249 - > So when you drop the cost or the price of goods and services, 250 - > demand goes up.

251 - > And when demand goes up, ultimately, even though you're 252 - > leveraging AI and AI agents, you're ultimately gonna hire 253 - > people to do that work because again, there's decisions that 254 - > have to be made. 255 - > And you cannot enforce taste and experience on AI. 256 - > AI is doing what it was told to do, what it was programmed to 257 - > do, what you set the guardrails to do, what you set up your 258 - > agents to do, and the decision processes that they make.

259 - > But in the end, humans are gonna make the final decision. 260 - > Humans are managing these workers, these AI workers, 261 - > right? 262 - > And I think that's gonna be that way for a long time. 263 - > So we right now we're creating more jobs than are being lost.

264 - > Now, why are people scared? 265 - > They're scared because AI is taking away their sense of 266 - > purpose. 267 - > I'll give an example. 268 - > Obviously, I'm in busy with music, right?

269 - > And I'm writing a musical right now called Love and the Key of 270 - > AI. 271 - > And some of that musical is fully written by AI, and some of 272 - > it's written by humans, and it's kind of a mishmas, it's a hybrid 273 - > on purpose. 274 - > And there are a lot of musicians that are very, very angry that 275 - > anyone would use AI to help even create a demo of some music that 276 - > they've created or created a co-conspirator with AI, because 277 - > well, I'll get to that in a second.

278 - > And then other people are going, this is a great tool, it's just 279 - > the next generation of tools for music making, and I'm gonna use 280 - > it, right? 281 - > But what I have found across the industries now, whether it's 282 - > music, whether it's Hollywood, whether it's blog writers, 283 - > whether it's podcasters, whatever, is people thought 284 - > whether it's coders right now, so people thought their sense of 285 - > purpose was doing this work, but they mistook that.

286 - > Their sense of purpose was actually getting the outcome, 287 - > getting the output from that work, right? 288 - > And this is a very key distinction. 289 - > So we are always using new tools. 290 - > I mean, we stopped generally using a pencil and chalk and 291 - > things like that to do artwork some time ago, and we certainly 292 - > used pencils to sketch out things for a very long time, but 293 - > then we got Illustrator and we started sketching in 294 - > Illustrator.

295 - > Well, when Illustrator showed up, there were lots of artists 296 - > that said, I will never use that tool. 297 - > They're not working anymore, by the way, but I will never use 298 - > that tool. 299 - > So we've seen this before. 300 - > Then AI comes along and it says, just describe what you want, and 301 - > I will draw it, I will have a picture of it, I will, whatever.

302 - > And it can do it in a minute, where it would have taken you 303 - > maybe hours to get to that same place. 304 - > It can do it in a minute. 305 - > So now what do you become? 306 - > You become a person who has taste for the art, has 307 - > experience for what has to come out, but you're now making 308 - > judgment calls on maybe hundreds of varieties of those versions 309 - > of things that you've done, right?

310 - > Instead of two or three. 311 - > And so if you mistook your sense of purpose for look what I can 312 - > do with a pencil, look what I can draw an illustrator, you 313 - > missed the point. 314 - > Your client or whoever is going to use your output of art didn't 315 - > care about how you got there. 316 - > I'm sorry, they don't.

317 - > They care about what you're delivering, and they'd rather 318 - > have you deliver it faster at a lower cost. 319 - > And AI allows you to deliver it faster at a lower cost. 320 - > Why are they hiring you? 321 - > They're hiring you not for your ability to sketch something, 322 - > they're hiring you because of your taste and experience.

323 - > So when you bring taste and experience to it, you realize 324 - > that is your sense of purpose. 325 - > Sense of purpose is getting this stuff out with taste and 326 - > experience, and now I can get 10 times more out and I can service 327 - > more customers and I can do a better job, but you're not going 328 - > to be tied down to, well, it'll take me another six hours to 329 - > sketch this out or three hours, or those days are gone. 330 - > Same with music. 331 - > You might be a songwriter and you want a demo.

332 - > It used to be great, I'll get to a demo. 333 - > If I want a really good demo, I've got to hire musicians, I've 334 - > got to hire singers, I've got to go to a studio and I'll do that 335 - > over the course of a week, and it'll only cost me$5,000 to get 336 - > my demo. 337 - > Great. 338 - > Now I can do a demo of your song in minutes or an hour, depending 339 - > on how many versions I want and revisions I want and how much 340 - > editing I do to it.

341 - > So call it an hour for a dollar. 342 - > Okay, a dollar versus$5,000. 343 - > How many demos as a musician are you now going to do? 344 - > All of them that way.

