Cohort 18 - Day 1 - 8/19/2025

take this thing off. Okay. All right, we are officially recording. So welcome, guys, to day 1 of 2, um, of this AI course, How to Build a Business in a Weekend with AI. Really excited to have you guys here. Um, as I was just saying, there's a few people that I, uh, I've known for a while. Chris definitely wins the— takes the cake on going back the farthest. He and I have been through a heck of a lot together. And excited to share some of that. That's kind of some of the point of this course, right? This— some of you may have taken AI, other AI courses, or paid $29 a year for this thing or that thing, or watch some YouTube videos, and now you're here and you want to really kind of step it up. And one of the things that will be different in this course is it is really focused on entrepreneurship plus AI, right? It's not just, hey, here's some cool tips and tricks, here's a couple of cool platforms. My goal for each of you, and you— so many of you are here for different reasons, but my goal for most of you, I guess we'll say, is I want you to be making more money by the end of tomorrow, right? And so that means you have to be selling your product, you have to be reaching out to people to sell your product, you have to actually be an entrepreneur, right? You can't just sit back and let AI do a bunch of magic and then you know, just expect money to drop into your, you know, into your bank account. But, you know, some of you guys are here. Corey Summerhays is the CEO of an awesome company. He's partnered with Jeff and there's a few other people from, from a great company we've done some work with. And Corey, you know, Chris, I guess, as well, right? So some good, some good connections in here. Corey and his team are based in, based in Vegas. And they— if you need something painted, guys, Corey can do it, although he usually likes bigger jobs. He's more of a commercial side of things. But, but anyway, so Corey and team Unforgettable Coatings, you guys, you know, you're obviously here not maybe to start a business now. I hope, I hope, Corey, that you have the same type of approach that I do as, you know, when I'm CEO, I want my employees to, you know, to be awesome at their job, but also to have some side gigs and things like that, or at the very least have a personal brand website that you're growing and building. And everyone has this side project they want to do.. And, and I hope that this enables you to just whip that thing up, you know, almost instantly. There's so many different little things that I want to do, and I'll talk about that as we go on. And AI and all of these processes and workflows and things that I do have enabled me to, to whip up so many things that have been on my mind for many, many years. So really excited for you guys to be here. I want to do a very quick intro. Everybody gets— each person gets, you about 23 seconds to introduce themselves because I know there's a lot of us. There's 13 of us. I guess there's going to be 12 of us since one couldn't make it. And then Taylin's not here yet. We'll have him introduce himself once he gets back from taking his daughter to school. So what I want is first name or— yeah, just introduce yourself, your name, where you live. I think that's a cool thing for everybody to know. And then what do you do for work? Right. And so just kind of your title, right? Like CEO or, you know, head of marketing or head of whatever. And then one more thing, do you know how to code, right? I know the answer for some of you, but I want you to just answer that question. Do you know how to code? Spoiler alert, I don't know how to code. We're going to be doing a lot of coding today, which is fun, right? Chris has been my CTO since I've known him, since 2016. And so I'm excited for him to be taking a course today to learn how to code from me. So that'll be a lot of fun. Now, it's not all about coding, guys, but it is definitely something, um, that's, that's very interesting. And so I'm excited to, to have that go. So I'm just going to call your names out because it's in the order that I'm seeing it. And you know what, it is in alphabetical order, so that's what we're gonna do. Um, so, uh, we'll go Corey, then Chris.

Great. Corey Summerhays, uh, started a paint business out of college 20 years ago, Unforgettable Coatings. We've got a few different brands acquired a new construction business, Final Touch Painting, up in Utah. Does, uh, we're on the airport up there doing big projects. I started a residential business as well and then started franchising here a couple years ago and bringing the brand Go Painting across the country. I am always excited about employees that are going on to entrepreneurship and learning and always looking to invest in them. And so, today, I brought some others that are in different departments, different businesses within our corporate umbrella of businesses, and each one is looking for different ways to solve their projects. And so, I think I've prepped them well, Jon. If they want to open up their own business as well, then yes, we We love that. Love to see people go on and be entrepreneurs. No, I do not know how to code. Worked with Jon in several other projects. We've got a project going that's on my mind as we go through this course and looking to grow that.

Look forward to it. Awesome. Thanks, Corey. Chris, we'll get to you in just half a second, but I just have to throw in one quote that I love. You know, there's— I don't remember who said it, but there was a conversation between two executives and they said, what if we invest in our employees and they leave us? And the guy responds and says, what if we don't and they stay? Right. So I love that. Yeah. Corey, I already know you're awesome. So glad to have you on here.

Chris, go ahead.

Yeah.

So Chris White, Jon and I worked together off and on for better part of, what, 8 years, 9 years, something like that. Um, I'm the CTO at an organization called Adminify right now, where it is about as deep into AI automations and services as you could possibly be at this stage in the game. Uh, so a lot of fun stuff there. I'm also, uh, the founder of Invoice Blue, uh, and a few other companies, um, that are kind of just hanging out there. And I do know how to code. Oh, by the way, I'm from Vineyard, Utah. I've been coding for 12 years now, I think, and I don't not expect to learn how to code from Jon today. I very well will probably learn a lot of cool stuff, but I don't know a better entrepreneur than Jon. So what I really hope to learn today from Jon is how to do better with what I have in order to just make my, my, uh, biz dev and connections and networking and all the things that I don't do well in the entrepreneur space better. Excited to be here, Jon.

Awesome. Thanks, Chris. All right, we'll jump into Denise and then Emily. And you're muted, Denise.

Sorry, I'm not muted now. There we go.

Denise Summerhays. And I am one of the many Summer Hayses on here. Anyway, I work for Corey Unforgettable Coatings and has— let's see, my job title has changed so many times. I just say I'm the gopher girl and I go for— I've done a lot of different things.

Right now I'm its office manager.

But anyway, I'm looking forward to this class.

This is really exciting and I do not know how to code.

But looking forward to it. Awesome. I love it.

Thanks.

All right, Emily.

Hey, good morning, everyone. My name is Emily Lindsay. I'm also with Unforgettable Codings, and I could not be more proud to be part of their team. I'll just preface with that. I've been here only a brief time. I think today is my 7th day. Super excited, fully embrace Corey and his vision for the entrepreneurial spirit of his employees and those that are on his team. I live also in Las Vegas, proud to be a Las Vegas person as well. I do not know how to code. I couldn't tell you where to begin with that, but I'm super excited to leverage the tools, Jon, that you're going to be sharing with us today to hopefully broaden my horizon on the utilization and just what's out there with various AI tools. So happy to be here.

Very good. Thanks, Emily. We'll go Jeff, then Lisanne. All right.

Jeff from Las Vegas, part of the Corey contingent here.

And I'm excited to be here.

We've been working with Jon on some different projects, and so we touch base with him, it seems like, every few weeks on different things. And this is, this is one of the most exciting things that we've been a part of. So I'm looking forward to I'm getting involved with this today. I am the director of branding and communications at Unforgettable Codings. So working on the marketing and branding side, really enjoy that and looking forward to learning about some ways that, uh, the AI can help us there. Um, I've been doing some research on it over the last few weeks and it's been really exciting. So looking forward to learning more about that. Uh, my coding ability begins and ends with, uh, DOS. Several years ago for those of you that were alive then.

Anyway, that— but no, I can't code.

That's the long and short of it.

Very cool. All right, Lisanne.

What's up, everybody? I'm Lisanne. I come out of Orem, Utah. I unfortunately don't work with Corey. I feel like a little bit of FOMO here on this call. So, Corey, maybe we'll have to have a conversation because it seems like—

ready to start a painting franchise.

I know, I'm in, I'm in. I don't even know what it takes, I'm in. Um, I am the founder of Lumina Growth Lab. I am a marketing and sales consultant and an agency, um, that works with a blend of HI and human intelligence and AI, artificial intelligence. Um, so I'm really excited to be here. I don't know a thing about coding, um, and, and so I've been leveraging other tools, but I have Really big dreams to create communities centered more around thought leadership. And so this is kind of my learning to dive more into some of the things that I'm actually more passionate about. And yeah, just here to learn. I'm really excited. And oh, I'm also the host of the Marketing Matrix podcast. So it's Jon, you're going to be on the podcast in a couple days, so we're excited to have you. Um, but yeah, that's me, really excited to be here and learn from everybody and learn with everybody. So thanks for having me.

Very cool. Thanks, Lisanne. Um, Michael and then Mitesh, and welcome Taylan. Taylan, we're just doing intros right now. Name, where you're from, what you do for work, and whether or not you know how to code. So Michael, then Mitesh. Hi everybody, I'm Michael Summerhays. I work with Corey on the new construction side of his company, Final Touch Painting. I'm based up in Salt Lake City, Utah. I work on project managing, but also kind of transferring into the estimating side of things right now. And I don't know how to code, but I have been kind of playing around with AI and I've been getting more curious with that stuff. So I'm excited about today and excited to learn. Love it. Mitesh.

Hey guys, my name is Mitesh. I'm from India and I'm a manifestation coach here. I've been doing this for about 22 years now and we've got about 20,000 students in our community who learn manifestation with us on a regular basis. We have a big— not a big, but a kind of a medium-sized YouTube channel with about 7 to 8 million views every month. And I'm looking at understanding how I can use AI to grow the online coaching business a little bit further to scale things much faster. I have zero coding experience, wouldn't understand anything. I think the best introduction I heard was started and ended with DOS. Same here. I think I can vouch for that too. I'm looking forward to how I can use AI and leverage AI to, as an entrepreneur, grow my business faster.

That's very cool. I never learned how to program in DOS. I learned how to launch like Doom and Commander Kane. Things like that back in the day. But I was just a gamer. I was what, I don't know, 10 years old, 8 years old, 7 years old when those were coming out anyway. Very cool. All right, Sharon, then Stephanie.

I am Sharon from the UK. I don't know how to code. I've been around lots of people for a long time that do. So I'm hoping that just by, I don't know, environmental impact, I've got some talent towards it. I do understand it and I've sold it, but I don't do it and I'm not capable. But my enthusiasm far outranks my ability, and that seems to be a little bit dangerous, but it is where I'm kind of at. I run a brand called Engage and Prosper. It's just me, my professional background, marketing and fellow of the CIM, which probably doesn't mean anything to anyone in the US, but it's quite high up in the UK in terms of, in theory, lots of knowledge around marketing things. But that happens naturally when you get to a certain age, especially with the amount of people that do marketing these days. So I have a very clear proposition that I really need Jon Cheney and anyone else's help in making it work, because as I said, my ambition on Replit and other software is far greater than my ability to make sure it works properly. And obviously I want a product that is fit for purpose. So very pleased to be here looking to share and learn and, and see where I can make it take me. So nice to meet you all.

Wonderful. Thank you, Sharon.

Stephanie. Hello, Stephanie Ellis. I'm in Indianapolis, Indiana. Um, I learned to code in my early 20s because there was a really nice IT guy where I worked and I gave him lots of brownies, so he taught me how to do really basic coding. So that's my extent of my knowledge. Um, I just went out on my own the last 4 years for my own content creation agency, and I have, uh, my LLC is The Heart Magician because I'm a heart-based writer. And kind of as everyone was talking, I was thinking like, when I first started, what kind of shaped who I am now is that I had a boss in corporate America and he told me, you'll never make it in my world, you've got too much heart. And I spent like my whole career thinking I'm doomed, but now I get to use that and I get to have like the most amazing clients who all come to me who are like operating from a place of heart. So I'm excited to figure out how I can integrate AI into that with obviously the human-centered component being a big piece of that.

Yeah, I love it. Stephanie and I are working on a cool project together that we're really excited to share more about at the right time. I'll probably drop a couple hints during the call on it and just some very, very high-level stuff since they're starting to drop a couple hints on their side too. But we'll save that for a little bit later. But Stephanie is an amazing person. I've just started working with her over the last month. And anyway, lots more to come. We're just at the beginning, Stephanie. So anyway, Taylan, last but not least.

Awesome. Sorry I'm late. I'm Taylan White. I grew up in Utah. And I've been a software engineer for about 10 years. So I do know how to code. And I'm excited to learn more.

So yeah, Talon's amazing. Talon is Chris's brother. Um, Chris and Talon and I— so let me introduce myself a little bit and that'll tie in all this stuff here. Chris, Talon, and I, um, were in another startup called Seek, um, that started back in 2016. Chris and I were there from the very, very beginning, and Talon joined not long after. And then we built something pretty awesome together. That was a really, really fun adventure. And then we went off and did some more things, and we're still, uh, we're still kind of in a lot of those things. And so anyway, really excited to have this group here guys. So, for those of you that don't know me, by the way, I think this is awesome. We've got multiple people in the UK. We've got India. We've got Utah. We've got Indiana, Vegas. We're all over the place. I love it. This is a global— I love Zoom. I love that we can all be on the same call together. I think it's very fun. And I also love this small group format, guys. You know, everyone knows everybody now, at least a little bit. I know that takes an extra 20 minutes, but it's worth it, right? Some of the relationships that you work with here and the people you get to know. Really, really make it worth it. And we've had lots of people take this course together and then go to like work on some things together as well. So definitely keep those opportunities open. And anyway, I'm excited to learn from all of you as well. Anyway, I live in West Mountain and West Mountain, Utah is— I'm staring out my window. I can see the Y over there on the mountain. So I'm not too far from Provo, about an hour south of Salt Lake. We have a little 5-acre ranch out here that's mostly filled with weeds now. It's beautiful. My wife's amazing at gardening, and we have so much food that just pops out of the ground. She's, she's, she always is like, I'm just gonna plant a few extra, right? Now we have like 900 jalapeño peppers, and I don't know what to do with them. So anyway, we've got all kinds of fun stuff. We've got about 60 chickens, and anyway, um, got a horse, and we used to have goats, but those things are terrors, and so we got rid of the goats. Um, they're fun, they're really fun. We had fainting goats, so those are even more fun. I would go out and scare them all the time. They just fall over anyway. But we sold them to some Mexicans on Cinco de Mayo, so they ended up probably as dinner. But anyway, we don't have those goats anymore. And anyway, so yeah, as, as many of you mentioned, you know, I've been an entrepreneur for really my whole life, you know, definitely varying levels of entrepreneurship. I've worked for other great companies as well, right out of BYU. I'm a BYU fan. And I'm wearing my BYU football shirt because I'm meeting with the BYU athletic department today. I'm a huge Cougar fan and football season is like this Saturday. And like, anyway, I watched, I went back, I was watching some reruns last night just because I was like, man, there's no football games on. And anyway, so anyway, I'm a big, big football fan, excited for that. But, but really just been an entrepreneur. I love, um, for me it's kind of like a video game. It's like trying to like get to the next level. Like I come up with an idea and the next level of that is, okay, well, can I, can I show something? Can I make something actually look like a business? Right. And so what do I need? I don't know, a website or maybe a product or whatever it is. Right. And then now the next level is, okay, I need someone to buy it. Right. And so then I do whatever I can to get to the next level. Right. And I always just feel like it's a video game just saying, okay, how do I get to the next level? And even when I have customers, okay, great. How do I now automate this and build a factory? That can automate this thing, right? Chris and Taylan and I have had a gaming day or two, and they know I love factory games. And, you know, so I'm not so much the, like, you know, I don't know, Fortnite and Halo and those types of whatever. Um, I— they're fun, I think, but I just— I don't have a lot of time for video games these days. And when I do, I like to play a factory build building game for some reason. I don't know why. I just like to, to see how pieces fit together so that I can create more efficiency and automation. I don't know why that's fun or why I like to do that, but that's just what I like to do. And anyway, so when I discovered AI, so I started a company in 2016 called Seek. We raised $13 million for that. We ended up selling it a few years ago. Back in 2023 is when I left the company, put another CEO in place, and then it sold. We sold it, kind of did the final sale last year. And we took the valuation of that company all the way up to $130 million. And then we went back down and we sold it on the way down. It didn't go all the way to zero, but we definitely had a good outcome in terms of being able to learn a lot and do a lot of good in the world. We had our customers were— Seek was an augmented reality company, right? And so, we helped work with companies. I mean, our customers were Lego, Nestlé, Walmart, I don't know, Six Flags, all kinds of companies out there that had all kinds of different things where they wanted to use 3D and, and, and be able to deploy it for different reasons, mostly e-commerce related, right? Being able to see that, you know, try the shoe on or, or, you know, see that table or that couch in your living room or your kitchen or whatever, using your phone and doing all that. And again, Chris and Talon were— Chris was the CTO. He was founding chief technology officer and very critical to its success. And But anyway, that's really, that was the first kind of, I don't know, big thing I did and then sold that and started another company called Velocity Access and that one's still going. And then I kind of accidentally founded this GenAIPI thing and that's why we're here today. Back in February of this year, I post on LinkedIn every day and Facebook, whatever, it's usually the same thing. I come up with something and I post it, right? And by the way, I don't use AI to write my social posts. I do occasionally, you know, use it for some ideas, but I actually like to write. It's something that I just enjoy doing. I enjoy creating. And so every day I sit down and I say, okay, what do I want to share today? And it was February 27th, a Thursday. And I said, okay, I've been working with this AI stuff for a while now. I've gotten pretty dang good with it. I'm going to give myself a challenge and I'm going to do it publicly, right? Hey, I want to see if I can build a business, start some— start a new piece of software today, Thursday, and get my first customer by the next day., right? Can I start building some software and have a customer the next day? Which means I had to finish the product and then sell it to somebody, right? And I actually had people during that challenge, I posted it and lots of people followed, 100,000, you know, impressions and whatever on those posts and things like that. But I just kind of, kind of live blogged it, whatever. And, and I even had people saying, hey, I'll buy it. I don't even care what it is. I'm like, actually, I don't want you to buy it. I want to actually see if I can sell it to somebody that doesn't know me that actually buys it because of the value. Can I actually create that value? And, and make it happen. So anyway, I dove in, worked all day Thursday, started about 9 AM, worked until I think 10 PM, like almost nonstop. I didn't get, get up from this spot right here. And, and by, by 10, I was like, I'm, I'm dead. My brain is toast. And so I went to bed, got up the next day and started about 8 AM. And this time I went till 3 AM. Because I was like, I'm so close, I'm so close, I got this. And by 3 AM, I was like, you know what, I haven't integrated payment yet. I've never set up Stripe before. I hear it's kind of tough. I don't know if I can do that. I think that'll take me a few hours tomorrow. I'll just do it the next day. So Saturday, got up. I had a service project at 7 AM, which is really hard going to bed at 3, but I jumped up, went and did that, and hung out with my family. And then by about 3 PM, I finally sat down And by 3:00 AM, I was done. I finished it, right? I had integrated Stripe. It was way harder than I thought it was going to be. And all these other things that kept popping up and testing it anyway. So long story short, ended up taking me 3 days instead of 2 days to build it. And I pushed go on Facebook ads that I wanted to run towards this thing. And, um, and so now it's Sunday, you know, Saturday night at 3:00 AM, Sunday morning, whatever. So I get up, I don't work Sundays. I went to church, just kind of hung out and said, all right, I need to let my brain reset. Monday, got up and checked my ads, and I had about 130 people had clicked on the ad and landed on the website. I was like, that's pretty good. Um, so obviously something's, you know, something's working with the, with the messaging of the ad. And, uh, and then I was like, oh, I wonder how many sales I have. And so I jump into Stripe and zero, nothing, only my test transactions, a bunch of dollar transactions. Anyway, I, um, so I was like, man, what's going on? And so I went and I tried to buy something and there was a payment failure, something I had set up wrong. I had made a change that had broken something, and, and that was my own fault. I didn't test everything after I had made that change. So anyway, I took an hour, fixed it, and I said, okay, I'm a sales guy, I can make this happen. So I get up Tuesday morning and, and I just jump on LinkedIn and I start just reaching out to some people, some CEOs and people in the area. I was like, hey, you know what, I think these guys might be, might be the right, um, you know, type of fit for what I'm doing here— training, corporate training and certification, things like that. That was what I was trying to do.. And, and I had someone, I had someone agree to meet with me. Johnny Hanna, for those of you that might know that name in the area, he's a really, really sharp guy. Started some big companies. Homies, one of the ones he started. Entrada is another one. Huge companies in the area. And now he's on to another company, which is taking him. His couple of people in his family have taken this course as well. So anyway, got him to agree to a $15,000 contract. Right. And so from from Thursday starting my first line of code to 5 days later, $15,000 contract. That's a real contract, training, all this stuff. And, and it ended up turning into lots more. And then 30 days later, I had already passed the 6-figure mark in revenue and it just grew from there. And so again, I didn't really start this, this company to do these trainings like I'm doing with you now. I actually just wanted to see if I could start something fast and test my entrepreneurship and AI skills together. And, and that's, that's how we ended up here. So, so anyway, guys, really excited to have you on here and to teach you about the process that I did. There's 100 different things that went into making that happen. And I had 500 tabs open. I probably still have most of them open. I always have too many tabs open, guys. I probably need to close them all. And so hopefully that doesn't overwhelm you. But, but by understanding how to put these things together and these different pieces together, You can do a lot, and I hope that by the end of this, these 8 hours, which I promise will fly by, that you feel like you know how to orchestrate with AI, right? One of the really important things that I want to just say now, and then I want you to pay attention to, and I'll bring it up many times, is you are the boss when it comes to AI. If you're getting bad outputs, it's because you're giving it bad inputs, okay? Just in the same way, if I told Chris, if I said, hey Chris, I want you to build something for me, right? I want to— I want you to build me a weather app. And Chris was like, great. And I was like, okay, get to work, right? And that was all the instruction I gave him. Well, I mean, who knows if he's going to build something. He's going to build something cool because I know Chris does good work. But is it going to be exactly what I was thinking? I'd probably get on and be like, oh, but you forgot this and you didn't think about this. He's like, well, you didn't tell me, you know. So AI is exactly the same, right? You got to be a good boss. You got to give good instructions. You got to be there. To support your AI employees. But hopefully we can show you guys how to, how to have a swarm of employees working for you to allow you to launch those side businesses, to supercharge your existing businesses, everything that you guys are doing right now. So, all right guys, with that being said— oh, and also, I, like I said, I don't know how to code. I don't know how to write one single line of code. Now luckily I've been around really smart people like Chris Taylor for a long time and other great— we had another CTO in there at one point called— or his name is Thane, and and he was— he would speak in words that I didn't understand usually. And but I would ask him, I'd be like, tell me what you mean. Explain to me what that system is. Draw this on the whiteboard. Help me understand how it fits together. And so I did that a lot of times. And so I understand what I'm looking at when I look at code, when I look at things, I understand what's kind of happening, even though I can't write it right. And so that gave me a little bit of an advantage. And we're going to do a little bit of a coding boot camp, about a 10-minute coding boot camp later today. For those of you that aren't familiar with some of the things that will help you as you dive into coding. And again, guys, we're gonna do a lot more than coding. We're gonna do coding, but when you're building a website, when you're building a business, you need strategy, you need images, you need video, you need marketing plans, you need all kinds of stuff, right?

