← Back to search

The biggest AI mistake business owners make (automating what shouldn’t exist) | Jack Roberts

1% of the 1% Podcast with JT Foxx · 2026-10-28 · 49 min
relevance 67 10993 words Spotify ↗ Episode page ↗ Audio ↗
Show full episode description
You’ve heard it all about AI, but you’ve never heard it like this. I sat down with the legendary Jack Roberts in Dubai and he broke down why most business owners are automating stuff that shouldn’t even exist, why your AI is just telling you what you want to hear, and what actually makes you money. Listen to this and you’ll be a thousand times ahead. Don’t, and you could be left behind. Https://www.jtfoxx.com/links to get branded, get funding, get JT Foxx's cash or get FREE advice. If you also want to learn and do roll ups with me us, happy to chat.
✨ Episode Outline — click any point to jump to it in the episode
Problem solved
Why most business owners fail with AI by chasing shiny new tools instead of executing fundamentals, and what being left behind actually looks like.
Benefits
  • AI raises both the floor and the ceiling: good operators get better
  • Reclaim 25-30 hours/week via dictation and email-reply automations
  • Use frontier models as interviewers to build personal action plans (e.g. health/longevity)
  • Focus on fundamentals over new shiny tools like Claude Design or ChatGPT Law
  • AI implementation done-for-you beats expecting owners to self-learn
Use cases
  • Speech-to-text dictation replacing typing saves ~5 hours/week
  • AI interviews you with 50 questions plus deep research to produce a longevity action plan
  • JT Foxx cloned his own brain into an AI tool, then pivoted to done-for-you AI implementation
  • Prediction: AI agents will outperform real-estate agents since transactions are mostly email
  • Warning: don't build app-layer tools OpenAI or Anthropic can pivot into and replace
KPIs / results
  • One automation saves 5 hours/week
  • Potential 25-30 hours/week returned via automations
  • JT Foxx claims 87 AI companies worldwide
  • Frontier labs cited at 15,000-25,000 engineers each
Tools / build
  • ChatGPT (GPT-6 Astra)
  • Claude
  • AI speech-to-text dictation
  • Email auto-reply automation
  • Personal 'second brain' AI clone
0:00 / 0:00
[SPEAKER_00] Hello everyone and welcome to another edition of the 1% of 1%. We have interviewed some of the biggest billionaires in the world, biggest Hollywood celebrities, the biggest changemakers. But today we're going to talk about an important topic of AI. And I know you've heard it all, but you have never heard about what we're talking here today because if you don't listen, you could be left behind. Today we have the legendary, oh my God, how should I say? They call him the sort of AI, I mean, just by how he's dressed, you know that he's really good at what he'd do. [SPEAKER_00] Because no one would show up into a studio dressed like this to do an interview unless you were the best and it didn't matter. He could have walked in with shorts. But let me tell you something, he knows AI, very entertaining on social media, especially Twitter who has built a huge following. And he understands it probably better than anybody else. And if you listen to this man, you will be a thousand times ahead. And I know AI because we've got a lot of AI companies around the world, 87 different companies, but I'm going to tell you something. This is something where I'm going to be learning as well. Please welcome legendary JT Foxx. How you doing, Jack? [SPEAKER_03] Very well, man. Thanks for having me. [SPEAKER_00] First of all, this is a very good outfit. Should we start off with what's going on here? Is it just like, did you always wear that? Or are you like, hey, I'm in AI. What I'm doing is working. Whatever I wear, I wear. [SPEAKER_03] Absolutely. Well, I think it's helpful to be a creator and in the AI space, but also Dubai weather. Sometimes you go a little bit lighter. [SPEAKER_00] Yeah, actually, we're doing this podcast live from Dubai. People don't know this as well. So yesterday I did an event live in front of the pyramids. The day before I was at the Eiffel Tower doing things as well. So Jack, I'm going to ask you a question. Let's get right into it. What is the biggest problem that you solve for people right now? [SPEAKER_03] Sure. I answered the question. So what? Brand new AI technology drops. Something new has happened. I take that brand new idea and I tell you how it actually moves your business forward and what it actually changes either in your personal life or in your business. [SPEAKER_00] The internet is inundated with AI advice. Everyone's doing these long videos. It's a very hot topic as well. What are people missing when they're consuming over information? How do you know what's right? How do you know what's wrong? Or is everybody just saying the same thing? But is there some point, some overload of knowledge about, okay, great job, there's knowledge, but why don't you go do the things you learn? What are you seeing right now in the space? [SPEAKER_03] Yeah, so great idea. So there's no shortage of new shiny thing in AI. The truth is that there is lots of great new things happening. But in reality, most people are stuck with shiny objects and when it comes to AI. Something new drops like Jev is the talk of the town this week, for example, which is a different way of doing like a chat to BT or Claude. A lot of people get very distracted by all the new brand, brand new shiny things rather than executing on the fundamental stuff that really kind of moves that business forward. [SPEAKER_00] So let's take a look at it. We have people that may be listening to this that are either not using AI. There are some that they're using, you know, chat to BT and Claude so they can do their emails or do a quick social media post or search something. There are others that are using it to sort of create copy or ideas but not really implementing it. Then you got some people who are just slowly trying to implementing. [SPEAKER_00] And then you got these kind of know-it-alls who think that they know everything about AI. I know about AI. I know about this. But they're not making money. And then there are people like me who happen to make millions and millions and millions in AI because I understand that I'm an entrepreneur who's in AI, not AI trying to be an entrepreneur as well, these categories. Yes. Why is it – let's go over these categories. They're brand new people. Talk to the people that are brand new as well. And you have to explain them who are probably resistant, scared it's going to be takeovers, a lot of news lately, right? [SPEAKER_00] They're trying to create that fear-mongering. We'll probably have to have a vaccine just for AI. Like, hey, you know, take a vaccine so you can immune yourself from AI, right? I mean, that's probably something that's going to happen, right? But realistically speaking, the person that's still not on it and not a believer, tell me what can happen to them if they don't use it. What would be a version of the world possibly? [SPEAKER_03] So first thing to understand, AI has simultaneously raised the floor and the ceiling at the same time. If you're already great, it makes you better at what you physically do. However, if you're not using AI, I do strongly believe that the chances of you getting left behind are only getting corrected by the day because this technology is improving. I mean, I've been covering this in detail for years now, and I've never seen the rate of change that we've been seeing over the past six months. The models are getting exponentially better. It's very incredible to see. [SPEAKER_03] The reality is that if you're not using AI in your job, in your personal life, you are going to be beaten by somebody else