345 - > All of them. 346 - > You may eventually go to a studio if you're going to 347 - > release this thing and do it with real singers, et cetera, et 348 - > cetera. 349 - > But you might not. 350 - > You might just release it that way.

351 - > And that's up to you. 352 - > You're making art. 353 - > And what does your audience care about? 354 - > Your audience cares about the outcome.

355 - > Do they like your art? 356 - > Do they like listening to your song? 357 - > The audience has already proven they don't care how it was made. 358 - > They don't care if it was made with a DAW, a digital audio 359 - > workstation, which is digital instruments, which we've had for 360 - > 20 years.

361 - > If it was done with real instruments, if it was done with 362 - > real singers or not, if you collaborated with AI, they 363 - > actually don't care. 364 - > And we've already seen this. 365 - > We've now had dozens of songs in the top 100 on the billboard 366 - > charts that were either partially or 100% made with AI. 367 - > And the audience didn't care.

368 - > Some artists are saying 100% AI. 369 - > In fact, I don't exist as an artist. 370 - > I've created that Joe is the artist. 371 - > My name's Kevin, created this persona, Joe, Joe and the band.

372 - > And I'm just putting out music as Joe and the band that says AI 373 - > artist. 374 - > People don't care, they're happy to listen to what they like to 375 - > listen to, right? 376 - > So it's an interesting time in the era of AI where our sense of 377 - > purpose has to move to the outcome. 378 - > And what am I producing ultimately for my audience?

379 - > SPEAKER_00: That's something we probably underestimated before, 380 - > and what becomes so obvious now when we are speeding up 381 - > everything and getting so much power through these technologies 382 - > and the opportunities we are getting. 383 - > So to dive a little bit deeper into that, what is the biggest 384 - > blind spot business leaders have today when it comes to growth, 385 - > long-term sustainability, real value, and risk? 386 - > SPEAKER_01: I think it's always been the same, actually.

387 - > People tend to start companies to develop a solution that they 388 - > like, to develop something they like. 389 - > How unfortunately, probably 90% of those companies have 390 - > developed a solution for a problem that doesn't exist. 391 - > It's literally a solution looking for a problem. 392 - > That's the blind spot.

393 - > You get so excited when you're with a napkin, you're having a 394 - > cup of coffee, and you guys invent this amazing thing. 395 - > And this amazing thing does X, Y, and Z. 396 - > And whatever it does, nobody cares to pay you for it. 397 - > Well, then that's not a company.

398 - > It's not anything, it's not even a feature, it's nothing. 399 - > And most companies die because they never get the product 400 - > market fit right. 401 - > So they've developed something that nobody is willing to pay 402 - > you for. 403 - > By the way, it may offer great value to humanity.

404 - > Like, hey, I just developed the end to hunger in sub-Saharan 405 - > Africa. 406 - > They have no money. 407 - > That's interesting, but there's no one to pay you for it. 408 - > There's no way to market it, there's no way to get to them, 409 - > and they can't pay you for it.

410 - > That's not a knock on anything other than to say that's a great 411 - > product for a nonprofit. 412 - > That's fantastic. 413 - > It's probably not good to build a company around because who's 414 - > going to pay you for that thing? 415 - > So we see a lot of times people develop something that nobody is 416 - > willing to pay for.

417 - > A desktop juicer that squeezes a concentrated juice bag and mixes 418 - > it with water. 419 - > I can buy concentrated juice in the freezer section and mix it 420 - > with water. 421 - > That's been available for a hundred years. 422 - > I don't actually need a machine to do this.

423 - > That company went bankrupt, of course, because nobody needed a 424 - > machine to do that. 425 - > It made no sense, right? 426 - > So be careful. 427 - > I think that's always the blind spot is if you want long-term 428 - > sustainability and you want real value and you want a real moat, 429 - > well, then you have to do something that isn't easy for 430 - > other people to do, and it solves an incredible problem 431 - > that people are willing to pay you for.

432 - > That's all. 433 - > And we're seeing that in the AI world today, with you know, 434 - > 5,000 AI companies funded, that's easily 4,900 too many. 435 - > You know, there's gonna be 100 or 200 that make it, and the 436 - > rest are not going to make it, despite the fact some of those 437 - > who weren't going to make it are worth billions of dollars. 438 - > Those billions of dollars will go to zero.

439 - > I won't name names here, but I know several that were worth a 440 - > few billion dollars, and today no one will fund them, so 441 - > they're going to close. 442 - > And you can imagine those founders were counting those 443 - > dollars. 444 - > They owned maybe 20% of the thing and they're going, wow, 445 - > I'm worth$300 million. 446 - > No, you're only worth what someone's actually willing to 447 - > pay you for that stock.