That's different.

You know, the coding, the website is just like one little piece of it, but then there's a lot of other things that have to go into it to actually turn it into money. Right? That's why we're doing it. That's why we start a business, right? There's lots of reasons to start a business, right? But ultimately they all come back to, hey, it's not going to work unless we're making money. Okay? So that's what we're going to work towards. It's always pushing towards your first customer. Okay. All right, guys, to start to kick today off, we are going to do some, some fun coding, right? That's what we're going to dive into. So we're going to start with— I'm just going to jump into here and I'm going to click share screen here. And for now, guys, a few times I'm going to— often I'm going to share my screen. Sometimes I'll be okay, just watch for a minute. And then I want you to do. Okay. And so today, right now, we're going to start with the x.com and this nice little crab. I don't know what that is, but we're going to go to SuperGrok. Okay. And I had you guys sign up for, for Grok, for ChatGPT as well as Replit. And we'll be using those 3 platforms as well as a lot of others that you can, you know, you don't have to sign up for today, but you certainly can at the right time. I'm going to show you, introduce you to different platforms over the next 2 days. There's, there's a really important reason to use different, different platforms, guys. One of the things you have to be careful of with AI is sometimes AI just makes stuff up, right? They call it hallucinating, right? And, um, it's happening a lot less these days, it really is, but it still does happen. And you, and you want to— what I like to do is I like to use AI to help me better use other AIs, right? And so if I want to get ChatGPT to do something, right? I might ask ChatGPT to help me do it, right? And have it help me write the prompt that I need to then give it. Or I might actually use Grok to say, hey, I need ChatGPT to help me do this. Or, hey, I want to do this. I want to give it to another AI. Help me write instructions. And I'll bounce back and forth. I'll fact-check back and forth. And I really like Grok. I think Grok is incredible. I also love Claude, right? Anthropic Claude is amazing. It's amazing AI. ChatGPT is amazing. They're all good. They all have their strengths. And their weaknesses, right? And, you know, you just kind of learn those quirks over time. One other thing that is also important, guys, is some of these AIs have bias, right? Just like when you go to Fox News, you're going to find a more right-leaning bias. If you go to CNN or MSNBC, you're going to find a left-leaning bias. And that's okay, right? It's just we're humans. We have opinions. And when we talk and we express things, they're going to come through that prism. They're going to shine through that prism of what we believe and think, and that's what's going to come out. And so for me, I like to examine all sides, right? I like to say, okay, what are they saying over here? What are they saying over here? I try to find, you know, everyone's like, what's the perfect centrist truth, right? And I don't think that really exists, right? I really don't think it exists. And so for similar reasons, I'm going to use ChatGPT, which might be a little bit more left-leaning. I might use SuperGROK, which might be a little more right-leaning, right? And they're not super, you know, usually they're not super overt in their biases towards things. They're really trying to give you good answers. And quite frankly, you could say, hey, give me a right-leaning answer, you know, or give me a left-leaning answer, and it will do that, right? It's what it's really trying to do is please you. All of these ARs are trying to please you. Okay, so that's just a little bit about how to select different models and what you're what you're going to work on there. But okay, to start off here, here's what we're going to say. Using p5.js, create me a mind-blowing visualization. Simple prompt. By the way, spoiler alert, this is a bad prompt. We'll talk about why. Now, p5.js, guys, is a— we'll get into front-end versus back-end later, but this is a front-end language, kind of a fun JavaScript variant. It's very simple. You would— I kind of think of p5.js as what is used to make arcade games, like the old ones, like Pac-Man or something like that, right? Now, I've just asked it, hey, using p5.js, create me a mind-blowing visualization. So it knows, okay, I'm going to write some code. It's going to think through here and it's going to do this. Let's see what it says here at the end in order for me to run this thing. Okay. All right. So it says copy the code into this. I don't like how it just gave it to me here, but anyway, it's giving me a website here and then I should be able to just copy this. Whoops. Hold on. I did this wrong. I think that's where I start. Paste, play. Wow. Well, okay, so look what happens here, guys. We have a syntax error. We're going to— it's probably that I didn't copy the top line here. Let's see if that fixes it here. It's okay if you guys don't know what's going on here. You're not supposed to yet. Yep, it's not working yet. So I'm just going to give it this error here. I'm literally going to screenshot this, this thing right here. I'm going to drop it in here. Got this error, give me this. I can just copy the code and paste it in the p5.js editor. Okay, all right, so the first time I actually got some code to work with AI, which was a long time ago. It was a couple years ago, but I felt so empowered. I felt so cool, right? I had— I don't know how to write code, and all of a sudden I was able to just describe it using natural language. Hey, here's what I want to do. And then it would just write the code, and then I could just build stuff, and I could tell it to change things. And I was just like, man, this is so cool. This is so fun. It's so easy, guys, that my kids know how to do this stuff, right? So if I was going to make this prompt a little bit better, the reason this is a bad prompt is because just like that prompt prompt I said earlier with Chris with, hey, just make me a weather app. I didn't give it any instructions, I didn't give it any details, and so I'm leaving it up to AI to just fill in the gaps. Okay, so I just copied all that and we're going to paste and click play. Okay, so here's my mind-blowing visualization, and I don't think it's that mind-blowing. I don't even know what it is. What's it supposed to do? I can click on it and it just keeps going more and more. It's like a fractal. That's kind of cool actually, guys. It's cooler than I thought. Look at this, I'm just clicking and it just is infinite zooming. Yeah, that's pretty neat actually. I wonder how, how long I can infinitely zoom. I'm, I'm shocked that it went that far. That's, that's actually pretty cool. So again guys, I didn't give it, uh, an explanation. It's also eating my memory by zooming in this far. Yeah, it is. It is. That's— I think that's called a fractal. Yeah. Draw fractal. Yeah, that's what that was. Really neat. So anyway, I didn't give it any instructions. I just said, hey, create something in p5.js and, you know, whatever. Okay. All right. So, so that's fine. That's great. So now one of the couple of things that I did that I want to just point out for today, guys, to make your life easier, because we're going to be bouncing back and forth between different AIs and using different things. I want you to recognize this little copy button right here. Okay. Instead of having to do like I did earlier, I had to like select and grab this. It's because I asked it, I gave it a bad prompt and it gave me— anyway, it was kind of weird. But you just push this little copy button that copies everything that's inside of here, inside of this little box. Okay. And then I'll take it over to here and paste it. Okay. So let's do something a little bit better. Let's give it a little bit more instructions. Okay. Now, Create an interactive model of the solar system. Make the design neon and fun. When I click with my mouse, it should shoot a shooting star out of my cursor. The planets should be— let's see here, we're going to say an interactive 2D model source. So planets should be rotating around the sun in their proper orbit. Yeah, let's just say that. We'll just start there. Okay, so, so this is a little bit better prompt. Now I'm giving it a few more details. Now it kind of understands, okay, I need to create something. And by the way, guys, I did not— I want this to— I want to point this out. I did not tell it to use p5.js, right? That's because it understands the context. It understands, okay, now I know what this person wants. They want to be able to— they want me to give it to them in a format where I can just copy it and paste it over here into p5.js. Okay. And now they want to model the solar system. So one of the cool things is if I ask Taylan again, picking on the developers here, if I ask Taylan to go build this for me, Talon would say, great, I can do this, no problem, right? But he would have to go and he'd have to jump on Google and he'd have to look up, okay, what are the orbits of the, of the planets? You know, what, how far away from the sun is it and how fast does it need to go? And Talon's more than capable of then figuring out that math and setting it all up and building this thing. He could do it. Absolutely. But, but AI obviously can do all that, you know, in a split second. And, and it does all that research for me. Right. And so it's just going to figure out here's the background stars. Here's some planets. It figured out the Earth radius, and here's the period, size, color. She figured all that stuff out in, in seconds, right? And I think it's just amazing that it can do that. I love, I love the power of AI. Think about how much time is saved right there, all that research that any of us would have had to do. I don't, I don't know if any of us just knows those orbits off the top of our head. Maybe there's someone out there that does, but I'm not one of those people. Okay, so now I can copy this and I can grab this right here Command A, Command V, and here we go. Here's my solar system. Okay, so obviously something isn't working right. Something's wrong with the background here. It's probably something that is pulling the background. I don't know why it's blue. All I see is a blue screen. Okay, so I just talked to it, and I would say that, by the way, also to Taylan. If Taylan delivered a product to me and I turned it on I would say, hey, Taylor, and I just see a blue screen. He'd be like, oh, I'm sorry, let me check on this, right? And there might be some z-index problem or something, who knows, right? So he's, he's going to work on that. My developer, my developer is now going to work on that and fix this. Um, and so yeah, the code sets the background to black, so blue screen doesn't match. It's trying to figure out what's going on. By the way, Groq used to be faster. I could switch this model. Um, so it's on expert right now, could switch it to quick responses. And actually for this for this thing, I might actually, I might actually do that. It doesn't need to think that hard about this. We're not giving it hard, hard tasks here. Um, but let's, uh, let's see. Current date, user might be reporting a BSOD. Yeah, it's, it's thinking way too hard about this. Searching the web, just searched 20 different things. This is pretty funny. All right. Um, okay, there's a couple questions we'll answer while this is happening. So, um, So, Lisanne, yeah, Grok is, um, Super Grok and Grok are, are both good. The thing, the thing that Super Grok can do that Grok can't do is really kind of some of the deep research stuff. What I really love about Grok, guys, one of the reasons why I like Grok versus ChatGPT or even Perplexity or other things is Grok has access to Twitter. It has access to X and what everyone is saying in the world. And so when I'm doing market research, I can do that with Grok. Grok is my preferred market research when I want to figure out what are people going to pay for, what are people talking about in this arena, right? Yeah. Now it thinks I'm seeing a blue screen of death. All right. We're going to just stop you right there and we're going to do this. No, the p5.js editor is rendering a blue screen. Stop. Go. Okay, so, you know, it's not super smart, but anyway, guys, um, the Groq system is amazing for doing research. We're gonna be doing that later. Okay, and then listen, what is considered a hard task? So a hard task is going to be something that is pushing the limits, right? That's kind of where I'm, where I'm defining a hard task, pushing the limits of what would be really hard even for a human to do. To set up a, you know, to set up a basic, you know, solar system in code should be super, super easy. This is something that's very, very basic that beginner developers should be able to do. This is very, very simple stuff here. And so, so that's why I'm saying that's easy for us. It's hard, right? For us. Yeah, I can't do it. I don't know what's wrong. Right. And so I have to sit here and wait for the code. You know what I'm going to do? I actually am going to switch it to quick responses so we don't have to sit here and wait while we're doing this stuff. So yeah, great question, but for things like Replit, do we even need Groq or GPT-5 for building stuff? So yes, we do, but for different reasons. I'm not generally using Groq or GPT-5 to build code. Unless it is something simple like this. I'm trying to— if I'm trying to do something fast or I just need a quick demo or I don't need to save it, for example, then I'm going to use this. But if I need to save it, I need it to do something, I'm probably going to jump into Replit and just have it do that. But I am going to use Groq and ChatGPT for other parts of the process here. Okay, guys. So while this is continuing to work, I don't know why it's taking so long. We're going to just do this. I'm going to say just rewrite it, rewrite the code better, and I'm going to have it do it on fast. It should be able to do this like almost instantly. Groq is going really slow. See, that's how fast it should be able to go. So we're just going to stick on fast for now. Um, what I want you guys to do is I want you to go ahead and, um, and I want you to use this. Go ahead and go to Groq. I'm going to give you a prompt here. Using p5.js, create a mind-blowing visualization. You're not going to actually do that. You're going to be create me a, whatever you want here. Put whatever you want inside the bracket or don't put it inside the brackets, delete the brackets, put that create me this, maybe a Tetris clone or a model of some system that you're trying to come up with or whatever it is you want to. Show me where I can deploy it. I want to just copy the code and put it into the P5.js editor. Okay. And then give it all the instructions you want, right? Add details. Does not have to be a lot of details. You saw how many details I did, guys. Just start there and do this. Okay. So I just dropped this into the, into the chat so you guys can come and do this a lot today. I'm going to give you a prompt. You can copy it, paste it into yours. And then fill in the gaps where you need to or change what you need to change. Okay? But I should be able to copy this now, come over here, and this should work. Play. Come on, man. Roc is really struggling today. I've never had to take this many, this many attempts. I think it's because it gave me something not ready for that. P5.js anyway, so it's going to fix that. Okay guys, go ahead and ask questions too, and you don't have to just ask questions on the chat. You can come on and be like, hey Jon, I've got a question. Okay, so feel free to, feel free to do that. I'm going to keep building mine here until I get it to actually render me a nice solar system. Yeah, Chris, you mentioned, um, you know, it's great for generating prompts to pass to Replit. I don't find myself doing that a lot. I actually find that it gets me in more trouble when I do that, and there's a couple reasons for that that I've, that I've discovered. Um, it, it works for some systems, but for Replit specifically, like, I just trust clawed better. And so I actually do my planning inside of Replit, and there's specific ways to do that. Um, I'll show you guys, I'll show you guys that stuff later. All right, let's see if we can get this to actually work now. Hey, I have solar system! I don't know why that took so long, but now I have a solar system working. Everything's rotating. If I click, it shoots a shooting star. So I know, great, I've, I've created something that actually works. Pretty fun. Okay. All right, guys, I'm going to ask each of you, and it is voluntary. I won't force you to, but I want you to share what you make. So create something and then I'll have you share your screen. This is actually secretly also a test to see if you can share your screen in Zoom, if you know where the buttons are, because we're going to have you share your screen a couple times throughout the class. Kind of fun. All right, any questions? Lisanne, have you built something yet?

Yeah, you want me to share?

I would love you to share. Go ahead and share your screen.

So I asked it to do, uh, penguins, some jumping off of an iceberg, some swimming, some congregating. Nice.

They're not exactly— they're like kind of jumping up and down. So obviously what you could tell it here is you can be like, the penguins are just kind of randomly floating around. Can you make them more specifically jump and slide and whatever. Anyway, but again, just in the same way we would say, hey, Taylan, here's what I want you to do next. Just don't think about it as like you're coding. You're just talking to your developer. Hey, here's what I want it to do. See if you can update it and see if you can make it even more powerful. That's cool. Awesome. Remember guys, as you're discovering, as you do this, you're not going to make Call of Duty with this. These graphics are not supposed to be high level. They're very, very low level. All right, who else has something to share? I have a question.

Yes, Stephanie, go ahead. Um, do you keep— when you first go to p5.js, it says function setup, create canvas, function drop, and then you paste below that?

No, I would just select all, delete, and then just paste. So yeah, you're just going to replace the whole thing.

Yeah, good question.

That's what I did and it didn't work. Yeah. If you look at the very bottom on the left, there will be a little error thing. It might say syntax error or something like that. Go ahead and share your screen, Stephanie, for just one second so we can just show you how to do this.

Hold on, I'm navigating to a different screen.

Good. No rush. Mitesh, did you say yours was ready? Maybe while Stephanie's working on that, you can share yours.

Go ahead, go ahead, Mitesh. Sure. So I created two. First, I said create me an interactive web page to design a funnel. With an interactive view. And it did create something. I'll share my screen with you. And then the second one was about creating a Tetris game with words of emotions on that. So I'll share. Oh, nice. The game is going on. So this was the Tetris game with words of some kind of words in that coming in automatically. Yeah. All right.

This was—

and the preview came directly here.

I didn't have to go paste the code. Correct. You can do that. If you go directly to Groq, You can just preview the code right there instead of having to go to the p5.js editor.

Very good. Yeah, so that was good. On the top here, I asked it to create this. Yep. Yes. So the preview here, I don't know, I don't think it really worked. So it didn't really make it interactive, it just had some buttons in it.

That's why. Yeah, it was just very, very high level. Yeah, that's, that's what I experienced. I mean, you can have it create, you know, really, really, um, you know, intense things inside of Groq. Groq is very capable, but it's not, um, yeah, it's not gonna touch Replit.

So, all right.

Awesome.

Thank you. Thank you, Mitesh. All right, Stephanie, are you ready there or you still working it? You're okay. All right. Who else?

Taylor, what you got? I have, uh, It's building right now, so just one sec.

If this works. Okay, that's cool.

Corey, what you got?

I'm ready whenever you are. Okay, Corey, we'll have you go next. Stephanie, go for it. Okay, yeah, we're seeing your Replit screen instead of the Groq one. Yeah, there you go.

Okay, yeah, so go ahead. So here's what it looks like when I dropped it in. Yep.

It's not doing anything. Yeah, so click play, click the play button on the top left. Okay, perfect. So It looks like there's an error at line 19 there. If you scroll down, see the console? Sorry. If you just look at the console there, scroll up to the top of that. Then push that little flag that's pointing out there so you can see what it says. Yep, right there. Then just copy that and then drop it into Groq and say, hey, I got this error. Make sure you copy the whole thing. There you go. Yep, perfect. Yep, just copy it and literally you don't even have to say this is what I got. Just paste it, enter. Perfect. So it will, it'll say, oh hey, let me fix this. So it'll write you some new code and then copy it, paste it back in. You should be ready to roll. So this should not take long at all. Wait, I'm sorry, hold on one sec. So go back to that. So keep scrolling down. So one of the things that AI does, guys, the way that an LLM works is it is— it's basically trying to guess what the next thing should be. So it's literally writing it live. So you have to wait till it's done before you push copy. So go ahead and scroll back up and copy now. Now it's done. Now that it's done, you know, with the actual code part, keep going up. Yeah, that part right there. Go and push the copy button and then the top right and go back to the p5.js, select all. And then paste it and play. You should be able to just Ctrl+V or Command+V and then push play. Awesome. That's cool. Tell us, what did you build?

What is this? I wondered if I went a little too detailed. I said I wanted Pegasus flying in the moon with fairies and lots of glitter and lit pink laser beams shooting from the moon.

Well, there you go. You got what you asked for. That's cool. Is it interactive at all? Like, can you click or move your mouse or do anything, or like your keyboard to make anything happen? No, it doesn't look like it is. You could obviously ask that. You could say, hey, allow me to have control of one, you know, with my WASD keys, and, and then with that, right, you can just start iterating and building on it. But nice job, you did it. Thank you. Very good. Okay, we will stop your sharing and then— oh, unless someone else is going. Are you ready or do you want Talon to go? Talon, go for it.

All right, he's still working hard. Oh, I haven't granted settings.

I'm gonna have to leave and rejoin real quick. That's part of why we do this.

Chris, what about you? Yeah, let's see if I can share.

Let's see. It work? Yep. We're seeing it. Mouse hunt. Move your mouse. Oh, nice. Yeah, super cool. See, Chris is a developer. He knows how to talk to it, right? He knows he's got the shortcuts.

I got these eyeballs chasing me around.

Oh, I lost a life. Yeah, you lost a life there. Oh yeah, you have 3 lives. How many prompts did it take you to get this, Chris? Just one. Very cool.