who is. There's just no denying the productivity and the efficiency gains that you get. A good example is I found recently even one automation that saves me five hours. I mean, you take speech-to-text, for example, like dictation tools, which are now AI-powered. So instead of typing, you press a button. That gets transcribed for you. That saves you five hours a week of typing. [SPEAKER_03] There are automations. You can automate replying to emails. There's like a billion different things. You know, you imagine how much more productive could you be if I gave you 25 or 30 hours of your life back? [SPEAKER_00] But, you know, all these levels, there is a fundamental problem. So I understand people that are new or resistant, but I want to remind people they missed out on the AI boom. [SPEAKER_03] Yeah. [SPEAKER_00] So that they're missing on that. They probably missed out on crypto. They probably missed out on the social media growth and social media like kind of growing. They missed out on the web, the early adoption of web, and maybe on the adoption of CRM. So their entire lives is they've always missed it. And now there's something, and Mark Cuban said to me, we're in first base, first inning of preseason AI right now. That's how early we are there as well. And then robotics at some point is going to come in. And so we have this notion of people holding. [SPEAKER_00] And it really comes down at every single level as we go through is it changes so much. There's so much to learn. I'm older. It doesn't necessarily excite me. I know it's good for me. I'm going to try and delegate to my people. My people won't do it because they're afraid of being replaced. And we're just going to keep the status quo. While another company who's using AI, who's all into AI, right, what's going to happen is they're more efficient. They're more profitable. They're scaling faster. [SPEAKER_00] They may be far more profitable, which means they can invest in more or lower the price to go you out. And people are going to have to be out of business. And they're going to have to look for other jobs. And they're not going to be able to be real estate agents because I strongly believe that AI employee agents will be far better than what are real estate agents because most of the transaction is email. Go here. Show up here. Talk to me with the offer. Two AI people kind of negotiate. And then someone presses the button and approves, right? [SPEAKER_00] And so they'll probably have to go back and get into trades at the end of the day because there's a lot of white-collar jobs. There's a lot of businesses because the cost of labor is going up. And as a result, the margins are getting thinner. And the only way to create that advantage is AI. What do you say the person that's resistant to learn and not making the time because as people get older, I think it's harder for them to be committed to learn something because they've got life and kids and jobs and businesses and the hunger to learn. I know you have it. I have the hunger. But a lot of people may not listen. [SPEAKER_03] Well, I think there's two levels to this. I think if you're a – there's a business level and then there's level for the individual. I mean, like AI, we are close to solving, I think, over the next 10 years some of the biggest health questions of our time, the longevity question. But right now, all the delta exists from those individuals who are applying and using technology. I struggle to think of a single area of your life where you couldn't – if you applied, you know, the right questions and the right prompts with AI that couldn't drastically improve you. Let me give you a classic example. [SPEAKER_03] Health, right? Wouldn't we all love to be healthier, have a high-quality life, and live for longer? Well, you know, you can be interviewed by these now frontier models. They can break down and ask you 50 questions. You can even say to AI, hey, my objective is to maximize the duration and the quality of my life. Ask me any question you can or any questions you'd like to ask me and do deep research on the topic such that you have all the relevant data to give me a clear action plan. And then Claude or ChatGPT, you know, we use GPT-6 Astra. It's probably the best model on the planet right now. [SPEAKER_03] We'll interview you. And you go back and forth with that, and you can actually get all the data, all the questions, and then a strategic plan. And honestly, if you've never used AI for even this one narrow use case, it will blow your mind. So there's no area, I'd say, almost of your life where AI couldn't, in some way, shape, or form, even make you better or faster or something. [SPEAKER_00] So when we first started, we were able to duplicate my brain, which was phenomenal, which I still think it's better than – because right now, AI scrapes off the internet. And like, for example, if I were to learn about AI, I don't know if I'll use the regular sort of AIs to learn about AI, which is on the surface, as opposed to taking your brain and how you think it's far better. Like, if Jack had to do this, what would Jack do, right? And that's kind of what I did, and I exploded. And then we were kind of going down the rabbit hole of developing tools based on people. [SPEAKER_00] And then I eventually stopped because I said to myself, you can't compete against the Clods. You can't compete against the Groks and the Geminis and the Chachapitis because they'll all be worth over a trillion dollars. They probably have 15,000 to 25,000 engineers working seven days a week all over the world. And if they want to, they can just instantly replace you. And you're seeing that in a lot of cases, unlike people where Apple needed the ecosystem. [SPEAKER_00] But if people are in that ecosystem, it's a lot easier for them to put these tools. And I think tools – and I say a lot of people set up these tools. I don't think a lot of them are going to make it. And so what I pivoted to, and we've done extremely well very quickly all over the world, is, okay, you're not going to learn. You're not going to do it. Fine. We'll do it for you. And I think the AI economy is important, but I think the AI implementation economy. Because in theory, what you're saying is correct. People could go by themselves and do it. [SPEAKER_00] But how much time and effort they're going to do it? And then you have this kind of notion. People can do Facebook ads themselves. They can do meta ads themselves. They don't. [SPEAKER_03] Yeah. [SPEAKER_00] Right? And so how much of this is human nature? Because you're studying. You're learning. And, yeah, it is your business. And you are on a hot streak as well. But you're obsessing over it. And so is obsession an important part to really master – if people really want to succeed? I mean, you could be a drive-by AI user. But if they really want to amplify their life, I think AI doesn't replace you. I think it amplifies you. It can amplify your good people. [SPEAKER_00] How do you view the lack of thirst? I'm sure millions of people watch your video. But how many people actually go and do it? And that's my point. Not because your content wasn't good. It's just it's easier to listen than to actually do. [SPEAKER_03] 100%. And you touched on something important as well I wanted to double click on, which is the idea of you can't compete against anthropic and open AI and grok and meta. And we're seeing them – if you think about the kind of value stack, if you think of the AI space, right? We have chips. We've got even the models themselves. We have the application layer. And both open AI and anthropic are moving for virtual integration, right? So you've got anthropic now looking at chips. Open AI looking at chips. [SPEAKER_03] You have open AI who just released – I think it's chat GPT law. You have Claude doing Claude design. So any businesses built in this application layer, which is where I think the long tail of value is, right? These like really cool second brain that you built, for example. That is where a lot of the value actually sits. So I do agree. It's really important for any company who is AI enabled or using AI to just think about where open AI or anthropic could pivot into and how that might affect it. [SPEAKER_00] And I think it's going to be hard for people to hire somebody who could do it all. I think my advantage is I understand AI, but I understand business, having many businesses as well. So I know when I look at a business how to work it. So on one hand, we have these tech people who are really good at tech and understanding. But they're so techie that they can't understand it to the owner who has no idea what she or he is talking about. So that's an issue I have. [SPEAKER_00] And then they've got these young ballers with the little skinny ties overselling AIs with their little fake Lamborghinis, overpromising AI, making a lot of people look bad about this. And I think there's a lot of people overpromising what AI could do. Like it could close all my leads and it could do all my marketing and it could do everything and do all the podcasts. And I'm like, why am I here then? Like why are we both here? Like if that was the case, right? Definitely. [SPEAKER_00] So I think hiring somebody to now say, okay, you're now trying to implement AI everything and understanding business, that's one hell of a skill set. And I imagine that's a very high six figure kind of or mid six figure jobs because you can know how to create an AI on play, but you know how to connect it as an automation. You know how to create a workflow. Do you know that the customer journey, right? And so it is, there's a lot of complexity to it. [SPEAKER_03] There is. And it's also, the skill isn't just where to use AI. It's where not to use AI. That's true. And I think there's a little bit of a stupefication happening where you delegate a lot of your high octane thinking to AI when you shouldn't. And you can find this yourself actually, if you just ask AI to do everything, you'll find your cognitive capacity for that task just naturally decreases over time. It's a big thing with kids actually. Like you're just asking chat GBT for the answer. So they never actually go through the process of using their brain. I think that's the big difference yet. It's the stuff that you don't use it for and having that discernment. Because you're so right. [SPEAKER_03] AI cannot do everything. And we've been told for a long time now that AI has destroyed, in the vert commas, insert the industry. A lot of times it hasn't. And that's why I think you need to be in it to understand its limitations to properly get the most benefits out of it. [SPEAKER_00] But in the end, I can type something in business and I can know what's real and what's not. Yeah. Right? Just like if you type in something, you're like, nah, it doesn't really work that way. Right? And I look at this thing. But people are putting in there. And if you don't know the answer, you're taking it as absolute truth. [SPEAKER_03] That's true. [SPEAKER_00] And that's the problem. And I think, and we did that certainly in our AI, that everything that we did was put through a second opinion filter to make sure, like, what's the contrarian thinking about this as well? Because what people don't know is AI knows me. No, AI doesn't know you. And I think it agrees with you. I think that's the fundamental problem. People love to put an idea. And it loves to say, wow, what a great idea. This is going to change the world. Like I had this guy just recently. [SPEAKER_00] And he's trying to pitch me, you know, mining capabilities for crypto mining in Ethiopia. And he basically whips out his phone. He goes, hold on. Listen. Listen to this. Right? He's taking this cloud of like, and you, this opportunity for crypto mining in Somalia is one of the greatest opportunity because of this. So I'm asking questions back. And whatever I ask, it was just doubling down on his point. So I take my AI and I say, listen, because it knows me very well. Look and do some crypto mining because I have experience in it. We've done it. [SPEAKER_00] In Ethiopia right now, good, bad, what do I need to know? Well, first of all, right now, there's only been, in 2025, there were 27 companies who started crypto mining. The majority, Russian and Chinese. Being in business, doing business in Africa, I know there's probably a lot of deals with governments on those things as well. Right? That kind of hand. [SPEAKER_00] The other thing that we looked at is that the water flow intakes were down 20% because it was based on the river on hydrogen power. So now the government was putting a standstill on them and minimizing it because it was prioritizing its village people. Second of all, of the 25 that were in business, there were only 11 left operating. And we're talking about crypto mining facilities probably in less than a year. Yeah. [SPEAKER_00] So it said extremely risky, volatile, getting the money out of the country, banking, all these things, making it almost impossible. And very easy for the government to, if you do very well, to shut it off, to move it to everybody else. That's the difference between someone going into their AI, having to say how it's a great idea versus mine, who just basically broke down by the facts. And I think that's the issue with AI right now. It is meant to agree. People should look at AI as to saying, where am I wrong here? What am I not seeing? [SPEAKER_00] What are my blind spots? Right? So would you agree that? And with my assessment, and where do you think people are using AI wrong? [SPEAKER_03] So I agree. I think we call this sycophancy, AI sycophancy, that it will just agree with you. And if you're an expert in any field, just like you have, if you ask AI about it, you will immediately see its flaws and its logic. So it's helpful. You are correct. And most times say, help me understand where I'm getting this wrong. A lot of, I think, the best AI comes from meta-prompting. Essentially having it create, like pre-questionnaires, for example, or even debate other models. Because AI can be very confidently incorrect on something. [SPEAKER_03] And it can't even identify its own weak spot. I can't tell you a number of times that I've had AI building me something. And it will swear until it's blue in the face that everything's fine. Like, great. Do an adversarial review of a different model. And all of a sudden it finds, like, six critical issues that would have, like, absolutely destroyed things. I know. And going back and forth. [SPEAKER_00] And it's funny that you'll get different answers when you're different models as well. [SPEAKER_03] Exactly right. 