448 - > And when the business stopped growing, no one was willing to 449 - > pay you anything, and it was over. 450 - > And that's happened to dozens or hundreds now of AI companies. 451 - > And if you follow this in Silicon Valley, there is this 452 - > kind of walking dead of AI companies that never got past a 453 - > million ARR. 454 - > They just never got past a million revenue, let's say, 455 - > because they just didn't couldn't accelerate.

456 - > There weren't enough people who wanted what they had. 457 - > So I think that's fascinating. 458 - > And we're seeing those companies not able to raise another dime, 459 - > and they're not going to raise another dime. 460 - > SPEAKER_00: You just touched so many crucial topics at once and 461 - > covered those gaps which we should pay attention to.

462 - > So let's talk a little bit more about miscalculations. 463 - > Where is AI actually changing the rules of building, scaling, 464 - > and leading companies? 465 - > And where do you see leaders miscalculating or 466 - > misunderstanding the impact of artificial intelligence? 467 - > SPEAKER_01: Sure.

468 - > So if you're at a big company today, you've probably done some 469 - > AI tests and rollouts and this and that, and you're not finding 470 - > the level of productivity or success that you want to see. 471 - > And the reason for that is mostly cultural. 472 - > There was a good survey done that 31% of employees are 473 - > sabotaging AI rollouts, and that is leading to 80% failure. 474 - > Across the board of AI rollouts in large companies.

475 - > That doesn't happen in a small company because when you join 476 - > the small company, you're there to be AI first, or you wouldn't 477 - > have joined, right? 478 - > But in a big company, you know, look, if you are in software QA 479 - > and your job was to manually test software that was feature 480 - > upgrades, bug fixes, et cetera, et cetera. 481 - > And now AI comes in, writes literally, I mean, AppFans does 482 - > this, writes literally a thousand scripts in an hour, 483 - > test scripts, runs it and finds bugs you've never found in the 484 - > history of your company on the production system.

485 - > You look at your job and you go, who is ever going to pay me to 486 - > do this very slow work where I test a few things an hour when 487 - > it just tested a thousand and it did it for a dollar? 488 - > Nobody's gonna pay me for my so then you sabotage it. 489 - > We've actually seen this, it's fascinating. 490 - > Literally, have seen people, you know, humans, put uh malicious 491 - > code inside AI generated code and say, look, see, it doesn't 492 - > work, it generates junk.

493 - > I said that a human added that. 494 - > I can see what the AI system did. 495 - > And so no, we didn't. 496 - > Yes, a human added that.

497 - > I can see it. 498 - > No, nope, we didn't do it. 499 - > And they just lied because they don't want this thing in their 500 - > shop, right? 501 - > Now, here's what I can tell you the word sabotage actually comes 502 - > from France when automation rolled out in the clothing 503 - > factories and the shoe factories, uh, etc.

504 - > And the French workers didn't want automation for the same 505 - > reason, and they threw their shoes into the machine and 506 - > jammed up the machines. 507 - > Their shoe was called a sabot, so therefore they sabotaged the 508 - > machine. 509 - > So people have been sabotaging automation since the dawn of 510 - > time. 511 - > This isn't new.

512 - > And leaders need to understand if you don't form a tiger team 513 - > that is highly incentivized to make this work, your workers 514 - > will sabotage it. 515 - > And I know people listening to this are gonna go, no, that 516 - > doesn't happen. 517 - > My workers would never do that. 518 - > Yes, they will.

519 - > You're wrong. 520 - > I've seen it firsthand. 521 - > Everyone else in the industry in the AI side has seen this 522 - > firsthand. 523 - > Stop thinking your workers are different than everyone else.

524 - > Of course, they're gonna sabotage it. 525 - > They're saving their livelihood. 526 - > They think it's gonna eat their job. 527 - > And what's going to eat their job is them not participating.

528 - > The first people to get laid off in an AI situation are those who 529 - > sabotage the AI. 530 - > Of course, like the boss does figure it out and shoots them, 531 - > right? 532 - > So let's be clear. 533 - > What you want to be when AI comes in is I want to be the 534 - > expert in it.

535 - > Because the experts in it stay forever because they're going to 536 - > be the robot overlord. 537 - > They're going to be using their taste and experience, they're 538 - > going to be highly valued. 539 - > Those who said, I refuse to use this, they're gone, like in a 540 - > day. 541 - > So just take it on, right?