How detailed was the prompt? I can show you it in a second. And I actually don't have Grog, so I used That's fine, that's fine.

Oh, and you're dead. That's pretty cool, that's pretty cool. Yeah, show me what you— what you did. Please render me a p5.js targeting for mobs or chasing, enhance it with some type of scorekeeping over time, multipliers with points as they don't get hit, maybe like 3 lives, more mobs load in as the game continues. That's pretty impressive.

ChatGPT-5, guys, is Really sucks.

Yeah.

So one thing that it does, um, is it— this renderer, just to save time for people, is it doesn't like, um, these headers, which is the— this part right here, right? So if you're gonna— rather than have it rewrite the whole thing, because it does take time, if it renders it with this part then you just get rid of this part and get rid of this part, which is like— it's hard to explain, but basically just the inner pieces, and it'll load.

Is that— yeah, that's why, that's why we use AI to do it, is because we don't know to look for that, Chris, right? Right. Developers— for you, for a developer, right? Obviously a developer is going to have a huge advantage using AI instead of— there's so many developers out there that are just I don't want to use AI. AI is stupid, right? And I think that's just crazy, right? That's like saying I'm going to be a mathematician and now the calculator exists. I'm not going to use it because I want to use my brain, right? Yeah. Okay. I mean, that's cool. That's good for you. And I think it's a great exercise, but everyone else is going to use the calculator and they're going to go way faster, right?

So there is a catch here though, is that I didn't know that that renderer did that until I watched everybody else have the same problem. Yeah. And so I asked ChatGPT what was causing it, and it told me. And so I learned. So even as a developer, I didn't know that, and I used ChatGPT to figure out what the answer was to save myself time in the future. So there's definitely opportunities. Uh, like, AI is a great equalizer because it lets you not just, like, figure out what like not just build these cool things, but also like learn why behavior happens and how to like save time. And I mean, eventually you could actually learn to write code pretty well just by watching and learning what ChatGPT and Anthropic, like whatever their models do.

So 100%. Yeah, I have learned a huge amount, guys, about coding by doing it a huge amount over the last year. Right? Like, I just— I mean, AI is doing the work, but I know what it's talking about now.

Corey, you got something to show? Yeah. So I've got my code here. I don't— where is it that you're cutting and pasting that?

However, yeah, we're dropping into p5.js. Go ahead and share your screen. I'll walk you through it.

All right.

Well, we'll do it together here. No, I don't have permission to do that. Sharing screens.

I'm getting that right now. You're good. Yeah, you should be able to just push the share button and access. Yeah, share it. You— it might require you to restart your Zoom.

That's possible.

Open system settings. Let me figure this out here. Yeah, so that's okay, Corey, we'll do you in a minute. Taylin had to just do the same thing. You have to grant the settings and it'll say, okay, I need to restart Zoom to make this take effect. Right. And so you'll just leave the, leave the meeting and join it again.

You want to share? Absolutely. I didn't see Chris's, but mine's probably cooler. I guess the audience shall be the judges. Galaxy shooter. And I can shoot back. Wow, is it chasing your mouse? Yeah, that's funny.

Ours were very similar. His was like eyeballs chasing him.

I think yours is a little cooler though.

Uh, I will concede. Jalen, I'm impressed at the quality of the flying saucers.

So I actually—

anything there?

I did say make the UFOs higher detailed, so my original prompt was Using PG, I spelled it wrong, but yeah, it still figured it out. Create me a galaxy system with UFOs, and then it kind of just went from there. Make the UFOs higher detailed, make the galaxies way more detailed with planets, and then it had an error, so I pasted it in.

Yep.

And then have the UFO shoot laser beams at the cursor, and then I decided to have the— make the laser beams more impressive. And then my last one was, how can I make it so the cursor can shoot back and blow up the UFOs. Yeah, that's where— mine, I only used one prompt, so—

ah, D charged more, but he got a lot more done. No, you know what, um, I love that, guys. I think that's super cool. Thanks for sharing. Um, all right, uh, who has not gone that wants to share? Cory's back. Maybe you can share now. Anybody else? I won't make anybody— I won't make everybody do it. Um, Corey, I'll share something real quick. Yeah, go ahead.

So with the retro gaming theme that everyone seemed to be doing, um, I did a—

do you guys remember the game Police Quest? Police Quest? I don't know if I know that one.

Yeah, maybe you're not as old as me then, Jon.

Um, so it was just—

it was a one of those kind of point-and-click adventure games where like they give you prompt and then you decide what to do. One of the things that AI has been interesting to me is, is using it for like role-playing games, other scenarios, having it create different things, and then you walking through kind of a text-based thing. Anyway, so that's what I created was a little, uh, like police quest thing. Let me see if it'll let me share. Um, no, it's just giving me errors on everything.

I don't know why that's the— oh, there we go.

Zoom whiteboard docs screens.

Okay.

Oh, that looks like I have the same thing that everybody else is.

Yeah, just click through the settings. You'll have to— it'll ask you to restart your Zoom and just restart it. I'll let you back in. Doing it now. Sweet. Okay. Yeah, guys, this is part of why I do this right now is because once we're, once we're in, we don't want to have to, you don't want to have to do this later. Sharon, can you share your screen? I'd love to see your— I'd love to help you get something up and running.

Okay. I'm going to go on and off of mute just because of background noise while I'm doing it.

Okay.

But, um, you're fine.

I can share. Yeah, 487 lines of code, like you said, there should not be too much. It's, it's very capable of dealing with that. I built some of these where I took it up to like 5,000 lines of code. I just like kept iterating and iterating and had it do, you know, create start menus and storage and all kinds of good stuff. Um, all right, Sharon, whenever you're ready to click share. Oh, why am I, am I getting a ding?

That shouldn't be coming through. Okay, sharing now. Can you see it? Yeah, I've had multiple windows and I keep stopping and starting anew because I don't quite know where I'm putting it. So do I delete all of this? No, you—

Or do I put it in between it? Yeah, so delete all of that.

Delete that whole thing. Yeah, I did that once already. And then I just basically paste it in return and then nothing happens.

Yeah, so go ahead. That's okay. Push the play button in the top left there. That's how you run it. Gotta run the code. Okay, so then an error came up at the bottom. You see how on the bottom there it says console? Yep. Yep, so just copy that whole thing there. And then just take— yeah, just the real important thing is that type error right there. Okay. So, just copy that piece right there and take that over to Groq. Do a new chat or just put it in? Yeah, just right there and push—

yeah, just Command+V or Control+V. This might be a different— prompt though, because I've already tried to fix it myself several times.

That's okay. That's okay. All right. So it's going to fix that. Move that body tag. Okay. Here's the updated. Here's the updated thing. Okay. Let's scroll down and see how it's doing.

It gives it to me in HTML and then it also gives me something for JavaScript.

So yeah. Okay. I see some of what's going. So yeah, it looks like you're trying to do it in JavaScript instead of p5.js. Scroll Scroll to the very, very top. I want to see your very first prompt. Well, that might be— it is right at the top now. Yeah, if you just scroll to like the very, very first question you asked it.

Well, the first question I asked it was on ChatGPT, and then I brought that over to Groq and used a super prompt from ChatGPT to tell Groq to do something. So— got it. Might be too clever, maybe.

And it's already gone off the track. Yeah, you might be working a little too hard here. So here's what I want you to do. I want you to go ahead and just tell Groq I want you to just type this. Say, I want you to rewrite this in p5.js.

I did tell it to be done in ChatGPT-5.

Yeah, just do that here. Rewrite this in p5.js and give me something I can copy and paste into Um, p5js.org. Okay, and just send, right? So it— yeah, you're, you are trying to bounce back and forth between these two things, but we're getting ahead of ourselves just a little bit there. And it looks like it might be doing the, the hard thinking, and so it'll, it'll think through this for a second. But you should be able to, um, you should be able to copy paste this over into the p5.js click that play button and it should work. It should work after, after getting this done. So it's going to think for a minute. We'll, we'll have you stop sharing your screen here for a second while it's doing that. And, and then when you're ready. Yeah. All right. Anybody else want to share? Corey, did you get yours working?

Yeah. You should be ready to share there.

And you know what? I— okay, we got it. Exterior paint color picker. Are you trying to rebuild what we've built with you guys?

I love it. I spent most of my time figuring out how to make sure I was ready to share my screen. I did not put much of a prompt in there, Jon. That's all good.

Basic, but— So what happened? Is it interactive at all? Like, if you click on anything, does anything happen? No. Yeah, not yet. So yeah, guys, for those of you who don't know, we, uh, there you go, there you go. Yeah, that's cool. Yeah, so there is some sort of a color picker that you can do to change things. That's kind of cool. You can repaint. So I guess you select it and then you choose the color. Amazing. It was kind of fun. That's kind of fun. Um, I love it. And guys, this is all supposed to be like very, very, very basic, very simple, right? So don't feel bad if you're like, ah, this doesn't look good, or I'm not doing this well. Um, that's what's supposed to happen here, okay? I want you to struggle just a little bit on this one. Um, all right, anybody else have something they want to share?

Um, Jon, uh, I'm gonna give it a go. I'm not sure if the Unforgettable group is just having all the same issues, so I'm going to go ahead and try to share because mine actually turned out kind of cool in my opinion. Um, let me, let me give it a go here, okay? Okay, so the idea was—

is it sharing?

Yeah, yeah, yeah, we can see it. So the prompt was basically create a Tesla Cybertruck that's painting commercial buildings. So that was just kind of where it started. And it This is kind of what it turned up with just the little dots of color. But I'm actually a little proud of myself because I came in with zero knowledge.

So, Jon, you're doing great and just happy to continue to absorb the knowledge that you're sharing. I love it. I think it's so cool. I love the colors and all that stuff. Can you control it at all? Can you like make it drive around or whatever? Anyway, it's super easy, guys, to add that in, to add that type of a piece in. Where, where you just say, hey, okay, now that we've got this, let's make it interactive. And when I do this, then make it do that, and it'll just start updating, right? And again, you're just talking to it in the same way you would talk to somebody that you hired to do that. Okay. All right. Sharon, do you want to try to share? You don't have to if you don't want to, but, but if you want to try to share what you got, we'd love to see it. And we're about done with this, with this kind of first activity here, and we're going to dive into the next thing pretty quick here.

Stephanie, did you share something? I, I just finished the code, so I'm only just pasting it in now.

Awesome. Okay. Sharing. Yeah. Share it in a second.

I had a question, Jon, because while everyone's been doing this, yeah, I went between ChatGPT and Groq and put in the same prompts.

Yeah.

And it's amazing to me how very different both of them are and it, I keep updating it. So is it like in the world of code? It's not like if you're specific with a sentence in writing, there can still be vast interpretation between systems.

Absolutely, Stephanie. Yeah. So Stephanie, you and I, you know, you're a writer, right? And I'm a writer. We both love to write. And I guarantee you, if we were given the same topic, even if we were given a 4-page brief, and then we said, okay, let's write the first chapter. We're not going to be close. We won't be in the same room. That makes sense. Right? That makes sense. And it's because, and we're going to talk about this in just a second here when we get into Replit, it's because of something called chain of thought reasoning. And it's really cool. It's a really cool thing that when you understand how it works, you understand why we're getting so many different things. I could give the exact same prompt to Grok 10 times and I would get 10 different answers. Absolutely. Wow. New conversation, new prompt. New conversation, new prompt. It's going to, it has to fill in the gaps, right? And so it's gonna fill in the gaps kind of randomly and it's gonna make decisions and then it's gonna, you know, operate off of those decisions and then keep going. And so there's just 1,000 different ways that it can go or maybe way more than 1,000 when you give it some sort of prompt unless you're just ultra, ultra specific all the way through.

So code is as creative as writing basically.

Absolutely. Absolutely. Okay. Code is absolutely a creative task. Okay, Sharon, how are you doing? And you are muted, but hey, we can see something.

That's cool. I love it. There it is. And it's not blurry. In theory, I asked it to create a compass which had 4 quadrants and it has done a version of it, but I'm not excited by it.

Let's say, okay, a lot of work to do. No, you know what? And don't think that you have to do a lot of work here, guys. This is not supposed to be pretty. Chris and Taylan made theirs look pretty somehow. They just got some good results out of it. But bottom line is, the real goal here, guys, and you did it, you've made something, you wrote some code and you made it deploy. You could now copy this and put it onto a website. You could actually deploy this. That's not great, so you're not going to. But now you can code and you couldn't an hour ago. That's cool. AI has the ability, guys, to help you do anything. I could ask AI to help me build a nuclear reactor in my backyard, and I believe it could help me do it. Now, I'm going to run into some constraints when I don't have the materials I need for sure. But in terms of how to do it, it can walk me through it, right? Where do I start? What's the very first thing? Okay, what's next? Now what do we do here? How do we make this work? How is this piece going to work? Right. And so I think it's, I think it's awesome, guys. Thank you all for sharing. And, you know, this, this activity has taken us about an hour to do, maybe a little less than that. But intros plus this, we're right on time in terms of our schedule of what we want to get done today. But this just gets you over the hump. You build something, it's cool, it's fun. Building a game is super fun and easy. Now you've gotten past this hurdle of, okay, can I actually build something that deploys and works? The answer is yes, you absolutely can. If it messes up, you get a little error, syntax, whatever it is, Great. Tell, tell the AI and it'll fix it for you. Hey, it looks ugly. Great. Let's fix it. Hey, do this again. Upgrade the flying saucer. Whatever it is, you can have AI jump in and fill in gaps very, very quickly. If you were trying to learn how to do this on your own, right, you know, without AI, it would take you months. It would take you months to get to something where you could actually have anything that we showed today actually working. Okay, you know, like one of the first things you learn in, in a lot of coding classes— I haven't taken a lot of coding classes, but I know this one at least— is Hello World, right? It's like, okay, let's just get a website to display the phrase Hello World, and you know, it's just a few lines of code and boom, hey, I've deployed something, right? Now your first thing you ever did was an interactive game. That's crazy, right? The, the jump between Hello World and interactive game is several orders of magnitude of difficulty, right? I mean, we're 100, 1,000 times more powerful than, than we were a few years ago before we had AI to help us out when we're learning new things like this. Okay, so you guys can all throw those projects away and you never have to look at them again. Although it is actually pretty fun to, to do this with your kids, right? All of my kids know how to build little things in Grok or ChatGPT and JavaScript. JavaScript or p5.js or Three.js or whatever it is that they want to mess around with. They know how to use it all because I've taught them. And all they have to know is how to ask with normal language and how to read. You just have to follow the instructions, right? Hey, I did this, I just did this. And if you were careful and you actually read what Grok wrote you or ChatGPT wrote you, it walks you through the steps step by step on where to go and what to do. Okay. All right. So now we're going to take it to the next level. And this is back to watch and, watch and learn on my side here. So you guys can all pause for a second. Just sit back. Welcome to Replit. Replit is an amazing tool. We're not— you're not going to get into Replit yet. Okay, so I don't want you guys to go jump in and jump ahead. There's a lot of things we have to do. In fact, I'm going to show you the wrong way to use Replit first, just like I showed you the wrong way to use Grok first as well. Then we're going to go in, we're going to make some awesome strategies and plans and business plans and all kinds of fun stuff like this and dive into the really business side of things here. So there's, I've built dozens and dozens and dozens of apps inside of, inside of this. And it's because every single time I teach any group, whether it's a course like this or whatever, I, I come up with a new idea, right? So maybe you guys can help me come up with an idea. I'm just looking around for some sort of business that I could do. I've had some ideas. I had an idea the other day and I forgot what it was. I was like, oh, I should build that in the course. And then I just forgot it. I like lots of things. If you look around in my office here, I've got, I've got 3 pianos, got a piano over there, one right here, got one over there. So I like music. I've got Lord— I've got Aragorn's sword on the wall over there. I like movies also. And I like music. I like kayaking. What can I do with kayaking? There was something with kayaking I was actually thinking of the other day that I wanted to build. All right, here's what we're going to do. So I like to— yeah, I like to kayak, whitewater kayaking in rivers. And one of the most important things that you need when kayaking is water, which sounds obvious, but just because there's a river there doesn't mean there's always water in it. Right. And so I need to be able to know when a river hits a level that I'm interested in. Right. What I want to do is build something that will allow me to— or that will like notify me when a river hits like the level that I want it to be at, right? When it's at the right CFS, cubic feet per second. So let's build that. So build me a website or a web— let's say a web app that will notify me when rivers hit the right levels for me to go kayak them. Okay. All right. So now we're just going to let it go. So right now what it's doing, guys, so Replit is— we're going to talk about chain of thought reasoning for a second here. And this is, this is really just still blows my mind when it, when it, you know, dives into this. There's a lot of different platforms out there, guys, that can help you do this. Now this is more of an agentic system. This is built to be able to build websites, to build full software. I have built incredible, incredible things using Replit. All of GenAIPI, everything you experience with GenAIPI, all of this was built with Replit, 100%. Every functionality, the email systems, the payment systems, the courses, the, the, you know, predictor tests, all of these things. I don't know if you guys have tried any of this stuff on the website, right? But all of this was built with Replit. Okay. And so we're going to go ahead and approve plan and let it go. So what it's doing here, guys, the way that these agentic systems work is they do something called chain of thought reasoning. And while we sit here and watch this, it's actually designing the visual preview. It's writing real code. That isn't just a funny animation. It jumps in here and it's just going to start building something and it knows, hey, you know what, I'm probably going to need a menu and I need— hey, that's not kayaking, that's scuba diving. I like scuba diving too. But anyway, I don't really have to wait for the ocean to hit the right level when I'm scuba diving. Okay. Emily, we're going to pick on you. I want you to walk us through. Just tell us how you get ready in the morning.

Jon, that might go through the duration of this call.

I don't really know if we want to go through all that, but that's okay. Give us the high level.

What are the kind of— brevity of it?

Sure.

So generally wake up, hit the gym. Is this the kind of way you're going about? Yeah, yeah, yeah. Okay, so, um, go to the gym for about an hour, um, come home, get some protein in some fashion, whether that be a protein shake, some eggs, something like that. Um, hop in the shower, obviously get ready for work, uh, check my phones for any last minute updates, emails, calendars, any types of updates like that. Um, this is a very brief version, by the way. And, yeah, head out the door, take some calls in the morning commute maybe, and ready to rock and roll.

So let's dive in on one thing that you didn't mention, but I'm sure you do. Brushing your teeth, right? So I want you to talk us through— pretend I'm an alien and I just landed on planet Earth, and the alien says, what, what is brushing teeth? And you have to explain to me step by step, how do I brush my teeth, right? You don't have to explain to me why, right? I just want the how.

Walk us through step by step exactly what you would do to brush your teeth. Okay, so you, you take what's called a toothbrush, which is an object that has bristle-like, bristle-like appendages attached to a stick that would— you would hold in your hand and with some force, little to medium force, scrub your teeth. Prior to that, you put some toothpaste on it, which is a cleaning agent to brush your teeth, hopefully rinsing with water upon completion of brushing your teeth.

