Welcome to the Particle Accelerator. I'm Ben Johnson, founder and CEO of Particle 41. Back in 1995, my guest looked at the early internet and asked a simple question: where are the people? And then answered it by inventing the thing billions of us now take for granted, live chat on a website. He built that into LivePerson, took it public, and ran it for the better part of three decades. But that's the old story. Today he's arguing something that cuts against the entire trillion-dollar AI industry: that the future doesn't belong to the giant model trained on all of us, but belongs to individual AIs trained only on you, that you own it, you control it, and you get to monetize. He raised $10 million to build it. He's got thousands of people waiting to get in, and he's pointing to the same idea at our kids with a device called Kid. Rob Lasco, welcome to the show. Say your name the way it should be said and then tell everybody who you and you are AI is.
Yeah, Rob Locassio. And uh, Ben, thanks for having me on the podcast. Yeah. Um, I started URAI actually about two years ago, and uh, with the simple premise that individuals should have their own AIs instead of the aggregated knowledge that we have. Each of us should be in control of our knowledge ourselves. Our AIs with the ability to share and monetize our AIs, um, instead of people monetizing our data for us, which is what's happening with the large language models today. So simple premise. Um, we're about to launch the platform, about a week away from launching the platform, and uh, kind of excited to, you know, talk about what we're doing.
Awesome. Well, 1995, the early web. What did you, what did everyone else completely miss that you saw?
I think that when I look at the original concept of the internet, which kind of holds true today, when e-commerce came into play, people's perspective was that the website would be like a vending machine: you put products up, people put the credit card in and they would buy. And especially in those early days, people weren't comfortable with that. And I just saw the idea of putting a human in the forefront, having a way for a customer to interact with the person, ask questions. That's how you build trust, you know, versus like especially back then, like put my credit card, maybe it gets stolen. And so it hasn't really changed when you think about it. There's still people like to talk to people and they want to have a conversation around products and services before they buy. Obviously Amazon went in a whole other direction, which is great, but for the majority of brands, they still have to communicate with those clients one way or another.
Yeah. Yeah. If you could talk to one person's individual AI, living or dead, who would it be?
Um, I would say, if I could talk to one person, uh, it's interesting. It's like I'd probably talk to Jesus Christ. Yeah. Yeah. I go there first myself.
Go. Right. I was just gonna say, just go straight to Jesus Christ and have a conversation with him. If we want to be more modern day, I mean, I think someone like, you know, Muhammad Ali or Martin Luther King, or um, actually there's a bunch of people on the platform that I like talking to their AIs also. And uh...
But uh, yeah, the idea of getting, having a really deep conversation, a very personal conversation with someone who's not here could be a quite fascinating opportunity.
Yeah. Your AI once recommended you a Robert Frost poem. It was never trained on. Coolest thing it's done or the spookiest?
Uh yeah, it was quite interesting because I was putting together, I write all my LinkedIn postings out of my own AI and I do a lot of my communication out of my own AI. And um, I was asking it for what did it think like one of the more, you know, what quotes that would kind of fit me to do something I was posting on LinkedIn. And it brought this Robert Frost quote. And what was interesting about it is I never told it about that quote and why that's a favorite of mine. It's, you know, it's not like you dump everything into it. It's you put a base level of knowledge inside of it and then as you use it, it gets to know you. And I have a framed picture of that poem that actually it's in storage right now. I left it in a box. But it was the poem that I have normally above me. And so it was just so fascinating because it knows me really well and then kind of just figured, I don't know, it guessed at what it would be, but it was interesting. I never looked at it as spooky.
I always look at it as like wow, my AI is really getting to know me, you know?
It's spooky. It's spooky if it's another AI like guessing things at me. But when you feel like you've given it a lot and it's yours, you kind of feel like, okay, it's just getting to know me well.
Yeah. It's surprising me with its understanding of me. Yeah.
Yeah. Building for the end of life with your AI and for your four-year-old kid model. Which one keeps you up at night?
Um, I mean there it's it's just really if I if I think about the broader things I'm trying to do is I think the AI world we have one narrative and that is that AI, you know, is going to take our jobs. It's better than us and all this stuff. And it's creating a world of distrust. Um, and what keeps me up at night is really about, you know, making sure that we're being good stewards about what we're trying to do as a company. And my vision is that everyone owns their own models and their data sets and you should. It makes no sense to aggregate knowledge. I just want you to know that like I get why Claude and ChatGPT are really important today, but they and running processes they do an amazing job at especially Claude, but when it comes to really deep knowledge in many different areas, I want an expert. I was just talking to somebody on our platform who is like one of the best interviewers in the world. I mean this guy has interviewed everyone in the world and I mean Muhammad Ali and all the presidents and Robert De Niro, I mean everyone, Leonardo DiCaprio, just for Esquire magazine. He has a methodology on how to interview people. He actually and even how to do a podcast and ask the right questions. It's his methodology. And the best of ChatGPT and Claude can never do what Cal can do being on our platform and talking to him and giving access to that. So I think there's just such a depth of human. We haven't even seen the beginning of it, but I see a world there. So that's what kind of keeps you. I just want to make sure we get to that world.