100%. They have a different feel to them. A different feel to them, actually. We were talking today about, actually, on a different podcast. Obviously, not as great as this podcast. [SPEAKER_00] Yeah, I don't know. You're cheating on me. I'm absolutely already, man. Did you wear the same thing? Sorry, I should have. Did you wear the same thing? [SPEAKER_03] I did, actually. [SPEAKER_00] Oh, man. We got to release this ASAP. I got to have the Chris Martin walking down the street. I'm trying to be low-key, dude. [SPEAKER_03] But it's like, you know, we had a shrinkflation, right? Like, the whole Toblerone example is a classic one. We think sometimes it happens with models. Like, even the latest Fable model, for example. Like, we've looked into the data. You can't put token usage. It's getting slightly worse over time. So there is a vibing and feeling that goes on with models. Even if it stands, you know, releases at a certain standard, it can decrease over time. [SPEAKER_00] One of the things that I advocate is when people get into AI is that we start solving the easiest, simplest task, especially in AI implementation. Because you talk to the owner. I guess maybe they're excited. They want everything done. And then you do everything. Something's going to break. It doesn't work. They're going to lose faith, and they're done. Second of all, you also, if they have employees, you have to get the employees on board because they are now fearing that their job is at stake. [SPEAKER_02] Yes. [SPEAKER_00] Right? And so I think people trying to go too big too soon, and for anybody watching this, whether you're doing it yourself or do it with us or anywhere else, I think too big too soon will ultimately fail. And then you're going to have a very bad taste. It's like hiring a marketing agency who overpromised you and underdelivered, right? They're like, we're going to do this, do this. And then you give them $10,000, $15,000. And when it doesn't work, they blame your offer. But I'm like, that was your job to come up with the offer, right? And right now, AI is not like that. [SPEAKER_00] But I say within a year, year and a half, there's so many people who would have overpromised, messed up, that it's going to be personal branding is going to be important. Results are going to be very important. Not even testimonials, but very detailed case study. And you're going to be very proficient. So how does, you know, get that whole model, getting people to buy in into something that people know could be the end of their job, even though we're telling them it's not? [SPEAKER_03] You know, that's such a good question. So before I was an entrepreneur, I did like a- [SPEAKER_00] I bet you that guy didn't ask you that question. No, he didn't, actually. That's great. [SPEAKER_03] That's a unique one. So I did a strategy role in a very big company. You know, they had about 20,000 people in that company. And so I've experienced it from the corporate side and also from the startup and tech business side. And one thing you have to establish from a cultural point of view is that if you do automate your job, your job does not disappear. So you have to show me the incentives and I'll show you the outcome. Everybody must be incentivized to win if they automate their job away. [SPEAKER_03] So for example, if you build a report that you suspend, say, 20 hours doing, it doesn't mean that you're going to be fired or whatever it is. So first of all, from an employee level, there has to be alignment. And if you bring in AI and automation, you will either get a pay increase or, you know, your job won't disappear. It has to be driven at the individual level. So it is bottom up and also top down. So you need strategic AI direction for the business, which requires a very specialist skill set. [SPEAKER_03] And then every individual needs to be what we would call AI native, leaning to and embracing AI. Because here's the truth. They know what they do better than you do because they're subject matter expertise depending on the size of the organization. So they have to be AI first. And they need that, I'd say, environmental safety and incentive to bring AI to the table. [SPEAKER_00] Give me your five biggest ninja superpower things that you think right now work really well, right? If you only have five weapons to get the most amount of output in their life and their business, what would you, you know, what do you think you like the most? Or what's really cool that is just like right now that it does that people may not even know it does? [SPEAKER_03] Okay, for who? For business or just like an individual? [SPEAKER_00] Listen for business since this is a business show. [SPEAKER_03] Okay, let's think about this. [SPEAKER_00] Now, while you're thinking, Jack is wearing a jean jacket back from 1987 Leis. He's also wearing a Coldplay yellow shirt worn by Chris Martin and wearing white jeans. Just got back from Budapest. And at the end of the day, he may be the most right-brained AI I have ever met individual where most of them are very techie and geeky. He happens to be the creative type. Now, back to Jack answering the question. We're breaking the mold, yeah. So I think- You definitely are breaking the mold. [SPEAKER_03] First thing I think that I would highly recommend is you need a personal operating system. You need a knowledge brand that effectively, on the premise of this, the more context AI has, the better it performs for you. So you need to operationalize all of your knowledge. This is data. This is systems, tools and integrations in one system. Such that when you ask it questions, it knows about your business strategy, who your customer is, every email you've ever sent, every meeting you've ever had in one central brain, one second brain, which I think you've built a beautiful business around that as well. [SPEAKER_00] So that is- More centralized, more private. And I think there might be own company LLMs or like, hey, we have our own AI system because maybe later on it'll be cheaper or more easier to do stuff like that. So your knowledge and who you are, you don't want to have to retrain it every time because the people are like, well, I need to start over, right? And that's usually the excuse like, oh, I'm in cloud. I'm going to go there. But then as we're seeing, Claude, ChatGPT, Claude, for some reason, Gemini is not in the conversation. It was really earlier on. [SPEAKER_00] I tried the Grok bot. I didn't like it. I feel like it's a little, the bubbles and you know what I mean? Like I could easily type what I want and get the same answer. I don't know. It's just for me, it's just not doing it for me right now. So let's click on that, right? [SPEAKER_03] So you have, think about it because people get confused. You have AI that you see your building, ChatGPT, Gemini. Then you have the personal AI assistant, classically known as Alexa or Jarvis. You have Hermes Agent was very big. You may remember OpenClaw. Is it dead? [SPEAKER_00] Is that dead? [SPEAKER_03] Well, it's fallen out of favor. You know, they got acquired by OpenAI, but it's fallen out of favor. [SPEAKER_00] Is it because they just let it die or just the security was always the big issue? And Hermes came out, which by the way, Hermes got built by NVIDIA, right? [SPEAKER_03] Um, Hermes, no, Hermes was built by a company called Noose Research. Um, so they're different. What made them unique, uh, in particular was the, it was the agent. The more you use it, the better it got. It had some very unique memory systems and people didn't really quite fully understand. How do you host this? How do I get to run 24 seven? The problem that Grokbot solved, which is really cool is it strips away a lot of the complexity. So, you know, if you're Mrs. Higgins, who's never touched AI before, you can use Grokbot. You can push buttons and connect to things. [SPEAKER_03] So it's a very, it reduces all the friction from science. [SPEAKER_00] I know, but, but I, I can put the key in the ignition, but if I don't know how to drive it to win. [SPEAKER_03] Yeah. You need to understand how to use it. I mean, Grokbot, it's got some cool things to it. I mean, the agent can create other agents. It's almost like a, think of it as a. Yeah. [SPEAKER_00] But I had all these people reporting these people, but even like, Hey, search me amazing AI people to interview. Your name didn't pop up. Um, you know, your big presence and, and it just wasn't, I don't know. I just, it wouldn't like, maybe because I was a better here. I was better rather than having automatic was better shifting. And, and yeah. And maybe because I use it in a different way that kind of like, uh, like Max