Boom. All right, perfect. So there's— and there's obviously tons more steps, right? You could say, hey, you know what, you know, you know, first you're going to turn the water on and you're going to like pull the toothpaste out and you're going you're going to open the cap and you're going to put a little bit of toothpaste on and you're going to close the cap, then you're going to put it back, right? There's all those little steps. And in, in all of the things that you told us, you go, you work out, you come home, you get some protein, you do these different things, right? Each one of those is just a very high-level task. But our brains, we've got quantum computers inside of our brains, right, inside of our heads. And so we don't have to, you know, step by step figure out what that actually means, right? But AI is actually thinking through, okay, I've been given a task of, you know, brush my teeth. And so, all right, in order to brush my teeth, I'm going to need to do these kind of high-level steps in order to do this. I'm going to need to do this and I need to do this and I need to do this. Right. And so it builds these kind of little task lists that allow it to accomplish this. So right here, what it's doing is it's building a website and it says, okay, well, a website obviously is going to have, you know, some sort of a menu. There's going to be some sort of a hero image and, you know, a title. And then there's going to be some interactions and, He wants me to build something related to rivers. And so there's probably going to be multiple types of rivers. And so it's just thinking through, it's using that same chain of thought reasoning where it takes a primary task, breaks it down into subtasks and breaks those subtasks down into, okay, I'm now ready to write a line of code that does that, that supports this task and that all of these tasks support this task. Right. And it's amazing. It's amazing. This is actually a relatively recent development in AI. A year ago, it could barely do this. 2 years ago, it could not do this chain of thought reasoning. And it's so incredible to watch the speed at which AI gets more and more intelligent. Now, what we have here is what's called a visual preview. I can't actually do anything with it yet. I can't click the buttons yet. But this is pretty dang good for a first shot, right? Building a custom website that's going to send me the, you know, hey, it got a couple images wrong, whatever, right? We can fix that later.. But aside from the images, this is, this is almost a usable product, right? Like, hey, I need to search for the rivers. Here are my rivers. Here are my active alerts, optimal conditions. 5 of them are ready to go. Here's, here's a few different, you know, kind of these river cards we'll call these. Yeah, it might be a little bit too high. This is really, you know, let's, let's stay out of the Merced. These are actually real rivers that are fun to kayak. That's pretty cool. But anyway, bottom line is, I, I don't know, are you guys impressed by that? I'm impressed by that. I am still impressed by Replit every single time it whips up a custom site for me. Um, has anybody here ever built a website using like Wix or Squarespace or anything like that? It takes forever, guys. If you've ever taken the time to go do that, Wix is supposed to be, you know, um, just come in and, you know, click a few buttons and you have a new website ready, right? And here is— this is one of my websites, you know, my— for my Flow State Leadership. And I actually totally turned it off and I rebuilt it with Replit because even though I can use templates and all these things, there's so many things to set up. Like, it's crazy how much you have to set up. And I don't even know— like, here's the site, edit the site, and I probably spent— and the site looks terrible if it'll actually pull up, I think. I never, I never like really published this, but, but I wanted to build something for my consulting agency just called Flow State Leadership. And I wanted to have something that talked about, you know, it was all flow state. I wanted to talk about flow and it's terrible. Like, it's, it's terrible, right? And I'm not kidding you guys. I spent like 3 hours trying to get this to work better and I couldn't get it to work right. I started working on some other content. For one of these pages here, this market-based validation. I wrote a bunch of content that was really nice, but, but man, the design isn't engaging. It's just terrible, right? And so now I'm just like, I'm done with Wix. I'm never going to use Wix ever again. Wix is supposed to be easy, but no, it's way easier now to build a completely custom website that can do anything I want. And that's true to a certain extent, right? If you're building something that, you know, Shopify potentially could be easier in some ways, although setting up a Shopify site, I'm not kidding you, takes days to get right. It takes a long time to do that. And now this can just kind of whip this up for you really quickly. So Sharon, you asked, is this real live data behind or is it just a visual example? This is going to be fake data for now. We're going to fill in data later once we learn about APIs and how to connect all that stuff. Right. Okay. Now let me answer another question here. Agentic. Agentic basically means multiple agents working together, right? Instead of having, you know, just kind of, hey, I'm going to ask a question and it's going to answer, it thinks through these multi-step processes and many times is actually pulling up multiple agents to do lots of work together. Okay. Good questions. Good questions, guys. And then I did see the question about Replit versus Lovable. Replit wins all day long. And the short answer as to why is Replit is the more complete system. Um, you don't have to leave Replit to do tons of stuff, whereas with Lovable, you have to integrate a bunch of external things to make it work. And so especially for a beginner, um, that doesn't want to have to go and learn Supabase and all these other things, um, Replit's going to be— Replit's going to be better. Replit actually is the bigger company in terms of like market cap. I think they're about twice the size of Lovable. Lovable is great. I think they're doing good things, but I personally am not, I'm not a fan in the same way, or it's not even close for me. So look at this. I can now click on a river and I can go back. I can say, you know, I can enable alert, right? None of this stuff works yet, right? This is all fake data for now. We'll have to connect it to some APIs and things like that. I can click on my rivers, maybe. Yeah, that doesn't work yet. Looks like the only thing that works right now is these cards. But yeah, it's got the difficulty, it's got a USGS station. Last updated, current level, water level gauge levels falling by 0.1 feet per what? Right. Anyway, so pretty cool. This is pretty cool, guys. I'm, I'm, I'm impressed. Okay, so now let's build it the right way. This is good, but we're going to build great. We want to build great. Okay. And so we also haven't figured out this is, this is a, you know, an important question here. How am I going to make money? With this website right here? I don't know. Replit doesn't know. I didn't tell it. It didn't even come up with a way like that.

There's no—

I can't even think of how I'm going to do that, how we're going to make money with this game. So we are going to stop with Replit and we're going to dive into ChatGPT. Okay, so this is the time to pull up ChatGPT in a minute here. You can watch me for a minute still and then I'm going to give you guys— everyone's going to— we're going to get about a 15-minute session to just kind of work on your project. Okay. So you all came hopefully with some sort of an idea. If you don't have a specific business idea or a specific tool you want to build or whatever, build a personal brand website, build it for yourself, build, you know, something. Maybe you have a book you're wanting to write. Maybe you have a message, a charity you want to do, whatever it is you want. I want you to know, guys, that I don't want you to stress too much inside of Replit. You can have unlimited projects. Okay. Building this first project, you work for a few hours on it during this class.. But now you're gonna, but after this class, you're gonna be unlocked. You're gonna be able to go in and whip up things in a, in a matter of minutes into hours, right? And so, I have, I've had some people come through this class, by the end of it, they have 4 working apps. Okay? So don't be like, I don't know if I'm gonna pick the right thing. Like, don't stress about it. Pick something and just go. Okay? All right. So I'm gonna take the exact same prompt that I gave this, that I gave Replit. I'm gonna take it over into ChatGPT here. I'm just gonna paste it. I'm gonna say, We're just going to change a little bit. I want to build a web app with Replit that will notify me when rivers hit the right levels for me to go kayak them. Okay, and I'm just going to tell it that. Okay, so now we're in ChatGPT. So ChatGPT is a brilliant, brilliant system. It sometimes thinks a little bit longer now, and that's fine. You can tell it to do other things. It used to have a bunch of other, you know, models that you could choose from. Fast, Thinking Pro, legacy models. There used to be a lot more things in here. It's much, much simpler now with ChatGPT-5. Some people like that, some people don't, but that's what we have to work with. Okay, so we have ChatGPT-5. And so, all right, here's— so now it's kind of putting together a little project for me, and that's fine. So a lot of other things we'll talk about here in a minute. While this is thinking for a second, I want to talk about one other thing before we dive get into deep here, and that's something called context. Context matters in AI, guys. It needs to understand what you mean. Okay, if I were to just dive into a conversation with anybody, with any of you, and I just started talking about, you know, the design of a kayak or the design of some system of some marketing plan or whatever, but I didn't give any details around it, you'd be like, what are you talking about? Can you give me some more context, right? And so understanding context is important for humans. It's even more important for AI. If you don't give it context, what it will do is it will make it up for you. And now all of a sudden you're building what it wants to build and not what you want to build. Remember, you're the boss. You have to keep— you have to— you have to keep it informed in the right way. We're going to show you how to do that in a really nice way. So there's 3 levels of context that I want to talk about inside of ChatGPT. And actually, I did something a little bit wrong here, but we'll back up and fix it here in a second. We'll copy this little thing here because we're going to do this again. So first level of anything is this chat. Okay? If I were to just keep having a conversation with this chat, or if I went back here and I said, hey, you know what? I have a question about Oh, let's see here. Misleading zoning claims. Email. Earache relief in Cabo. All right. So my wife had had a really bad earache when we landed in Cabo, and it just turned out it was just something that was left over from the plane and it resolved itself by the next morning with an antihistamine. So no big deal. But, but we were having some conversations about it. If I were to say, okay, you know, the earache went away, Okay, so it— I don't have to tell it what I was talking about. It remembers all of this stuff. It remembers the context of, hey, this is what's going on. She had a bad earache, right? This isn't working. And all these different questions and things that we went back and forth about. And then it's saying, okay, great, enjoy Cabo. And it doesn't know that we're not in Cabo anymore. No big deal. But bottom line is it understood the context because we were still in that same conversation. So I can go back even to an old conversation. Dive right back in, pick up where I left off. It understands what's going on. Okay. Just like we did earlier when I was asking it about p5.js and then I said, hey, now build a model of the solar system. It knew to use p5.js. Okay. It understands the context of the conversation itself. There's a second level of conversation, and that is at the system level. Okay. And so I can click here in settings and look at my personalization. It has this cool thing called Memories, right? I can click Manage Memories. So it, uh, it knows all kinds of things about stuff I've done, who I am. Like, you can— it's hard to see here, it's kind of small, I'm sorry— but it says Jon is fluent in Spanish and Chinese, um, he's a class 5 whitewater kayaker, Jon's a member of the Church of Jesus Christ of Latter-day Saints, Jon led and founded Seek, um, Jon currently serves as CEO of this, Jon is accomplished pianist, whatever, right? It knows things about me. Okay, this is important. Like, what about my religion here? What if I was Jewish and I ask it a question that's spiritual, right? I don't want it to give me a Catholic answer or a Hindu answer. Like, none of those are bad, they're all great, but if I'm Jewish, I want a Jewish answer, right? And so by helping it understand who you are, you are going to get better answers that are going to match more closely to what you want it to respond with.

Okay.

And you can obviously, you know, go outside of that. If I said, hey, What do Jews believe about this? I could obviously talk about that. I mean, it's not going to give it to me in a, you know, the Church of Jesus Christ of Latter-day Saints lens necessarily. But if I just ask it without giving it any details, well, it's going to have at least the context that it knows that about me. Okay. If I ask it about kayaking, it knows I'm a class 5 kayaker. So it's not going to teach me about, okay, what's this little— what's a paddle, right? Or what's an eddy, right? I know those things. It knows I know knows those things. It's going to say, oh hey, well this thing, and you're going to go here, and here's how you run this rapid, here's how you get down. Since you're an experienced kayaker, you're going to know this, this, and this. This is what you want to bring. You know, it just knows that stuff. Okay, so that context is important. But one thing that I love in here, and a few AI programs or systems out there have, have this capability, but it's this project thing here. Okay, and this is actually what we want to use for the, for the river app. So kayaking, kayaking river app. Okay, we'll just create the project. And the project is really cool because you can add other things to it. You can add project files. I can just click here and drop a bunch of files in here, right? And I can just upload, you know, PDFs and, and Word docs and, you know, whatever the heck I want, I can drop in here and it can then reference those things. Okay, then I have instructions. I can actually tailor the way that the GPT is going to respond to me, that this, that this project is going to respond to me. Okay. Let's give it a few instructions here. We're gonna say, I want you to know that I am the best kayaker in the world. And I want you to address me as such. Okay. Whoops. I clicked. I meant to press enter. I clicked. I did push enter. I meant to click shift enter. Let's do that. Oh, come on. Give me more. Why will it not let me go down? Oh, it's not letting me do like extra lines. That's okay. We'll just keep typing here. I want you to also act like an expert startup advisor. Think like Steve Jobs. We want to create a category-defining product Yeah, okay. Don't agree with me just for the sake of agreeing with me. I want you to actively improve my ideas. If I have a bad idea, tell me. Don't just be a yes man. Help me build something special. Okay. All right. I'm going to copy this prompt. You can drop it into your, into your thing when you, when you get to that stage there. I just dropped it into the chat so you can use that in a minute. I'm going to click save. So I've started a new project. I've given it a little bit of instructions here, and now I'm actually going to give it that prompt from earlier. Kayak Alert app plan. Right top, and we're going to copy this. Kayaking River app. And okay, so now I'm inside of my project, Kayaking River App. I've started my first conversation. Okay. And now I've told it what I want it to do, just like we did earlier. But now it has a little bit more context. It knows I'm the best kayaker in the world, and so it's going to treat me as such. And so that's great. Now, obviously, I'm just, I'm just kind of joking here on that. I'm not the best kayaker in the world. Yeah. King of Class 5. Love this. Okay, so it you know, whatever, it's, it's going to appease me, right? Remember, it wants to please me. That's what the AI is actually doing. So it's going to help me figure out this is— these are the types of things we're going to have to do. Okay. You know what? This is great. And it's going to come up with some cool plan. But I want— I want it to be a little bit better than this, right? It's telling me all kinds of fancy things that we don't need to understand right now. Okay, guys, by the way, we will take a break here in a minute. And by a minute, I mean probably about 20 minutes or so. It'll be our first break. It'll be our only break today. It'll be like 10 minutes long. We got lots to get to, not enough time to do it. All right, so stop writing. I don't, I don't want you to write all this stuff. I mean, it's cool that it's writing some USGS polar information, but I'm going to have Replit do this stuff later. I might just click stop. See, it's just thinking too hard. AI is a little bit overeager sometimes. All right, so it gives me now— pulls USGS, wire notifications. All right, now we're going to stop. Okay, we're just going to have it stop doing this. Okay, so what I really want to do is I want to be in charge. It's just going off and, and like setting up SendGrid mail stuff for me. I, I don't need it to do that yet. Okay, and I'm not going to do that inside of, inside of this conversation. What I'm going to— I'm going to do that over in Replit when I'm ready. What I first want to do, guys, remember, we're the boss and AI will fill in the gaps where we don't fill them. And so instead of just letting it build this, like, it just— I gave it one sentence and it's just building my plan for me like it knows what I want, but it does not know what I want. It has no idea what I want because I haven't told it. And you know what, guys? Humans are meant to create. That is our core superpower. We can create humans, we can create art, we can create music, we can create jobs, we can create memories, we can create so much. I want you to try to think right now about something that you've created, whether it's a nice recipe or a garden design or whatever it is, and I want you to think about how you felt while you were creating that. And I bet you can't come up with a creation memory that's sad. We are happy when we create. We all create different things, we all create differently, and different things make us happy. That's fine, right? Um, you know, variety is the spice of life. That's great. We're all— we all bring our unique things to the table, but we all find the most joy in life when we are creating. Okay, and so We can't let AI take that from us. We can let AI help us, help us do the work, right? And, you know, get, you know, execute on the mundane, right? But we want to keep the creation to ourselves. And so here's how we're going to do that. So I'm going to say this. I want you to interview me one question at a time about this business idea. I want you to ask me about the mission and vision, the branding, the name of the company, the USP. USP is unique selling proposition. Why are we unique? Why are we, why are we important? Why are we going to be here? The USPs, maybe even the monetization strategy, and anything else you need in order to help me build a business plan, a solid business plan. Let's just say that. Okay, that's it. So I'm going to go ahead and click enter here. I'll drop this prompt into the chat for you guys because you're going to use that in a minute. Okay, you're going to tell it you're just a high-level idea and then you're going to stop it from making everything up, right? So now we've done this. We've— so if I were to ask it, hey, help me build a solid business plan, right? Ask me some questions on it. It would give you like 20 questions right away. Okay. That's a really inefficient way to use AI. Our brains don't do great at answering 20 questions at once, but they are really good at answering one question at a time. Okay. And so if you let it break it down for you, this is a trick I use every day. I shared this on LinkedIn yesterday, actually, this little— this kind of little idea. I want you to interview me one question at a time about how I'm feeling about something in life, about my business strategy, about my new sales strategy. How do I overcome this objection?

Whatever.

Help, you know, and what it does is it, it forces you to extract information out of your brain and give it to the AI. Okay, so what's the mission of this business? What's the core reason it exists in the world for kayakers, potentially beyond? So I have to, I have to think of my answers here, but let's just say kayaking Kayaking is something that matters to me very deeply. I want— I don't want to miss the best opportunities to kayak, and when the river is right, it's time to go. Okay, so that's, that's my, I don't know, this kind of feeling behind it. All right, and so now it's gonna, you know, it's gonna say, okay, I'm gonna process that, I'm gonna store it. And then it's going to ask me the next question, or theoretically it should. The one thing I don't like about ChatGPT-5 is it thinks a little bit longer. And so it just— I like speed, but that's okay. All right. That's powerful. You're tying this to urgency, timing, and the soul of kayaking. Don't miss the magic moment. Let's sharpen it. A mission should be short, memorable, inspiring. So it's even helping me turn it into a mission, right? To be sure, kayakers never miss their perfect river day. Awesome. That's a pretty good mission. Now it helped me kind of pull out of what I said and it turned it into exactly kind of what I wanted, but, but wrote it out as a mission. Now did, did ChatGPT come up with that on its own? No, I guided it. That is my thought. Did it help me express it a little bit better? Absolutely, right? And then I can choose to tweak it or keep it and be like, hey, that's great. All right, now what's the vision, right? If this idea goes big, I want every kayaker in the world using this. This thing. Okay, awesome. Now I don't know if I'm actually gonna launch this app. Sometimes I, I push farther with some of these ideas, but, but I'm just gonna kind of act like that. Okay. Love it. That's bold category king energy. Awesome. Okay. Now let's get to the identity branding. What do they, what do you want 'em to feel? And it's just gonna keep going. Now let's say, I don't know. What do you think? What do you think is a good approach for this? So even if I don't know the answer, I can say, you know what, give me some ideas, right? So let's break it down like Steve Jobs would, right? What emotion do you want to own in the paddler's head? Core paddler vibe, trusted outdoors data platform, playful chasing flow brand, right? Start with number 1. Yeah, I agree. Yeah, I agree. Let's go with number 1. Okay, next question. What name ideas are already bouncing around your head for this app? I don't have any ideas for this piece. Give me 10 I can choose from. Okay. So you can ask it to help you with these things. You can have it, you know, work through some of these pieces. Okay. Flowstate. Hey, Flowstate. That's kind of funny. Since that's my company already.

Stoked.

First drop. Floodgate. Anyway, So I can keep doing this, guys, and I can go through about 15 or 20 questions. So what we're going to do now for the next 15 minutes is you're going to do the following. One, set up a project. Okay. Two, so remember when you click that project, you'll just name it and then you'll set up your, your instructions, tell it to act like Steve Jobs, or maybe you don't like Steve Jobs. Maybe, maybe there's another designer or maybe a writer or someone who you want to help kind of guide you through this. I like kind of throwing some feel things like, hey, I want you to help me build this like this website or whatever it is. Right. Give it a few instructions and then just give it your top-level one sentence, right? And just say, I want to build a web app with Replit that will— right? Or I want to make a website with Replit that will— boom. Okay, whatever it is. All right. And, um, that's your idea. That's your idea you're going to do. And you can be as complex as you want with this one, right? You want to build your, your paint project, um, Corey, go for it. You could try. I can warn you, some of the things that, that we've done in that are pretty technical, but, um, you can still try it. You can still do whatever you want. Okay. All right. But anyway, guys, um, dive in and you have until 11:06 my time. So 10:51. I'm gonna, I'm gonna just keep typing through this and answering my questions. But if you have questions for me, guys, you know, chat with me, speak up, whatever. Okay.

Hey, Jon, quick question.

Yeah.

If I may, the first prompt that we give this is the one where you had stated, I want to build a web app with Replit.

Is that where we kick things off? Yep.

Yep.

That, that, yeah, that's how you start your, yeah, once you've built your project and you're like, all right, I'm ready to start talking to ChatGPT about this. Just give it a 1 sentence, 2 sentence, whatever, 3 sentence, 5 sentences. Just tell it what you want to do, right? But you're really going to get into the details after, after you get through this, this interview process. Yeah. So, or as you go through the interview process, you don't have to be super detailed in the beginning. Just kind of say, hey, I want to build— this is kind of the high-level project I want to work on.

Okay, great.

Thanks.

Yep.

So, Jon, we don't need at this point to say I would like a verbal identity. This is my color palette.

None of that stuff right now. Go for it. You can do that too. I actually like to do that. I like to say, hey, here's the color palette I want to work with. I want a fuchsia color and I want this and I want that. And, you know, please provide some, you know, some good, you know, complementary colors that go along with that for secondary and tertiary colors or whatever. You can tell it all kinds of stuff. You can ask. One of the things I asked it in my prompt here was, you know, branding. I said branding. And, and colors are absolutely part of the branding, right? It should ask you about it. If it doesn't, you can tell it also. You say, hey, by the way, I want these colors in the branding, right?

Yeah, great, thank you. So if I'm clear about the project, do I still need to start with ChatGPT, or would I rather just go into Replit and just put in the ideas there?

No, no, you absolutely want to do this in ChatGPT. Replit is not capable of doing some of what we're going to be here. There's a lot of work here you're going to want to do. You want to use ChatGPT to plan and to strategize and to do all that. Now, there is some of that you can do later, but what we're building right now, Mitesh, which is just kind of skipping ahead a little bit here, is we're building guidelines and a framework. We're going to take this framework and we're going to hand it to Replit. Okay? And then Replit's actually going to be like, like, it's going to be like giving Replit a 10-page prompt instead of a single sentence.

Yeah. I see. So it's whatever project. So for an example, if I can share my screen with you, I want to kind of build an app and I've kind of got all my requirements in a way written down on one page. So this is what I would kind of give to ChatGPT saying, okay, this is the kind of app I want to create. These are the requirements. Now convert this into a prompt for Replit and give it to me.

Not quite. Not quite. What I want you to do, what I want you to do here is I want you to say, hey, here. So copy that whole thing, right? And then I want you to go into ChatGPT, start your project, right? And I want you to say, hey, I want to build an app that does or that helps users with these goals or whatever it is, right? Correct. And then and say, here is, here's kind of my outline that I have right now. I want you to interview me one question at a time about this. So that you can understand exactly what I want you to do. I want you to go through this interview. It's going to pull stuff out of you that you haven't thought of yet. Then we're going to use it to build this.

We're going to use it to build a business plan. Yeah. Interesting.

Okay, that makes sense. All right, perfect. Thanks. Yeah, very good. Sorry, guys. I know every time we share a screen, it takes over your screen.