Yeah, that makes sense. I mean, um do you still find that there are times when uh you know, a pen and paper notebook and a good night's sleep is better than um you know, your your human life model?
Yeah, look I I it's human first. Our whole technology is about human first. So anything that you do as a human, as an individual is first and then the technology is here to support you. So I do both, you know, it's like I do a lot with my AI. I feed a lot into my AI. It helps me with a lot of things, but I take notes and I journal. Um, and because I think there's a process of writing that your brain is a different process than typing or speaking and I want that muscle always. So I just think yeah it's a I you know there's a dance going on between human and technology in my world. It's like a tango and that tango is what's important to me. So humans have to do what they have to do as individuals as unique individuals.
Yeah. Well, make it make the core argument for our audience as the big models get smarter off all of us. They get smarter for all of us. Why does the advantage move to the individual AI you own? And you may not want to share this, but I just had a curiosity as you were talking. Is there, are you guys using an LLM to drive just the conversation and then the human stuff is kind of your RAG or you know is there something complementary there that you're doing or tell me a little more about it.
Yeah. So there's when you're talking to your own AI it's just your human life model. So we have this human life models. You put data in, we break it into seven categories and those are like identity, values, goals, all of that is mapped and that's unique to you. We do connect a large language model for you to have world knowledge and you can use any one. We just get the tokens. We give you the tokens to get world knowledge and then we have a real-time feed coming in about everything that's happening in the world, the news and everything that's happening because your AI at all times should be aware and present of all information including its own. When you share yourself with somebody, it's our model. So think of it as your own large language model called a small language model that is being shared. So they're only tapping into the box of your knowledge. So our AIs, they don't hallucinate and do things or make things up. They'll say, 'I don't know.' And then they speak. They also you can hear them talk and they sound like you and I mean you could talk really low, you could talk agitated. It depends on what how you're communicating with your words. But that's how we deliver a full experience right now.
That's cool. Um you call it your X factor containerized in its own model. Uh what can an individual AI do that I can't get by just dumping my notes in the cloud?
So the big part is we have a thing called services and you can create a service around your knowledge and your AI will actually look at your data sets that I recommend. Like Ben, it's like here's how you would do a digital consulting or here's how you do a podcast or in my case it's like building a company and raising money and stuff like that. You can build out a service. Um we have a tax accountant who has built services on how to create a tax return and that service can be memorialized in an AI that then you can share and you can monetize and you get scale. This gentleman Cal who's on our platform, this writer, he built out a service around interviewing and how you ask the right interview questions and how do you do it? And so you can only get that from Cal and he's memorializing that as an AI, his AI on the platform, and then you could take that in a very structured way and learn from him. So yeah that's cool.
Yeah that's cool. The human life model captures you mentioned seven dimensions of a person: identity, worldview, story, mindset, mental models, habits, growth. Why those seven and what makes it you instead of a chatbot?
Yeah. Yeah. So it's identity, world, story, mindset, drive, pattern, growth are the seven. So we just found as we did the science on clients and especially as we started to look at people's personal stories, we all have the Joseph Campbell hero journey story in us. I started from nothing and I'm now here and I went through a lot of failure and success and learned that story inside of it has a lot of embeddings around these seven areas. So we look at data from two perspectives. One is your personal data set, life story. The other one is the facts, business processes, the facts of your life. Those things are separate in our world, but they all make up the human life model and then we look at you from SE. So we found this is the best way to manifest a human on the platform to be feel truly unique and individual.
Yeah, that's cool. The business model is the provocative part. People set a price, share their AI, and monetize it. Even a future market where individual AIs are bought and licensed. Walk us through that world.
Yeah, I what we're really doing here and working with clients is it's a transformation from the current world and a lot of that I'll call it the web 2 world that they're operating in. So they're using the tools of web 2, could be whether it's social media to promote themselves or they're using CRM systems like Salesforce. All of that is our present past. The future that we are creating is that they take all of that and that becomes an AI on the platform in which they can just sell access to. They create the price that they want to sell access to. And then they can scale that across many people. Now it's a new thing. It's totally a new thing. It's the first to be done. But it really allows these small businesses and professionals. We have doctors, lawyers, accountants. We have creators on the platform like writers and coaches, a lot of business coaches or personal coaches and it allows them to take all of that knowledge and then make a business out of it, but make a true pure AI business, not one that's kind of stuck in the past.