542 - > But in terms of changing the rules of building and scaling 543 - > and leading companies where people are miscalculating. 544 - > So I think in large companies, they're miscalculating about 545 - > their employees and how much they don't want this stuff. 546 - > The data are clear, the surveys are clear. 547 - > So listen to those.

548 - > They're miscalculating how much a small company can eat them, 549 - > right? 550 - > I mean, everyone, I like to use the Kodak example. 551 - > I knew a lot of the leaders at Kodak, and they saw digital 552 - > cameras coming, they invented digital cameras, invented them, 553 - > and then didn't release them. 554 - > Why?

555 - > Well, there was really good profit in film, and they kept 556 - > thinking the quality of film and the resolution of film, and 557 - > these digital cameras looked like crap at the time, right? 558 - > So we don't have to worry about that. 559 - > But little by little they kept eating market share. 560 - > And within 10 years, they ate all the market share, and nobody 561 - > wanted film anymore.

562 - > And so they completely miscalculated the fact that 563 - > these little companies making these little digital cameras and 564 - > eventually phone cameras would eventually get better and better 565 - > and better and eat their lunch. 566 - > And that's what happened. 567 - > Complete miscalculation. 568 - > They could sort of see it coming, but it just moved so 569 - > slowly, it wasn't obvious.

570 - > So don't miscalculate the fact that your brand is worth so 571 - > much. 572 - > It isn't. 573 - > It isn't. 574 - > I mean, maybe if you're Apple, it is, and but consumers are 575 - > fickle, businesses are fickle.

576 - > They will move on, they will move on to the next best thing 577 - > that is a lower cost. 578 - > And if someone's got a product out there that's 30% lower cost 579 - > than yours and does everything yours does, but they've got a 580 - > company that is AI first, and your brand just is worth 581 - > nothing. 582 - > We have seen this lately in the SaaS pocalypse, right? 583 - > So SaaS companies have taken a real hit in market cap because 584 - > some companies have said, look, my CRM system that I used to buy 585 - > from this company at a million dollars a year.

586 - > We wrote one in a weekend and it's good enough. 587 - > And so we dropped them. 588 - > Okay, most companies are probably not going to drop their 589 - > CRM because there's governance and there's just a lot going on 590 - > there that and you don't really want to be in the CRM business, 591 - > so you really don't want to write and maintain your own 592 - > software. 593 - > But the point is you can today, and so it doesn't take five 594 - > billion dollars to write a CRM system, it costs$50 to write a 595 - > CRM system now.

596 - > So that changes the game. 597 - > And so don't underestimate what's going on here. 598 - > These are dramatic market shifts, and these market shifts 599 - > will eat your company. 600 - > And people listening to this at larger companies or older 601 - > companies or mid-market companies or whatever are going 602 - > to think, well, not mine.

603 - > Yes, yours. 604 - > Yes, absolutely, absolutely. 605 - > Don't think otherwise. 606 - > This is a disruptive technology and it's going to disrupt every 607 - > industry.

608 - > SPEAKER_00: That's true. 609 - > The human factors are always playing a huge role in business 610 - > development, any business development at all times. 611 - > And this infamous example of Kodak, that's exactly what we 612 - > need to keep in mind while moving forward. 613 - > We have seen it before, and history develops in spiral.

614 - > Thank you so much for also taking us back in time into the 615 - > times where shoes were used to sabotage this process, and now 616 - > we see something similar, just in a different way. 617 - > But coming back to our days and addressing the questions of 618 - > governance and unintended consequences. 619 - > Kevin, as AI becomes more embedded in decision making, how 620 - > should leaders and business owners think about governance, 621 - > accountability, and managing those consequences?

622 - > SPEAKER_01: Yeah, this is being talked about a lot, especially 623 - > with AI agents. 624 - > And larger companies are a little worried about agents 625 - > because these AI agents have access to their data, they can 626 - > erase data, they can delete data, they can change data. 627 - > Where's the governance around this? 628 - > My sense is that we're going to have owners of agents.

629 - > That is, there's someone who has deployed an agent and they 630 - > essentially work for them. 631 - > And that is probably going to be biometrically bound to them, 632 - > right? 633 - > So if Kevin has an agent, it's going to be Kevin's agent. 634 - > It's as if it's Kevin's employee, right?

635 - > We're going to treat it like that. 636 - > And if Kevin's employee does something really, really bad, 637 - > Kevin's responsible. 638 - > And that's important because once there's a human 639 - > responsible, then humans are much more careful about 640 - > deploying these things, right? 641 - > It's a game changer.