And again, Jon, you like ChatGPT for all—

for this interview versus Grok? Yes. Okay. We'll use Groq in this process in a minute here.

But we're gonna, we're gonna, our project manager is ChatGPT. Got it. So I'm just taking your prompt about, I want you to interview me.

I don't need to say Replit to it right now. No, you can, you can tell. Yeah, I mean, that prompt where it says like, yeah, I'll be building this in Replit, right? You want it to know from the beginning what you're— right, where you're going to build it and how you're going to build it, because it just helps. It helps with a couple things down the road.

Perfect.

Thanks. Yeah, more context. Hey, me again. Sorry, Jon, for the interruption. So I went in there and did the first prompt, like saying I want to build the web app using Replit. And then when I went to copy and paste your next prompt in there, it's almost like it's blocking me from moving further. Like it doesn't want to take any more prompts. Am I incorrect in thinking that's what I was to do next, or?

No, no, no, you should, you should be able to do that. You should be able to prompt, prompt it, no problem. Um, go ahead and share your screen.

Let's look at it real quick. It's the same thing happened to me.

Just refresh the page and it, and it should start. Yeah, you might be on a web browser occasionally, guys. Some of these things crash. They're, there's, it's software too, and something weird stuff happens.

Yeah, I did refreshing fix that for you, Emily. Um, it did. I believe it did. It's just—

it's—

you just gotta step back and copy that and then go again. Yeah, perfect.

But I think we're good. Thank you for the help. Yeah, awesome. Guys, as you continue to work through your plan here, I'm, I'm now in some conversations with it where it's talking about monetization. We've figured out how I'm going to make money with this thing. I've figured out how I'm going to market it, how am I going to get my first 1,000 users. All right, if it's not asking you that type of stuff, or if that's not important to you, whatever, that's fine, but If it's not asking that stuff, you should, you should prompt it. You should say, hey, let's talk about how we're going to get our first users.

Jon, can you scroll up to the beginning one more time?

The very first, there you go. Yeah, and I dropped that one. Maybe I didn't, maybe I didn't copy that one in there. I'll drop that in the chat just so it's there for later. You guys take another about 2 minutes. Remember, this is being recorded, and so you'll be able to go back and spend more time on some of these things later if you want.

Hey, Jon. Yeah, if it's— if you're answering the interview questions but it says no search results found, please try again with a different search engine or search query.

What does that mean when it's like you're not— it's on ChatGPT.

Yeah, it just said no answers found. Yeah, no search results found.

And I was just like responding to a question. Can you share your screen really quick? Sure enough. I've never heard of that before. I just want to see what it looks like. Yeah, I have no idea. I would just refresh your page and then do it again. Okay. Yeah, I have no idea what's going on there. It's weird. Okay, guys, take one more minute. Just answer a few more questions. You should have gotten through at least 5 or 6 questions, if not you know, 12 to 15. I'm not sure how many I made it through. Yeah, I did about 15 questions or so. Now, while you guys are finishing up answering a couple more questions, When I did this for GenAIPI, I took 3 hours doing this part. I really read every word and thought through things. I spent a long time on the brand. One thing that I quickly did, you guys might not have been watching, but I came up with the name Eddyline for my river app here. Then I was like, you know what, that actually I think is a kayak brand. I jumped on Google really quick and I did a search, Eddie Line. And I was like, oh yeah, it is a kayak brand. And so then I went back and I said, hey, we need to come up with a better word and— or better—

a better name.

And I came up with a better name. But you want to check for things like, hey, is maybe the— is the domain available, right? So I should go to— I should go to GoDaddy. Actually, you know what, um, I should go to GoDaddy and check this. I'm almost 100% sure that it's not going to be available. I switched to High Water. High Water is a pretty common phrase, um, and so it's oh, highwater.com is available. That's crazy. It's $600,000, but I'm, I'm shocked that it's actually available. It's pretty cool. Um, now there's obviously other, other, uh, you know, highwater.boatsolutions.com, dev, cloud, pro, highwater.pro. That could work. That could work for me. Highwater.club. Really cool if I just do highwater.app. Yeah, someone's got that. The pro one isn't bad. Highwater.pro, kind of cool. Info, shop, life. Highwater.life. I actually like that life. It's kind of fun. Highwater Life, that could be like a phrase. It's only $1.99 for the first year, then $51.99 a year after that. That's crazy, guys. But yeah, Highwater is actually a pretty cool name. And yeah, anyway. I like, I like that a lot. And I actually like highwater.life, highwaterlife.com. I can also do highwaterapp.com. I can do High Water whatever, right? This is part of building a business, guys. You got to come up with your name. You want to look at, you know, the Facebook availability. Does Facebook have, you know, if I, if I search for High Water on Facebook— oh, you know what, I don't want to be viewing this as Jenny IPI. I'm gonna be viewing it as me. High water, what comes up, right? I'm going to search for these things. This is something I'm always doing. Uh, people, pages, high water. Okay, so there is something here. Uh, it's an apartment building. There's a musician or a band. There's a producer. So nothing that's really like significant, nothing that has like thousands and thousands of followers, and definitely nothing that I'm seeing in the, in the kayaking and the whitewater in the river world, right? So I'm probably pretty dang clear with using this, this brand Highwater. I think it's actually pretty good. So I might do highwater.life if I'm buying that, um, so that I can connect that later. We'll come back to that at a later time. Okay, on the, on the domain. Okay, so I went through, I answered my monetization, we figured out all these things, my mission, my vision, my branding, name, USP, monetization, target, go-to-market focus. Now, I didn't, Stephanie, to your point, I didn't actually go over colors. It never asked me about my colors. And so I could certainly, if I wanted to, say, hey, give me, you know, let's, let's focus on the colors here. Sometimes I'll actually do that a little bit separately. But anyway, what we're gonna do now, guys, now that you're done with your interview, and I know you might wanna go back and do it more later, and that's fine. But, but now what we're gonna do is we're gonna say, say this, and I'll, I'll copy this prompt for you because you'll use it word for word. Word. Um, okay, now turn this into a 2,000-word business plan. Use the Canvas tool and based on my answers in our discussion. Okay, copy, send. All right, so the Canvas tool, guys, is a, um, it's a, it's a neat tool inside of ChatGPT that allows you to like have a document Right. And so it's creating now a document in here that will act kind of like a Google Doc. You'll be able to edit it after it's done. I can't edit it yet. I can't, I can't type yet, but as soon as it's done, I can come in here, I can edit it, I can copy it, I can save it. It will just sit here in my conversation so I can come back to it later. And so it's created a nice executive summary for me here. HighWater is a category-defining web app built for global kayaking and river sports community. Its mission is simple. Yet powerful. Never miss your perfect river day by leveraging real-time river gauge data, weather integrations, and unique travel time-aware alerts, which is super cool actually, right? Like, what if I'm going to a river that's 2 hours away? Well, then I need to make sure I'm hitting the peak in that exact level, you know, and I need to leave 2 hours early for that. But anyway, it's really cool. HighWater ensures that paddlers never miss the magic moment when their favorite rivers and sections are flowing just right. Vision is bold. Every river user in the world relies on HighWater to decide when and where to paddle. Okay, really awesome, right? Like that is very good. Now again, if I'm working on my own business and I really want to nail it, I'm going to read every word of this and I'm going to get it just right. For today, guys, I want you just kind of roll and we're going to do our very best. We're going to build this out. Okay. Mission and vision, brand identity. We did— now I didn't talk specifically in the brand identity piece about the, about the colors, and that's fine. So we might want to work on that a little bit. Here's my USPs. Here are some of the things that it does. Here's maybe some of the future stuff. Here's my target market. I'm going to go after some of the kayakers that are really, you know, pushing it. Here's my monetization strategy. We're going to have a free tier, then $8 a month. My go-to-market strategy. I'm going to start with influencer seeding and then it's just going to share and grow. And then here's my roadmap, exit strategy, financial projections, risks and mitigation, conclusion. Okay, so high water, never miss River Day. So this is awesome. This is pretty awesome. And you all have your own plan now. You should, you should put in that prompt and turn this into a 2,000-word business plan based on my answers in our discussion. Right. So now again, I can come in here, I can type in this if I want. If I want to edit it, I can. What I'm actually going to do is I'm just going to click copy here. I'm going to copy. It says failed to clipboard on this on the desktop app and they haven't fixed that bug. But if you click it again, it works. If you're in the, if you're in the web app, you should be just like on the website, you should be just fine. Okay. So I'm just going to copy this. And so what I'm going to do now is I need some feedback. I need some market feedback on this. Okay, so I'm gonna jump back into Grok. Okay, I'm gonna jump into here. I'm going to say, please, um, do some research. And by the way, I don't want this to be fast, I want this to be expert. Please do some research on the market, on what users are saying on X, um, and everywhere else online about this business plan. Okay, we're going to give it a little bit more about the following business plan. Actually, we're going to say the following business plan. Give me 5 ways I can improve it. Focus less on the actual business plan itself and more on the idea and market. Otherwise it will be like, oh well, a good business plan has a team section and you don't have a team section. Okay, I get it. I don't care, right? Yes, that's right. But I want you to actually, for now, I just want the idea threshed out. Okay, so I will copy this prompt for you guys in Grok. Okay. Mitesh, 2,000 words is kind of right at the limit of what the Canvas is good at doing. If you try to do 10,000 words, it can't do that. And if you do 1,000 words, it's not enough detail. So I've just found that 1,500 to 2,000 words is a really good spot for the Canvas tool. Okay. So then I'm just going to say go and it's going to start thinking and it's going to do a month's worth of research for me. It's going to search 150 different websites. It's going to search all of Twitter and/or X or whatever this thing is called and going to figure out what's going on here. Okay, so there it goes. It's starting. Yeah, thought for 14 seconds. I'm amazed that it went that fast. Normally it goes a lot longer than that. What was my prompt here? Very interesting. Wow. Broaden integration with JSON activities earlier. I don't want to do that. But anyway, guys, what I'm doing— so go ahead and do this, right? Go ahead. You don't have to sit and watch me do it. I'm going to read mine. Go ahead and do your research. What you want to do is you want to bounce it off of Super Grok and see what kind of stuff it says here.

Hey, Jon, quick question here. When you're taking what you've copied from your business plan, in ChatGPT, you paste that into Groq for, and then post the prompt that you had pasted into the chat?

Yeah. So paste the prompt into, into Groq and then go hold Shift+Enter or Shift+Return and then paste your business plan. Yeah. Okay. Thanks. Yeah. If you just push Enter, then it will just send the prompt without doing that. So if you do Shift+Enter, then it will do it without sending it. Mix. So guys, the reason I wanted to do this is, um, I want it to— I want it to dive in deep. Feel like you didn't do that. Like you didn't. What? I have a little conversation with Grok for just a minute here, guys. Ask it some stuff, maybe ask it a follow-up question, have it research a little bit more. So Grok used to have a feature that was called Deep Research, and it would go and it would search, you know, the internet far and wide for all this stuff. ChatGPT can do this. You can go into a research mode with ChatGPT, and it will take about 10 minutes to do it. So we won't do that today, but just know that there are other tools you can use to do this. The thing that I really have liked about Grok, at least in the past, is, is it, is it can search. It has access to X, whereas no other, no other platform has access to X posts. Yeah, I give it a better prompt and now it's searching for tons more things. I'll give you my extra prompt here just for fun. Listen, I'm not sure why you're having trouble with, with ChatGPT there. I don't, I don't, I've never seen that happen.

Maybe use a different browser if you're on Chrome. My screen really quick.

I just noticed something.

Kind of weird. Seen it before. So, in my project, it's just like not responding. It's like my queries are just disappearing. But then I noticed they're like being dumped into separate chats over here.

Weird. And not sure. Have you seen that before? No, this is okay. This is ridiculous. I'm not sure what the heck is going on here.

Sorry, I don't want to, I don't want to distract from the course.

You're not, you're not. Do you have, why don't you try just switching to Safari?

Okay, sounds good. Just, you could also try clearing her cache. Yeah, that's why I'm a little complicated, but we could walk it through it.

Yeah, that's why I'm saying just switch to Safari for now. That'll accomplish the same thing. Okay, great. Yeah, yeah, that's really weird. Okay guys, so now that you've got some feedback, hopefully, hopefully you're to that point now. You might not be there yet, and that's okay if you aren't. Um, I know you might, you guys might be just a little bit behind. When we do our 10-minute break, you can, um, you can catch up a little bit if you want. Okay. So we've got our updated market research summary. It's going through, it's figuring out, okay, here's, here's how we really, you know, focus on, on some of the things that we can improve here. So what I'm going to do now is I'm going to copy the feedback that I got from Groq. Okay. And I'm going to go back to ChatGPT and I'm going to say, please create an a new canvas called Improved Business Plan and integrate the following feedback. Then I paste what I got from Grok and I drop it in there. So now it's going to create a new business plan that's going to improve on what we've done. So I really like, guys, I really like going back and forth between different models. And apparently sometimes some models, you know, break and they have problems. And so having multiple available to you is really, really helpful. You'll find that some are better than others. Some are better at creating a document like this. Some are better at doing research. Some are better at whatever. But between ChatGPT and Grok, and I do use Claude and I do use Perplexity sometimes too, but but I'm not even— I don't use Perplexity very much. It's probably the least— the platform I use the least. I use Claude and I use ChatGPT and Groq. Those are the ones I use the most. And they're, you know, they all have their strengths. Claude has— Anthropic has a great project management tool as well inside of it. I'm a fan of that one as well, but I just use ChatGPT more. And so that's where all my projects are. But I have tons of business ideas and tons of projects and things that I'm working on.. And anyway, so, so I like to, I like to bounce in here. Like if you look at my general AI proficiency, my GenAI folder here, I have hundreds of conversations, tons of documents, tons of things. And so anytime I have something related to my business, like, hey, I need to do this, I need to add a new page, I need a new functionality, or I need to build a proposal for someone, whatever it is, right? I can just jump in there and it knows everything about my project. Okay, so now guys, I have a really solid business plan. Okay, the business plan, guys, is more important than you think to the development, to the development stage. Okay, um, the development, um, if you, again, if you just go into Replit and you say, okay, here's what I want to do, and even if you give it a detailed development plan, right, when it, when it has to do its chain of thought reasoning it, it often looks for a why. It wants to figure out a why. Why does the user care about this? Why is the user asking me for this? Oh, I think he wants it because of this, right? If you tell it from the very beginning, I am building an app that is— whose purpose is to do this and it makes money in this way, then when it goes to fill in its gaps, which it's going to have to fill in the gaps, we're not going to— we're not able to give it 100,000 lines of code. And so, when it's filling in the gaps and trying to make decisions on structure, on different things, on different features and whatever, if you guide it, if you guide it properly, then it will push, it will develop with those goals, those outcomes in mind, right? So, Chris, you were talking earlier, you're like, "Hey, you know, I'm here, Jon, because I want to, you know, you're a great entrepreneur and that's awesome. I want to learn. I know how to build the stuff, but how do I get more customers?" Right? I think, I think you could improve your overall process simply by, by applying this, right? Just by saying, okay, I need to think about the end customer, why they are paying me money, and whether or not they're even willing to pay me money before I've even written my first line of code, right? And every single feature and everything points towards that happening. Now, a lot of people do that already in app development, web development, everything they're doing, right? They're saying, hey, you know, we need to get them to the checkout screen and we need to make sure it's easy to get there and whatever. But, but the bottom line is AI can fill in so many of those gaps for you and build extra features you didn't think about and extra make it, you know, just super obvious in every place where and how to get users to actually sign up and pay for this and use the product for what its intended purpose is. Okay. So we've now, guys, we've now built a business plan. I've got my HighWater improved business plan, right? I just said, please create a new canvas. I didn't copy that one for you guys. I'll give you this one as well. Copy, paste. Okay, so please create a new canvas. And all I did was drop my X analysis into here and then it created a brand new plan. It still, you know, kept everything that we had here. And then it says, hey, here's some recent analysis, here's what we're doing, here's how we're going to make it a little bit better. It's going to be stronger because of these different things. We got the exit strategy, financial projections, all of this different stuff. Okay. All right guys. So from here, we're now gonna build one more document. Okay. And then we're gonna jump into Replit and then we're gonna take our break. Okay. Or maybe we won't take a break. I don't know. I know guys that I've kind of been going for a while, by the way, if you need to get up and go to the bathroom, grab water, whatever you need to do, just keep doing it. I feel like time just flies here. I wish we had 2 hours longer than what we have planned today, but it is what it is. It is what it is. And you guys probably can't sit as long as I can. Um, all right, we're gonna go to— we're gonna go to the next stage here. So this is, this is, uh, not a tough prompt, but, um, I want you to build a new canvas called Feature Lists. This, um, this will be a complete list— whoops— complete list of all frontend and backend features that we will need for this app. Or yeah, for this website to work. Yeah, that's all we need. Copy. Oh, you know what? There's actually one more thing I need to do here. Don't specify the technologies or systems that we need to use to, or for these features. Just list the features. Okay, that's a better— that's going to be a better prompt. Let's send this off. Okay. Okay, so now I've got my user onboarding, my dashboard, river detail pages, alerts, notifications. Now what it didn't do for me here on the front end, which I don't love, is I need a— I need like a marketing landing page. Um, please update that document to explicitly include a landing page for marketing. And yeah, yeah, maybe not. Maybe I just get them right into it. Maybe I just let them use it right away instead of that. Yeah, you know what, I'm gonna do that. Sometimes, guys, I want a landing page like like GenAIPI, right? GenAIPI, this is the landing page here where I'm going to sign out, where I'm selling the product right here. And then when I want to actually get into the app, then I'll come in here and I'll log in, right? And then I get into everything else that I have inside here. And now I don't need the landing page anymore. But for this one, I want people to be able to land there, find river information, and then if they want to engage more, then they can sign up and do that. But I want them to be able to use it it without having to really sign in. So that'll be something that I'll think through. Okay, so I've built now 2 documents or 3 documents really. We built the original business plan, Highwater Plan here, right? Then I've got Improved Business Plan and I have my feature list. Those are my 3 documents that I have here. Okay. Or sorry. Yeah. So I only need these last 2 ones now. I didn't, I didn't go over the branding and stuff like I wanted to talk about, but I'm going to do the branding kind of manually later. So I'll do that that way. I'm ready now to get into Replit, guys. And so if you want to just, you can keep working along with me. You can watch me for just a second here. So what I'm going to do now is I'm going to jump into Replit. I'm going to go back home. I'm going to just throw away my old project and, oh, you know what I'm going to do? I'm going to ask ChatGPT. I'm going to say, please write me a 3-sentence description of what we are building. Okay, do, do, do. Okay, building High Water, the definitive river utility app. Okay, copy, back to Chrome. What I'm gonna do is I'm gonna paste this. So when you get into Replit, guys, um, it'll just say describe an app you want to create. I'm just going to paste my 3-sentence description and I'm going to push Shift+Return. Don't send, don't start building yet, okay? We got to do some other things. And then I'm going to come back to ChatGPT and I'm going to grab my documents here. Okay. So the way I'm going to do that for this feature list, I'm just going to click copy. Okay. And I come back over to Replit and I'm going to paste it and it's going to paste. It's going to look like this. It's going to look like an attachment even though it was just pasted text. Okay. But it's a big amount of pasted text. And so that's why it, that's why it looks like that. Oops. And then I'm going to come back to ChatGPT and I'm going to grab my improved business plan. And I'm going to try to bring— I'm going to try to click on it, close out. Their app needs a little bit of work. I should be using the— I should be using the website, but whatever, I'm in here already. So I'm going to copy that and I'm going to come back to here. I'm going to paste that and say, be sure to closely adhere to the guidelines and, um, and attached. Okay, all right, so now I have a prompt that's really ready to go. Okay, now Replit knows what to do in a much more complete way. So again, to review, you should have in Replit, you should have a 3-sentence 'here's what we're building,' you should have this 1 sentence of 'be sure to closely adhere to the guidelines and plans attached,' and then you should paste your business plan and paste your feature list. That's it. Okay, so once you're there, you're ready to click Start Building. Okay, so I'll go ahead and click that and we'll get going. So when it starts going, it's going to make an implementation plan, um, and it's going to ask you some questions. Just click Start Building. Just Just tell it that the prompt for the 3-sentence portion, Jon, is just write me a 3-sentence description of what we are building. Yeah, please write me a 3-sentence description of what we are building. Paste. Okay, so back here it's reading all kinds of stuff.

Well-researched MVP plan. So within the same project, every time you want to create a new document, you say create a new canvas for this and it creates kind of a new document. Document, and then you can keep updating that document. Yes, kind of.

