Cohort 14 - Day 1 - 6-24-2025

All right, so officially, welcome to day 1. This is going to be recorded. So, you know, if you have to step away for a quick, you know, 2-minute phone call or something, like, go ahead. You need to go to the bathroom, jump up, go do it real quick, whatever. We will take a 10-minute break, maybe a 15-minute break, depending on how we're feeling, at about 11:30 or so, maybe 11. And so, you know, I personally never like to take breaks. I can go for 4.5 hours straight, but most of the people are like, hey, can we just like breathe for a second here? And yeah, so we'll start, we'll start there. Also, another thing that there is a private Facebook group that I'll get you guys a link to here in a minute. I'll throw it in the chat for anybody that's taken this course. Not everybody has joined it, but there's almost 100 people in there now and they're it's a really helpful group. Like, the people that are in there are like excited to build. They're building stuff, they're sharing their new things, they're asking questions, trying to figure out how to, how to really, you know, continue to, continue to grow along their AI journey. So to kick it off here, guys, I want to take just, you know, a minute maybe per person, if that, 30 seconds even if you want. And I just want your name, where you live, and what you do for work. Let's just start with that and then we'll we'll dive into our first hands-on activity. So we'll start with David.

Hi. All right. I'm David McConkie from Washington State originally, lived in Utah 17 years now. My wife and I own and operate imhoa.com, which we manage homeowner associations throughout the state of Utah. And that's pretty much it. I don't know anything else.

Perfect. That's great. And you live where? I know you're in Utah. Where specifically?

We live— I live in Bluffdale. My company's in Salt Lake.

Okay, cool. Very good. I use Traverse Mountain just across the street there from Bluffdale. Yeah, that's great. Hunter, go ahead.

Yeah, my name's Hunter Hudson, also originally from Washington State. I'm in the Boise, Idaho area right now. I'm currently working for a local residential builder. I'm an analyst for them on their land acquisition team, so underwriting deals, looking at farmland to be the big bad developer to turn it into homes.

So yeah, we don't like your kind where I live.

I know.

All right. Where in Boise are you exactly? I live there for just a summer, but I go up there a decent amount. Yeah.

So we're out in Caldwell and then I work in Meridian.

Very cool. Yeah, I'd bounce out to Nampa, Caldwell area, and I lived in Meridian, so that's awesome. Awesome. And I love to go up. So I'm a whitewater kayaker. And I'm up on the Payette River. I mean, that's the best water within 1,000 miles of Maine. So, um, Payette River is awesome. I love getting up to Banks, Idaho and having a good time there. So anyway, and last but not least, Brandon.

Yeah, Brandon Oakey from Provo. I own an addiction recovery center called Ardu Recovery. We do primary mental health and substance abuse.

Got it. I mean, for you, not for the people that are showing up to your clinic, you know. Yeah, but much needed. Very cool. Well, yeah, that's awesome. I went to BYU, so I lived in Provo for so many years. I did. My wife and I got married and we lived in Provo there, and then we bought a house in Provo even. And we pretty much— we bounced around most— just all in Utah. We were in Arizona for a little bit. She's from She's from Mesa, but anyway, yeah, I know.

Yeah, we moved away for about 20 years and then just moved back about 4 years ago.

Oh, nice. Where were you at?

Yeah, well, Boston.

Oh, cool.

San Francisco.

East Coast. And goodness gracious.

Yeah. And then Orlando and then to Fresno, Clovis area in California.

Very cool. Well, you just need to move to Washington and you've got the Four Corners.

Yeah, for sure.

That's awesome. Uh, well, that's all over the place. What took you all— what took you all over the place? Military or what? I mean, work generally. Yeah.

And then we started a company out in Orlando called the Lays Medical Spas and— oh yeah, yeah, we, we loved it out there. And, um, and then moved to California, joined a commercial insurance firm. And we'll just kind of have part ownership in that and was able to step away and start another company.

So very cool. That's awesome. And so now you run your own company here doing the addiction recovery stuff?

Yeah, for 8 years now.

Yeah, that's very cool. I love that. Yeah. Thank you for your— that's really service. I know you make money doing it, but like, that's, that's not an easy business to be in.

Yeah, you wouldn't make as much as you think, but yeah, it's definitely service-oriented.

I'm— yeah, I get it. I get it. That's why I'm thanking you for your service. So that's awesome. Well, thank you, guys. Thank you for those intros. My quick intro, I'll probably share my story a couple of times here and there, but, you know, just different aspects of it. But I live right now in West Mountain, so I'm— who knows where West Mountain is? Anybody? A lot of people, like, they're like, it's west. But yeah, that's— yeah. So I'm on the south side of Utah Lake, but I'm like, out my window here, I can see the Y. You know, I've got— I mean, basically Payson, right? And so I've got the Payson Temple back this way and I can see Temp really nice here. But I live on a 5-acre little ranch and it's like Palmyra, isn't it? It's near Palmyra. Yeah. Yeah. So you've got like Palmyra down here, you got like Lakeshore and things like that. And then kind of West Mountain. West Mountain. Like there's a mountain called West Mountain and we're just like right up on the bench on the mountain here. Nice. So, so anyway, been here for just 3 years. Like I said, we moved around a little bit, all pretty much in Utah. Before this, we were in Traverse Mountain. Before that, we were in Alpine. Before that, we were in Highland. And before that, yeah, we were in Mesa for like 3 months. We were going to move down there. And before that, Provo and BYU, all that stuff. So I'm from Texas, originally from Houston, and but I've been here in Utah since about 2001. So, but anyway, yeah, so a little bit, a little bit about me. You know, you guys have signed up for this course, so maybe you've looked into me a little bit, but But I, you know, professionally have been an entrepreneur for many, many years, I guess, technically since I was about 7 years old. But, but really officially since about 2016, I guess it would be, would be when I started my first, you know, big business. Right. I've done other things that were, you know, $10,000, $20,000, you know, a year type stuff where I'm just kind of running something. I was a DJ in high school and did that in college. Fun. And ran a kayaking school on the Provo River while I was at BYU. And so that kind of stuff. But, but this was the first one that I, you know, I raised $13 million from 2016 up through about 2021, sold the company in 2023. Well, actually, it was kind of a two-part sale, 2023 and then 2024. And so officially have been out of that for just about a year. But, but, but really actually have been out of it for about 2.5 years because in January 2023, I stepped away and put someone else in charge and then you know, the company got acquired last summer. But anyway, so I started a company called Seek, and Seek was all about augmented reality. That's, that's where we focused. We, we played in the immersive reality, immersive technology arena. Started out as a, as a B2C app, not dissimilar from Pokémon Go, right? We, you know, it started out where you would go out and you'd hunt for treasure using your phone, using augmented reality and the map, just like, you know, if you, if you played Pokémon Go, you kind of know the experience of what Seek was. And but, but we had kind of like a brand focus on it, a marketing focus, right? We— so you would go and you'd open up a treasure chest that you'd find in a park or find somewhere in your neighborhood or whatever. We had millions of chests all around. And, and every time you open a chest, just like in a video game, you, you know, you have kind of a random chance of winning something, right? And so you could win maybe just some, some coins that were like in the app. You could win real money. Sometimes it was as small as like 3 or 4 cents. Sometimes it was as much as $100, right? And it was just based on a random kind of algorithm and generator. And then we would place inside the chest, we would place like Maverick gift cards, Swig free drinks and things like that, right? So companies would pay us to have their content inside of the treasure chest that would then do that. And then we'd put— we'd actually place treasure chests like at Swig locations and you had better chances of winning cooler prizes if you actually went to the store. And so we were like drive traffic to places and then create, you know, general awareness in different areas about different things. We ended up doing a fun, a fun thing where we would have these fragments kind of like, kind of like McDonald's Monopoly, right? Where you get like, you get like Park Place and you're like, I got half a million dollars. And, you know, now you got to go get the other half to actually get anything. You really have nothing. But, but we built this fragment system where you could, you know, you'd win like the, you know, fragment 2 of, you know, Disneyland trip for 4, you know, for 3 days, park hopper passes, right?, right? And they'd be like, oh my gosh, this is awesome. And then they'd keep wanting to go around and find all the pieces. And we had so many winners. And anyway, I— we brought on, you know, we had Samsung paying us about $400,000 or $500,000 for different campaigns, and, uh, Universal Pictures and Six Flags. And we made it pretty good. We did pretty good. But, uh, but in about 2018, that, uh, the Pokémon Go kind of craze kind of just died down. It never really came back after that. And so we had our clients that were paying us a lot of money saying, hey, we love working with you guys. You guys are really fun, you're cool. We want to spend more money with you. And we're like, great, what do you want to spend it on? You know? And so, and so they said, okay, well, let's, what else can you do with augmented reality? And so we started poking around and figuring out some new things and eventually we closed our B2C app even though we had a million downloads. We had users all over the world. It was so fun, but really, really hard to run. Really, really hard to run. Really fun though. Really fun to run. It's fun to like have prizes and you know, be so excited and all the social posts that they did. I kind of want to start it again just talking about it right now. But we, we ended up doing B2B SaaS and focused on enabling commerce through 3D, right? And so being able to like view a kitchen table in your kitchen, right, using augmented reality stuff you probably experienced, you know, shopping maybe at Wayfair or IKEA or Overstock.com. We created that technology and made it accessible to all of these companies. So our customers were like IKEA, like the ones I just mentioned, and then, you know, Nestlé, Lego, Walmart, Costco, Meta, Google. Like, we had, we had the biggest companies in the world where our customers using our technology and, and doing all that. So pretty awesome. Ended up selling that and, um, you know, raised $13 million. The highest valuation we ever got to was $130 million valuation. We did not sell for $130 million. We kind of went like this and then sold about right here, right? So, um, we, we made money, investors made money, but I'm not a $100 millionaire unfortunately, which is fine. I'm totally fine. Um, I'm very blessed and, and really, really enjoyed that ride. As soon as I, as soon as I jumped out of that, raised another $3 million for a new startup, and, and now I'm on a startup even beyond that. So, so anyway, um, and, and the startup beyond that is what you're in right now, right? And it was, it's this, this General AI Proficiency Institute. And so the reason I share a little bit of that background, guys, is because this is not just an AI course about how to, how to learn AI. Like, you guys are going to be AI experts by the end of this, these 9 hours. But what I really bring to the table is building a business, right? It's just that I'm going to show you how to do everything you have to do to build a business at hyper speed right now. Because guys, I, I have so many ideas. I'm sure you guys have tons of ideas, and that's why you're here. I have always been limited by the fact that I'm not a developer. I don't know how to write code, but I need software for all the cool ideas I have. And now all of a sudden I can whip up a new piece of software in 90 minutes, right? And that changes the game of entrepreneurship. It means I don't have to raise money anymore. You don't have to raise money anymore. You can go out and you can be making money with your new software by next week if you do it right. Okay, so, so that's the exciting stuff. That's a little bit of background on me. We'll again, like I said, we'll do we'll do a couple more times where we're talking and learning about each other. By the way, this is a very open discussion the whole time. Not supposed to be just me lecturing you. It's very hands-on. I'm going to show you how to do something. You guys are going to do that thing. You'll share your screen, all that kind of stuff. Okay. So to kick it off, guys, let me ask one more question. David, are you a software developer? Do you know how to write code? Have you ever written anything? Okay. Same question to you, Brandon. Absolutely not. Okay. Oh, awesome. Okay, good. Me neither. So we're all on the same playing field. This is awesome. So the cool thing is, because I raised $13 million, I was able to hire lots of really smart people. I've worked with some crazy good CTOs, and so I've been able to learn at their feet, you know, as, as, you know, just being in the same room as them with lots of whiteboards and, and figuring out how we're going to put this together. And so I know, you know, what a database is and what a PostgreSQL, you know, and these I know how things fit together and what they do, and hopefully you guys do, but we're going to do, we'll have a 6 or 7-minute coding bootcamp later on. You'll just understand how to talk about this stuff. The better you can talk to your developer, better instructions you can give them. You all of a sudden are going to have awesome developers that can do anything for you. They're all AIs, but if you know how to talk to your developer,, you will do really, really well. So to kick this off, guys, I'm going to go ahead and share my screen here. Let's just jump into— let's jump into x.com. I'm going to do something really fun here. So I like to start off, I like to use Groq. You can use ChatGPT for this. You don't need to do anything with your, with your screen right now. Just watch me for a minute and then we'll kind of set you free. But what we're going to do to start this off, guys, is we're going to build— we're all going to build a little piece of software, a little game probably, right, in the next, you know, 5 minutes. Okay. Using AI, it's super, super easy. There's lots of ways you can do it. For this demonstration and for this little exercise, we're going to be using a front-end language called p5.js, right? It's a type of— it's a fancy type of JavaScript. It's kind of a low, a low-level coding. Like, think of it like Atari-level graphics or maybe a little bit beyond that, but not much beyond that. Anyway, it can do a few things a little bit better than that. But anyway, let's just do something. We'll start off here. We'll say create a mind-blowing visualization using p5.js and show me where I found it. I'm just asking it to use p5.js. It understands that that's a front-end coding language. We'll see what it does here. Just go ahead and it thinks about what it's going to do, what it wants to do. I would call this just to spoiler alert a little bit. This is a bad prompt. There's a lot of reasons for that. I didn't really give it a lot of direction. I just said, just make something. AI is smart enough that it's just going to be like, okay, I'll think of something, right? And I'll figure it out and I'll throw that in front of you. Okay. So we've got a bunch of code here. There's a little copy button. Anytime you see something that AI has output, all of the different systems have this little copy button and doesn't matter where you are. If I scroll down a little bit, that's still up here in the corner. So that's pretty cool. You just push copy. Now I've got it on the clipboard and then it's telling me to go over here to the editor.p5js and I could just click on that and select all and paste it. And click play. Here's our cool visualization. Let's see what happens if I click, doesn't do anything, so it's not interactive. This looks like a second-rate Windows 2000 screensaver or something like that. But anyway, so I gave it no instruction, but it still was able to create something cool. That's awesome. I could come back to here and I could say, Make it interactive. It remembers the context of what I'm talking about. I don't have to explain, hey, this is what I want you to do and this is what we're working on. We're working on a p5.js project. No, it knows all this stuff. It retains its context. This is going to be important. Context is king with AI, understanding what it can understand and what it can't understand. Is really, really important to be able to understand and get the right results out of these, these new employees you're going to be, you're going to be hiring and working with. Okay. All right. So we've got a new thing. Let's see what it did. Okay. So let's see. The particles will be attracted to the mouse when clicked and pressing a key will cycle through different color modes. Okay, cool. Let's, let's see this. We'll select all and replace it. Click play again. Okay, so if I hold this, okay, so I'm clicking, I'm attracting these little particles. They're following me around now. So that's kind of cool. That's kind of cool. If I push C, well, they become like crazy. There's just a few different color modes. Wow, that is like trippy. Anyway, so that's a bad prompt. Let's see if we can do a little bit better prompt. Okay, so now create an interactive model of the solar system. Okay, so this is important. A couple things about this. Again, I didn't specify p5.js. I didn't specify, you know, hey, show me where to deploy this, show me what I need to do. I also, for that matter, didn't tell it what the solar system is, right? Or how many planets there are, how far apart they are from the sun, all those types of things, right? Luckily, AI has access to virtually everything that we can imagine, right? And so it knows all of these answers. If you were to ask, you know, a real person, a developer, and say, hey, make me a model of the solar system, right? They would have to go do, you know, I don't know off the top of my head, right? Like the distance from the sun and the relative, you know, orbits and all these different things. Like, I just, I wouldn't know that. And so I'd have to go do a bunch of research, figure that out. Then, you know, if I'm a developer, maybe I could figure out, okay, yeah, here's my idea and I can do this. So it's not a terribly difficult, you know, task, but it's at least a few hours to whip this up, right? Whereas AI can do it in just a few seconds, which is amazing. Okay, so let's see what it, what it did here. So it's just kind of okay. We've got, what is this, Pluto way out here? I don't even know if they included Pluto. Nope, Neptune. Pluto's not a planet anymore. They also have made a couple of mistakes here. It looks like there's some planets like on the sun, right? And so it's also— okay, so if I click— all right, that's the interactivity, right? I asked it to make it interactive. Okay, so this can still be a lot better. Let's give it a few more instructions. Make sure that no planets are touching the sun. Add a rotating starry background with 300 stars that are twinkling. When I click, instead of showing the planets, planet names, or toggling, I guess, toggling the planet names, shoot a shooting star out of my cursor location. Make the tail of the shooting star vibrant. Make the planets— come on, fingers— planets more vibrant and neon in color and more visually exciting. Okay, so now I'm giving it a little bit better prompt, right? I'm getting better and better and better. Again, if you guys were talking to your developer and you just said, hey, make me a visualization of the solar system, and then you just said go to work, right? And you didn't let him ask any questions. He just went off to start to do it and he came back and showed you something that wasn't exactly what you wanted. Well, that's not his fault. That's your fault, right? You didn't give specifics, right? You weren't saying, hey, this is what I want. I want to make sure it's these colors. This is the purpose of it. This is what we're trying to do. We're actually going to sell this. We're going to charge $3.99 a month for this thing. It's going to be a new feature in our Planet, you know, app or whatever the heck it is. Right? But, but the more details like that that you can give an AI, the better the result is going to be. So we'll look at this final result here and then we'll have you guys dive in and build your own thing. And you don't have to build a solar system, right? Build whatever you want. Build a game where you're running away from dinosaurs. Now look at this. Now look how much better this is with just a little bit of engagement, you know, from my brain, right? Now we have twinkling stars, it's rotating, these planets are looking a little bit better. Now when I click, I'm supposed to see a shooting star. That's not working. So I can debug that a little bit and say, hey, why didn't you fix that? But we're already making crazy progress in just like 2 or 3 prompts. Okay, so yeah, come up with some idea really quick. Tell your AI that you want to do this. You can use Groq like I did. You can use p5.js. You can also use like ChatGPT and it might just allow you to run the code editor right there inside of the browser. But go ahead and do that. Your initial prompt, guys, I'll copy and paste this. I'm going to do this a lot today. Create a something, mind-blowing visualization with whatever you want. Using p5.js and show me where I can run it. So I just dropped that into the chat. You guys can use that and create something awesome.