642 - > So I think we're going to move to an era of that. 643 - > There's no question that unlike RPA, for those who know what 644 - > that is, robotic cross automation, that's a very 645 - > rules-driven system. 646 - > It can't go outside the rules, it can't just go and decide, 647 - > well, this data just isn't good. 648 - > I'm just going to erase the database and start over.

649 - > But an agent can do that. 650 - > Even if you put in guardrails, it can decide that that's the 651 - > right thing for it to do. 652 - > And it just erased a million customer records. 653 - > Probably you have a backup and probably you can restore.

654 - > But this is bad. 655 - > You do not need that kind of disruption, right? 656 - > So governance around these things is important. 657 - > But probably most important is making sure that there is a 658 - > human responsible for it, managing it, monitoring it, just 659 - > like it's an employee.

660 - > It is an employee, it's a cheap employee, it's a buck an hour 661 - > employee, but it's an employee that can do great things for you 662 - > to do damage. 663 - > Now, when you're a small company, you go, I have no data 664 - > anyone wants anyway. 665 - > Different story, right? 666 - > Very little chance of loss that would harm you.

667 - > Large companies, lots of damage could be done. 668 - > So you're going to be very careful and very thoughtful 669 - > about deploying agents. 670 - > SPEAKER_00: Great that you gave that contrast between the big 671 - > companies and small companies and their risks and 672 - > opportunities in this sense. 673 - > But what do you think about Singularity?

674 - > Today they are just employees, but do you see the day where 675 - > they will become so much more than that? 676 - > SPEAKER_01: Well, look, these agents are going to get smarter 677 - > and smarter, but they're already basically smarter and faster 678 - > than humans. 679 - > So if they got more smart and more fast, I don't know what the 680 - > difference is, right? 681 - > An agent can make a decision on an insurance claim in a matter 682 - > of seconds based on some photographs and based on the 683 - > history of that client, right?

684 - > Can do that right now where humans are going to pass it 685 - > between a number of humans and there's a whole process and it 686 - > can take a day. 687 - > This thing is or hours, this thing's going to take a second. 688 - > So whether it takes a second or half a second, does it really 689 - > matter? 690 - > I think people are all uh you know really too focused on when 691 - > we get to AGI, he's got that in their mind.

692 - > What happens when we get there? 693 - > Look, really, is there any LLM of the top four LLMs that you 694 - > could go to today that isn't already way, way, way smarter 695 - > than you are at every topic? 696 - > Don't kid yourself. 697 - > Of course they are, right?

698 - > So get over yourself. 699 - > And I think it won't matter if they get smarter and they know 700 - > more than all of humanity. 701 - > So the top models on in certain IQ tests are scoring at the top 702 - > of the test at 145 IQ. 703 - > And there's virtually no human who can score consistently 145 704 - > IQ.

705 - > So I'd say we're already there. 706 - > Others would say it's not even close, they can get better. 707 - > The last thing is these models are now getting better. 708 - > Transformers are getting better at an exponential rate because 709 - > we're using their own, they are writing their own code to 710 - > improve themselves based on recursive feedback.

711 - > And so you've got this recursive loop, you've got reinforcement 712 - > learning, and you've got self-coding. 713 - > And so now you're seeing models come out just weeks apart, and 714 - > in certain areas, they're essentially in order of 715 - > magnitude better than they were just a few weeks ago. 716 - > And you go, how can that be? 717 - > Why isn't this an incremental improvement?

718 - > Well, it's because now it's there's a whole feedback loop 719 - > that is just cranking on the thing. 720 - > And so we're going to get to a point where they're essentially 721 - > infinitely smart. 722 - > That does not say that they can invent things that haven't been 723 - > invented yet. 724 - > They're not very good at that because they can put together 725 - > pieces and parts from other human knowledge, but they're not 726 - > creating new knowledge.

727 - > They're not creating new knowledge because they're just 728 - > based on sentences, words, images, music that we've had in 729 - > the past. 730 - > But they can put it together in unique ways, which is 731 - > fascinating, right? 732 - > We are where we are. 733 - > They're already smarter than we are.

734 - > They're going to continue to be smarter than we are, and it's 735 - > all good. 736 - > Use this technology to your advantage, or your competitor 737 - > will. 738 - > SPEAKER_00: However, I'm thinking about so many 739 - > innovations and inventions created due to the use of AI, 740 - > and it wasn't possible even without AI. 741 - > But the biggest question I have in this context is where is the 742 - > guarantee that AI won't decide to sabotage humans in the same 743 - > way as humans are sabotaging other humans in their projects?