Wow, that's more organized than just the chat. Yeah, yeah, yeah, I like, I like it because you can't edit the chat, you can't put your own thoughts in there. I like having those documents, I think it's really useful. So here it just says here's the initial version, whatever, and I'm just going to click approve plan. Okay, don't worry about any of the stuff it's saying, um, just kind of know that we're still going to be using this plan to to guide it as we, as we push forward here. But now, guys, look, I mean, think about how much better of a plan we have here, right? We've got— we're building, you know, HighWater. We've got this. We know exactly what we're building. We've got an executive summary that has our business plan, our monetization, our target market. Who are we going after? Yeah. And the branding is not great right now. I need to work on the branding. And so I'll probably tell it later to do a branding facelift and I'll show you how to do that. I want to show that in here. That's why I did I didn't put it into the prompt, but you certainly can try. It's usually not going to get it right in the first shot anyway, and so you're going to have to do some tweaking anyway. But, but hey, for now it's, it's now building this out. And is it going to be similar to what it built the first time? Yeah, it might be a little bit similar because the core functionalities or features are going to be there in terms of, hey, we need to be able to check a river level and everything's like that, right? But this is going to be a little bit more complex as we, as we dive in. It looks— it used that same scuba diving picture here, and it picked the New River again, although the New River is actually in California. So not sure what West Virginia is doing here. That's okay. Or maybe there's a New River in West Virginia. New River, West Virginia. There is one. It's cool. I didn't know that. Can you kayak it? Kayak New River, West Virginia. It's interesting. Anyway, so, all right, guys, so this is where, once you've gotten to this stage, this is where I want you to take your 10-minute break. Okay. So we will take a quick 10-minute break and we'll just kind of crank through fast. So it's 11:34 right now on my clock. We'll go to 11:45 and then we'll, we'll come back, stand up, walk outside for a second, but just get it going. This takes about 10 minutes. That's why it's a good time to do this. Um, so just go ahead and take a break for a minute and we'll see you back in a few minutes. I'm going to, I'm just going to leave this up, but I'll turn off my camera and my microphone. See you guys in a few. All right, guys. Welcome back. Oh, a 10-minute break, standing up, stretching a little bit. Helped.

Okay, let's look at this chat here. Hey, Jon, I might have to hop out. So where would we get the recordings to watch this later?

Is it by email or the same course platform? Phone. It will be email. Yeah. So I will— I'll send— so at the end of this session, it'll take about 20 minutes for it to process the video on my computer because it's 4 hours long. Then I'll upload it just to an unlisted YouTube link and I'll send it your way. So I'll send it out to everybody. I'll just reply all to the email from this morning, except for Denise's email, who I got wrong on there. But I'll update that and I'll send it out. Perfect. And then we'll see you guys tomorrow morning, same time. So yeah, if you, if you miss out on a little bit, you can always go back and catch up later. Okay. Thanks. Thanks, Mitesh. So, Steph, you done on my business plan yet? So yeah, please create a new canvas. Yeah. Integrate the following feedback. Yeah. I want you to build a new canvas called feature lists. Complete list. Yep. That's good. Yeah. You'll want to use— I mean, those will all be separate prompts, Stephanie, but yeah, those are the ones.

That's this kind of process you'll go through.

Yeah.

Sorry, it pasted in as a jumble.

That's okay.

That's okay.

I was trying to figure out where the 3-sentence description we were building fit in. It's right there. Yeah, that's at the very end when you're getting ready to come into Replit, right? So you're in order to get into Replit, you're going to paste a 3-sentence description and then you're going to paste your business plan and you're going to paste your feature list, right? Okay. Yep. Thank you. And that's what gets Replit to where I now am at this, at this point. Okay, guys. All right. Very good. So yeah, anyway, does anybody's app starting to come together? Anything look good?

Anything worth sharing your screen on where you're like, wow, look what I just built?

I've built a couple, but I'm building a third. I love it. Fair enough. I personally love this stage. This is the first time in the process of building a business where it's starting to feel a little bit more real, right? You have an idea, you've talked about it, you want to build something cool, and then all of a sudden you're like, hey, I finally can see it. It's good. It's starting to be real. And that's awesome. Now there's some things here, guys, inside of Replit that, you know, work well and some that just, I don't understand why it can't get some of these things right. The design is really, really kind of ugly here. There's some really obvious design mistakes. For example, I can't see my menu and I guarantee you if I select this, I'll be able to see— oh wow, they don't even have the menu items. There were menu items before, so I'm not sure. I'm not sure what happened to it. But look at this right here. You've got this view demo button that's white on white, right? So I can't even see that unless I scroll over it. So that's kind of weird. I don't even know where the menu went. So there should be a menu somewhere there. It had a menu in the— oh, that's cool. It's got light and dark mode. It just built light and dark mode for no reason., but I guess that's what it decided to do. So, um, that's actually kind of cool. I actually kind of like it better in dark mode right now, but we're going to keep it in light mode. Now what I'm going to say is I'm going to tell it— I'm going to teach you guys a couple things really quick here. Um, so there's a little thing when you're— once you get inside of here, the next thing you're going to do, and really something you're going to do for the next, you know, 6 hours of work on this project probably, right? Might take— you might, might move faster than but you're just going to start clicking on things and pointing at things and then saying, hey, this doesn't work, or I want this to be better. So for example, right here, I'm going to click this button, right? And I'm going to say the text on this button is white on white background in light mode. I can't see it. I can't read it. Read it unless I scroll over. It. Okay, so it will fix that problem. Okay, um, let's see here. Sharon, you're already building a couple there. Are these other ones what you built in the past, Sharon? Oh, this is just right now. Okay, awesome. You got a couple things going right now. Well, very cool. It says page not found, so you might not have published it yet, but that's okay. Yeah, both of those are page not found, um, as you can see right here, but that's That's all right. We'll get, we'll get to where I'll help you be able to share it really, really well. So it fixed that right now. I can see this now. I can, I can scroll over and it does still change a little bit, but it was able to fix that. So now it's not a problem anymore. Pretty cool. OK, so now I just fixed one of my problems. If I click on Gully River, a new river, Potomac River, it's not doing anything there. There's no, there's no images here. It was cool before when there were images. You've got kind of just like an OK, you know, simple, honest pricing. I don't know. Oh, this isn't great messaging. And so I'm just going to start going through and just starting to, to, you know, upgrade it, right? I can also come down here and click on like mobile app. Yep, that doesn't go anywhere. API, that doesn't go anywhere. Contact us, that doesn't go anywhere. And so I have to just start kind of making this thing start to come to life. Okay. Um, Replit. Okay. Built user authentication with Replit Auth. I don't love any of this stuff here. All right, guys. So before we dive into, you know, doing a bunch of stuff inside of, inside of this app and inside of Replit and starting to bring it really to life, um, it's also possible there is a menu in mobile. Nope, there is no menu. They just got rid of the menu, whatever, for whatever reason. Um, which is interesting. Okay, let's move this over here. What I wanna do now, guys, is I wanna take you into a folder I have for all of you. Um, let's see here. My Drive, Build a Business in a Weekend with AI. So this folder right here, you'll, you'll get an email with this tomorrow after we're done with the course, but it has a bunch of folders and kind of guides and workflows and different types of things that you'll want to look at here. Like, for example, this financial one. I think this is pretty fun. Building GenAI API would have taken about $3.2 million to build without AI, right? Instead, I built it for a few hundred dollars. Took me a few days, right? But I've done it. I've raised money. I've raised millions of dollars. I've paid, you know, Chris and Talon and, and all of their employees and all of my employees and everybody that we had to make that happen. And it takes time, takes tons of time to build stuff out like this in the old way. 5 years ago, it would have taken $3 million to build GenAIPI. Right. And I did it for a few hundred bucks. Right. That's a, that's a, that's an improvement in business process of 99.99%. Like, it's crazy how much of a difference that makes to our world. Okay. And so there's a few other things here in this, in this, you know, some overviews, some software and tool things. We're going to go over coding quick start guide in a minute, but we're going to take a little break for a second here, guys. I want you to all just kind of breathe for a second on, on building. And I want to share something with you. We're going to have just a little discussion here. Okay. I want to share a video. Let's see here. I got to do type here and I got to pull this up. Give me just one second. That's not the right one. Oh, I know where it is. Sorry guys, I should have had this pulled up already. I normally do and I did not, and that is my fault. Um, okay, but I got it here. Okay, so let's make this full screen, and then I need to come back and share my screen again. Okay guys, so we're gonna watch this little 2-minute clip, 2.5-minute clip. This is Eric Schmidt. He is the former CEO of Google, um, or chairman of Google, I'm sorry, and very, very smart guy, multi-billionaire, um, many times, many, many times over, and really smart guy that's really quite plugged in. Now this video, um, this little interview or this little clip is from about April 20th, so everything he's saying is already 4 months out of date. Which is like a lifetime in AI, but it's, it's still spot on.

So I want us to just listen to it and then we'll discuss for a minute. 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 one year.

Okay.

What happens in two years? Well, I've just told you about reasoning and I told you about programming and I 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 there are 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, 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. 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? Where's my Zoom? Just stop share. Whoops. On annotate. There we go. So let's just pause on our working for a second. What do you think about what Eric Schmidt just said there?

What, what kind of thoughts do you have?

Open discussion time. Part terrifying, part amazing. Yeah, that's, that's a good reaction. It is terrifying. What does it, what does it mean? Now, Corey's company is actually decently immune for a little bit longer because he's got a lot of painters, and that's a pretty physical process. AI is not replacing painters yet. It certainly is very easy for any of us to put together the machine that can do it, right, in our, in our mind, right? And smart people with robotics will be able to replace, you know, human painters and that's great in some cases at least. But for a long time, the physical trades are going to be— are going to be a lot safer. But if you're Chris—

Chris, what are your thoughts watching that? Well, I would say that if you ask any software engineer how much they love, um, you know, huddling over a computer and writing code all day, they might lie to you and say they love it, but But personally, I'd rather be outside, like enjoying the weather. So I like it. I think that there's an overestimation of how many people are going to be able to do this because it's not just about writing code. And you, you kind of said it, Jon, is it's actually about human creativity. And I think we're seeing a downward trend in human creativity because people rely almost too much on AI for everything, especially like there's some studies done on students in school and, and apparently it's, it's getting a little bit, um, out of control where their, their creativity levels are just dropping dramatically because they rely so heavily on AI for everything. So there's a little bit of a shift, and Taylan and I have— we're already seeing it where we write less and less code as new newer, more intelligent models are coming out, right? If you grab anybody off the street and you try to take my job or Taylor's job, even yourself, Jon, at this point you couldn't do it because the stuff that we're doing is so architecturally complex that it's just not doable yet. Uh, in the future it will become more doable as the models gain better context windows, high-level intelligence stuff. But the problem is you actually need a really good product manager or project manager systems architecture in order to do this stuff, because AI is not going to have all the answers. If you give it some really simple prompt, it's not going to be able to give you exactly what you want because it's not going to think through all the scenarios that you can go back and forth with it. But again, the amount of people that do that, like the people on this call. I think the people on this call are different because they're looking at not just having AI do everything, but growing with AI using their, their own kind of creative lens and, and becoming the director of their own movie.

So yeah, it's kind of an interesting, uh, trend that we're starting to see. Lots of thoughts in there, Chris, from some of what you said. One of the things that was interesting that you— that I wanted to point out is you said a lot of developers or a lot of people will become more product managers, right? Instead of having to hunker over and write raw code, right? You won't have to do that anymore. And that's really kind of what we're doing here, guys, is we're sitting here building in Replit, is we're saying, okay, here's kind of the business I want to do. Now let's try to make this look and feel like I want it to, right? I want to make this feel like something that I want to hang out in, right? And, um, That's, that's, that's an interesting shift. Alright. So what else? What else, guys? What are the thoughts you have as you, as you listen to, is AI gonna come and replace everybody?

Go ahead, Sharon. I've listened to a few other people with similar approaches and, and information and things. And when I look back when I was young growing up, which is quite some time ago, The technology and innovation gap that existed then was me being able to schedule the video recorder and my mum not being able to. And that gap has just widened and widened and widened, not just between my mum and I, but society generally. And it does blow my mind to consider how big that gap will become. And so whilst we are excited and fearful at the same time, of the potential and the tragedies of some of the AI implications and the super intelligent side of it. I also wonder, just will half of it be even able to be used if some people just can't even understand what is sometimes being proposed?

Yeah, I think that all technologies have a gap in terms of usability. One of the things that's that's different with AI, Sharon, that I think is quite a big leap forward for most of humanity is the language needed to interact with AI everyone already speaks. The language needed to interact with code like Chris and Taylor know how to do is complex. It's like learning a foreign language, right? But AI understands those languages, and it understands English and Spanish and Chinese English and Arabic and Morse code and whatever you want to speak to it in. Heck, it understands dolphin. I don't know if you saw that project where Google's working on being able to literally speak to dolphins, and they've decoded what dolphins are saying because it can process ridiculous amounts of data much, much faster than humans can, and it can find— it can identify patterns, right? And so, and so I think that the barrier is actually probably lower than you think. Think in terms of being able to communicate with these crazy powerful platforms, right? Like, people hear AI and they're just like, oh my gosh, super advanced technical, like, but it's not.

It's just, hey, fix this. That's not what I mean, Jon. I don't mean it like that. I agree with what you've just said. I mean it from the point of if some people, and I know it's growing number all the time, but we're all on a different trajectory or different journey. We've all got different, you know, so it's just that that you still have to— people still have to be able to understand what they're using and buying and the value and the all those things. And if it becomes so too advanced for them in their own understanding as a customer, then that— all I'm trying to— I'm not— I haven't got the vocabulary to explain what I'm trying to explain it yet, but it's just that it becomes so— you know, it would be like trying to take, um, the internet to the caveman. That's what I'm trying to explain, that I think could be a problem, that the caveman would just— it's, it's so far evolved for some people. I don't— I'm not saying people are cavemen still, but you understand the concept I'm trying to, to explain in terms of the video recorder scheduling and my mum not being able to do that. And then now where we potentially could become, where there's a big part of the population that just cannot understand even what's being proposed with some of the things that potentially do. Now, whether that, that exists anyway, it's just they're never normally aware of what's under the bonnet and what's happening. Yeah, I don't know, I just, I'm just aware that because I don't even consider what the potentials could be because they're so gigantic.

Yeah, Jon, I'll attempt to, because what Sharon's saying is resonating with me as well, and that I think it's tough to articulate this. It's, it's this gap between, uh, while you're saying that this technology is going to available to the masses. It's this inability, perhaps, for the masses to be able to be of the intellect required to put in or to query the way that you have made this look so simple. But the engagement to intelligently utilize and get to the point that you're producing the results that add value. It really does take an intellect, a patience, and real work and real curiosity that I still think is limited to not the masses, but nevertheless, the tools to erase that gap and to get the masses to be able to draw value and utilize it will be interesting to see the evolution of all of this.