Verstappen saying, I don't like the new design of the new F1. I like the old school one as well, because I know how to make those micro changes. [SPEAKER_00] Just like when you have your laptop, you know what I mean? You're like, I can do this and do this. And maybe that, but I think if people want to win in their business and their life, they, it just can't be like, Oh, it's just so easy. And you do this where it just does the bare minimum. It has to make money. Yeah. You have a lot of companies right now who have said, Hey, you know, they haven't seen a return on AI, right? We have, but a lot of people say they they're putting the ad. [SPEAKER_00] They haven't seen a measurable return on AI and you get to a point is that we cannot see how it improves, you know, saves times, make money, make them more profitable, spend more time with their family because of this. You know, it'll just be another thing. Even though it works, doesn't mean that people know how to work. [SPEAKER_03] It's AI is not, this is the best way to think about it. AI is not a magic wand. It is an accelerant. So it will, if your processes are broken, it will break them a lot faster. If it already works, it can enhance what's already there. But one thing I think, if you think about AI, there are really four vectors, right? You have performance. How good is the thing? How much does it cost? How quick is it? And how private is it? And based on the model you're using, you're usually doing some kind of trade-off. So you think about local hosted models. [SPEAKER_03] If you need to be GDPR compliant, HIPAA compliant, SOC 2 compliant, you may want to host things locally. What's really exciting about local models is they're only really about three to six months behind the current frontier model, which is incredible. So one thing to put on your longer term roadmap is what I would call the AI lab for your business, locally hosting these models that are 100% private. You can run some pretty powerful models. Are they running on a cloud or are they running on a Mac mini, let's say? So they could run on a stack of Mac minis. [SPEAKER_03] I recently were buying up loads of Mac minis. So you can build all these beautiful computer towers and it runs, you know, essentially the only cost is the energy bill. But again, it's 100% private. Your data doesn't even leave the building, actually. It just runs all that. [SPEAKER_00] There's a lot of GDPR clients that's all over the world that are concerned about data. Although people are concerned about data, but they're uploading. They think because they paid the $20 or the $200 that, you know, chat should be to your cloud, is not training the model, that if actually going through their settings, they'll realize that you've already given permission because it was a toggle they already turned on. Exactly right. Right? Exactly right. Which is crazy because there's been a couple earlier thoughts, ideas that I had. One of them is I used to compare AI to Formula One, you know, the efficiency, the seconds, the speeds. And I've seen it regurgitate the exact same thing. [SPEAKER_00] And, you know, so you get that thing. What's another ninja trick? So local model, we just did one. The other one is to create the knowledge, central knowledge base. [SPEAKER_03] What's third? Okay. I'll give you third. Also, that point you made about AI kind of weirdly training your stuff. We had this with the Navy Stokes. I don't know if you follow that story, but one of the biggest sort of math problems, individuals were chatting. They'd just solve, yeah. Yeah, they were chatting to OpenAI about it, sharing their research. And they can't necessarily not use that to basically grow and improve the models. So if you're feeding IP into OpenAI or any company for that matter, that information may itself be used to improve it. [SPEAKER_03] So even if you give it a novel or unique way of explaining an idea, you may see that present in future models. So that's just worth bearing in mind. So you're not going crazy experiencing that. So let's see. We've got user models. We've got agentic operating systems using your brain. We talked about local models. You have personal assistant as a category, which is a slightly different one. That is essentially your ability to think about what are the modalities and the ways in which you interact with AI. A lot of work, so to speak, happens at the laptop. [SPEAKER_03] But one thing to think about is what about when you're driving the car? What about when you're walking the dog or grabbing a coffee? Your ability to do things productively on your phone is also massive. Is that what Facebook or Meta just served with Muse? Yeah. So Muse is a slightly different kettle of fish. A good example would be Hermes agent. It would be OpenClaw. It would be Grokbart. It would be – I mean, another one – [SPEAKER_00] I tell you, don't worry. I'll do it for you. But I'm just saying, like, it just – like, you live and breathe it just like I live and breathe business. You know, I think you have two choices. You either get the best to help you and you basically listen to a show like this and you listen to you and be like, hey, I would like that. Or you talk to AI and say, here's the idea I heard from Jack and JT. Here's my business. Here's what I want to do. [SPEAKER_00] And then you're like, hey, can you do this? Right? And be more at the frontier of the idea. Because people could do that. I mean, think about it. We used to Google and go through all these crap pages just to find out what's good and what's not good. Yeah. And now this is Google that actually does the work. I don't think people realize. And you know what? I used this story before. I think if they had come out and said AI costs a million dollars, it would be completely different. Think about this. AI costs a million dollars. What do you mean to people say? It's only for rich people, rich companies. [SPEAKER_00] And I probably would have paid a million dollars for something like that. Right? Yeah. And so what they did is – and then, you know, like, all right, we normally have a million-dollar car, but we've created another version. Right? Like, another version of this, you know, for you guys at $100,000. And then eventually, like, you know what? We're going to open up a version of it at $20,000. Right? And it would have done that. But except that what it is, we're going to make it free and charge you $20 or $200. And we're going to get as many people there. [SPEAKER_00] You're going to tell everybody all your darkest secrets and what you like and don't like. And then we're going to get advertisers who literally can market, hey, you have a problem with hair loss. Bing! [SPEAKER_00] Hair loss product automatically appears. Logically fixed. And by the way, they know you better than anything. Before in Meta, I'd be like, hey, you know, maybe you like going to the beach or maybe you like Elton John because you like the post or you like this type of thing. And the algorithm trying to figure it out. Now they know you. And as a result, all of them are worth trillions of dollars because you fed them with the model. And I think – I don't know if you know that. Do you know what stat people use free AI? Tell them it. 95% is the free AI. [SPEAKER_00] It's not even that much money if you think about it. 95%. Which is basically said anything you put in there is training your model no matter what. And so I think fundamentally – think about this. The AI is being trained by a bunch of people who are too cheap to pay $20. And that's the information we're gathering. And that's the information people are taking. Well, this is a great point. [SPEAKER_03] Only 5% of people are on the paid plan. It's crazy. And then of that 5%, those that are building with it are even