Yeah. We've had many people come to us and want to build something like this. So, the fact that you're democratizing it, standardizing it for folks is really cool. Um I've heard a couple of people like Hermoszi and a couple of other folks that they've put something like this together. Um and I don't know what kind of I mean I think it's a great idea for your brand owner, right? You're if you're a founder brand type of model where hey, you could talk to the founder. It might be my AI, but you can get all my information and we can work in real time, but for me it's asynchronous. Like I already put all my stuff in there. Um so it's you it feels like a second brain. Uh you own the model and it pulls your quotes. Um what's the onboarding process look like? Like what does it take to get set up?
It's pretty straightforward. You come on the platform, you can record your voice that takes about 30 seconds and just through 30 seconds we really can replicate the range of you and your tonality and all that. You can take a picture of yourself and you can decide whether you want to have your face or a normal picture or an avatar or something like that. And then you basically can connect yourself to your social feeds like LinkedIn and all that. We start pulling data in. You can upload data, but you can start talking to your AI and it'll start asking you questions and all that. There's also the marketplace. So you can go straight, you can also look at the marketplace. You can see services that are being sold. We have things called thoughts, which are really ways for people to take small sections of their knowledge and promote them on a feed. And through that feed, you can see pictures and video and you can talk to the other people's AI. So you could talk to the AIs on the platform. If you're talking to your own AI, you can have a conversation. It will bring in other AIs. It will recommend other things on the platform that can help you. So instead of you going out to a large language model and getting your answers to your questions, we look for the knowledge on the platform and that knowledge can help you, but that's always provided from the individual. It's not like we just put a block of text. We put it with the individual and you could then chat further with that individual. So we have this guy Jordan Tom, he's like a great coach for children, speaks at schools, and he's got a million followers on social, but he's built his AI out. So if you have kids, he's a great AI to talk to. I talk to his AI all the time. Ask him questions. My kid does this. What do you think? He's amazing to talk to, but I get access to him immediately. His AI, you know, it's his AI.
Yeah.
Yeah. Yeah. No, that's really neat. Um, one of my other podcast guests is building a similar thing as an assistant for the elderly. So a couple of different persona choice options, but to figure out like cognitive decline, health, just making sure that they have, and it sounds odd to call it companionship, but they have somebody to talk to and tell their stories. And then all of that's being collected and recorded as a digital artifact of your elderly person, you know. So really cool application in the elder care space to give that kind of companionship. And you're talking about a coach for your kids, an AI coach that your kids can talk to when they're challenged by their sports or want to create a new habit or whatever. That's really cool.
Yeah. And on just connecting your previous thought, a lot of people are trying to build things and we call it like a Frankenstein of stuff. So there are some people like your friend who's trying to I'll call it make a platform out of it and we're making a platform so you can make it available to many people and then a lot of professionals I find they're on their own spending 105,000 trying to cobble together a voice model, an image model, a video model, a RAG, a thing. And then they're sitting there with Frankenstein or somebody sells them to do this and they're sitting there with Frankenstein and we just want to make it affordable so people could get the power of a very powerful AI but in a way that's affordable. So yeah.
Yeah. No, it's really cool. I mean the data ownership is the whole pitch. Private versus public data sets, nothing bleeding into a public model. How do you guarantee that? And why can't the big platforms credibly offer it?
It's not the business model. I mean they have they're making billions of dollars off of aggregated knowledge and for if tomorrow they said oh we gave up on our business model and they our terms and conditions if you took them and you can and put them into a ChatGPT you'll see it'll say like what's unique about this is you own the data it's not so you can if you want to summarize our terms and conditions you'll see that it's all about you own your data, you own everything, we don't put your data into other models, we don't train models with your data. Even stuff that we don't use, like we use third-party models for video jam. You can make a video of yourself, make a promotional video, make a movie out of something. People are doing that out of their books. We use third-party models for that. But we have terms and conditions with them that they cannot train based on your data set. So everything with us is all about having a closed system when it comes to your data, data privacy, data security.
Yeah, that's cool. You showed me how your AIs are friendly collaborating. You even said one of your AIs literally emailed you a business idea. Where do you think AI agent-to-agent collaboration goes and what breaks first?
I think the agent-to-agent stuff is really amazing. So on the platform you have your AI that you're talking to. I can bring in someone else's AI. My AI will talk to their AI and then I'm part of that conversation. You can put multiple AIs in. And I find it quite powerful because you're able to, your AI and another AI will work together on an idea and I think you get to really great outcomes when you do that. So I think the agent-to-agent piece is what I see as a huge unlock on the platform.