But I think that's what some of what companies like Apple are working on, right? Like Apple is notoriously behind You know, they're, they're terrible at AI right now. Anyone would tell you that, right? Like, Siri sucks. It's just bad, right? Their attempts at AI are terrible, and they've recognized that, and they feel like idiots right now, and they've expressed that, and they know. They're also looking right now at completely redefining it, right? I just saw that Tim Cook just, just committed like something like $160 billion to AI over the next few years in terms of like that we are going to win. At AI. And the way that they're gonna win, I think, even though no one really predicts this right now, or at least relatively few people do— I'm not— not no one, but most people have just kind of written them off. But what Apple is good at, and they've always been good at, is taking what seems like really complex technology and making it really, really easy. And I think part of what will make AI more accessible is the voice, right? It's where you get into the you know, hey, I can just talk and it does it right. Like, for example, I had, I had a magical experience the other day with my Amazon Echo, right? It's an, it's a, I have the one that's like on my wall and not a lot of people have this one. It's the, you know, you've got like the little devices and you've got your Amazon Echo devices, but, but I have one that like sits on the wall in my kitchen and they've rolled out early AI access to that and I can have a normal conversation with it.. And when it was like the Amazon, I don't know, Prime Day or whatever back in July or something last month, I was messing around with it and there were a bunch of TVs for sale and I wasn't really buying a TV, but, but it was, but they were so cheap and I was just like, hey, I'm shopping for a new TV and I'm, I'm interested in one that's about 65 inches, maybe 75 inches. Do you have any that are in the $500 price range and which one do you think is best? Right? That type of a request to like Siri or Alexa, as for most people, it just, it just dies. But it completely understood me, pulled up 3 TVs on that screen there in my kitchen while I'm doing the— I was literally doing the dishes and talking to it. Okay, pulls up these 3 TVs and is like, hey, I think, you know, here are 3, but I think you should do this one, right? And it's like, great, can you just throw that in my cart? And, or, and, and it did it, right? And so it took additional instructions, additional things. But that was the most natural experience I've ever had, ever, with AI, with shopping, with integrating it into something. And so I think that the big companies like Amazon, Apple, things, you know, Microsoft and plenty of others are going to integrate. Their job as product managers is to integrate AI in a way that doesn't feel like AI, right? It's just going to be, I'm just going to talk. And that's what I kind of meant earlier when I was saying It understands you. I could have had that same conversation with Amazon Alexa in Spanish and it would have done great and wouldn't even skip a beat, right? Because it can speak every language. It can speak our language. And so it's crazy. It's absolutely crazy what is possible when you remove those barriers. And Corey and Sharon, I'm so glad that you brought those up and articulated that well. I do agree that what we are learning today that we've spent the last 3 hours learning so far, and we've got another hour today and 4 more hours tomorrow, is ahead of the curve. And you're ahead of the curve. You're paying me $1,250 to be here so that you can be ahead of the curve, so you can then go win in business and, you know, have an advantage over your competitors. Right. Or bring things to life that you've wanted to do. I think that's part of what you're trying to accomplish here. And I think that's awesome. All, all money in this world, almost all money in this world is made because of a knowledge gap, right? I'm going to call Corey because he knows how to paint better than I do. Straight up, his blue wall looks better than my wall. It does. Okay, so if I'm going to paint, well, I can try to figure out on my own, but I'm probably gonna do better just call up Corey and be like, hey, can you come do my wall for me? Because he knows how to do that. If I want to plant a garden, I'm not going to try to do it. I'm going to talk to my wife. And if anybody wants to have the best garden in the world, talk to my wife. Hire her to come build your garden for you. If you want to get your appendix removed, you're going to go to someone who has more knowledge than you about how to safely do that. Okay? If you want legal help, you're going to go to a lawyer. If you want— and so again, we pay for this knowledge gap, right? And so when you have a knowledge gap, you have an opportunity to make money because you guys are applying AI to whatever industry you're in. You're going to have this knowledge gap and ability to go and be ahead of the curve. The changes that are coming because of AI are very, very significant. You know, we talked about painting and we talked about coding with Chris. But what if you talk about a lot of entry-level type jobs, right? Right now, the market is already feeling that in a big, big way. What about a call center, right? A call center that employs 200 people. Okay, you're gonna have a few managers in there and then you have 200 people that are just answering answering calls and doing basic tasks that an AI can easily do today. Now, right, you can replace that 200-person call center with an AI and maybe leave 5 or 6, you know, higher level, you know, hey, let's escalate it to a human type thing so that you can, you know, but for most tasks, for 90% of the tasks, and this is easy, just math to figure out, for most of those tasks, AI will be able to speak, execute those, those tasks, They'll be able to take care of things better, faster. And by the way, they can work 24/7, 365. They don't need Christmas off, right? And they cost a fraction, a tiny fraction of what that other employee costs. And whether you like it or not, the business owner is going to make that decision. They're going to switch to an AI customer service center. Right? They're going— and this is going to happen in almost every job you can imagine. Every company in the world, the estimations are by 2030 over 400 million jobs will be replaced by AI. That does not mean— I want to be clear here— does not mean that you have to go out and panic, okay? You have to adjust. That's what, what has to happen here. If you adjust you can win, you can come out on top. The place where I worry legitimately, this is why I started Gen AI PI, is I was talking to people. I live out in a farming community here. Now, there's a lot of very successful people that have their ranches out here, you know, way more successful than me financially. They're doing great. But there's a lot of people that really are just farmers and they do, they do great. They do $10 million a year in farming, but, but they're not tech people at all and whatever. And it's just kind of a different style I love them. They're amazing. I love my neighbors. I know my neighbors better here, even though we all have a minimum of 5 acres. I know my neighbors here better. I could walk into 50 people's houses, just walk in without even asking because it's awesome. I love my neighbors, but I've talked to them about AI. I was interested before I started GenAIPI. I've always just been talking about tech and using AI for a long time. And I'd be out to dinner on a double date or something, and I was talking to a construction owner here, a construction company owner. Owner here. And I was like, so how much do you— what do you use AI for in your company? He's like, I haven't tried it yet. I'm like, what? This was in like January of this year. And, you know, AI's been out for a couple years at this point, and he hadn't even— I was like, you haven't even downloaded ChatGPT? He's like, no, I, I mean, I haven't— I've never tried it. I was like, okay, get your phone out right now, download it. Okay, what's something you run into all the time that's hard and difficult for you to deal with? He's like, I got it, I got an idea. Them. You know, it's really complex calculations for, you know, how much, you know, for flow through pipes, calculating how big of a pipe I need and thickness and what types of materials and all this stuff based on these different variables. And it's like, there's no way AI can figure this out. Like, yeah, it's going to be a piece of cake for them. I said, so ask it a question that you would, that you would normally spend, you know, maybe an hour figuring out. And he's like, okay, great. And he asked it and it comes back with this perfect explanation, all these, all these equations and everything, and he just reviews it and he's like, this is exactly right. Like, that would have taken me an hour to do manually, and that just did it in 30 seconds. And I, I saw him about a month later and he said, I use AI all day, every day. I ask it 40 questions a day now, like, on every job site. I've showed my managers how to use it. I'm showing these people how to use it. It's crazy, right? So obviously, guys, there's ways to use AI at every, at every level level, from the tech to the plumbing, right? Um, and the people that are learning and becoming proficient with AI are going to be the most resilient to this new future. And it, it's a scary world. It's a scary world, guys. Um, where, you know, right, right now I can tell you right now that I know for a fact Burger King, McDonald's, Taco Bell, all of these fast food restaurants They are testing one-man stores, right? And it's going to be done through, um, robotics and AI, right? You apply robotics, the robotics are probably going to get it right, right? They're going to place the order, you're going to do this, it's going to come out, it's going to be perfect, right? And you just have one person there that just kind of feeds the machines, okay? This is not the first time this has happened, by the way, guys, right? This happened when we went you know, when the Industrial Revolution happened, when electricity became available, when the internet came out, you know, there's been big jumps and leaps forward. This is the biggest. This is the first one ever in the history of mankind where the new technology could think for itself. That's a big deal when it can reason for itself. Now it has to be pointed towards something by a human. But when you start to get into what Eric Schmidt was talking about there, recursive self-improvement, where at that point, 10 to 20% of the code— I know for a fact now that it's now 30 to 40% of the code is being written at Google by AI. And that's only going to continue to improve, right? It's going to get better and better and better, and it's going to make itself better. It's going to learn how to make itself better and better and better and better. And obviously, you know, there's plenty of movies where you can go explore the, you know, the bad sides of that. You know, Arnold Schwarzenegger acted in a couple of them. And amongst others, right? So you can go and enjoy those, those fun scenarios if you want, the doomsday scenarios. The one thing I still do believe in, guys, is the power of human creativity and the power of human adaptation. I really do believe that no matter how powerful AI gets, humans are going to figure out how to adapt and use it. And I think that society is going to progress at a very dramatic rate. We're going to cure cancer. We're going to have phones or phone batteries that last for longer than a day. We're going to have, you know, we're going to, we're going to have warp engines figured out. We'll be able to travel to Mars in, in a, in a 2 hours, whatever it is, right? Or a day or whatever it takes. Right. But we're going to have improvements throughout every industry you can imagine that are going to make your life better. The thing that we haven't figured out yet, the thing that we haven't figured out yet, and it scares a lot of people and I understand why. Is what is the transition going to be like, right? How are we going to get— like, if you look ahead 20 years, right, we've all changed. There's new jobs we've never even thought of. We're in space now. We're doing all these crazy things and all this stuff happening just 20 years from now. But what is going to happen? What is going to happen when in 2 years from now 100 million people lose their jobs that year because AI and robotics have been installed now in those McDonald's and in those call centers and in those bank teller positions and whatever it is, all these, all these positions that are going to be done. What's really another thing that's really interesting that wasn't predicted. If you were to go back to say the year 2000 and say, hey, you know, AI is going to come out. What kind of jobs do you think it's going to replace first? What are going to be automated first? And a lot of people said, oh, it's going to be the blue collar, all the repetitive stuff, right? And we're going to have machines, you know, doing mining, and we're going to have machines doing plumbing, we're going to have machines building roads and all this stuff. Nobody, nobody predicted that the artists were going to go first. No one predicted that. That is a crazy outcome that no one really predicted. But now AI can write music a million times faster than I can, right? I've got all these neat tools. I've been doing it my whole life. I am an expert professional, top 0.1% in the world musician, and AI can do it better than me, faster than me. Now, better is obviously, is obviously a subjective, you know, I think my music's awesome, right? And I love it, and the people that listen to my music love it, and the people that play and buy my piano books and all that stuff, that's awesome. But, um, but the reality is AI can put out a million new songs a day on Spotify, and they do. It's about what's happening right now. There's about a million new songs that go out on Spotify every day created by AI, right? And, you know, I see Stephanie like, ugh, like, it's not even attractive. You're not even excited about that, are you, right? Like, I'm not excited about it. I don't want to listen to AI stuff. I want to listen to the Beatles I want to listen to the next Beatles. I want to listen to Benson Boone or whoever it is you like, right? I want to listen to people that I can really connect with and relate to and watch and see on stage and, and, and all that stuff. And but, but as a musician professionally, how the heck am I supposed to compete? I struggle to write 20 songs a year, right? That's hard work. That's like 1 or 2 albums a year. I'm barely keeping up with that some years. And some years I don't release an album because I just didn't have time, right? I have 10 or 15 albums out now and that's great, but that's over a lifetime. And so graphic artists, user, you know, user experience designers, website developers, you know, all of these things we've shown now how we can, you know, I didn't have to hire Chris or Talon. To build GenAIPI. Now, is it a super high-tech something? No, but whatever. I'm making money with it. I'm making hundreds of thousands of dollars and millions of dollars with it. I've turned it into a real company and I didn't have to go employ developers to do it and everybody else in between. So guys, my goal is hopefully not to scare you. I want you to simply be aware. Being aware of these dynamics that are going to be present, they're starting to show, they're there, there are signs now, but it's going to get bigger. And if you're aware of it and if you know how to use it and if you're already way ahead of everybody else, you're going to be great. You're going to do great. Okay. So that's all. That's all. I just— I like having this discussion and listening to Eric Schmidt talk about that. And again, if we were to have him, you know, jump on this call right now, 4 months later, he might actually adjust down his, his estimations on when we have AGI and ASI. He said 2 to 3 years and 6 years for AGI and ASI. Artificial general intelligence, where one AI is smarter than every human on the earth on everything, right? Or than any single human on the earth. And then ASI is smarter than the collective of all human brainpower in 6 years. I think it comes faster than that. What happens when that happens? We don't know. I think about it and talk about it every day. I don't know. It's very hard to predict. But I do. I do know that we are meant to create and we are meant to— and we're capable of adapting in just incredible ways. So I'm excited about the future. I really am excited about the future. And I'm trying to do my very best to help others be ready for it as much as possible. Okay, so guys, we've got about 40 minutes left. We're going to dive in back into Replit here. And I want you guys, because what you're going to do over the next maybe, I mean, some of you work on this today, you're going to want to dive in and make this better. So first of all, I'm not a huge fan of just the design of this. And so I might, I might do that. You saw how I did a couple of things, but inside of, inside of Replit here, there's some things that are important that I want to explain right now. And some of that is inside of this coding document. I'm going to go over this coding being kind of boot camp, I'm going to take about 4 minutes to do it. Okay. So this is a long document you can look at later and you don't have to read all of it, but I want you to just have a couple of words that you just have at the tip of your tongue. Front end versus back end. Okay. Front end is what we can see. Okay. It's the visible part of your application, right? So it's everything we click on, everything we type into, the formatting and the colors, the buttons, all that stuff. Okay. So as an example of that might be a login form. Where users enter their email and their password, right? So on GenAI API, right? That's all this stuff, right? You've got all this. Oh, why did it do that? Oh, it must have been a delayed sign out. So anyway, I've got my username, my password. This is all the front end elements. When I click on any of these things or I type into this stuff and then click submit, it's then going to talk to the back end, right? And it's going to say, okay, let's check and see if this person should be let in. So the back end is the behind behind-the-scenes logic that processes data, handles requests from the front end. Okay, so you've got all this data living there behind the scenes that makes everything possible, and then you have what you see and what is displayed. Then there's then a bunch of rules that determine how all those things talk to each other and work. Okay, so the backend, it's processing data, handling authentication, login, security, communicating with databases where all the data is stored, right? So for example, checking a user's login credentials against against the database and then returning an authentication token, right? Hey, is this a valid username and password? Yes, great, let them in. No, hey, display a message that says incorrect username and password. Simple. That's how all things work, right? You've got a frontend and a backend, and, and that's all you need to know, okay? Databases are where— and we don't need to go into like the types of databases for today. We're going to be using the SQL database, um, inside of Replit that is automatically going to be set up for you. But a database is where data is stored. Okay. So when it goes to check that username and password, it's going to check the users table. It's going to say, does Jon Cheney exist with this password next to it in the users table? Right. So there's— it looks like an Excel document, right? That's what the— that's what a table, a database table is going to look like. You're going to have all these, all these tables, rows and columns. Columns of here's information and, and it's all organized by type. Okay. And you'll have lots of different tables to talk to each other. That's really all you need to know. Then you have the ability to query that data, right? So you can ask it questions. You can say, how many users do I have? How often does that user log in? How often does this happen? Now, in order to, to, to know how often that user logs in, well, you have to have data on that. You have to store something in your database that says, Hey, every time you log in, put a timestamp and save it on user logins or something like that, right? So you'd have to have some sort of data, some sort of measurement inside to do that. So as you're thinking through building out, you know, maybe most of what you're building is going to be pretty simple, but as you start to think about more complex relationships, you can start to build some pretty cool stuff. And again, AI can take care of all this for you, but if you just understand how some of these things talk to each other and work together, you you can start to let your brain put the puzzle pieces together. Okay, APIs, Application Programming Interface. Okay, this— an API in simple terms is a connector, right? It connects one thing to another. Okay, so for example, back to this, connecting this sign-in button to a specific table, there's a, there's a little API call. Something that says, hey, is this happening? There's a little connector between those two things. Now you can also connect APIs outside of your own app. For my app that I'm building here with my river thing, well, I can go and set up little sensors by rivers so that it's measuring the level of rivers as I go through. Or I can say, hey, I wonder if someone else has already done this, right? And yeah, they have. I can go to the USGS, US Geological Survey, and yeah, let's say river data. Right. And here's the current water data for the nation. Okay. So someone else has gone in there. The U.S. government has all of these different stream flow conditions and they also happen to have an API that I can plug into so I can pull from that data and power my app. Right. I don't even have to go out and spend any money on that one. A lot of APIs do cost money, right? So if I go to like weatherapi.com, I am going to integrate Weather API in a bit here in my app. And this allows me to have access to all kinds of weather data. I can know what the temperature is and what, you know, what's coming, you know, down the pipeline. What's the forecast? Now, here's the cool thing. Here's my pricing. Most APIs have a free tier, right? This one I can call up to 1 million calls per month for free. I'm at $0 if I'm building a weather app, right? And I can afford $0 a month. Once I'm at 3 million calls per month, I can probably afford $7, right? Because I'm gonna have users paying me for my weather app, or I'm going to have enough users that I, that, that I have, you know, enough traffic that I can actually generate revenue with ads, right? So APIs, guys, allow you to connect to other people's data and pull it into your app, and then it's up to you. You're the creative one, right? That's what we've been talking about a lot. You're the creative one that gets to pull it all together and build a unique combination of data and connections and things that then turn this into a really awesome application and usable system. Okay, we don't need to talk about the different types of APIs. An API endpoint is a good thing to just know about. So for example, maybe you'll set up Stripe later in your app because you want to set up some sort of payment portal. Stripe is a great option. There's other options out there, but Stripe is a very, very popular one, and it's going to have an API, and it's going to have specific kind of, kind of like a web address, right? Where you're saying, hey, I'm going to point at this and ask it a question. Right? And so these endpoints is where you're, you're getting to those APIs, it's the access point, okay? Um, another couple, another, a few words that are very helpful to understand in, um, in APIs. Basically the entire internet works with 4 words: GET, POST, PUT, DELETE. That is all that happens. How many of you guys have ever posted something to social media. Yeah, that's actually a technical term, right? Sorry, I had to sneeze. Um, so that's a technical term. Um, you, you post something to social media, and what that does is you add it, you add a post to the database with a reference to an image that's stored somewhere, right? And then when I click on Denise's Facebook profile, then it's going to get, or it's going to retrieve all of her posts, and then it's going to organize it according to the instructions of how that website is constructed and built. Right? That's how it works. You're just getting data or you're posting data. Maybe you're updating it. Maybe you made a comment, you're like, maybe I shouldn't be that harsh, and you edit it and you replace it, whatever. Okay. Or you anyway, so, or maybe you delete something, you say, you know what, maybe I shouldn't post that and you just delete it. Okay. But everything on the internet, every news article, Every, you know, the weather being displayed, it's getting, posting, putting, deleting. It's doing something there. It's updating something. It's getting something from an API, from a database somewhere. Everything on the internet operates on these four principles. Okay. You can also use this CRUD functionality. This is kind of related to, you know, if you're adding, maybe you create an admin system, right? So let's see if I have, I have an admin system here in GenAIPI. Right? Maybe I want the ability, let's go into course, no, live courses and let's go to reviews, right? And I'm gonna manage reviews. So I have CRUD functionality here. I can create, I can add a review, right? I can edit a review, I can delete a review and I can obviously read the review, right? So you could say, hey, you know what? I wanna add a new admin section in my system that has CRUD functionality for users, right? I can add a new user, I can see the users that exist, I can edit or update them, or I can delete them. Right? So CRUD functionality is a great way to just say, hey, I need an admin system for users with full CRUD functionality. Okay, don't have to remember all this, guys. You will have access to this document, but it's just good to hear it and just kind of know that it's there. Now, authentication, this is the last thing we'll do, I think. Yeah, I think that's it. That's all we need to do here. Yeah, we'll talk about that other stuff later. But if you need people to be able to log in, so I need people to be able to log into my River app because they have to be able to save favorites, right? They need to be able to have, hey, I like the American River and I like the New River and I like this river and whatever. I need to be able to save that. So I need people to be able to log in and then there's going to be a database that's tied to logins. I personally in Replit like using what's called JWTs. JWT, okay, instead of what they call Replit Auth, okay. Replit has built their own auth system that then allows you to use Google to log in, Facebook to log in, whatever. And so I'm actually just going to say, um, yeah, please completely replace the Replit Auth system with a standard JWT system. Let's see here, database. I'm just going to look on something. I do have a database. Okay. Okay. So I like JWT. I'm not going to go into why, but I just don't like Replit's system. The main reason I don't like Replit's system is because it doesn't allow me to edit it a lot inside of here. And so I'm just going to have it rebuild that. That will take a few minutes for it to do that. Okay. So that's just kind of the basics. Quick, quick guide. Hosting deployment. We'll talk about hosting and deployment tomorrow. We don't need to dive into that right now, but you will have access to this full guys, there's a lot more stuff in here. We'll go over security, we'll go over some other things like that. It'll just kind of at a high level allow you to see this. Now, I want you to understand something, guys. In order to build GenAIPI, I already told you the story. This took me about 35 hours to build. Okay. So we obviously can't do it in a 9-hour course or an 8-hour course or a 12-hour course, right? It's going to— it takes time, right? What I'm hoping to do right now is help you get started and then show you how to work on it. And then you're just going to kind of repeat this process over and over some of these things to just refine and make it better. Right. So for example, right here I can click View Demo.

Okay.

Well, first of all, do I even like this? It came up with this design. I also am not a huge fan of the design. Okay. And so there's something that we need to do now that everyone needs to do in their Replit project. Okay. As long as you already have your first Replit piece pulled together, then, then we're ready to do this. So over here on the left, there's a little place that says Open Files, right? And this allows you to see all the different files that were created in order to, you know, put this together. By the way, setting up just your basic file system when you're putting a new website, a custom website together, is already a big task, and it's awesome that AI does it for you. Okay, so, um, yeah, Sharon, you said you've done this with the database, found the code, just went in circles. Yeah, you have to be— you to be careful if you're doing, you know, in order to— there, there are huge amounts of tips and tricks in order to get this done right. Don't use Replit's auth, um, just use your own database. That's the only thing that I don't like that they've done. They've done so many other amazing things. But right here, guys, in your files, in your attached assets, you have two files here, right? You've got your business plan and your feature list. Okay, so what we're going to do is we're going to take a prompt from my Replit Tips and tricks. I'm just going to copy this and paste it in here and I'll read it to you and explain what we're doing. So I'll say, please use the attached business plan in attached assets. By the way, I'm going to drop this into the chat for you. Okay. And the attached feature list in the same folder. Create a README.md file with the business plan and create a progress.md file with the feature list. Don't write any code. It. Just create these documents on the feature list. Create a checkbox system that shows us what we have completed and what we still need to create. Please examine the app and make sure the progress.md file is up to date and add any items that might be missing from the list that we've created. Also add references to these files and their purposes to the replit.md file. Okay, now there's a lot of stuff in there you might not understand. Bottom line is, guys, an MD file is just a document. It's called a markdown document. Okay, very simple. But what we're doing is we're taking these guidelines, these business plans, the why, the why behind everything that we've done, and we're, we're handing it over to Replit in a way that it can then use it and reference it ongoing, right, as it's learning what we're, what we're trying to build. Okay, so we're going to have it just write two extra. Right now these documents are just kind of stuck in attached assets, and attached assets is not not part of your app. Okay? We need to get it into the app, into documentation format. Okay? So I'm just gonna go ahead and click go. Right now for me, I'm adding to queue because it's working on replacing the JWT. Okay? But what I need to do is I need to get— I need to take control of Replit again. Right now it's just building stuff. Right? Without, without knowing what I really want it to do. Even though I gave it that initial thing, it, it, it's forgotten already. Right? It just forgets that. You have to remind it constantly. And so in a minute here, after it's built those plans in, I'll be able to say, okay, I want you to redesign the app, the visual side of the app, and I want you to follow the guidelines from my README.md file, right? I want you to follow my guidelines for my business plan. This is what we're trying to accomplish. Here's how we're trying to do that, right? I want you to really follow what I said, and it's going to go through and it's going to make some adjustments. It won't change everything. It'll reuse what it's already built in some ways, but it will change some of that stuff. And then I need to give it some instructions, obviously, on design and see if we can get it to look prettier. Um, but, but anyway, guys, go ahead and do that. Go ahead and send your attached business plan. And then when it's done, you should be able to click on it and look at the MD file. This is something that Replit did not used to do, and it's something that I actually taught the Replit team that is a critical element to why I find success in building with Replit. I've been working with the Replit team now for months, and they've been telling— they've been basically using me, um, as a free resource to help them make their system better because I've been able to build so many good things and teach hundreds and hundreds of people how to use their system in ways that are actually turning into businesses that are making money. A lot of people get into Replit and they finish their app 40%, 80%. They don't get it all the way to where it's like, okay, hey, I'm actually making money with this now., and that's not very helpful. You want to get it all the way through to where you're actually making money. Okay, so I just have to wait on my agent here for a minute.

Any questions about any of the coding stuff we went over or this, these MD files and what they're for? Go ahead, Sharon. I didn't catch how you got the files. I understood the process, obviously, on the business plan, and I've done all these things just not in the right order. But how did you get them into the files? Did it automatically just add them to the files when you've uploaded it?

Or because I just normally copy and paste and things. Yeah. When, so when we, I didn't upload them, I copied and remember these were pasted content, right? Yeah. Okay. I just copied and pasted them in on that initial prompt when it was like, what do you want to build? And we wrote 3 sentences, pasted that. And so it's just still sitting here, right?

Okay.

So I just know it's sitting there, right? And so if you haven't done that, you can paste it again, right? You could just say, hey, use the attached files, right?

And paste them, right? And build that, right? So you just— So could you, so could you create a sequence of documents separately and then upload it and then refer it to that and basically say to always refer to that?

Because that was some of the advice I was given from Replit, I think. Yeah, absolutely. And like I said, Replit is learning that from feedback I've given. I'm sure other people have done that as well, but they've added this replit.md document and they've taught their AI, their prompt, to refer to the replit.md file so that it knows what's going on. And so that's why I added, I actually modified my prompt a little bit to say this right here. Also add references to these files and their purposes to the replit.md file. Right, because they've added that in. But then when I have these guidelines in here, these— I like to think of them in my mind as bumper lanes, right? You're bowling, and instead of going into the gutter, going off some crazy direction, I've got these bumper lanes. And so if I tell it, hey, work on this thing, be sure to reference the businessplan.md file or the README.md file to understand why, and then it will read it and review it and be like, oh, I understand why I'm doing this now, and then it will build in the right way.

Okay. So basically always give it the context and the parameters that it's got to work within unless you explicitly are changing those. Could you just remind me on the API side of things because that can get very complex, particularly if you don't really know what you're doing. I understand the concept of API and things like that. I've worked with it with developers, but I'm just a bit concerned about how I would make sure that that would be reliable and what the best process is to make sure that I was able to access. So I've already done Stripe connections. I've done, you know, follow the secret keys and all that that sort of stuff. So I follow instructions quite well when they're simple. But the API piece, I think I just don't know if that's a reliable thing to bring into a solution when it's somebody else's API.