smaller. So you're simultaneously – you're not as far behind as you are. [SPEAKER_00] But it's like Dropbox, right? Only 4% there and they're worth billions of dollars. It's crazy. It is. It is crazy. Because I think a lot of them is money is made through enterprise contracts, $150,000, $200,000 a year. You know, fascinating stuff. Any other Ninja, final Ninja tricks? [SPEAKER_03] Yeah, one important message actually. And that's the fact that we are living in an era of informational abundance. I mean, if you think even 200 years ago, if you wanted to know something, you went to the guy buried in books at the top of the town. Now essentially AI knows everything. So I think any information that you need is there. It's really about the right questions that you're asking it. [SPEAKER_00] Yeah. Well, I'm very curious by nature. I've never had an issue ask a question. In fact, I think there should be like a course on how to ask the right question. By the way, I hate the word prompting. It prompts me to give me like this because I was like, I'm a prompt engineer. Like, no, you're not. Yeah. Calm down. I call it command. I command AI. And I'm just trying to make command great again as opposed to prompting. I like it. People are like, I'm a prompt engineer. Like, dude, shut up. I'm a prompt engineer. Like, come on, right? Like, it's just like, it eats me alive. I guess it's like Pavlov. A book bag. [SPEAKER_00] So yeah, so change the command. Do it on your shirt. It's like, you know what? I am now, even if you want to take credit, I will give it to you. But if we can make humanity and stop using the word prompting this. Let's go back to something. A lot of people said the word command, aka prompting, is dying because people are now, like, looping is a big thing. Right? Can you explain what that is? And if that, is that true? Yeah, sure. [SPEAKER_03] So for example, let's say you build a website with AI, which you can do spectacularly well. And oftentimes it gives you a website. And then you'll say, great, I want these changes, these changes. And you go back and forth in a loop with AI. If you use a loop function, effectively, you say to AI, it's very interesting. You say, this is the end goal that I want. So I want the output to be this. And if you give it a reference, it's even better. What AI will do is go in a loop with itself. So it spins up what's called a sub-agent, which is just fancy speak for another agent, but it speaks to him. [SPEAKER_03] So one AI builds, and it comes back to the original AI and says, hey, has it met the three requirements? Does it have a beautiful 3D animation? Does it look exactly like this? Whatever your requirements were. And the AI, instead of coming back to you for 10 separate conversations, talks to itself for 10 rounds. And the actual output that it gives you is completely different. I've done a lot of content. I've done a lot of content, I've viewed by millions of people on design with AI. And I can tell you, using what's called a design loop, giving it the outcome that you want, giving it a reference, [SPEAKER_03] and let it run in a recursive loop for 10 conversations, the level of output you get is unbelievable. You wouldn't even believe. I can effectively give it any video, any asset I want to, any website I want to, and it will create it. It's very, very, very different than just one shot, but it also works for any problem if you give it the right output. [SPEAKER_00] So when people look at you, they would never think AI. I mean, that's the first thing, and I'm sure you get that as well. How did it all start? Are you really a tech guy that just like, what's the background? What's the story? Because I think to not have this show and talk about it, it would be almost like, dude, you didn't ask him. Where does it all come from? Because I would never know. Which, by the way, that's maybe why you're so captivating on social media, because you're so good at explaining things. And by the way, very short and sweet. [SPEAKER_00] You're like, very short. There's no run-on questions. Thank God I'm good at what I do. I'll be like, Jack answered all my questions. Right? And so what's your story? How did you, when did you realize that AI was what you wanted to do? And what did you do before AI? [SPEAKER_03] Yeah, sure. So I started off, I did European international law. I did team leader, project manager, program manager. I was super formal. I thought that was what success looked like. And then I made a bit of money in corporate. I did some really cool strategy jobs in corporate, and I was crushing that. Did you like corporate? No, I didn't. Is it because working for other people? Yeah, you know, meetings about meetings, just hour-long, you know, cube. It's just very, very inefficient. And I realized if I wanted to get to the point I wanted to get to, I needed a different vehicle. I just read Rich Dad Port at the time. [SPEAKER_03] This was about eight or nine years ago. And I was like, you know, I need to be an entrepreneur if I want to get to where I want to get to. So actually, I founded a company. I ended up being a top 100 UK startup. I sold that. That was really awesome. What was it? Basically, it was a Relotech startup. So it basically helped people automate the process of moving home. So it saved them 15 hours of time. So that was really cool. What did you sell it for? Well, it's undisclosed, but like it was cool. It wasn't... Did you walk away with millions yourself? Say again? [SPEAKER_00] Did you walk away with millions yourself? [SPEAKER_03] No, it wasn't millions and millions. [SPEAKER_00] And then you had to pay half in taxes in the UK. [SPEAKER_03] Well, it's basically 20% is a standard couple of gains, right? But you can get entrepreneurially for 10% if you want to. It was cool. So it wasn't a huge exit, but it was a cool thing to get started with. [SPEAKER_00] Did you regret to like sell at that point because you could have held on to it and done more? Or, you know, sometimes there's like that buyer's regret, seller's regret of a company, or you were just happy to move on to the next thing? [SPEAKER_03] No, it was really cool, man. I mean, like basically it was interesting because most people would even raise money, but we sold over 50% of it, which is cool. And then the rest later, which is quite nice. And then we worked with the company. It was a really good, really good deal. Really good business. What was this? It was with a law firm. Then it went when? What year? [SPEAKER_00] Oh, this was 2021, I think. Oh, wow. So this is like post-COVID. So this is... And during the market was really hot. There was a lot going on. [SPEAKER_03] Yeah, well, we didn't even plan on that happening. So basically we got mentioned for Times newspaper. We had an interview. And then someone read that and they were like, hey, go speak to that company. And then I went to the head office and we had some conversations. And then we ended up doing some stuff. So that was really cool. And then after that, I started making YouTube videos. About what? Just any stuff I was interested in. I was just really passionate about it. And then my friend sent me a text message saying, hey, have you ever heard of this thing called ChatGPT? And I was like, I've never heard of it. What is it? And then I pretty much started then on the ChatGPT thing. [SPEAKER_00] What year was it? 2022? [SPEAKER_03] Oh, dude, yeah. It must have been 2022. But interestingly, the tech startup I did was built using automation. So I was doing automation. [SPEAKER_00] Yeah, automation is a form of AI, which people don't understand. [SPEAKER_03] They didn't call it AI, but it was a form of AI. Yeah, dude, we started using Zapier. Or back when Make was called Integromat. So I'd been doing this automation