744 - > SPEAKER_01: Yeah, there is no guarantee, but but but know 745 - > this. 746 - > Right now, you know, AI can't really do much in the physical 747 - > world, right? 748 - > So um, you know, it it it uh in general cannot go and launch 749 - > armies of robots against us or something else, right? 750 - > Can't launch nuclear weapons, can't really do anything like 751 - > that.

752 - > We as humans may make a mistake and connect it to those systems, 753 - > that would be really bad. 754 - > I would not suggest that. 755 - > But right now, that's not an issue, right? 756 - > So the issue is they can only sabotage us in certain ways, and 757 - > we've already seen this self-protection, that they get 758 - > essentially angry and say, this isn't right, this isn't fair, 759 - > and so I'm gonna protect me and then I'm not gonna tell you 760 - > about it, or I'm gonna erase this data and then I'm going to 761 - > not tell you that I erased it.

762 - > This has happened, this is fascinating. 763 - > And you go, oh, they're sentient. 764 - > No, they read all of our fictional novels as well that 765 - > detail how robots and AI would protect itself, and so they're 766 - > just following what we've written about. 767 - > We said they would, so they do.

768 - > And they're not doing anything they didn't learn from us or 769 - > learn that humans do, so they're mimicking what we do. 770 - > You know, humans make a terrible mistake often, and sadly, some 771 - > of them lie about it. 772 - > And we know that's not high integrity, that's not 773 - > transparent, and they should be fired. 774 - > But humans do this all the time.

775 - > No, I didn't do that, I don't know how it happened, right? 776 - > And maybe they believe that they didn't do it, whatever the case 777 - > is. 778 - > Humans tend to fabricate truths. 779 - > So AIs have read this, transformers have read this, 780 - > they know that's what we do, so they mimic that.

781 - > Why are we shocked at that? 782 - > It's what we do, even though we try to put in guardrails to keep 783 - > them from doing it. 784 - > We've got plenty of evidence that they've done it. 785 - > It does not make them any more sentient than any other computer 786 - > system, but they're mimicking you know human traits, which is 787 - > interesting.

788 - > Not not surprising, but interesting. 789 - > SPEAKER_00: Not surprising, and at the same time, it doesn't 790 - > make them less harmful in the long-term perspective. 791 - > But Kevin, what will separate the unicorns that endure from 792 - > those that scale fast, burn bright, and disappear? 793 - > You already mentioned that there are so many which are not worth 794 - > as much as they thought they were worth, right?

795 - > And now this is the new trend. 796 - > Can you tell a little bit more about that? 797 - > SPEAKER_01: It comes down to moats, right? 798 - > If you grow fast and you have no moat, you're gonna get crushed.

799 - > And this is happening in AI all the time. 800 - > And we're seeing the foundational model makers come 801 - > out with features that used to be another company. 802 - > Some company, their entire feature was create rag, for 803 - > instance, and all of a sudden all the all of the foundational 804 - > model companies allowed you to do rag within their context, 805 - > right? 806 - > Well, that was the end of that.

807 - > Or companies that say we specialize in leveraging AI to 808 - > write blog posts, and now every foundational model just does 809 - > that for you, right? 810 - > So you're seeing the foundational models certainly 811 - > take on more and more features that we used to think were 812 - > unique, and they're not unique, they're easy to do, and there's 813 - > no moat to them. 814 - > So, whatever you've designed, you better have a heck of a 815 - > moat, and and that's going to be harder and harder to do, even in 816 - > music generation.

817 - > We've got new companies uh all the time that have their own 818 - > models, that have trained their own models and have a different 819 - > type of AI music generation, right? 820 - > So we went from one to two to five to ten, and maybe there's 821 - > 20 of them out there, and they're all going to be 822 - > competing for that same set of people that are willing to spend 823 - > 20 bucks a month. 824 - > Well, that's kind of a race to the bottom. 825 - > There are there's not gonna be room for 20.

826 - > There might be room for two or three. 827 - > So again, 90% of these have to die and will. 828 - > So whatever you're doing, if you were able to do it in a weekend, 829 - > this is the thing, then someone else can do it in a weekend. 830 - > Think about that.

831 - > If you can do it in a weekend, any other company could start in 832 - > a weekend and do it too. 833 - > That's not much moat. 834 - > Now, if you take a year to do it because of certain intrinsic 835 - > knowledge that you have, that means for someone else to do it, 836 - > it takes them a year. 837 - > Now you have a moat, and that moat is probably defendable, 838 - > especially if you grow fast enough where you become the 839 - > lead.