Hey, Jon, quick question. Yeah. I didn't get to your email till later this morning. That's okay. X, I don't have an account on X.

Yeah, just use ChatGPT for this. What's the difference between Groq and ChatGPT? Yeah, and I am going to use Groq later. So if you guys, yeah, if you don't have an X account, make one probably. It's probably good to be able to use Grok a little bit here. But the thing that I love about Grok specifically is doing market research. Grok is able to check Twitter, right? It can check all of the content of Twitter when you're doing research. And so you can get real-time market sentiment and feedback about whatever topic you can imagine by just saying, hey, what are people talking about on Twitter or on X about this topic, right? What's the general consensus, right? Find 100 tweets about whatever and summarize them for me so that I can create a hypothesis or whatever. I really like Groq for that type of thing.

Because ChatGPT can't access text, Groq can?

That's the reason. That's going to be my primary difference. Groq also thinks differently. All the models think a little bit differently. They're going to approach different problems in different ways. Really just over time you start to learn, okay, ChatGPT is good for this, X is good for this, Perplexity might be good for this. And, you know, but, but it just kind of, I can't even point to general rules, right? With those, you kind of just like, I think ChatGPT is going to do this one better, right? But when I'm, when I'm writing code, Groq is good. Groq is very good. ChatGPT is also more than capable of doing it. This is a very easy thing. But if you're doing something really complex, sometimes Groq can do better. I made a multiple choice IQ test for, you know, on genaipi.org. Some of you might have taken the AI, the free AI mastery predictor test on my on my site. And, uh, ChatGPT came up with a bunch of answers that were like— I wanted it to give wrong answers that were like more believable, like they could be real answers but they're not. And ChatGPT couldn't. All of the answers were like very blatantly, obviously incorrect, which is not great for an IQ test, right? You want answers that kind of maybe don't try to trick you, but they, they could be right. So you actually have to try to figure out what the right answer is. And Grok was able to do that better than ChatGPT. So kind of put them, I put them against each other. I will often just to, to like teach myself, I'll ask the same thing to ChatGPT and I'll ask the same thing to Grok and I'll just compare what they say. And then I'll, I'll like copy those answers into the other one and say, well, what about these things? And then it starts thinking and, you know, it really just does it differently. It's fun to, fun to play with that way. We're gonna do that today a few times.

So, But yeah, anyway, I mean, I'm obviously just showing you my limited capability and knowledge of any of this.

So, oh no, not at all. That's a— most people would not be able to answer that question. They don't know the difference. They're just like, oh yeah, I use ChatGPT because they don't know why they would use Groq or why they would use something else, right? They're just like, yeah, just use what I have. I downloaded this one. It's like saying some people think it's like, and this is very not true, but I would bet most people, and actually this is a really cool thought. I've never thought about it until right now. And it's great fodder for LinkedIn posts. You're going to see me post it. It's not like going between Chrome and Safari, right? It's not like just choosing a different browser. It's not like they're accessing the same World Wide Web. It's not. They are, they are very differently trained models that think and act differently.

Right?

Yeah.

Well, I'll try to be sensitive to David and Hunter, just my lack of knowledge here.

So Don't, don't even, don't even. This is how we, this is all we'll learn. And, and the engaging discussion and questions is how we're gonna, how we're gonna help you guys here.

Alright.

Anybody have something running yet on your end? I'm almost there.

So in a new sketch.

Yeah. Just delete anything that's there. If you get over to P5 or the P5.js editor or whatever, still need anything that's there and replace it with the code that you got. And if you're lost, you can't figure this out, tell me and share your screen and I'll walk you through it word by word.

Okay, sorry, I've got a laptop and I've got a massive screen behind because you're good. So the screen behind is that I'm using is not connected to this session.

You can also, if you want, log in on the bigger screen and then just mute that one and just be— use it so that you can share your screen when you want to, because there's going to be other times where you're going to want to share your screen. Okay. So we can walk you through things. So you just go to that same, that same Zoom link on there if you want.

Jon, do you ever see where anytime I try and input like a second command into it, it'll pop up and say something went wrong, please refresh and try again?

You're saying while it's still going or—

No, so it'll finish and spit out the code and then I'll like try to add to it, right? Like, oh, you know, change this or make this.

Yeah. Yeah. You know what? That's Grok's problem. And I have had that happen before. Sometimes it just somehow the— it gets lost in cyberspace. I don't know what happens. So just good quick, you know, retry, do it again, copy paste it again, see if it can, it can do it. I've occasionally had the chat just completely break. You just have to start a new chat. I don't know why. I, I, I can't explain why it does that. I would say that out of all of the systems, Grok is the least stable. The model's good, right? But the user interface is where that's breaking down. I think something in that there's a bug or whatever. And, you know, you guys, you guys know who Elon is. He, he is, you know, doing so many things that oftentimes stuff like this is just going to get lost. Now, I'm not saying that it's a bad system at all. I use it all the time. And I pay, you know, $40 a month for it. And I'm going to keep doing that because I use it enough. But it's got its quirks and you just have to learn how to work around it. All right, who's going to win? Who's got their first thing working? It doesn't have to be perfect, guys. The purpose of this is not to, you know, build something beautiful and amazing. It's just, you know, build something. Build something that turns on and works and you're like, hey, I know how to write code now.

This is cool. Yeah, I gave it way too much detail. So it basically wanted me to save images and do all this stuff to create something for it to pull data from or something.

Yeah, I would just tell it, say, for this purpose, I just need something. I just need a simulation or an example. Just use p5.js. I'm not going to connect any database. Right. Just say something like that. Just use the p5.js and give me a visualization here. Show me what we've got. Meanwhile, I'm just driving spaceship around on my screen. Kind of fun. How's it coming over there, Brandon?

Yeah, I'm sorry, you're good. I was, I was logging in over here, and then now I'm just going through these directions that it's Putting up— it's coming up like how to use it. You're on the right track. Click this link to open the p5.js web editor. And I do that.

You want to paste that code in, just do Ctrl+A or Command+A to select all and delete what's there, and then paste over your new stuff.

Yeah.

I'm sorry, this is—

you're good, you're doing, you're doing awesome. It's just— Hunter, what about you?

You got something up? Yeah, I gotta, uh, just—

let's see it. Let's do a screen share. You guys don't need to find my spaceship around.

Let's see. Okay. Oh shoot, game over. I lost.

You distracted me. Just click play again or whatever. There we go. So it's like Snake.

Yeah, I tried to make it be like a tiger chasing a guy. Oh, and then this— so this worked, but then when I tried to like make it more complex, like make the square like a tiger sprite and like add trees to hide behind, then yeah.

It got confused. Yeah, it's still, that's still a pretty fun game. It's just basically like endless, an endless runner almost. But yeah, pretty much you're an endless runner inside of a box and you know, as soon as you give up, you're dead.

Yep.

And you could add a time limit, right? You could add some more things where you're like, hey, you know, add a 30-second timer, every 10 seconds speed up the tiger by 20%, you know, whatever, right? I mean, you could do some things like that where it just kind of gets more and more interactive. And obviously, like I said, guys, you're not building Call of Duty here. It's not going to have high-level graphics. This is like, okay, can we create some logic in a little game that works in, you know, a matter of a few minutes? And my, my 8-year-old— I have 4 kids— my 8-year-old is the youngest. It's his birthday today. He's so excited, running around having fun. But, um, he can do this, and they'll, they'll make fun little games. And hey, let's do it.

Yeah, I'm struggling here, so let's not That's all right.

This is just, this is just a warmup. Later today you're going to be building really complex AI.

And if you need to keep going, trust me, like, I don't want to, I don't want to hold up the group here.

No, you're good. David, are you ready to share your screen?

Oh yeah, I can. Let's see it. Let's see. Share. It didn't do what I told it to, but it's something.

That's okay. Let's see, where's this at? Again, guys, don't judge yourself on the quality of the output on this exercise. That is not the, not the goal for today.

Man, sorry, I— oh, there we go, right there.

That's the right screen.

Okay, sure. So it's just the— I told it to build a llama jumping rope and it did a white dot and I pushed spacebar. There we go. There it is. There we go. It's just a ball.