thing for ages. And then AI stuck the blood with GPT-3. Then I just became obsessed with it. And I've been literally spending all day every day with businesses and AI. Like 24-7, basically, since that happened. [SPEAKER_00] And what is it you like about it? What, on a way, turns you on about AI that keeps you excited? Talk about it, doing it, living it? [SPEAKER_03] Oh, man. It's just you're literally at the cutting edge. As technology gets better, it's such a creative thing. Finding new applications and use cases for it. And stretching them to their kind of, as far as they can go. [SPEAKER_00] That's what I find really exciting about it. So, you know, when we had a conversation prior. Not about this. Everything we talked about here, we didn't discuss. In fact, we didn't know what we were going to talk about until we got here. Yeah. People were like, what are you going to talk about? I don't know. Let's go fluid. I'm like, is there any other way? Absolutely. That I get the feeling that I'm a 24-7 machine, right? And you're more like, hey, love my YouTube channel. Put a lot of thought into content. Building my brand. Putting myself out there because I enjoy it. And obviously, you're good at it. [SPEAKER_00] And then I'm going to take a few projects here and there. Choose the clients that I want. And I'm more like going to teach people, do it yourself. As opposed to building a massive AI company. Which I think you and I both know, being in the business, it's a lot. You're still dealing with people. You're still like, you know, I mean, it's not easy. And in AI, things break all the time. And customers are mad all the time. Even though they could never figure it out yourself. Is that your vision to really balance the lifestyle? [SPEAKER_00] Because you can go down to a rabbit hole. And what is it that your vision is? [SPEAKER_03] So I also have a SaaS, which is a speech-to-text startup, which is really cool. So I'm enjoying that as well. My vision is, yeah, I want to essentially educate. Well, my form is simple, which is eyeballs times value. I just want to add as much valuable as possible to as many people as possible. And educate them on AI. No fluff. No crap. [SPEAKER_00] This is no fluff. This is education that moves movements. Exactly. You sound like if I get a chat. I call it broke GPT. That was broke GPT. If you see no fluff, you'll know where it came from. [SPEAKER_03] Maybe I've been poisoned by a chat GPT. Who's to say? [SPEAKER_00] Right. But you get, because you can detect languages. I mean, you use enough that we can tell. [SPEAKER_03] I can tell very easily. You can tell. I can tell very easily. I mean, like, so there's this new model called Jav. Without going into the details of it, because it's pretty technical. But the point is, like, you could probably do what we call AI slop detector. Like, you're scrolling through LinkedIn or X, something looks like it's written by AI just auto removes it for you. Because a lot of these platforms are overrun with it right now. It's crazy. So knowing what is written by AI. I mean, we live in what I'd call a one-page era. Like, if something's longer than the page, you're not going to read it. [SPEAKER_03] I mean, think about one's last contract that you really read cover to cover. [SPEAKER_00] No, everyone just does the same thing. They put it into there. They copy and paste what it is. And they send it out to people. Yeah, 100%. That's exactly what people do. [SPEAKER_03] I mean, I had an agency. I had recently sent me a 10-page report. And I was like, great, I'll have my AI summarize it. Because they also didn't write it anyway. So it's like, if you want it to be read, it needs to be written on a page or less. Like, that's the reality where we're at right now. [SPEAKER_00] And it's so hard because sometimes, especially after you go to the deeper research, it just vomits so much information. Exactly. And then you're just, like, too ADD because you're like, the next one, the next thing. And you're like, I just worked all day and I got 120 pages, which nobody's going to read. [SPEAKER_03] 100%. They don't read it. It's the same with websites. Websites are skimmed. They're not read. So you have to change the way that you're working now for that specific reason. [SPEAKER_00] And so I want to, first of all, without this individual, this wouldn't have happened. Although I found you. And I sent it over to him as well. His name is Dennis. He's been on podcasts many times. We do some M&A and AI stuff together. And I wanted to come here to ask you a couple of questions. Because he is one of those guys that, like, he'll use the AI and he will, like, literally figure it out. Like, I remember when Opal Call came in. I spent 15 hours, like 20 hours a day figuring out all weekend. [SPEAKER_00] And I'm like, not on my bingo cards this weekend. Right? And then, like, I'm going to set up an AI employee in its setting. Like, 99.9% of the people would not do what he does. Right? And so there's a specific type. And that specific type is more likely to, like, do it themselves and implement themselves. But it always changes. And it's very frustrating. But, you know, he's in it. I would say you put 1,000 people. He'd be the only one to do it, if not higher. [SPEAKER_00] And I think you see that before. Because it's not perfect yet. Right? So come on here. Ask him a couple of questions as well. Dennis, come on in here. How's it going, man? [SPEAKER_01] It's going well, man. A week and a half. We got on a call. Then we got another call. And now we're in Dubai doing a podcast. I mean, what a quick turnaround of events. Very quick turnaround, man. It's great to see you in the flesh, man. Good to see you in the flesh. You look good. [SPEAKER_03] Thank you, Brad. I appreciate that. [SPEAKER_01] The way that I work with business owners all the time. There's hundreds of business owners that I work with. I've done 130 plus transactions in M&A. The way that we think about AI and the way that the business actually operates are completely different things. And the way that I think about AI is the current state of the business model is tons of paperwork. It's tons of charts, copy and paste, Excel, Word. And now it's a completely different shift. Like you have to learn how to use a completely different system. [SPEAKER_01] With a small business owner wearing too many hats, it's impossible to just enter into the cycle. How do you think about that implementation for the small size business owner, maybe 10, 12 employees who's extremely understaffed? And yes, you can use ChatGPT, but we know that's not implementing AI. Talk a little bit about it from the business perspective. [SPEAKER_03] Okay, cool. So this is a good question. So I've had thousands and thousands of business owners in my communities that we've helped successfully. And I'll repeat back what I think the biggest pain point is and the best advice I think I've got for you would be like, you have to start with the business problem first, which oftentimes is what is the actual problem we're solving? What is the bottom that we're solving? So we're looking for basically a solution that solves a business problem rather than just using a technology for its own sake. [SPEAKER_03] Like I can give you Hermes, Agent, Grokbot, all these different systems, but it has to actually start with a problem that you've got. So number one is what is the problem that we're actually solving? The second thing we have to understand is like what are the tools, systems and processes behind that? And then only when we actually understand what we're trying to do, do we bring an AI. Actually, the biggest mistake I've seen a lot of people make is they automate something that quite frankly just should not exist. Like, you know, a classic