And then they could change that rule or that flexibility or that cost like that. And then— Yeah, they can. And that's part of that. I mean, that's a reality, right? Because again, I'm not gonna go out and and put those sensors in those rivers. Not going to happen, right? And so I have to rely on other people's APIs for my weather, for my, for my river data, for my— I'm not— I'm also not going to go build a credit card processor right now. The cool thing is APIs are swappable. If you don't like Stripe, you can switch to another one. If Stripe goes out of business, if they change their prices, easy fix. You're just going to swap some API keys, tell Replit to update it to the new documentation, whatever. Now, yes, you do have the risk of when you're building your own solution, your own software, you absolutely have the risk of someone else updating their API and you go and you're like, hey, my thing's not working, right? And then you go in and you figure and you're like, hey, my thing's not working. Oh, hey, something was updated with the API, right? And so then you have to go in and update that. And that's part of the responsibility you take on when you're responsible for development, right? And we're going to talk about this concept a lot more tomorrow, guys. We're going to talk about it. It'll be the first thing we talk about tomorrow, right at 9 AM my time. And it'll be a discussion around something called surf versus swim. And it's really understanding how to work with other people's technology and how to get ahead in a market like with AI moving so quickly and making really smart decisions around that.

So All right, let's, let's keep going here. So, oh yeah, go ahead, Stephanie. I don't know if it's just me. I mean, I've been following along until we really got to putting the business plan in Replit, and then I'm seeing what you're saying.

I'm watching. It just feels really over my head. Yeah. Where are you? What, at what step?

Like, where are you stuck, I guess? I don't know, I kind of got lost with the files that you were adding, and maybe it's because I'm literally not mirroring you because I still don't have my business plan.

I want to take my time and chat. Yeah, but I'm just— so what I recommend, Stephanie, and thank you for saying that, what I recommend is going back and watching this again. Just take it back to that spot, right? 2 hours in or wherever it is that we started doing that. And and just, you know, watch the step and pause it. But really, all you have to do, Stephanie, it's so, so simple. My first prompt, my very first prompt is right here. Okay, that's where you— that's right where, you know, you go to the, you know, create a new app and, and hey, what do you want to build today? You give it your 3-sentence description, right, that ChatGPT can help you write. You copy in a business plan, you copy in your feature list, and you type, be sure to closely adhere to these guidelines and plans attached. And go. And now all of a sudden you have an app that works, okay? Or at least first version, right? And then from there, it's all about tweaking and tweaking and tweaking and tweaking and tweaking, right? And so we're gonna talk about that now. So, okay, the app is not running. Get it working. Okay. So at this point, now that we've given it— now that we have— and by the way, look at this. You've got your README.md file. This is my executive summary that ChatGPT wrote for me. This is what we what we uploaded, Stephanie, right? That's this straight out of ChatGPT, right? Just word for word copied and pasted, and then it threw it into a document for us, right? So that it can store it over here. Then I've got my progress one, right? And the progress one's cool, right? It's got all these little check marks here. We built this, we built this, we haven't built any of these things yet. We built a few of these things, right? So super cool. We know what we've built, what we haven't built, and more importantly, Replit knows what we built or what we haven't built. Okay, so really, guys, this is quite simple. And I know that sometimes you move— we move a little bit fast in this, which is why I wish we had a couple more hours, because I would give you 20 minutes to do something instead of 10. Or, you know, sometimes that helps. So we've got it working again. But at this point, you become, as Chris was talking about earlier, you become more of a project or a product manager, right? All you're going to do now is say, all right, I really want this to feel and look better. So I'm just going to give it some instructions and I'm just going to say, I really am not a fan of the app design. Can you please refer to the README.md file and take a new and, I don't know, better approach to the design? Based on what you see there. We are trying to build a category-defining app here. Okay, so I'm just talking to it in the way that I would talk to anybody. Now there's a few tricks and things that I'm doing here. I'm having it refer to the README file, asking it to— a couple things that you can do here related to design. Is you can also, instead of saying, hey, category-defining app, things like that, giving it some big words, so it's like really tries to go for it and do cool things. You can also tell it, hey, I really love apple.com, or I really love this website or whatever, and it will go to apple.com, look at what it's done, and it can just apply all of that to, to what's happening here in, in Replit. So it's really very capable of doing so many things. So it's going to work on this. It's going to build it for Paddleworks. More grit, authenticity, utility, focus. All right. So it now knows what it, what it needs to do here. I actually want to go into the replit.md file and just see what it does here. Yeah. So it added into the replit.md, it says here's the business plan and strategy, here's the development progress tracking, and then the technical architecture, the Replit one. It's pretty cool. Any time, guys, so you can inside of the agent, let's talk about a couple other things while this is working. You can click new chat. You can't create a new chat while one is in progress right now. Okay. But if I click new chat, it comes up with a brand new chat. It will have forgotten everything that I've ever said to it. It does not remember. It doesn't copy the context from the old chat into the new chat. Okay. As long as I'm in this, this single chat, it remembers everything that I've done here. So you can also see how that might be a problem at some point, right? If we have hundreds of pages of text I've talked to it about, first of all, your costs are going to go up every time you're talking to it and asking it to do something. It's costing money. So far with this app, I've spent $3.78. So that's pretty cheap. Not bad at all. But if I keep talking to it in this, it's going to consider everything that I've said to it every single time I ask it to do something., and that's really inefficient. So you want to be able to do new chats. But if you're not, if you don't have something like this progress file to keep track of what's going on, or this README to tell it why we're doing this, and this technical architecture one that it's already figured out, then it's going to be lost, right? Every time you click new chat, it's like hiring a new developer. The new developer is capable of getting up to speed, but it's going to take you a minute to figure it out. The AI can instantly up to speed if you keep these files updated and you keep it working off of these files constantly. So even when I'm done with something, I'll say, great, please update the progress.md file to reflect what we've done. Awesome. And then we'll go in and we'll add a checkmark or add some notes at the bottom, whatever it does. Okay, so it's working through its problems now and it might take some of these, some of these steps. And this is part of why it's difficult in a 3-hour class or 4-hour class to show you all of these things is you have to sit, you gotta wait, you gotta wait for it to do it, right? It's going to take 10 minutes, 5 minutes, 3 minutes, 15 minutes if you give it something really, really complex to do. Okay, um, another thing I'll show you right now is I'll show you how this, uh, how the system works. Okay, um, in terms of pricing, so you all signed up for maybe a monthly plan, you paid $25 or $30 or $35, whatever it is. Um, so I've got my monthly credits, I've already used all my credits, it resets in 7 days. I use my credits pretty dang fast. I also have a ton of credits. I have 2,250 credits just chilling here because Replit's awesome, um, and I'm working, I'm working closely with them, so, and I teach a lot with it, so they basically gave me credits to be able to teach with. You have all this crazy stuff. You don't really need to learn a lot about it, but what I want you to understand is that every time you use the agent, okay, my GenAI API, this is my main thing. This is— I spent $64 this month working on stuff. Worked on another project called AI Advantage, which I'll tell you guys a little bit about tomorrow. Anyway, other systems in here, right? There's all these cool systems that have cost— have cost me money working on, right? And it's not a lot, right? Advantage, $30, $32, no big deal. Then you can look at all the other systems that are down there. You've got auto-scale deployments and storage, PostgreSQL storage, all this stuff. There's a cost to it, right? But it's very, very low. In 3 weeks, I've spent $2 on storage here. Here on compute and on all of these very active projects, right? And I'm at $22, right? The amount of money I'm spending here, guys— $138, 3 weeks into my 4-week, you know, stint. That's insane, okay? For me to pay one developer, right, I'm $20 grand a month for a senior developer. I'm $10 grand a month for a junior developer, right? I'm spending $138 to run a dozen apps. That's crazy, right? And by the way, that's covering— those are my AWS fees and all my things. Like, it's all inside of here. Replit does it all inside of here. And so, um, but anyway, understanding that, um, you know, there are different costs, but it all comes— it all comes down to this spot right here. You can click on this right here, this little— this little, like, um, what do you call that, speedometer-looking thing. Okay, so you can also look, um, at how much, uh, specific ask took, right? See, look, by the way, look how much better this looks all of a sudden. Like, I'm way, way more excited about it. Like, it even adds swears in there because it thinks that kayakers like to swear. So for whatever, that's great. Um, all right, never miss your— never miss your perfect river day. Um, but anyway, this one took 4 minutes, 10 actions, 245 lines of code, changed all this, cost me $0.82 to redesign my entire website. Pretty cool, pretty cool, guys. I think that's awesome. Okay, so from this point now, I'm just going to start, um, I'm just going to start clicking on things, right? So if I click on View Live Data, that takes me to sign up. I don't want it to do that, right? I want it to actually, like, Get Started, Join HighWater. Everything is telling me to get started, right? And so I don't know if I love that. Um, I don't want users to have to log in to view river data. I want them to have to log in to save their favorite rivers, set alerts, and do everything else that would require an account for. Um, let's redesign the flow a bit. And the buttons on, on the main page to reflect this. Okay, so I'm just acting like a product owner now, right? I'm just telling it, here's what I want it to do. Here's the user experience that I'm wanting to have. I'm not using anything fancy, right? Stephanie, that's— I want that to be so clear here. I'm not a technical person. I've just— I have a little experience. Right? That's all. I just have a little experience in talking to it, and so I kind of understand how the pieces fit together a little bit. But the type of, of words, the prompts that I'm giving is no different than what anybody else could do. Anybody could sit here and, you know, if they understood my vision, say, okay, well, I'm clicking these buttons and they're not really going to where I expect it to go. First of all, view live data, and it just takes me— I don't want to have to log in. I want to be able to just get to my River data right away. Way, right? That's easy to explain, right? That's easy to explain in plain English in a couple sentences there, right? If I want to start free, it's going to do that. Start Pro Trial, it's all, it's all taking me just to that same spot right now. So I need to just kind of work on my flow. What do I want that to be? Sometimes it's good to go get your whiteboard, right, and write it out and be like, okay, I want people to go from here to here, and in order to get there, you know, sometimes it's nice to write it out. I'm pretty good at just kind of visualizing it in my head, But anyway, I want them to browse rivers. Now all of a sudden we've got a new spot, /rivers. Now it hasn't built the rivers thing yet. Great. Now let's build the rivers page, /rivers, which is— I saw that, by the way, guys, I saw that it was right here. I clicked on that, /rivers. That's how I know that that existed. Okay, so now it knows what it needs to do. It's wrapping up its current step still, and now it's going to start working on it. Okay. And now it's going to build out this next thing. So now we've added something here. We've got to sign in or create account, which I probably will condense those into just one button eventually. Browse rivers. But again, now I'm just working on details and I'm going to do this until everything just works as I expect it to. When I wanted to launch GenAI API in a weekend, or in a day. It was my first goal. I just built out the first thing and within an hour I had something that was honestly pretty dang close to what you see on GenAI today. It was pretty dang close to this right away. It didn't have all of these, you know, I added some more, you know, things, but, but it had kind of this, this kind of idea with kind of an example question here. I haven't changed this from day one. Okay. And then I've changed some of this content. I've updated it to be a little bit better and and it's a little, it's a little stronger, you know, more, more focused on enterprise and some different images that I've pulled in. And I'll show you guys how to do some of that stuff later. But, but the bottom line is it was about like this. But then I had to go through and I had to click every single button, right? It took me a couple of hours just working on the predictor test, right? I had to click on this. And at first I didn't have this kind of, hey, this is what you're going to do. These are the types of questions you're going to get. I didn't have this little little thing. And then I put it in front of some customers and they actually started using it. And people said, hey, I was confused when these types of questions came up. Why did that come up? And I said, hey, let me explain. And ever since I've added this little modal here, people aren't confused anymore, right, as to why this type of question was in an AI test, right? And so I can now click through these things and they work. The first time I clicked submit, it didn't work. Sometimes it would automatically select a question for me. Hey, or an answer for me. Hey, will you please not select a question or an answer automatically when I do this? Yeah, let me fix that. All right, so you just got to, you just got to work, right? You just got to, you just got to work through it as a user, right? And eventually you'll get to the point where you say, hey, this actually feels kind of nice, right? Hey, Riverflow data, perfect. So now we have no data, we'll have to pull in from APIs. We're going to do APIs tomorrow, guys, okay? We'll do that and we'll really focus on a lot of of fun APIs. We'll do the— we'll use the weather one, we use the river one. I'll show you how to do some emails and stuff like that, SendGrid, other stuff like that. But, um, but really your job is going to become, hey, I need this user experience to feel nice. If I click on this little thing here, does anything happen? Does this— does this work? Does this actually filter things? There's no rivers here, obviously. I can go into my database. One thing that I will tell you guys you want to make sure that you have a database in your, in your project. If you do not have a database, you can get there by adding— might you click this little new tab button and you just click on database. Okay, database is already open, right? But, and it will, it will say, you know, would you like to add data? But you can also just say, please, you know, please add a database. Okay, you need a database if you're going to have users, you need a database if you're going to store favorites, or preferences or paint color data, whatever it is, you got to have a database so that there's a place for it to go and refer to. And it's as simple as please set up a database for me. Okay. All right. And by the way, we'll also take a second to just show this. When you click on the database, looking at the different types of information that you have in there is actually really important. Like I can look at rivers right now. I can see what type of information is available. So we've got an ID that's going to be just like, hey, River 1, River 2, River 3. That's just kind of an internal ID. The name might be American River, location California or state California, right? Location San Francisco, wherever it is, right? Country United States, USGS gauge ID. Okay, so it's already contemplating we're going to be looking to the USGS gauge data, so it will fill that in. Then it will figure, okay, what are the— what's the difficulty, right? It's probably going to be range 1 to 5 or 3 to 5. Probably rating it based on river difficulties. Maybe a description, also probably pulling it in from the USGS, a latitude, longitude, so location. So there might be— that means we have the ability maybe to put a map in there, right? Hey, let's show this on a map. This is where it is. Okay. But because I know this is in the database, maybe my mind can start to think, okay, this is the kind of data it's imagining rivers have and that can be available to me as a kayaker, right? Is active.

Right.

This means is it even— are we displaying it or is it an inactive river that we can't see yet? It's not— it's not displayed. Okay. Created at, updated at, alerts, river confirmations, river flows. There's other stuff in there. I don't really know what all these things are. These are going to be, you know, connected river flows, river ID, flow rate, stage height, timestamp, source. So I don't know if it's going to try to store the flow, like if it's going to just update it and store it. It and then just display that and then update every 15 minutes or something. I'm not sure what it's trying to think about there, or if it's just going to pull live data, we might, we might not need that. Okay. River confirmations. I don't know what a river confirmation is. Oh, river confirmations. I know what this is. This was an idea that I had that said users can confirm the flow, right? If they show up to the river and like, oh yeah, that looks right, they can just click confirm. And that adds a cool kind of social feature that's really easy to just a one button. Yep, this is, this is right. The user's on, or the river's on, guys, come on out. Okay, but just understanding the types of stuff, right? You've got email, first name, last name, profile image. I'm going to need a profile, you know, something to have a profile image on there. What plan are they on? Are they on the free plan or the paid plan? What is their home location? I need to know that for mine because I need to know that I'm 2 hours away from this river, and so then it's going to do calculations based on my home as compared to the river flows and projections that it has in there. Okay. So understanding your database, guys, is— this feels technical, but it's really not. It's just about understanding the types of data you're using to put something together. Right. And a lot of your systems might be much simpler than this. You might not need all of that type of data. But anyway, guys, this is starting to come together, right? In just a few prompts. Right. Really just a few prompts and $5 of work. I have a pretty solid website with a good outline. I've started to work through some of these things. I'm almost ready to start pulling in real data, and I haven't tried creating an account yet. That's probably my next flow. I like to do this. Jon Cheney, we'll do my Gmail. We'll give it a password here and we'll create account, see if it works. Okay, we'll save it. Hey, it worked. That's cool. I have a dashboard. I have rivers. Obviously that went into the kind of the— it kind of got me out of my, my logged in kind of state, which is kind of weird. We've got alerts. So yeah, the Rivers isn't doing— it should kind of bounce like that nice and clean where it kind of keeps this header. The Rivers doesn't have that, that dashboard internal River piece. So we got to work on that and we have to figure out, okay, Rivers is accessible when you're not logged in and when it's logged in, then it's going to have this user dashboard up here. So just you're going to have to think through that. As a user and be like, oh, that felt weird. There's nothing technical about me clicking here and saying that felt weird. That's just— I didn't like that. What's different? Oh, I see that this thing disappeared and it shouldn't disappear. All right. Anyway, you just got to click the buttons, see what happens, talk to it, tell what you want. Okay, so guys, lots of information in day one in terms of just like trying to figure out how to think through this stuff. Tomorrow we're going to go deeper into APIs, deeper into Replit. We're going to do image generation and video generation, show you how to pull those things in. We're going to show you how to use all these tools together to do more advanced stuff with agentics, more agentics systems, and then how to do some marketing and sales using AI as well, how to really structure all of these prompts and things to help you kind of put that sales and marketing engine on hyperspeed and in some ways on autopilot. Okay.

So speaking of sales and marketing there, did it— this keeps on coming up— curious about, did it fabricate that stat there at the bottom, or how did that— the 98.3 accuracy and the however many users or whatever—

was that something that just generated on its own.

Which, which one? I was on my, um, it's on your— just the River, uh, site. Oh yeah, yeah, it mentioned a bunch of, uh, data or stats, right, where you just, you just put this together.

And so there's not— yep, right. How did that get, uh, that made it up there? Yeah, I just made this up, right? What it's always— it's— well, it— what it's doing is it's trying to show you what the end product looks like. Okay. It's trying to say, okay, hey, does this feel right? Then you go back in and you'll backfill with real data. You'll say, okay, hey, let's figure out how many rivers we really can get access to. And that's just— I mean, that's like a cool stat, but it's really just going to be how many rivers are in the USGS system, right? 0.9 thousand or whatever it is. Right. Or accuracy. We'll figure that out later, I guess. Maybe. And maybe we'll just remove that or maybe we'll just remove this whole thing. Maybe this isn't important, right? 12,000 active users, 50,000 alerts sent. Maybe this stuff shouldn't be here for now. Maybe we just hide it for now and we put it back when we actually have data to show. Right now we don't. Right. And that's totally fine. I could say used by Dane Jackson and that would be just as powerful because Dane Jackson is the best guy in the world and I have his cell phone so I can text him, be like, dude, use this. And he would post about it for me. Right. And so I can do that. So yeah, all of this stuff it's, it's trying to fill in the gaps. Like this data right here is fake too. Flow, stage, all this stuff, it's all fake. But what it's trying to show you is what is this going to look like when it's complete? Then we're going to go in and we're going to backfill it with live data, right? And we'll show you how that works. It'll be connected to the database, connected to APIs, and really flowing and building a nice, a nice system here. So, so anyway, yeah, good stuff, guys. We are 3 minutes over, which I'm totally okay with. Um, tomorrow, guys, um, we might go over just a little bit tomorrow, and you can, you know, stay, stay later or just leave on time if you want. But I just am feeling like there's going to be an extra 30 minutes tomorrow, just in my mind. Um, but anyway, hopefully you guys are, uh, are not overwhelmed. Um, go back you know, if you didn't get to the point, I want you all to come tomorrow at least to the point where you've got your, your first version of something in Replit, okay? Um, where you've done the business plan, you've done that, and you've copied and pasted it in, and you just click go and you have your first version. You can start to tweak it if you want and do all that stuff. And I think it's helpful to spend even 30 minutes just trying to tweak a few things and just learning how the system works. Um, one thing that I kind of showed you guys but you probably didn't notice it because you weren't, you weren't ready for noticing it is this little button right here, this visual editor, this edit button, right? Let's go back to here at the bottom. So let's say I want to edit this, you know, Corey, I'm like, you know what, I don't really, I don't really need this. So I can actually grab any one of these individual stats, right? And change it. I can click here and I can literally edit the texts, right? And then when I do that, I have to click save changes. Right. Or I can just say, you know what, I want you to actually just delete this section, right? Delete this section. Okay. So just get rid of the Trusted by Patented section altogether and okay, fine. We're going to simplify the website.

We're going to just make it a little bit stronger and better.

So show again where you grab that. Where was that? This little edit right down here. There's a little visual, like a little target reticle, right? Targeting reticle thing. Just click that and then that allows you to select anything here. And, and, you know, tell it exactly what you mean. Because sometimes you'll be like, hey, will you delete the Trusted by Padaware section? It would probably get that because it's a pretty simple app. But if I go into GenAIPI where I have 180,000 lines of code already written, um, it might like— it does— it might not know what I'm talking about, right? So if I can click the thing and say this is what I want you to do, then it knows the specific lines of code that it needs to delete or move or edit or change adjust, right? So anyway, okay, thanks, Stephanie. Um, guys, the link for tomorrow will be the same as today. Um, so we'll just see you same spot, same time tomorrow, and we'll let you guys go.

I'll send this recording out soon, and, um, have a wonderful day.

Thanks, Sean.