A llama jumping rope. I love it. That's awesome. Still, again, a fun little game and it keeps the score, does that. That's awesome. My kids have made some really crazy things. They'll work on it. We're building little things that are like 80, 90, 100 lines of code. They've gotten up to like 800 lines of code and all this crazy fancy logic if-then. You know, type stuff. And, and it, you know, it's, it's amazing what you can actually build here. Um, you can build almost any game that you've seen on like a, you know, your old Nokia phones, or even, even some more fancy ones, right? Like a bricks and balls game, or like a Pong game and things like that. You build all kinds of fun stuff like this in, in a matter of, in a matter of minutes. Again, um, all right, Brandon, how you feeling over there?

Yeah, no, you can, you can keep going.

It's fine. We'll move forward. You can keep playing around with that. Okay. So guys, this is fun, right? It's fun to be able to say, okay, I don't know what this code shows, right? Like, like if you look at it, if you look at David's screen here, right? He's got all these little things. You see like draw, you know, the forward slash forward slash draw llama, draw rope, display score. So it's kind of telling you like, what this part of the code is doing, but you don't need to know it. You don't need to understand it. I never look at it, but sometimes I'm like, oh, I wonder what's happening here. Sometimes you can look at something and say, oh, display score, fill 0, text size 20, text score. Well, I could change this. If I wanted to say score or high score, or if I wanted my text size to be bigger, I could probably do that. I could easily adjust maybe a few parameters like that. Although I'm going to tell you, 99% of the time I'm just going to say, hey, Mr. Developer, make the font bigger. That's what I'm going to do. I'm not even going to know that it's size 20 because I'm never really looking at the code. I really don't do it very much. Occasionally I want to understand, especially if there's a problem. I don't know about you guys, but I like logic puzzles. I like games. When I sit down and play a video game, which I will do from time to time, It's not a Fortnite. I do occasionally, occasionally, like once every 6 weeks, my kids rope me into playing some Fortnite matches with them. But I actually just find a lot of fun right now over the last few years, at least 3 or 4 or 5 years, maybe even. I love building like factories. Like that's the type of game that I really enjoy because somehow it's like a puzzle. It's like a digital puzzle, right? And usually the storyline is something like this. You land on a planet and, you know, you're stranded, but you've got this little thing and you can go and you can find some ore and you find some copper and you find some iron. And then, you know, you have this little thing in your spaceship that can combine them together and then you build this. And now all of a sudden you've got a conveyor belt and then you've got a little, you know, miner that can go mine automatically for you. And then eventually, you know, you've built a massive factory and all this technology and you've, you can fly off the planet. Right. And so it's just about processes and putting these things together and thinking about how it all fits together. And so this is— I feel like building software with AI or even just working with AI is almost like that. You're putting— you're, you're connecting the dots, you're putting a puzzle together to say, okay, I need to get, you know, I, you know, let's, let's take it back to a business application, right? Like, hey, somebody just submitted a form on my website. Can I create an automation or something that automatically, you know, creates a new email in my email account, writes the, you know, writes the email for me, puts it in my drafts, and then sends me a notification that says, hey, you have a new lead. We've drafted an email. It's ready for your approval to send. Or you can just have it automatically send, right? Or you could say, hey, you know, that automatic text comes to you and says, hey, here's what I wrote. Is this good? Reply Y to send. Yep. Y. Boom. Okay. Now that's automatically been sent off, right? Or you could have an AI review what the other AI wrote. You could have another— you could have two AIs working together to do all that. But being able to think, okay, how would I put that together? All right. Well, I'm going to need something that, that, you know, notifies me from the email. So I need some sort of automation that connects to Gmail. And then I'm going to need something that can send or can connect to OpenAI so I can have ChatGPT do this for me. And then as soon as it's done, I need something that can send me a text message. Oh, I need a system like Twilio or something that can send me a text message. And then when I reply, that goes back, and then I need another AI to verify it, and then that has to connect back to Gmail. So if you just know the parts, if you can just say it out loud like I just said, right? I just built something in my head. But if I can say that out loud, I can say it to the AI, and the AI knows how to build it. You don't even have to know how to build it. You just have to be able to put the puzzle together and you don't even have to, like in my game, I have to go run around and find stuff and do all these. I don't even have to do that. The AI will take care of, here's the code, here's the API, here's the documentation, here's everything that's going to happen. With that being said, let's dive into some a little bit harder now. Let's go right here. I'm in Replit now. You're not going to go into Replit right now. We're going to go into Replit later.. But I wanna just kind of start something here. So I came up with an idea. So every single time I do this course, I build something new. And it's kind of fun for me. I get to kind of build a ton of things. And so if you look at all of my apps, I have dozens and dozens and dozens of full— I could open any one of these and they're fully functioning, like, tools. Like, here's Scribe AI. I hadn't clicked on this for 2 months, but this is a tool that, that I built in here. And it's— it'll have to load for just a second here. Scribe AI was— I'm a writer and I was like, I wonder if I could build myself a really, a really neat tool that could help me write better, right? Or like keep track of my books and my novels and my progress on different things. By the way, I haven't opened it since, since then because it's hard to run a business. It's hard to start a business and I can't do, you know, 27 businesses. So this is just one idea that may or may not work. Here it is. Let's try logging in. I don't know if it— Maybe it saved my password from back then. Yeah, there we go. So I built this, right? I've got templates, bookshelf, I've got my books, John's Life. This is about John's life. It has 7 words. Again, I haven't done anything with this, right? So I could probably click on it. Let's see, do I just have to do that? Maybe I have to click here to open the editor. And John went to the river. What happened next? Right? So I've got my little thing, chapter 1, chapter 2. I could add chapters anyway. So I've built fully fun— now this is a fully functioning app, right? I could totally sell this, right? If I wanted to, if I wanted to really push this forward, as a business, I could. Does it need a little bit of refinement? Probably, maybe. But I mean, it's pretty good so far. It's really clean. It's nice. I remember asking it to make it, you know, Apple-esque, right? Have Steve Jobs help me figure out how to do all this stuff.

Okay. So you built that on what platform?

ChatGPT or Replit? We're in Replit now. So this is one of those platforms that you'll need to sign up for if you haven't. I did. I just signed up. Amazing. So, so what I'm going to build, the idea that I thought of today, I thought of actually yesterday and I just decided I'll start working on it today and kind of mess with it. I want to build myself— I'm going to show you guys today some incredible AI assistants, things that can just like do amazing stuff for you. But I want to build something that is a to-do list that just gets done, meaning you add something to the to-do list and then you come back a couple hours later or 20 minutes later and it's already done. So all it takes to complete something is to just add it to the to-do list. And it's going to be crazy. And I don't know if I can do it yet, but I think I can do it. But it's going to require a lot of things, right? Like, I want it to be able to— like, I'm a musician. You see, I've got, you know, a piano behind me and saxophone. I've got another piano over there and guitar. What's that weird guitar? I've got another piano over there too. Well, there's— yeah, so there's— I know that looks like a crazy weird guitar. There's actually a rainstick in front of it right there.

That's a rainstick.

Okay. And then there's an electric guitar behind it and then an acoustic and then I've got another one the other room anyway. So anyway, I'm a musician and I wanna see if I can, if I can build something that can, that could upload and like submit my song to like CD Baby or DistroKid or something, right? So that it's ready for distribution. But, but I want it to do the whole thing with me just like, hey, upload this song, here's the file. I want you to do it in my CD Baby account, whatever, and just like get to work. Right? So I want to see if I can build something to completely do that. There's a lot of different pieces that I've put together. It's probably like 25 things that I have to put together in terms of like different pieces of software and things behind the scenes. But I really want to see if it's possible. Nobody has done it yet. I have not seen it. I've looked around. I am very up to date with everything that AI can do. And I've tried all of the different, you know, agents and multi-agent systems. Nobody has built something that allows you to do an automated to-do list. Okay? So let's describe the app. All right. An app that— a to-do list that does your list, does the— what is it? Yeah, the tasks for you. I add a task to the list and it uses AI and automated systems to just do it. It asks the user, it pauses tasks and asks the user for input when needed and then resumes. Build the landing page. For this and the app itself. All right, go. I promise you, this is a relatively bad— not relatively, this is a bad prompt. I always like to start with what most people do. A lot of people use ChatGPT and AI, it's just like a fancy Google. Yeah, it's smarter and I can have a conversation with it, but no, it's actually so much more complex than that. Understanding how to talk to it, can be crazy, right? So anyway, AI-powered task automation system with the landing page core application. Innovative concept where users can add tasks and system attempts to complete them autonomously using AI with the ability to pause and request user input when needed. Okay, approve plan. So now I'm just going to let it get to work. So what Replit is, guys, Replit is a system that is using something called chain of thought reasoning. So while this is going, we're going to pick on, uh, David here. And David, I want you to tell us, how do you, how do you get ready in the morning? Tell us how you get ready in the morning.

What's your routine? I mean, you know, from showering to everything.

Just what you're— yeah, well, you get up and you get ready, right? Tell us what you do.

What's your, what's your routine? Yeah, I just get up. If I need to trim up any facial hair, I do it.

Then get the beard all nice and perfect.

Yeah. Yeah. Then, then take a shower. Get, get at least partially dressed and then, and then put hair product, beard oil and stuff on and then, you know, deodorant, cologne and then get, you know, put on my dress shirt or whatever I'm wearing for the day. And, you know, and that's pretty much it. And then I'll grab shoes and—

Oh yeah, let's talk about brushing your teeth for a second. Walk us through. Pretend I'm an alien. I just landed on planet Earth. Tell me how to brush my teeth.

Okay, I'm a little bit different because all of my upper teeth I had removed and I have a denture bolted on. So that's all right. So, so basically, I have to use a Waterpik and run it along the top edge of that so that it flushes out anything that potentially got between the gums and it, you know, And then, and then I get a toothbrush and I have an electric one and put some toothpaste on it, run some water over it for a second, and then, and then just move it around on all the lower teeth and then the upper denture to make sure it's all clean and—

Awesome.

I don't know. Okay, perfect. Spit out toothpaste.