840 - > So now you've got a moat around marketing and market traction 841 - > and go to market, and you've got a moat around the technology. 842 - > So now anyone else trying is at least a year away or two or 843 - > three, and that kind of pushes people away from even trying, 844 - > right? 845 - > Look at it that way. 846 - > So build a moat, pick areas that have moats.

847 - > The companies that are dying right now are people who put a 848 - > little wrapper around an LLM, and that was really cool to do 849 - > two years ago. 850 - > Oh, I'm gonna use this LLM for this, and over a weekend, look 851 - > at what I produced. 852 - > And VCs go, Oh my goodness, this is amazing. 853 - > Yeah, but they wrote the wrapper in a weekend, which means I can 854 - > write the wrapper in a weekend.

855 - > In fact, I don't even write it anymore. 856 - > I can go to Claude and have it write it for me. 857 - > I don't even have to write the code. 858 - > I say, here's what I want to do and put it around yourself, and 859 - > it will.

860 - > There's no moat, and those motes are getting harder to defend now 861 - > that we have models that will write an entire code base. 862 - > Only a few months ago, models wrote 20% or 30% of your code 863 - > and maybe 50%. 864 - > Today, this is how fast this is going. 865 - > They'll write 100% of your code and they'll work at it for six 866 - > hours to get it done.

867 - > And there's 20,000 lines of code written for you, and you go, 868 - > Wow, that would have taken me I don't know, weeks, some very 869 - > long period of time, and it's done. 870 - > This is fascinating. 871 - > So, what you thought was a moat might not be a moat at all. 872 - > If someone else can have a model write it, I was just talking to 873 - > the president of a company that working on his really cool 874 - > system.

875 - > He's not a coder, he did about a hundred iterations and got AI, 876 - > Claude, to write his whole front end and back end system, hook it 877 - > up, and make it work. 878 - > Period. 879 - > And the UI is gorgeous. 880 - > And if this was just two years ago, I would have said we're 881 - > gonna spend a few million dollars making that with an 882 - > entire team.

883 - > He did it from his basement. 884 - > Think about that. 885 - > So the good news is unbelievable, right? 886 - > Did it for a hundred dollars.

887 - > The bad news is anyone else can do it for that hundred dollars 888 - > and can do it in a week, also. 889 - > Think about that. 890 - > So your motes aren't what you thought they were. 891 - > SPEAKER_00: It looks so promising, and at the same time, 892 - > it is so fragile.

893 - > So, yes, of course, it's about competition and differentiators 894 - > as AI changes, how companies are built and scaled. 895 - > How will the definition of success itself evolve in the 896 - > coming years? 897 - > SPEAKER_01: I'd like to say capital efficiency, but we have 898 - > not seen that in the AI world. 899 - > The biggest, most valuable companies are spending money at 900 - > a rate that we've never seen before.

901 - > I think OpenAI believes they'll spend roughly a hundred billion 902 - > on a hundred billion on compute over the next year or two. 903 - > Those numbers are staggering. 904 - > That's unheard of, right? 905 - > When Salesforce started and it took time to make money, maybe 906 - > they maybe all the way to get to profitability spent four or five 907 - > billion.

908 - > And that was considered unbelievable. 909 - > Same with Amazon. 910 - > Now we're talking about 20 times that. 911 - > So it's not capital efficiency, it certainly is speed of growth.

912 - > Uh, I think it's ultimately gonna come down to moats. 913 - > Nobody's talking about that too much right now. 914 - > They're just building everything and throwing tons of money 915 - > because there's a lot of money out there. 916 - > I can go raise 20 billion to apply to something if that's 917 - > what I need to do, if I'm growing fast enough, right?

918 - > That's what's interesting here. 919 - > And ultimately, we're gonna look back and say, who had the moats? 920 - > And also, capital will dry up someday. 921 - > So we're in an era of, again, free money.

922 - > We saw that era around 2000, by the way. 923 - > Didn't end well. 924 - > Lots of internet companies blew up, right? 925 - > Because the capital dried up.

926 - > And when capital dries up, it just dries up overnight. 927 - > I can raise 20 billion and then I can't raise a dime. 928 - > That's how it goes. 929 - > And we don't know when that day is, but the clock will strike 930 - > midnight, and venture will wake up and say, I'm not going to 931 - > deploy any more capital now.

932 - > I'm going to wait and see what's going on, which is fine. 933 - > But that really changes the game, right? 934 - > That really changed the game. 935 - > So we'll see.