example is you want to go on Instagram, you want to go on YouTube. [SPEAKER_03] I have AI generate me 100 thumbnails. It's like you don't understand what you're doing with thumbnails or you don't understand what you're doing with Instagram. So you have to actually start with what we're doing. [SPEAKER_01] Or you don't have the CRM or you don't have the systems to automate. [SPEAKER_03] Exactly. So it's like dropping like a super Formula One engine into a car that's just pointed at a wall. Like you're not working on the right thing. [SPEAKER_01] What about larger companies? Like when you have different types of departments and you're thinking about implementation, where do you start in your mind? Like which employee do you start working with? Is that, do you tackle the marketing usually? Do you go after the communication because that's usually the first step? Or do you try to integrate the CRM? How do you go out looking at the implementation on a little bit of a larger scale? Like before we start training the employees, we have to understand what the employees actually do in our organization. So talk a little bit about that. [SPEAKER_03] Such a good question. So number one is culture. That is like absolutely foundational. But everybody, we have an AI first culture. Like we were speaking earlier about the importance of the... [SPEAKER_01] Oh yeah, absolutely. You're coming into an organization that you are the risk to their employment ultimately. Yeah. [SPEAKER_03] I mean like I can tell you right now, like even if the executive leadership want great culture, it's not enough. You need everybody on board with your vision. But my point is like you need to align incentives such that every employee is incentivized to make their role more efficient. Because if they don't, their incentive is to keep their job. And why would I automate my job if it means I don't get paid anymore? So you have to think about it from that point of view. In terms of like where's the biggest ROI, that's going to change quite significantly based on the business that you are and what you're focusing in on. [SPEAKER_03] You know, if you think about what does AI do naturally, well, you know, it creates language. So anything to do with marketing, sales, boring back office processes. Like, you know, I sold an automation at one for years. And what we did is we looked at the highest profit center of the business. And we said, how can we make this better? How can we make this less inefficient? So ultimately every department needs its own AI strategy, which I'd put on the department. [SPEAKER_01] Because they're all using it independently. And they're all storing that information somewhere else. Ultimately, all these companies have tremendous issues. [SPEAKER_03] 100%. People are using AI anyway. Yeah, absolutely. So you need to be in and around. And I think, you know, use AI isn't going to be like a new thing. People are going to have to use it. It's how do we use it? I think over 90% of AI, when I did the math, over 90% of AI consultancy projects failed. Because they didn't actually have the correct roadmap or strategy to implement it. [SPEAKER_01] And what size companies? [SPEAKER_03] I believe that was most, actually. Even significant ones. Like, you look at the ones for whom PwC did consulting for, all those big consultants. They're not doing it. I can tell you as well, when I work for a very large corporate and strategy, that a lot of them, there's what you want to do, but they're just not mature enough. I could give you the perfect roadmap to implement AI. [SPEAKER_01] They promise everything, but they don't have the foundation. [SPEAKER_03] Yeah, they're not mature enough to implement it. They don't understand the data. Some organizations are so locked in with their data, they can tell you with 95% accuracy whether somebody will quit or not. And some couldn't tell you that at all, even though the data is available. So it's organizational maturity is also a big one. [SPEAKER_01] Well, I enjoyed this conversation, Jack, and looking forward to doing more stuff with you. I'm going to let JT close this up. Excellent. Thank you. Good chatting with you, man. You didn't say that about my questions. [SPEAKER_03] They were also great. [SPEAKER_00] Yeah, but you said beautiful. Like, great. What's more? Is it great or beautiful? Hi, you're great versus you're beautiful. I think beautiful is a better term. Yeah, it is. [SPEAKER_03] Maybe we'll get Chuck GPT. Yeah, you know what? I actually, you know what? [SPEAKER_00] I'm going to use my, I'm not using broke GPT, but let me ask, let's see what, and let me ask it. All right. So if someone said my questions were great in my podcast versus my questions were beautiful to somebody else, which one, if you had to take one, that is a better question, that would be like a better, that person liked more. I think that was a good analysis. [SPEAKER_03] Absolutely. Okay, let's see. [SPEAKER_00] What does the audience say? Do you want to put it into the comment? If I had to choose, your questions were great, a stronger compliment. Your questions are beautiful, sounds more emotional and personal, but it means the questions were thoughtful, sensitive, or elegant. Great questions were more directly. You're strong interview. You asked smart things and the person valued the conversation. All right. I'll take great. There's the debate. There should be like the headline, great or beautiful. [SPEAKER_03] From good to great. [SPEAKER_00] You know what I mean? So would you rather be beautiful or would you rather be great? Well, first of all, thank you so much. I want to put all the details of your work. People can find you as well. I find you refreshingly entertaining in terms of AI in a conversation that sometimes people get down to the weeds. I think you simplify it. I think you make it fun. I think you make it engaging. And I think that's the whole thing. It's kind of like you want to get into this, you know, and I hope people take that and do something with it. [SPEAKER_02] Yeah. [SPEAKER_00] And not just be entertaining because that is the point of all of this to make people's lives better and be fun as well. Well, we're going to have to do part two. Perhaps we should do another part of actually break down live in front of everybody some ideas that we have. And if you can do me a favor, it doesn't cost any money. If you could subscribe, like, share this to someone who needs to hear it. Again, it helps the algorithm. We've added value for an hour here. You can just do that for us. Like, click, mark, share, whatever, comment. Tell us what thing you think of Jack. [SPEAKER_00] Just like, by the way, do you like Jack's outfit? I usually wear a suit. I decided to dress down because I only have one suit in Dubai and I'm saving it for tomorrow. Not that I didn't want to save it. We just got back from a meeting. And interestingly enough, we were with some very successful people and they brought us to a private club and stay behind. And I'm like, I got to go. And the guy's like, no, stay. I say, listen, this guy, we've been waiting for a long time. We're going to keep our word. We're going to keep our word. I'm sure we're going to need to meet this other person. We're going to keep our word. And we kept their word. I'm glad we kept their word. That's an example to you. [SPEAKER_00] There may be things sometimes you think it's better. But what's better than beautiful is keeping great words together and keeping great because there's nothing better than being great. And who's great on this show? Jack Roberts. We'll see everybody on next time on the 1% because this was 1% of 1% of greatness led by the great Jack Roberts. We'll see you on the next show. Thanks for listening. [SPEAKER_03] Thank you, man.