Don't swallow it. Yeah, I mean, unless you, you know, you like that kind of thing. But, but yeah, I mean, look, guys, all of you could have given, you know, thank you, David. All of you could have given your different kind of variations on that and whatever it is. But the bottom line is what's happening there is actually chain of thought reasoning, right? I said, David, how do you get ready? And then you broke that down immediately without even really thinking about it. You're like, oh, well, when I get ready, I do this and I do this and this and this and this and this. And then even And inside of that, like brushing your teeth, you had subtasks and subtasks of that, right? You have to, you know, you're going to get your toothbrush and you're going to grab your toothpaste and you're going to open the toothpaste. You're going to put that on. You're going to close the cap. Your body just kind of does this routine automatically by using chain of thought reasoning. And our brain does it really, really well. AI has only been able to do this for a few months, maybe a year, right? Maybe a little bit longer than that behind the scenes. But at least in terms of what we've been able to see, Being able to do really complex, amazing chain of thought reasoning is a newer thing. That's what's happening here. We've given it a task. This is the task. We said, hey, build me an app that does tasks for you. The to-do list that actually does. It then said, okay, well, I need to build a landing page. Well, landing page is going to have a header and an app button and a get started button. Hey, we're going to need some text. Well, I wonder what we should say. Well, this probably makes sense based on what we're doing here. Try it now, probably a demo. You've got a random hero image, how it works, add your tasks, AI takes action, get results. Amazing. See it in action, done doing to do. It's fun. It's crazy that it's thought through this and it's actually getting these types of things done here. But natural language input, Now, I would argue that the design of this is subpar for sure, but that's okay. We'll work on that later. Right now what it's doing is it's thinking through, it's using that chain of thought reasoning to break down tasks and then just come back and do it. Now, if you look here at all of the different things it's building, it's building the schema and storage and routes and app index, landing pages, components, everything that it's going to need. If you click on here and you open the files, you can see that it's made dozens and dozens of files already. If you just look at the components, right, button, calendar, card, carousel, chart, checkbox, all of these different elements here are, are, you know, have been created thinking through it's using that chain of thought reasoning to break down tasks and then just come back and do it. Now, if you look here at all of the different things it's building, it's building, you know, the schema and storage and routes and app index, landing pages, components, everything that it's going to need And if you click on here and you open the files, you can see that it's made dozens and dozens of files already. If you just look at the components, right, button, calendar, card, carousel, chart, checkbox, all of these different elements here are, are, you know, have been created from scratch to, to be able to come in and help, you know, put this thing together. And so it's writing— where'd you go for those files? Yeah, it's up. It's right here on the, on the upper left. Open files right there. Right. So, so we'll, you know, we're going to come back in here later, guys. And this takes 10 to 15 minutes for it to finish even just its first kind of draft. And by the way, when it, when it comes back. So if I try to click on try it now, watch demo, home app, get started, like it's nothing works right. Right now it's what's called a visual preview and then it will go back and it will just start making everything work. Right, so it just built a database. And yeah, this is going to be very, very complex. Um, now we'll, we'll come back to this, we'll come back to this, okay? But you guys get it, chain of thought reasoning. But isn't that— isn't this crazy? Isn't this crazy that you can, you can do this? This is not like— how many of you guys have ever tried to build like a Shopify website or a Wix website or something like that using like a template or whatever? I've done that many times. And it's, it's hard. Like, it takes forever. Like, it's supposed to be easy, but it's not. You get in and you have to choose your template and then you're like, okay, which of these 7,000 templates really is good and does it really have all the features I'm gonna need? And, and then you get in and it's like, okay, now you have to like, you have to write every, every single piece of, you know, everything, you know, every single thing. It's, they're starting to make some, some, some buttons do a couple things now. It's kind of cool. So Anyway, it did make a mistake here. It put white text on a white button, right? So it's still working through it. But now it's at least, you know, basically, basically figuring out, okay, we've got this. If I try to add a— I wonder if there's a database here. I'm curious. Nope, no database. Okay. So it's built a very, very rudimentary thing, right? Task description, you know, order groceries for me. Right, obviously a bad task description. Um, email, scheduling, research, communication, booking, other, priority, urgent, do it now, add task. If I go back, let's see here. Here's my tasks, pending, other, urgent. So we've got a few different little— in progress, no tasks yet, says add your first task. So it obviously isn't all connected together, but guys It built all of this completely from scratch. In order to do this in 2020, you would have to pay a developer. And even just to get to here— now ignore maybe a couple formatting issues, but even just to get to here where you have, you know, kind of a landing page and you've got an app and you've got all these forms and all these things, you're probably paying someone at least $5,000 to $10,000 to just get you to like a very basic functionality. Now, it might be a little bit better. And with maybe 3 or 4 more prompts, I take that price up to $30,000, right? I paid millions and millions and millions and millions of dollars to do, to do things. I built the entire genaiapi.org. Everything that you see on this website was built completely, hosted, everything happened inside of Replit. Very complex, beautiful website. You know, I can, I can log in, I can you know, I have admin sections, you know, my dashboard, test results, all these types of things, all of the live courses, places where you guys signed up for stuff like this, session dates and schedules and reviews. Because I'm logged in as an admin, I see this manage reviews button. I can add those reviews. Got all of these different, you know, live courses, test content, course content, user management, all of these different elements. I could make this user you know, an admin if I want, paid user, whatever I want to do. So it's very, very capable of doing very complex things and building sites that are capable of making money today. Right. So, so anyway, that is kind of your first little basic intro here. What we're going to do now, though, guys, is we're going to go back and we're going to start over. I'm actually going to throw this away, what I've just built with AI Taskmaster. It's fine. I could start working on it, but I'm gonna throw it away. We're gonna start over completely and we're gonna build a really nice prompt and we're gonna do this inside of ChatGPT. So we're gonna do some ChatGPT training right now about how to really get inside of, you know, the right, the right workflow here. And excuse my million tabs that are open. I was working on something last night and I probably should just close out of most of these. Let's just go, oh, you know what? I kind of have some research there that was related to this. I might just leave it open because I was actually working on this idea last night. All right, so first of all, let's get this out of O3 Pro. We don't want to be sitting around all day waiting for this. So ChatGPT, how many of you guys did not have a paid account before this course? Yeah, all of you guys. Well, that's interesting. That's interesting. That's awesome. So there's a lot of cool things that your paid account can do that your free account could not do. So we'll talk about a few of those things. We started talking a little bit earlier. I mentioned the word context, right? Context, you all understand in the context of just normal conversation, but in AI, it's not a whole lot different, although there's two meanings for it. The first meaning is, what do you mean, right? Like, you need to help your AI understand what do you mean. And the other one is like, what kind of background information is there, right? So for example, if I were to say, you know, ask it a religious question right now, It would say, well, um, because I know you're a member of the Church of Jesus Christ of Latter-day Saints, this is, this is the answer I'm going to give you. If I knew in— if, if I had taught it in my, um, in my prompt, oh, you're Jewish, then it would answer that question a little bit differently. And it would probably say, in the context of Jewish, whatever, you know, um, Jewish doctrine would say this. Or, you know, if I'm a Muslim, whatever. The point is if it knows you better, it will work with you better. Okay. So it has— ChatGPT has a cool feature. It can't remember the conversation, you know, all of the different conversations. Sorry, there's actually 3 meanings to the word context. The third one is there's something called a context limit, right? So if you were to take, you know, I don't know, the entire— I can— the thing is, it can probably do that. Let's just— it can't actually— it can reference it in a different way. But if you were to take a new book or a new something, right? You had written, you know, a million-word book and you decided to take that PDF document and just drop it in here into ChatGPT. Its context would be overloaded. It would, it would try, but it would really fail at a lot of things because it can't quite search it with this model. There's other models and other ways that you can do really, really big context. But has a limit of how much it can actually interpret at any given time. And so it does not remember every single thing you've ever talked about in every single, you know, chat here. It just, it can't remember that. It's impossible. So it has something. If you go into settings here, you can see data— no, not data controls, personalization. You can see that it has something called manage memories, right? And so this is memories that, you know, you're trying to do. Now, it already knows is building a scalable autonomous agent system. I'm using Replit. It knows that because I started talking to it last night about this idea, right? So I'm going to open up a project here. I'll show you guys in a second. But, um, you know, if you go down here, if you keep scrolling down, it knows I'm an entrepreneur and a composer and a consultant. It knows I'm a member of the church. It knows I'm a whitewater kayaker. It knows I'm fluent in Spanish and Chinese. It knows these things about me. It knows that I'm the CEO of Velocity Access and, you know, did these other things. Anyway, It knows stuff about me. Because it knows that, because it has that system-level context, it can direct me. When I ask it to do something, it knows that I love to build things using Replit. Whenever I'm asking, hey, I want to build a new thing that does this, it does it in the context of Replit for me, which is really cool. There's a lot of cool things you can do to customize your GPT. I can just say, I'm Jon, what do you do? You could be very specific. What traits should ChatGPT have? Right? Straight shooting, forward thinking, encouraging, Gen Z. You give it all kinds of stuff. Skeptical. You know, you can give it some prompts so that at a system level it's going to do this. Now I do this at a prompt level and at a project level. And so I don't have this filled out and that's okay. But Jon is fine. I can, I can do that. We can save that one. Okay. So we'll save that. And I've got my customization. One other little thing that you can play around with other stuff in here, but data controls. If you are worried about ChatGPT stealing your data or your ideas and you, you really care about privacy, which is fine if you do, you can turn improve the model for everyone to off. And then it will just keep all of your things encrypted and stored and no one will be able to see it. Okay, this is safe enough. I've done research on it. I would put my Social Security number and all my credit cards and my bank statements and things like that, and I can use it to analyze and help me prepare taxes. I am totally okay. To do that inside of ChatGPT. Okay, you make your own decision, do your own research if you want, but I am comfortable doing that because I understand how their data flows and they can't see it. They don't, they don't look at it, even if you have improved the model on for everyone. Although I might turn it off for that type of thing because you just never know what might pop through. But, um, it's, it scrubs data, it scrubs personal data like that, right? They don't want to see that, they don't want that liability. Okay, so they're trying to get away from that, um, or they're trying to stay away from ingesting your data. The other place that there's context, you've got your system-level context and you could talk to it. Then I've got this bump on my hand. I've got this random weird bump right here and I'm getting it removed surgically on Thursday. Right after this class, the next day, I'm getting this done. I've talked to it a bunch about it. I literally took a picture, circled it. I was like, what is this hand? I don't know what it is, this thing on my hand, what's this bump? And then I, you know, talked to it and I went to the doctor and then I figured it out. And now I know exactly what it is. It's this weird thing that only happens in, in people of Scandinavian descent, right? Which is random. And it's a genetic condition where fibrous tissue builds up under the skin of the palm, leads to thickened cords or nodules that can feel like hard lumps, can progress over time across the fingers, especially the ring and pinky to bend inward like that, right? Which I don't want to do because I'm a pianist, right? I'm a pianist. That's what I do. And so I can't have my hands messed up. So anyway, they just, they can't really do anything about it. They just kind of take it out, but it does hurt. It's painful when I touch it. It hurts even just doing that just a little bit. Um, anyway, so, but I could come back here. I could say, um, you know, I only have this on my right hand. Will it show up on my left hand? I can ask it a question about something. Because I came back to the conversation, I don't have to tell it again what I'm doing. Just like we were doing earlier in X with Grok, it just knows. It'll pick up right where you left off. It knows my memory, it knows context inside of any given conversation. Yes, 60-70% of people get it on both hands. That sucks. Really? I'm only 39. So anyway, that's kind of annoying. So maybe at some point in my life I won't be able to play the piano anymore.

That's fun.

But I won't, I won't worry about that right now. The other place you can do context is inside of a new project, right? So when you click on new project, you'll say, you know, example project. I'm going to, I'm going to just do an example one here and I'll delete it later. But you just create the project and then what you can do here is you can give it a bunch of context about something you're building. So Brandon, you've got a business that you've had for 8 years. I bet you have tons of documents, tons of documentation, tons of, you know, operating procedures and SOPs and all kinds of stuff that you, you know, that's about your business, right? And even maybe just a website, right? And, you know, all these different elements, maybe your customer lists, and you could drag all of that into the project files. Right?

Well, that's the issue, right? You're dealing with HIPAA compliance at that point, and it's like, okay, yep, like, would this be considered a HIPAA-compliant file holder?

You know, yeah, it's easy to— it's easy to find out. Is ChatGPT HIPAA compliant? Nope, it's not. To use it with protected health information, you need a business associate agreement with OpenAI, and even then it requires careful implementation. Implementation adherence data rules, right? So you're going to have to build something that's a little bit more internal, right? So enterprise and API versions of ChatGPT may be able to be used in a HIPAA-compliant manner. Okay, so when you're talking to your, your system, right, to Replit and ChatGPT, as you're building out something, if you're going to do something with that, then you're going to have to say this needs to be HIPAA-compliant. Okay, and that's the same thing, by the way, that you would say to your developer, right? This needs to be HIPAA compliant, and then it's going to look it up and it's going to say, okay, in order to be HIPAA compliant, we need this, this, this, and this. We're going to store in this type of database. We have to have this kind of authentication. We have to have this kind of security. We have to have this kind of whatever. We have to anonymize data. We have to hash this, whatever, right? And so there's, there's lots of ways that you can do that, but figure out what exactly is happening there. But if you do that with your own API and your own things, you'll be able to do it. Because when you're using an API, this is really important. We're getting ahead of ourselves a little bit here, maybe by 2 hours, but I just want to say it really quick. ChatGPT, so an API, guys, who here doesn't know what an API is? If you don't, that's okay. I can talk about it. We'll be very clear. An API is OpenAI API. An API is basically a connector. It's going to connect one service to another service or a button to a request. It's a connector. It's connecting two different things. What's it stand for? What does it stand for? It's App Protocol. I don't know. API.

API stand for?

I've— application programming. Yeah, so it, it, again, it's a connector. It basically— yeah, the interface can be thought of as a contract of service between two applications. This contract defines how the two communicate with each other using requests and responses. Okay, so ChatGPT is made by OpenAI, but it's actually just using ChatGPT is a platform, it's an application that uses its own API, that uses OpenAI's API to get access to an LLM, a large language model. So it's basically, and the LLM is basically a brain, right? And so if I send a request to the brain, the brain will answer back, but the brain isn't storing any data. Does that make sense? Whereas ChatGPT is actually storing data. How do I know that? Well, easy. Here's my chat history and everything in here. It's built to do that. It's built to store your data, right? And so ChatGPT will ask the brain a question and then store it for you so that you can reference it later. But if you build your own thing, that's, that's internal, right? It's, it's for Brandon's business only. Brandon's business eyes only. You can ask the brain questions and it will answer back, but it will not learn from you. Okay. ChatGPT is helping OpenAI learn, right? You're talking to it and doing stuff and it's, and it's figuring out how humans talk and what kind of things they care about, right?

It's learning.

Side question on this, Jon.

Yeah.

Have you seen companies use Replit or ChatGPT to eliminate their, their like SEO guy or their PPC guy or— Oh, yeah.

Developers and yeah, okay. Oh, yeah, all of the above. So I, I am building kind of as a challenge, GenAIPI. Solo. I'm not going to hire— well, that's not true. I probably am going to hire a couple people, but, but it's to run it. It's all with AI. I'm only going to do it because I don't want to be able— I don't want to have to talk to the AI as much as I do. But, but my goal is to be as small as humanly possible because, yeah, I can have my marketing team and my sales team and all these things all be AI.

And it just— because SEO, right, Google's constantly changing their algorithm. So if like your content is not reflecting what they're changing on their back end, right? Like your articles are getting pushed down, so you're not organically ranking quick enough.

Yep. So yeah, I just— What about this? Let me throw another wrench in that, Brandon. What if people stop using Google, which they have already? Really? A huge amount. They're using ChatGPT to search now. So are you— is your website and your content optimized for AI, right? If someone says, hey, I'm in Provo and I've got this problem, I've got, you know, drug addiction or my brother's having a problem and I need to put him in a facility, whatever it is, right? Yeah. Is ChatGPT going to say, oh, Brandon's place is where you should go?

Is it even— it only knows because you're loading the content within ChatGPT.