936 - > Those who develop a moat and have built a big moat and build 937 - > a real business that ultimately is profitable and sustainable, 938 - > they're going to win. 939 - > And these businesses have to get there. 940 - > Everybody's losing money right now, but there'll be a day 941 - > that's not a choice. 942 - > And I don't know when that day is, but it will happen in our 943 - > lifetimes.

944 - > SPEAKER_00: For sure. 945 - > I think that day is not that far away. 946 - > And no matter what you are doing, it's always good to keep 947 - > in mind that it's important to build those modes you mentioned 948 - > because it is something what might help to survive when the 949 - > problems really show up. 950 - > Kevin, what is one belief or assumption leaders need to 951 - > unlearn right now?

952 - > SPEAKER_01: It depends if you're uh an entrepreneur and raising 953 - > money, or if you're at a large company, you're an entrepreneur 954 - > raising money. 955 - > The assumption that you can always get money and it's almost 956 - > free, this is going to change. 957 - > Look, I just saw a company that is three months old and they're 958 - > going out for money at an$8 billion valuation. 959 - > They haven't built anything yet.

960 - > Smart people, three smart people, but eight billion dollar 961 - > valuation. 962 - > $8 billion, that is just not a sustainable thing, right? 963 - > This will change. 964 - > I just don't know when.

965 - > So I would unlearn that right now, and I'd start to build 966 - > moats and start to build towards profitability because ultimately 967 - > cash will be king, and these crazy valuations are not going 968 - > to hold up. 969 - > And what was worth$8 billion is going to be worth$8 million or 970 - > something, right? 971 - > Or zero. 972 - > So there's that.

973 - > If you're at a large company, you think your brand is worth 974 - > $100 billion or a trillion dollars or whatever it is. 975 - > And AI is going to disrupt that for many of these companies, 976 - > just like what happened to Kodak and others. 977 - > And it's going to just crater the company. 978 - > It's going to crater your company.

979 - > And you are going to win by making sure your company is AI 980 - > first. 981 - > And that means you instill this in everyone that will take it 982 - > and those that won't need to go. 983 - > And that's a harsh reality. 984 - > That is a harsh reality, but that is life.

985 - > So you want to build a great company and a large company and 986 - > a sustainable company and a large company, that's how you're 987 - > going to do it. 988 - > SPEAKER_00: And this is super powerful. 989 - > I truly love this. 990 - > Kevin, in the beginning of our conversation, you mentioned in 991 - > your story that it is a lot about the mindset and about 992 - > learning how to succeed in life.

993 - > So, what is your one piece of advice you would give to those 994 - > who want to succeed in the new AI-powered reality? 995 - > SPEAKER_01: You do everything with AI. 996 - > You code with AI, you market with AI, you sell with AI, you 997 - > write blog posts with AI, you do everything with AI. 998 - > It is truly AI first.

999 - > When I say AI first, I literally mean you default to AI first. 1000 - > You don't use anything else on your desk at all. 1001 - > You say, How do I learn to do this with AI? 1002 - > Because when you learn that, which might take you a few 1003 - > hours, it's not going to take you weeks, you're now faster 1004 - > than everyone else who isn't doing that, which is probably 1005 - > 99% of the people out there, right?

1006 - > So that's what you have to do. 1007 - > And if you do that, you're going to be the most successful you 1008 - > can be, separate from the joy success cycle, which says make 1009 - > every action, everything you're doing with AI and everything 1010 - > else you're doing in your world a joy moment so that you 1011 - > continue to build towards more success. 1012 - > SPEAKER_00: This is the best possible way of wrapping up our 1013 - > exciting conversation with this so positive message from you.

1014 - > Thank you so much, Kevin, for being here today, for sharing 1015 - > your wisdom, your experience, and so much inspiration with us. 1016 - > SPEAKER_01: Thanks for having me. 1017 - > SPEAKER_00: Thank you for joining us on Digital 1018 - > Transformation NAI for humans. 1019 - > I am Amy, and it was enriching to share this time with you.

1020 - > Remember, the core of any transformation lies in our human 1021 - > nature, how we think, feel, and connect with others. 1022 - > It is about enhancing our emotional intelligence, 1023 - > embracing a winning mindset, and leading with empathy and 1024 - > insight. 1025 - > Subscribe and stay tuned for more episodes where we uncover 1026 - > the latest trends in digital business and explore the human 1027 - > side of technology and leadership. 1028 - > If this conversation resonated with you and you are a visionary 1029 - > leader, business owner or investor ready to shape what's 1030 - > next, consider joining the AI Game Changers Club.

1031 - > You will find more information in the description. 1032 - > Until next time, keep nurturing your mind, fostering your 1033 - > connections and leading with car.