Well, not necessarily. It's just, is your website itself, right? Like, like if I go to my website to go back to my example, if I go here, is, is the code underneath, you know, everything, right? Is all of this, um, developer tools, is all of this code that makes up my website, is it properly tagged and structured in a way that Google recognizes it and indexes it properly? And does AI recognize it and index it properly, right? And so you have to make sure that you're building you're building everything in a way that it's going to understand that. You can't just teach OpenAI's model, you can't teach ChatGPT to become smarter and know about your business just by telling ChatGPT, hey, I have a business that does this, right? It does use your responses and your requests and everything to learn, but you can't like index it, right? It's just, you're just some random one single data point, right? There are— there's actually a company that just raised a bunch of money yesterday. I just heard about them. I think it's this one, Profound, right here. So Profound is— let's go try profound.com. I don't know what's up with their website, but yeah, get your brand mentioned by ChatGPT or Google AI reviews, right? It's a company that literally works with you And, and by the way, it's freaking expensive, right? I came in to like try it. I wanted to, I wanted to like, oh yeah, this is cool. This is awesome, right? Look, it's only $499 a month. It's crazy. You save 15% if you do an annual price, $5,000 a year. And so, but, but then, you know, so first you're gonna analyze how your brand and your competitors appear in AI platforms, and then it's gonna tell you, here's what we're gonna do to fix it. Here are the actions and insights and things like that. To make your website better so that AI will actually know you even exist.

So you think this is a good idea or a bad idea?

Oh, I think it's great if you can afford $4.99 a month. Yeah, it's just expensive, man. Like, like I could afford this obviously, but I don't, I don't want to pay that. It feels overpriced to me. It feels 10 times overpriced to me, especially because I know it's automated in the backend. Like, this is ridiculous.

Right. Well, I gotta tell you real quick, my company, we're paying almost $5,000 a month for two different AI systems that through API integrate with the software that we use.

Yeah, no, I think there's use cases for that type of a bill for sure. And by the way, I'm paying, you know, $250 for Google's AI and $200 for ChatGPT Pro and my My, my monthly AI bill is a few thousand dollars just for access to all the things that I use. But it's because, you know, I'm a superpower user and I teach it and I work with companies and I, you know, so I need access to everything. I need to know how to use everything. And I'm constantly playing with it and I'm connecting things together. But for one service like this to do one little thing, yeah, this is crazy. So you know what I did last night? I talked to— I found this thing and I jumped into X. And I said, okay, Grok, tell me, sorry, let's see if I can find it. Maybe it's over here. Maybe it's over here. Yeah, yeah, yeah, yeah. This is it. This is it. So I was like, what does Profound do to help companies show up in AI search? Tell me about their tech. Give me a solid understanding of what's going on there. I need, I need a, I need a game plan for this, right? And so I went in and I figured it out and I read what they're doing. Now I know exactly what's happening here and it gave me all these different answers. Here's how you're going to do these technical optimizations and stuff like that. I copied all of this and I took it into my Replit project. I came in here into apps and GenAIPI and I copied and pasted all of that stuff into here into Replit. I said, please read this article or whatever. Please read this and analyze the current state of my GenAIPI codebase and structure. Don't write any code. Just analyze and report back about everything you can see in this article, right? So after reading about that, here's what I figured out. This is what you need to do. And then I had it go through step by step by step, phase 1, phase 2, phase 3, phase 4. And I had it implement everything that Profound would have charged me $499 a month to do. Oh, and so I just saved myself $5,000 by asking AI to tell me What are they doing? Great, make a plan for me. Great, now do it and I'll monitor you and watch you. And I had to, I had to like follow along and, you know, but like proceed with the plan, you know, like it's not like I'm doing anything complicated, right? I'm just reading and understanding, right? And then, you know, I started doing some other things on, you know.

So it sounds like David, I mean, for David's situation, it sounds like he would be able to use this, but only if like he's giving the right prompts to do so.

Yeah, yeah. I mean, it's all about, it's all about just understanding what you want your outcome to be, right? All right, guys, we're a little bit off topic here, which is okay. This is what's fun with 3 people is I'm actually letting us go off on little tangents here and there because it's kind of fun. But I do want to get back to, I do want to get back to this ChatGPT element here. We're talking about context. We're building a project. We can add files, all this context. You're not going to be able to do that, Brandon, yet. So don't add any project files in. We're going to show you how to build your own custom and secure database that will be HIPAA compliant. I'll show you how to do that later. But right now, and I usually don't add files at this stage anyway, then I'm going to add some instructions. The instructions could be whatever, like always call me Bob or whatever. You can tell it to call you something. You could— Andy Bernard.

Or Andy Bernard.

Yeah.

But your voice.

Is equivalent to Andy Bernard from The Office. Only talk to me like Michael Scott, right?

No, like, have you ever heard that before? Because it was driving me nuts. I was like, Jon's voice sounds like someone.

And I was like— I don't know. I don't know if I've heard that. Andy Bernard? That's awesome.

I'll take it. Let's put in the prompt. Put in the prompt.

What does my voice sound like? Yeah, you would have to— I don't know. I don't know if it can do that yet. I don't know if AI can analyze your voice to see who you actually— I probably could. It probably could.

Aren't they worried about like, like infringement on like something, right?

Yeah, on copyrights. Copyrights, something like that. But right now, AI is the Wild West as it comes to copyright infringement. Um, there's so many things today that we're going to try and do where we can reference Disney stuff, we can reference, you know, obviously I've got NBC Universal personal stuff here where you can reference, you know, Steve Jobs or Apple or whatever. And, and it just, it can take stuff from you. You can tell Replit, you could say, I want you to build my, my tech website. I want it to look exactly like apple.com. Copy all of the fonts, all of the formatting, all of the animations and everything. Just change the words and the pictures so it's about my product instead of Apple's products, right? And boom, you've stolen all of the, the hard work that their, you know, million-dollar UI designers and UX designers and web page designers and all this stuff have done.

It's all yours, right? That's really the first time you've heard that?

Your voice sounds like— Yeah, I've never heard that before. That's funny. I've watched The Office. I've watched The Office all the way through, every episode, I think 16 times. I'm way up there. I'm a big Office fan, and it's surprising that no one's ever said that, but I'll take it.

And he's a big music guy too.

So he is a big music guy, although I can't, I can't sing. That's not, not that he can, but no, he's fine anyway. So only talk to me like Michael Scott. Actively try to improve my ideas. Don't just be a yes man and argue with me. Argue with me in a way that helps me improve my ideas, etc., etc. So you can tell it to do things like this. You could say, yeah, anyway, we'll just, we'll just do that basic prompt. That's kind of whatever. But now I've got somebody who's helping me, right? It understands a little bit better and say, okay, I want to sell more paper. Okay. And then I can hop in and Michael Scott will tell me how to do that. Okay. And that's what I'm talking about. That's the spirit of Dunder Mifflin. See, I didn't even tell it Dunder Mifflin. It just knows who Michael Scott is. Okay, so who's buying this paper? School district, law firms. Let's be real, paper is paper unless it's not. Okay, is it cheaper? Is it faster? Differentiate or die, baby. Hit LinkedIn. So it's going to— it's going to just start talking to me about whatever I've asked it to talk about. But the cool thing is now that I'm inside of this project, okay, I've got this project right here and I can come back to it, right? Let's go. Let's click here and then let's click back to the, to the project again. Sorry, I've got a lot of things here. Okay, so now I could start a new conversation and eventually I could have potentially, you know, dozens and dozens of, of, of conversations in here. Where are the rest of my projects? Please let me get them. Weird. Anyway, here's, here's my general AI proficiency institute project. I have so many conversations in here, although ChatGPT is not It's not scrolling down. I have all of my conversations in here, but they're not showing. Hold on, we're going to click refresh because that doesn't feel right to me. Where's all of my conversations? Okay, we're going to go into this and we're going to click on there. Now I'm just worried. Um, all right, well, for some reason they're not loading on the web interface. This— you can download the ChatGPT app That's what I'm in right now. This is just the, you know, it's, it's free. You just download it and then log in. But I can see all of the different conversations I've ever had with this project, right? And so it remembers the context of everything you've done here. Okay? So, when you're starting your project, and we're gonna have you guys do this here in a minute, and we'll stop for a second, talk about it. But go in, set up your prompts as you want, the instructions if you want. If you have any files that you wanna upload, then great. And then you're going to just start talking to it. So before we, before we kind of let you guys loose on the first thing here, you know what, we're going to go a little bit further and then I'm going to, and then I'm going to set you loose. Okay. So here's my AI to-do project. I've just started it. I literally just started having this conversation last night. Okay. And, and I basically said, all right, so I had some conversation with Grok and I copy pasted all of that into here. I was like, okay, I'm building this. I'm building this thing now. Okay. Then what I'm doing here is I had it. Now it understands what I'm trying to do, and I think I'm at a good spot where I can start a new thing here. I'm going to start a new conversation for this thing right here because I think we're into a new thing here. I want you to refer to all, to everything you have learned about this in this, in the context of this project. Normally you guys wouldn't have to put that line, but I'm going to just make sure that we say that. But I'm going to say this, I'm going to say I want to create a business plan for this AI to-do project company. Actually, we're going to call it a company. Company. And you could be a project, do whatever you want. Okay. If you're not building something new, that's fine. And we're going to talk about what you guys are building here in a second. But I want to create a business plan for this AI to-do company. I want you to interview me one question at a time about this business. We need to come up with a name, USP, unique selling proposition, marketing and sales plans, monetization. Strategies and everything else. You know what I'm actually going to do? I'm going to write branding in here and everything else that you need to know in order to help me write a great business plan. Okay, so I'm going to copy this prompt for you guys and we'll drop it into the chat here. Oh, there it is. And So you'll be able to use that in a second here when you start your own project. But, but now what I'm doing here, this is cool. Instead of having the AI do all of the thinking for me, I'm gonna have it pull out of my brain all of the things that I want it to know instead of just letting it guide the show, right? Be in charge. I wanna be in charge. I want it to help me and organize my thoughts. But now, okay, great. Who, who, who's your, you know, vision and target user? What's the core vision of the company? Boom, right? Now I can just answer this, right? So what do you want to, what do you want to create in the world through this AI to-do system? Who's it for? Founders, busy professionals, enterprise teams, creatives, all the above. So I'll just start answering. As soon as I click enter, go. I might say, I'm not sure how to articulate this exactly. Can you help me based off of everything you know from my other conversation? In this project? Yep. I have a solid foundation based on everything you've shared so far. Let me see. Let me, let me look. Right. Your AI to-do system, a draft of vision. So you're building a scalable autonomous agent system that lets users— when I get ready, I do this and I do this and this and this and this and this. And then even inside of that, like brushing your teeth, you had subtasks and subtasks of that, right? You have to, you know, you're going to get your toothbrush and you're going to grab your toothpaste and you're going to open the toothpaste. You're going to put that on. You're going to close the cap. Your body just kind of does this, this routine automatically by using chain of thought reasoning. Drop in any kind of task or goal from simple to deeply complex, and the AI gets to work immediately, only pausing to ask for human input when absolutely necessary. So boom, it already knows what I want. It's not a glorified task manager. It's these things. Your ultimate goal is to unlock human potential by delegating the busywork and execution-heavy tasks to AI. Vision framed. To create the world's first truly autonomous AI task agent and assistant that doesn't just help you think but actually gets in our brain, does it really, really well. AI has only been able to do this for a few months, maybe a year, right? Maybe a little bit longer than that behind the scenes. But at least in terms of what we've been able to see, being able to do really complex, amazing chain of thought reasoning is a newer thing, right? And that's what's happening here, right? We've given it a task. This is the things done for founders, creators, and doers who want to move faster than ever before. Pretty good. Approved. Okay, I like it. Okay, so now normally I'm not going to just do it that fast. I'm not going to give up and let the AI do it for me. I had already had 2 hours of conversations with my AI last night, and so it already pulled that out of my head a lot. And so that, that's really in line. But now I just say, okay, great, approved. Okay, vision's locked. Next, naming. What are we going to call this? Right now I dive in. Okay, so this is where I'm going to stop and you guys are going to dive in and do it. So we're going to first— first, we're going to do a quick little round robin again. We'll go David, Hunter, Brandon, and I want you to tell me what you want to build over the next 2 days here. Okay.

There's a lot of things that I'm going to use this learning for, but the first— the thing that I'm bringing to this that I want to work on is I started, uh, start— I built an Excel spreadsheet that I was starting to try to create formulas and all this stuff on basically things that determine whether a potential client or even an existing client is a good fit with my company. So basically a scoring system that will look at like how much deferred maintenance do they have, how well have they been funding reserves, You know, how old is the community? Because when a community gets to like 20 to 30 years old, roofs need replaced and other things. So, older communities, more work, you know, do they have lawsuits? Like a whole bunch of data like that, that then I came up with kind of some, a scorecard for each of those subjects. And then, and then some extra credit points too, where if I know they have certain aspects to the community then it adds revenue to it. So then it makes me want to overlook maybe some other issues that create extra work, right? So that's kind of where my mind is on this. And then trying to use AI to build something that looks good that I can literally just plug in answers to those things and pump out a score.

I love it. So almost a client scoring system. Right. Yeah. Or future. Perfect. Awesome. I have no more feedback. You'll be able to just describe that. So, so here's what you're going to do. You can get started if you want right now. Open ChatGPT, start the new project. Okay. Just click new project. Give it a name. Doesn't matter what the name is, right? You can just say scoring, you know, customer scoring tool, and then you'll figure out a cool name later. And then, and then just start telling it, just, just type everything you just told me. Just type it. You could also go on mute and you could turn on the, you know, the dictate. You know, you can just talk to it, right? And have it just, have it just, you know, do that. And then you're going to use the prompt that I just gave you that I said that says, you know, interview me one question at a time and, you know, let's build a business plan, right? And then, and then just start answering the interview. And then we're going to go, guys, after we all go through this, we'll go for about 15 minutes and you'll just kind of work on this. Okay, so this Go ahead and get started, David.

Hunter, go ahead. Yeah, so I have a few ideas. I think the one I wanted to try out here, um, seems like a big task, so hopefully it's not too much. Basically, I'm building something crazy today, so like, yeah, so it'd be like a, an AI-powered like land use feasibility tool. Okay, um, and so basically you could submit like a parcel of land and it would tell you like the zoning, how it could be rezoned, utilities on site, whatnot. And then have it also kind of tell you like estimated development costs, what the market is dictating the product, like ADP, average delivered price could be, and maybe spit out some high-level like ROI, IRR, financial metrics.

I love it. Um, all of that is very easy. The hardest piece in my mind is where is that data available somewhere? Like, if you want to find that data, like, where do you go? Is there a spot like, oh, I'm going to go to this website, I'm going to do this kind of query, and this is what I'm going to find?

Yeah, so that's kind of the issue. Like, it's available, it's just— so it's kind of what I do for work, right? But it's like I'm visiting 10 different websites and having to search. So if I had the ability to like put it in one place and then it spits out pulling from all these websites.

I love it. No, you can definitely do that. You're going to use the Perplexity API. So Perplexity is able to actually search the internet live for you and find that type of information. And so, um, you don't have to know how to do that yet, but it's very easy. It'll take you all about 30 seconds to implement. You'll say, let's use the Perplexity API to do this. And I'll say, great, grab your key, copy paste, and you're done. So we'll do that later, but I love it. That's really cool. Go and get started with your project. Tell them exactly everything you just told me.

It's awesome. Brandon, wait. Okay, two ideas came to mind. You let me know which one you think might be a little better on this front. So, you know, one component of our business is we've got humans that are answering the phone And, you know, humans err and flaw. In our industry, our, our employees are, you know, they, they've got their up days and down days. And so like when somebody's calling, you know, for help and assistance, it just really depends on their mood at that moment as how they're going to project in that phone call. So I don't know if there's a way that we could create, um, AI to answer the phone calls without any— I mean, with emotion, but without, like, emotion, if that makes sense. So that judgment without— yeah, no, exactly. Because, like, for example, like, I called Gusto, this payroll company, yesterday, and I, I just got somebody who was just irritated, you know, and then irritated with my question and concern. And so it was like they were just really blunt and short, and I just like, you, like Why are you such a horrible human being? You know, it's like I canceled my account immediately after that phone call, you know, just because of the way they're responding. So it's like there's got to be a way where, you know, and I don't know if we could create a, you know, a system like that where they're answering the phone call that sounds like a human and then create a system on top of that where you have another AI model monitoring that model to make sure that they're doing the job correctly and then give that AI model updates based upon how it's going.

You know, I don't know, like, so it is possible. Everything you're talking about is possible. It's a little bit complex to do this, the latter part of that, right?

Getting it to answer—

you've got to monitor, monitor— you're having— well, you could have it monitor and give feedback, but then to have it automatically, like, self- like self-improve, right? So it's like it's getting feedback and then it's telling the model, here's how we're going to change the prompt and we're going to get better and better and better at this. Um, it, it's possible. It's definitely possible. I'm trying to think. I think if you have—

and it will sound human-like and like—

oh yeah, we'll do that. That's the easy part. So I think that's a great idea and it brings the— the reason I think you should do that one— I don't know what your other one is, I'm happy to hear it, but it will allow us to teach the rest of the group a lot of cool voice stuff as we, as we work with you on it. Okay.

Well, the other one was just, I mean, you were touching on it where it's like, how do we eliminate SEO and PPC costs, right?

Yeah.

Somebody monitoring that and, and then using this information to look at all competitors' sites and see what it is that they're using so that we can like just constantly be ranking higher and Yeah, I don't know. Like, I don't know.

Like it's—

you could build that. That one's not terribly difficult. It's a scraper, right? You're basically just constantly— it's a, it's a cron job scraper. So maybe once a day it would go out, it would look at all your competitors' websites, see what they're doing, see what kind of keywords, see if there's any changes, new blogs, anything like that. And then come back with a daily recommendation of, hey, here's what your competitors are doing, here's how you can rank higher. And then you could take it a step further and then say, okay, great. Now, now do those things. Things for me, right? Write a new blog that fills in that gap and, or, you know, restructure this. And, you know, it'd probably be kind of like a— there's a one-time thing and then maybe an ongoing thing that you would do for it. So, but I think you'll— I think you'll have more fun building the first one today. Okay. So, yeah, awesome. Let's, let's, let's dive in. So go ahead and start your projects and tell it everything you just told me about, hey, I need— I want to build an AI voice system. That, you know, answers my phone calls. This is the type of business I am. And just start your project and then, and then say, all right, interview me one question at a time. Use the prompt that I shared with you guys in the chat. Okay, I'm going to create a business plan. I want you to interview me and let's go. So it's 10:37 now. Let's go to 10:50. So you got 13 minutes to build an entire business plan. That's all you get in the world of AI. Let's see here, I need to put my Do Not Disturb on so I don't get phone calls.

Jon, sorry, we're just, we're just creating a business plan.

Is that what you're saying? Yeah, right now we're just doing that business plan where you should be going through your interview, right? Get through about 10 questions, right? Going to ask you different things, you know, and you might change the prompt a little bit, Brandon, to like, I want to create a plan, a business plan for this project in my company. Right. So it's the same thing. A project in your company is kind of like a mini business inside of your company. So, okay. By the way, guys, when you're building a business, you got to come up with a name, right? Some of you might be building new things. Um, I always like to just kind of go through the branding exercise and then check GoDaddy, see if it's available. There's different ways you can do that. By the way, guys, when you're building a business, you got to come up with a name, right? Some of you might be building new things. Things. I always like to just kind of go through the branding exercise and then check GoDaddy, see if it's available.

There's different ways you can do that. I just want to mention, it's crazy the questions it's asking and how, um, it makes you think about things, how thoughtful it is.

Yeah, you're like, oh yeah, I should think about that. Oh, I do need to answer that question.

Yeah, it's fun. I just want to mention, it's crazy, the question is asking and how, um, it makes you think about things.

And how thoughtful it is. Yeah, you're like, oh yeah, I should think about that. Oh, I do. You need to answer that question. Yeah, it's fun. I had the best idea for a name. I wanted— I wanted it to be— well, Done would be a cool name, right? And then for the domain, like done.com is obviously taken, done.ai was taken. Then I was like, what about donewith.ai, right? Then be like, done with AI. Taken. Someone already did it, man. Dang, I was so sad when I checked that. You guys take just another couple minutes. How are you coming? Do you need a few more minutes? You want to extend this deadline a little bit?

Are you feeling— what, what point are we trying to get to?

You just want to get to where you've had a good conversation with it. You feel like it understands the concept. Now you should have a name. You should have maybe, maybe some branding. Your, you know, your vision and your mission. What are you trying to do?

It should understand those types of things. I tweaked your original prompt because I don't— it's a project, not a business. So, so I don't really need a name and marketing and all that, but it's I've been going through the questions for it to build out. It even— I told it I had an Excel spreadsheet with a lot of the data I was considering, so it asked me to include that. So I gave it to it, and very cool. Now it's asking questions related. Do you want a drop— do you prefer a dropdown box? Do you prefer to manually enter, or would you like it to automatically pull it from known client data? Like, it's— yeah, that's cool.

The questions it's asking are pretty crazy. Awesome. Well, it sounds like you're pretty far along, David. That's good. Hunter, what about you? Are you— I had the best idea for a name.

I wanted it to be coming. It's going good. We're getting nice and intimate with each other.

Uh, no, it's— You need a few more minutes or you feel like it's got— I wanted it to be, well, Done would be a cool name, right?

And then for the button. Yeah, a couple more minutes would be— I'm in a good spot, but it's— Yeah, yeah.

How about you, Brandon? I'm on a roll.

That's perfect.

Sorry, I stepped away to go to the bathroom.

So you're just, um— How far along are you here? Feel like you need a few more minutes maybe to work on this?

Oh, I have no, I have no idea. Like, like what, what, what, what should it look like at the end here? How many questions have you gone through in the interview?

Uh, 1, 2, 3, 4. The domain, like done.com is obviously taken. Done.ai was taken. Then I was like, 6. Okay, cool. I, I want you to get to 10. So guys, take another, let's say like, what about done with? .ai, right? Then be like, done with AI. I think let's say 5 minutes. Okay, we'll go to 10:55. Okay, we'll extend you just a little bit. And taken. Someone already did it, man. Okay, so sad when I— and just keep answering questions. Okay, if it, if it starts to get to a point where it's like, hey, I want to build— check that— the, you know, a development plan for you, like Don't just say, don't do that. Keep asking me questions about the business. Just redirect it, right? Ask me questions about the business itself. How can we clarify? Whatever, right? We'll get— we'll let— we'll let Replit— we'll let other AIs worry about that piece in a minute. And it here. Hey guys, just one more minute, then we'll move forward. Right, guys. So, all right. So you've been through this experience of being interviewed. Now, I use this this approach almost every day in something, right? Like I'm trying to figure something out and I'm like, hey, just interview me about this and what am I not thinking of, right? And so it will ask me the questions that will fill in the gaps of what I haven't considered yet, right? Like, why are you really going to be different? And what have you— have you thought about this, right? And I really ask the AI to be, you know, to be hard on me, right? I usually in my prompts are going to say, hey, I need you to challenge me. Me, right? Tell me when my idea is bad. Like, don't be afraid of your AI, right? Either the AI is built— its prompt is built in a way to— it wants to please you, wants to make you feel good about yourself. And so sometimes it will— it just be like, hey, great idea, Brandon. Like, yeah, you should definitely, you know, put dogs on the phone instead of people, right? Like, that's really innovative, you know. Like, no, no, it's just kind of stupid. I don't think you're really— be a good idea. So you need it to push back on your bad ideas. Okay, so now we're gonna do from here, guys. I've unfortunately spent like the whole time trying to figure out a freaking name, right? Um, and I tried so many different combinations over here on GoDaddy, and I just, I can't find it. I want to, I want to come up with like a name of something, like I want to ask Molly to do it, or Nick, or Tad, and you know, all these different things. And some of them are available, like Tad is amazing because it's only 3 letters, Tad AI. But You know, I'm going to have to pay for it. Oh, maybe Tad wasn't available. There was another one that was like $400,000. And anyway, oh, it was Nate. I think it was Nate. Tad was gone. Nate.ai is available. $90,000. So coming up with names is really, really tough. So, so anyway, I'm just going to say for now, let's call it— let's call it— let's call it Tad. All right. So, so we'll do that. So now what it's, what it's doing is going through and it's, it's, you know, you've had these conversations, right? So what I want you to do is, is this next prompt. So now you've gone through some questions. It could keep going and it could keep, you know, I normally would have gone through many, many questions here, but the prompt that we're going to do here that I'll have you guys do is this. You just say, please build a 1500-word business— word— business plan with everything we've discussed. Use the Canvas tool to do this. Don't focus on the actual software development side, just focus on the business itself. On the nitty-gritty details of the actual— okay, just focus on the business itself. Yeah, we'll just start with that. Okay, so we're going to go paste, give you guys that chat. And so when I do this, when I ask it to do this, it's going to open up a new tool. Okay. And that tool is called the Canvas. Right. So go ahead and paste that in. You'll see what happens here when you do it on yours. It'll open up a new kind of document. All right. And yeah, so just work on that for a second and then go ahead and take a minute to read through your business plan. Okay.

And I don't know why did it open up another document like it knew because of the prompt you're asking.

Yeah, it did that because I asked it to use the Canvas tool. So this is introducing it to a new tool called Canvas. And this Canvas tool is like a Google Doc or a Word doc. It's editable after it's finished. You can edit and go in and change things and then you close out of it and it saves that document kind of right there in your, in your chat. You can ask ChatGPT to go back and rewrite it or to change it or to do something or to tweak it. Now, your business plan, even if it's just a project, right, it's okay. You need a business plan for this project and how you're going to make, make this idea a little bit better. Okay, so let it build that out. And now you've got a business plan. Okay. So let me, let me have mine do it really quick. And it's going to build it out. It should know everything from my conversation yesterday to kind of fill all of this stuff out. I might have to fill in some gaps later, but this will be enough to at least get me started. ChatGPT is going really slowly today for me. I don't understand why. I don't think it's because I have too many things open. I mean, I've got a few, but I don't even close to what I've done in the past. It's okay. Problems, solution. The solution? Tad. All right. Podcasting. Same proposition. Truly autonomous. Target market. Identity. Revenue model. Unlimited tasks for $20 a month. I don't think that's gonna be what it is, but that's okay. Go-to-market strategy launch phase, 0 to 3 months. Nope, I'm gonna do all this in a week. I actually launch it— I don't actually know if I'm going to launch this. I'm going to see if I can build it. That's actually my, my first, you know, my first task. I'm probably going to talk about this a lot on LinkedIn. If you follow me there, you'll see me talking about it. Competitive landscape. It's doing a pretty good— it's doing a pretty dang good plan here. Okay, team and founder roadmap. All right, so when you're done with your business plan, guys, here's what we're going to do. What we want to do here now is if you have a Grok or an X account, right, this is where we're going to use it. If you don't, that's okay, but I want you to at least watch so you can understand what I'm doing here. If you don't have an X account, that's fine, but I'm going to copy the whole business plan. There's that little copy button up here, and then I'm going to come over back into x.com and I'm going to go back to Grok and I'm going to say, give me feedback on this plan. Okay, then I'm going to click this button that says deep search. Okay, what deep search is going to do is it's going to then look at, you know, all of the different articles out there and it's going to look at X, it's going to you know, all Twitter and all the different things that are out there. So I'm just going to click go. And so now I'm taking this business plan, I'm throwing it into a completely different model that thinks differently. I'm asking it to search and look at all of its databases and everything it has capability of doing. And it's going to run, you know, a month's worth of, worth of research in, you know, probably about 2 or 3 minutes here. So go ahead and do this, guys, with your plan. Go ahead and open Grok, do this. If you're not paying for Grok, you may or may not have access to deep search. I'm not sure. But you should still be able to ask it for some feedback. And it will be a different model that will give you some feedback on it. It's worth, it's worth bouncing it off of that other model.

So go ahead and set up an X account.

Jon, is it easy? I mean, signing up for an X account takes like 60 seconds, 40 seconds. Okay. Awesome. Yeah, yeah, it's free. Just set up an X account. Yeah, it's your— or yeah, it's your Twitter account if you used to have that. Yeah, I don't have social media. Well, there you go. Well, then I don't really know how to help you.

I know it's time to get on board, right?

Get on board. You know what? You can get on board or not. I just— it's hard for me to think about what my life would be like without social media because I do all of my business on social media. Like, it's just that's where I find customers. It's where I find connections. It's where I do all kinds of stuff, right? I go to— I go to in-person stuff too, right? Obviously. But like, Um, yeah, that's my, that's my network where I talk about stuff and where I get feedback and all kinds of things. So anyway, all right, so in a minute for me, a minute and 6 seconds to check 46 sources and says, hey, this is a good idea. Here's the market fit opportunity, here's your risks and challenges, here's analysis, technology feasibility, um, you know, competitive landscape, business model, go-to-market strategy. Great. Here are my, here's recommendations on how you can make better. Okay, so go ahead and do that for a second, guys. Take a minute to read everything that you've got going there.

Is it telling looking at the number of sources that it uses in the deep search? So for example, mine, mine used 10. Does that mean like it's a good idea because there's not a lot of stuff, or is that like a red flag because there's no validation?

You want, right? I mean sometimes, uh, did you click deep search? Yeah, yeah, yeah. If you click deep search, you got 10. Maybe, maybe there's not a lot out there. So that could be, that could be a good thing. Um, my guess is it's the latter because what you're trying to build is, you know, doesn't really exist. There's not a lot of resources. It sounds like every time you're trying to find information on land, it's kind of all over the place. And so you build a new process every time and you're trying to consolidate that and build a system that will help you do your job better. And I hope you're thinking about it, Hunter, in a way where you're like, hey, yeah, I'm going to make my job easier, but also what if I built this and sold it to other people that are working on projects?

You know, I want—

Yeah, for sure. I can make money and you can charge $1,000 a month for access or whatever you want to do. So however awesome it might be. Okay, so guys, after you've done this, after you've gotten some feedback here, One thing that I will do is I will actually unclick the deep search here and then I will turn on the think, which actually is going to use a different model. I'm going to say, give me 10 ways to improve this plan. Note that I will worry about the technical side outside or technical aspects outside of of the, this business plan elements. Okay, I'll, I'll, uh, yeah, I'll just drop that over. Okay, so now I'm just asking it to use a different model to think about it, to give me 10 specific concrete ideas on how to improve the plan. And, um, and now I'm getting some feedback. So I'm getting some research feedback, I'm getting some thinking feedback, some reasoning feedback. And then what we're going to do with this, guys, is we're going to copy the outputs. I'm going to copy this output, the first one, and I'm going to copy the second one. I'm going to go back to ChatGPT here. So I'm going to start doing it right now. I'm going to copy, I'm going to go back to ChatGPT, and I'm going to say— where did my thing go? Oh, you know what, I was doing it over here in the, in the app itself. So now I'm going to say, all right, Use this feedback and write a second version of the business plan and call it improved business plan. Use the canvas tool to do that. Here is the research and feedback. Okay, then I'm going to Shift+Enter, Enter, Shift+Return, Return to make sure I can paste this in. I'm going to paste everything I got from the first response. Then I'm going to come back to X again. I'm going to grab the second response here. 10 concrete ideas. Copy text. And then I'm going to come back to ChatGPT. I'm going to paste that also. So now I'm giving it this, use this feedback, write a second version, right? And here's that prompt in the chat as well. And then I'm going to just say go, right? And so now it's going to Create a new canvas. It's going to open this up, create a brand new canvas document, and it's going to rewrite the business plan considering everything we've discussed, everything. This is all the research it's done for me. And now I'm getting a nice refined business plan here that's pulled in a bunch of different information and started to, to work through this. And why do this? Yeah, because we're— I mean, we're just improving. You want a solid business plan, right? What you're trying to do here when you're building a business or a project or whatever using AI, you're doing everything at hyperspeed. And so if you were building a business plan without AI, you would say, okay, or a project plan, maybe you're just trying to build something new like you're doing with your phone system, right? Hey, I'm working on this thing. I want to think through this. I want to make sure I get the ideas right. But then you want some feedback on it, right? You'd hand that plan over to your co-founder or to your advisor, to your investor, whoever, and say, hey, what do you think of the plan? And, you know, a week week later they'd come back with a few minor thoughts, right? But here you can, you can have it bounce it off of other models, other AI models that are going to think differently, okay? And they're going to say, hey, you could probably improve this. And then you take all those ideas, and instead of having to manually go and throw it in, you just copy paste it back in. And now we have a new and improved business plan that's now been kind of bounced off of the market a little bit because you threw it against X. And you know, it's done a SWOT analysis now. It didn't have that first time. It's thrown some really great things in here. So now what I have is a fantastic, a fantastic business plan that's pretty much ready to give to an AI to start building something. Now there's some more things that we got to do, but before we do that, but now I've got a really nice business plan with some great feedback using AI to improve. So this is just a skill that I want you to adopt, right? Where instead of just asking one AI the question, question, ask it, ask for feedback, implement that feedback, right? Yeah. Premium Plus on text or premium? Yeah, so premium. I do premium.

What is the cost difference?

You're $400 a year or $85 a year. Yeah, I mean, I do monthly first of all, you know, so you don't have to go spend that right now. But, but yeah, I like— I do the Premium Plus. Because it just gives you better access to better models and you get all the new stuff related to Groq as it comes out. So you do what you want, but all right, so I'll give you guys a second to catch up to me, right? Get this business, this improved business plan. If you have extra time, you're already done, just spend a little time reading it, right? See what you can see, what you can figure out. Read through it, change it. Like, sometimes I'll go and be like, yeah, you know what, I actually don't really care about that. Um, let's just delete that. Yeah, it wrote my tagline. Just say it, Tad does it. Get Tad. Hey Tad. Use Tad.

I don't know if I'll Like that. Sorry, John, just coming back. Yeah, I've got Groq now, so I'm just copying and pasting the, the business plan from ChatGPT and I'm just throwing it in, or I'm asking— am I asking it to do something with this business plan?

Yeah. You're gonna, so you're gonna ask it inside of Groq. You'll use in the chat there. You'll say, you know, give me, give me some ways to, yeah. Sorry. Oh, I see it.

Okay.

I don't know if I, I don't know. You know what? Let me, let me grab the other one. I don't think I copied the first prompt over. Let me, let me do that. I think I just broke GoDaddy because I was trying a bunch of names. Yeah. All I said was give me feedback on this plan. That's my first prompt. Give me feedback on this plan and then paste. Right. Paste your plan in and go. Then you just paste it in, give it a minute to do its, its research, and then, yeah, see what it says. It should only take a couple minutes to do this. So guys, we're going to We're going to let you guys kind of let your AI kind of catch up for a second. But while we're doing that, I'm going to share a video with you guys and we're going to just have a little bit of a discussion here, kind of start a discussion around it. Okay. So let me, let me pull this video up, see if I can find it here. Right here. All right, so I'm going to share my screen again and click play. So this, guys, is, um, Eric Schmidt, who's the former CEO of Google and a big-time player in Silicon Valley, very keyed into everything that's happening there, still very active with Google and around it. So Just 2 and a half minutes, but go ahead and listen to this.

We believe as an industry that in the next 1 year, the vast majority of programmers will be replaced by AI programmers. We also believe that within 1 year, you will have graduate-level mathematicians that are at the tippy top of graduate math programs. So that's 1 year.

Okay.

What happens in 2 years? Well, I've just told you about reasoning and I've told you about programming and I've told you about math. Programming plus math are the basis of sort of our whole digital world. So the evidence and the claims from the research groups in OpenAI and Anthropic and so forth is that they're now somewhere around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement, is the technical term. So what happens when this thing starts to scale? Well, a lot. One way to say this is that within 3 to 5 years, we'll have what is called general intelligence, AGI, which can be defined as a system that is as smart as the smartest mathematician, physicist, you know, artist, writer, thinker, politician, I call this, by the way, the San Francisco consensus because everyone who believes this is in San Francisco. It may be the water. What happens when every single one of us has the equivalent of the smartest human on every problem in our pocket? But the reason I want to make the point here is that in the next year or two, this foundation is being locked in and it's not, we're not going to stop it. Gets much more interesting after that because remember, the computers are now doing self-improvement. They're learning how to plan and they don't have to listen to us anymore. We call that superintelligence or ASI, artificial superintelligence. And this is the theory that there will be computers that are smarter than the sum of humans. The San Francisco consensus is this occurs within 6 years. Just based on scaling. This path is not understood in our society. There's no language for what happens with the arrival of this. That's why it's underhyped. People do not understand what happens when you have intelligence at this level which is largely free.

All right guys, what do you think about that?

The world's coming to an end?

Is that what he's saying? Well, I don't know if you're all going to be replaced, but, but yeah, I mean, maybe a couple things along those lines.

Um, yeah, what else? Well, I mean, at that point, it like truth will become so distorted, it's like, what, what is truth, right? Like, if we're so reliant on machines on that, I mean, that's the danger of us like just jumping into this full force, right?

It is a danger, but as he said, he said we're not going to stop it. Why are we not going to stop it?

Oh, well, one, on the capitalism side, right? Like, we're going to eliminate how much of, of our costs so that we can use this information, right? Have, have AI do the SEO, the PPC, have AI— we don't have a, have a telemarketing We have AI calling people.

Yeah. What happens if the United States, we all say, you know what, we think AI is too dangerous. We're not going to mess with it. What do you think Iran's going to do and China's going to do? Oh, geez. Everybody else, right? Like they're not going to stop, right? If they're not going to stop, well, then we have to, we have to keep going so that not only can we stay in the lead, but we have to be able to defend against whatever their AI crazy stuff can do as well. Right? And so it's, it's like this AI arms race that is not going to stop. Everybody will keep going until, you know, AI then takes over the world.

But it's like you become irrelevant unless you're using this, right?

Like, that's—

well, what it made me think of is, is if I— in this, just throwing out a ballpark, let's say the AI replaces 50% of jobs in America in 6 years. Yeah. So then what do we do as a society, you know, and who is now profiting off of all that work that is, that has replaced people, you know, and I, it almost, it makes you question communism versus capitalism and everything when you think of of, okay, if there's potentially corporations making tons of money now and all these people are out of jobs because of that, you know, how do we as a society need to, um, find ways to tax those corporations to then help people continue to survive that no longer can work, you know? Or, or do they retool and find other places to work? It's just the— that whole aspect of a society, it could be dramatically altered in the next few years?

I think it's going to be a painful shift, guys. I think it's going to be— I think that absolutely we will figure it out. And if you were to fast forward 30 years from now, you would say, okay, we've figured it out. But the shift, the next 5 to 10 years, the shift is going to be crazy. And it might cause rioting in the streets.

Because people don't have jobs. It would have to be.

Yeah, no, for sure. Like, the estimate right now is that somewhere between 400 to 600 million jobs will be replaced by AI by 2030. That's 5 years, guys. Get rid of 500 million jobs in 5 years, take unemployment to 28%. What's going to happen? What's going to happen to everything?

Well, just hyper-regulation at that point, right? Because like, what's Google CEO's name again?

What's his name? Sam? Well, that guy was Eric Schmidt.

Yeah, yeah, Eric. Yeah, well, it's like it's all free right now, right? Like everyone has access to this. Yeah.

So who knows what the outcome is going to be when Everybody has access to it. Yeah. You know, a lot of people talk about universal basic income or maybe even universal, you know, advanced income, you know, things like that where, where just your housing and your food is taken care of and your transportation is taken care of. And then beyond that, you know, go innovate and do whatever you want to do. But, but everyone can eat now because, you know, automation— because take AI, that's one thing. Then you take automation and robotics and then We're not even going to bring quantum into the equation here, but you bring, you bring just AI plus robotics and you can replace almost everybody. Like, literally almost everybody can be replaced. There's very few types of people that can be replaced. Some of the most valuable jobs will be ones that are, you know, providing experiences, right? You know, maybe you own A resort. That's a great job, right? People travel real estate and, you know, real estate is definite. So, you know, it's going to, it's going to make a major shift. We've spent, you know, the last 30 to 40 years training 80 to 90% of our society to be super proficient with computers, to be able to do their job. And now all of a sudden, computers are going to take over their own job and you're going to be like, oh crap, now what do I do?

Right?

So we've done this huge shift to upscale everybody and all of a sudden we don't need any of it.

No.

Wasn't Google trying to get some of the employees from, uh, was it OpenAI for like $100 million?[No speech] And they're trying to poach Facebook.

Yeah, Meta. And, you know, they're, they're offering, yeah, $100 million signing bonuses.

That is nuts. Absolutely crazy.

But they're not leaving, which is— and they're not leaving— crazier. Yeah. What the heck? How is that even possible? I, I agree, Hunter. That's nuts. What do you think?

What do you think, Hunter? Well, I think it's interesting. One, I think you've been talking to Travis Chambers a little bit about the experiential, uh, resort type stuff, which I think is super fascinating too. But I like— you shared a few days ago on LinkedIn and Here I am quoting the teacher to suck up to him, right? But it, it's, I thought it was fascinating, the graphic that you shared about farming, right? And in the 1800s, you know, 90% of Americans were farmers. And if you would've said to them like, hey, in 150, 200 years, like, you know, 1 in 50 or less than that is gonna be a farmer. Like they just could not fathom that. And they'd be like, well, no, like society itself will crumble if that's what it's like. And I think we're kind of at that stage too. Like, I think there's jobs and industries and things that like aren't around or we can't even like conceive yet that will rise because of this, you know, huge shift into AI and automation. I couldn't tell you what that's gonna look like, but I, I'm not as worried about it, I think, as a lot of people, cuz it's like we're adaptable, we'll evolve and like something new will take its place. It's not like we're just gonna be sitting around like, hey, there's, you know, computers doing everything.

What are we gonna do? Yeah, I agree. I agree, Hunter, that that's the right way to think about it. I think that the long term is absolutely going to be okay. It's the short term that I'm worried about. I really am worried about. Yeah, right. There will be, there will be crazy shifts and people are seeing it, right? I mean, Amazon CEO just came out a few weeks ago and said, or maybe even a week ago, I think, and said, hey, we're going to eliminate a ton of jobs, we're going to fire 50,000 more people. Right? And we're going to try to fire more people. He actually said that. He's like, we are actively going to try and replace you. So if you don't want to be replaced, you better be the one that's figuring out how to use it, right? And so, yeah, right now, right now, our job security or our business security, right? The very survival of our business is dependent upon how well we are conversant with AI, right? Can we, can we talk to AI? Can we use AI to accomplish incredible things? Right? So stay on top of it, stay ahead of it. We can have this conversation for another hour easily. But I want to keep going and keep moving forward. We're gonna do— you guys take just another couple minutes. How are you coming? Do you need a few more minutes? You want to extend this deadline a little bit?

What are you feeling? What, what point are we trying to get to?

You just want to get to where you've had a good conversation with it. You feel like it understands the— one more thing. We're going to take a break in about 5 minutes and then. Yeah. So what I want to do here, guys, and we actually might go another 10 minutes because I want to, I want to finish this thing. So you've built a business plan. Okay, that's great. Or a project plan or whatever it is that you've called it. Okay. And business plan is fine, even if it's just for a project that's internal. Now we're going to ask it another thing. Now, oops, here we go. Now use the canvas tool to create a new document called features. I want you to list out every single front-end and back-end feature or capability ability needed to build this. Don't use, uh, concept now. You should have a name, you should have maybe, maybe some branding, um, your, you know, your vision, your mission. What are you trying to do?

It should understand those types of things. Oh, I tweaked your original prompt because I don't—

it's a project-specific, um, code types or services, just list out the features. We're going to drop that chat for you. What I'm going to do here is it's going to say, all right, here's what you're going to need. Mine's really complex and I'm probably going to ask it to do one other thing. Task input and control, user dashboard, account subscription management, support and help. Settings and preferences, Cosmo mode. What on earth are you talking about? Cosmo mode. It's pulling from a totally different project. You don't understand why it would bring up the Cosmo name. I'm going to literally ask it that question. Why did you bring up Cosmo in the feature? Please rewrite this. Without him involved. I'm curious as to what it— yeah, he's not even like answering. Just like, oh, sorry, I'll take Cosmo out of it. It'll just edit it, right? So anyway, go ahead and do this, guys. Go ahead and ask it this question, okay? Um, you know, to have it make this features document. So now we're going to have a business plan document and a features list document, okay? And actually, it should be called features list. I, I meant to say features list. And you can edit it later, you can rename it features list. That's what I want. I don't know if I can edit this thing. Features list. There we go. I edited the prompt for you. This only takes a second. What it's doing though is it's creating this frontend and this backend stuff and then it's going to—

we'll use this. [Speaker] Not a business, so I don't really need a name and marketing and all that, but it's I've been going through the questions for it to build out. It even— I told it I had an Excel spreadsheet with a lot of the data I was considering, so it asked me to include that. So I gave it to it and now it's asking questions related.

Please, in a second, I'm going to add one more little extra document that you guys don't need to do, but I need to do it for mine. Please put together a complete architecture But are you trying to share your screen, Brandon?

Yeah, I was just gonna— I, I don't—

I must have missed something on this. So yeah, say it again.

Go ahead and do it again.

Share the screen.

Yeah, yeah, go ahead and share your screen again.

You see it? Yeah.

Okay. What was that?

Where was I supposed to go? So you've built the business.

So this is the business plan. Do you want a drop? Do you prefer a dropdown box? Do you prefer to manually enter, or would you like it to automatically pull it from known clients?

Now you need to ask it for the features list.

Yeah.

Where would I go? Just ask anything.

Just ask it again right there at the bottom. Client data.

Like, it's—

yeah, that's cool.

The questions it's asking are pretty crazy. Awesome. Well, it sounds like you're pretty far along, David. That's good. Hunter, what about you? Are you— yep. So go ahead and paste that features list prompt that I gave you. You should be able to just Ctrl+paste. Yeah, there we go. Okay, so now it's making a feature list or features list document. And there's all your stuff. I feel like your ChatGPT is working faster than my ChatGPT. Oh, really? It's okay. All right. So you're good. Yeah, that's all you need, Brandon. You're perfect. Okay.

So I'll stop your share coming. It's going good. We're getting nice and intimate with each other. No, it's—

You need a few more minutes or you feel like it's going there for you and I'll come back to mine here. I asked mine to create an architecture plan. Services OAuth 3.0. It's not getting everything that I need it for. I'm going to have to go in. I don't understand why it's not. Accessing the rest of the plan. All right, I'm going to ask it to do this. I'm actually going to ask it to do it in my other thing where I really was going over everything. Okay, user interaction layer, Node.js, Python, GPT-4, AutoGPT. I'm going to have it just do that. Gmail. Okay, integrated via N8N. Hosted externally or API connected to Replit. We're going to API connected. Scalability plan. Yeah, so I know that you guys probably don't— you're not having to do an autonomous agent structure, so I'm going a little bit beyond what you would probably need to do right now. But I just want to, uh, edit just a couple of these things. Task status and inputs are stored in Firebase. I don't want to do that. I think I'm going to say probably, probably yes, Firebase is probably what I need here. Okay, we're going to do API connected. Okay, authentication handling. Includes authentication and I guess that makes sense. Playwright, I say strongly on replicas deployment VM. Yeah, we'll see what it does there. That's crazy. Headless browser control. Headless browser control, guys, means that you can have the AI agent use the browser for you. It can click, it can type, it can submit, it can do anything. And so that allows you to connect to any site instead of just something that has an API. You could say, hey, go to Megaplex and buy me some movie tickets or go to delta.com and buy me some plane tickets.

Yeah, a couple more minutes would be— I'm in a good spot, but it's— Yeah, yeah.

I'm on a roll. That's perfect.

So I stepped away to go to the bathroom.

So you're just— How far along are you here? You feel like you need a few more minutes maybe to work on this?

Oh, I have no— I have no idea. Like, Like, what, what, what, what should it look like at the end here? How many questions have you gone through in the interview?

Uh, 1, 2, 3, 4. Make me a reservation, whatever, right? You can have it, you know, search and do things on side of other— on other sides. And headless means I won't see what it's doing, right? There's going to be— it's just like a computer with no monitor, right? It's just gonna like go do stuff, but you don't have to see it because it I can see it without having to do any of that. All right, tech stack, Firebase, real-time state. All right, I'm looking— I'm feeling pretty good about this now. Okay guys, so what we have now is 3 documents. Okay, I'm going to copy this document. I'm going to come back over here and I'm going to say, uh, plan, business plan. I'm going to say this is the content I want. For the architecture plan, please replace 6. Okay, cool. I want you to get to 10. So guys, take another, let's say, let's say 5 minutes. Okay, we'll go to 10:55. Okay, we'll extend you just a little bit and just keep answering questions. Okay, if it, if it starts to get to a point where it's like, hey, I want to build documents with all of this exactly as I provide it. Below. Okay. So it will, it will now make a new document for me. It's the exact same document in here, but it's going to just override it. But what we have, guys, here while that's working, okay, we'll let it, we'll let it just kind of do its thing. You now have, you can scroll back up in your, in your chat here. Maybe, maybe I can. Why is my computer just not listening to me today? Yeah, it just would not let me scroll up. All right. Architecture plan is done. But, but now I've got, you know, my features list. I've got my business plan, my improved business plan right here. And so now we're going to go ahead and before we take our break here, we're going to start Replit. We're going to get it busy. All right. So this is where you can now go into Replit. All right. We're going to go just click home or whatever and you're going to go to a new place. Say, hey, Jon. Hey, Brandon, what do you want to make? Hey, describe an app or site that you want to create. Now what we're going to do here is we're going to go back to ChatGPT and we're going to ask, we're going to say, give me a 3-sentence description of what we are building. Okay, boom. Okay, and I will give you this little prompt if you want it. Give me a 3-sentence description of what we're building. So now here I am in ChatGPT. I, it knows my full business plan, it knows my feature list, it knows my architecture plan, it knows all this crazy stuff, but, and we know is we're going to call it Tad. Okay, Tad's going to do all of our jobs for us. So now I'm going to copy this 3-sentence description. I'm just going to paste it right here. Don't click Start yet, okay? Just paste your 3-sentence description right here, okay? Then you're going to grab these different documents and you're going to do it like this. You're going to go to like Improve Business Plan. I'll open this up. And I'm going to copy it, and then I'm just going to paste it. And when I paste it, it will show up as like this little attachment kind of thing, okay? Because it's a lot of text, it won't show the whole thing, it'll just show up as an attachment, okay? So do that with your business plan, then close that out. Find your feature list document, do the exact same thing. And then I'm going to add my architecture document. Okay, then I'm going to say, please refer closely to all attached text and documents. Okay, give you that little— whoops, did you accidentally click go? Yeah. Yeah. If you did that, it's okay. You can give it the documents again later. You can— No, I put the documents in. Okay, good. Well, that's okay. I was trying to give it that last prompt. Please repeat. It's okay. It'll work. If you got the 3 sentences in and then those 3 attachments, you're good to go. Okay. I just got 2 attachments. That's fine. Yeah.

Yeah.

Sorry. You did the business plan and the feature list. I added my other thing here. But bottom line is, guys, this is what you do. You click Start Chat. Chat. Then it's going to take a second and it's going to ask you— it's a little purple thing will come up like, should I— do you approve the plan? Right. And you just say, yep, go. Right.

And I can't find— I can't find those 3 things or 2 things or whatever.

They should just be in your chat. Do you want to— do you want to share your screen really quick, David?

We can sure search for them with you.

I feel dumb, but no, don't.

No, really?

Like, okay, by the end of tomorrow, you guys are going to be like experts at all of this stuff. All right, so you just lost your stuff. So just go to your project or your—

where's the project that you created? I thought I clicked an X that I thought was the canvas thing because it looked different than what was on your screen. Screen. Yeah. So I'm hoping I didn't just wipe out everything I've worked on.

I'm sure you didn't. Click refresh on this. Where? Where? Yeah, the top left, the refresh button. Oh, for the page itself. Yep, there you go. All right, client scorecard AI interview. Okay, go ahead and click on that chat right there. Hey, there's all your stuff. Oh, hallelujah. Thank you. Yeah, for sure. So yeah, go ahead and copy those things in, get over to your, you know, your chat. And so guys, I'll share my screen again really quickly here just so you can see this. And I know every time I click share screen, I take over your control. But, but yeah, just click approve plan. You can look at these things and, you know, okay, mock execution engine that simulates task completion with realistic delays. I don't want a mock engine. I want an actual engine that does that stuff. Show more features. Anything that says the word mock on it, I don't, I don't want it. Categorize, understand intent. All right, so, all right, go for it. Okay, so it's going to just start building this and, and it will— it's going to make an implementation plan really quick, approve the plan and get to work. Okay, so this is where we'll take a break, guys. It's 11:40 right now. Let's take a break for 15 minutes. Of course, you're welcome to sit here and keep working on things if you want, but we'll come back at 11:55, 11:54, whatever. And I'll turn off my— I'll mute and turn off camera and stuff like that for a minute here, and we'll be back. So go, go touch some grass or something, get a drink, do what you need to do. See you guys in a minute. But feel free, by the way, if you want to eat when you come back, bring food. Like, I don't care, eat it during the class. Okay.