Cohort 25 - Session 1 - 4-10-2026

Okay. All right, guys, officially welcome to day 1 of cohort 25. That's crazy, guys. Like, it's crazy to me that I've done this 25 times. And these courses— I said this in the email already, but I was kind of thinking about it last night— I used to do this course for 12 hours, and I found that 12 hours was— it was good, but, you know, it got a little long, right? And, and so, um, then I kind of tried, I tried going a little shorter than this and that wasn't quite enough. And so I've kind of found the happy medium at 8. But, but anyway, bottom line is I've taught this course for probably somewhere between 250 to 300 hours, like this exact kind of flow of how I do this. And that's just in these, these like specifically for, you know, this, this kind of 2-day thing. I've taught dozens and dozens of businesses this same kind of flow and workflow, and I teach this kind of workflow a lot. At schools and universities and, and all over the place. So I've, I've got a lot of experience teaching this. And, um, what's interesting, what I think is really cool, is what I teach in Cohort 25 as compared to Cohort 1 is actually really, really similar, even though AI has progressed dramatically. And I taught this first course about a year ago, I would say maybe 11 months ago, and, um, The, you know, the amount that AI has improved since then, all the tools we're going to use today, it's amazing. It is amazing how much AI has changed, but the principles that we're going to be going over today, guys, really have not. And that's really my approach here is we're going to be teaching, we're going to be going over, you know, a dozen AI platforms, maybe, maybe 15 or 16 or 17 of them. There's a, there's a lot of things we'll kind of hit some of them in more depth, of course, but, But, but I teach this concept, I teach these AI concepts around this idea of building a business in a weekend, right? And when you build a business, there's some just kind of timeless core principles that never go away, right? You gotta, you gotta, you know, build the structure of the business and you gotta come up with the idea and you gotta validate the idea and you have to build the idea, then you have to test the idea and you have to play with it and pivot it and massage it until it's done. And then you have to market and sell the idea and you have to have logos and you have to have, you know, video content and marketing. And so you have all these things that are— that, that we're going to be doing, and they're very general skills too. Even if you're not here— like, I know Tyson and Brittany, you guys are here as part of— these are, these are one of our clients, right? Um, here at, here at GenAIPi. And, um, they're here maybe to look at how they can learn some of these skills and apply them to their existing business, and hopefully some side gigs, side businesses, and side projects you guys are working on too. Because I, I really, you encourage. I really like people to have their own side projects, their passion projects, and hopefully this helps you launch those faster. But, but whether you're again doing that side project or here for a core business that already has, you know, millions of dollars, tens of millions of dollars of revenue, whatever it is, these principles should apply in just about everything you're doing. So anyway, really excited to have you guys here, and it is being recorded, so There will be times when you'll be like, whoa, Jon, you're going really fast, right? And please tell me that. Speak up and say, hey, Jon, could you just say that again? Or could you go back to that one spot? And don't wait like 10 minutes and be like, could you go back like 10 minutes ago? Just be like, hey, I really want to understand this. I had a really, really good CTO. So I, just a tiny bit of my background, I founded a company back in 2016, tech company, and I'm not a developer. I don't know how to code at all, right? Although today we're going to be doing tons of coding. It's going to be fun. But, um, I hired this amazing guy. His name is Thane Brimhall. Smartest guy still maybe I've ever hired. Uh, maybe, maybe there's one other guy that can kind of compete with him, but this guy was just a developer, super technical. And so when he would talk to me about some of the things we would do in my, in my last company, right? I, um, I guess a couple companies ago now, 3 companies ago now, I guess. Anyway, he would say something and I'd be like, I think I know what you said, but I didn't understand this word and this word and this word. And why do those fit together? What does this really mean? And he'd be like, hey, thanks for asking. Right? And he would just dig down and I'd be like, okay, now I understand this at a technical level, even though I couldn't code it, I would really understand it deeply and be like, okay, I really get that. If I didn't understand it back then, I really need to be able to understand this. I believe that understanding allows you to become more powerful, right? When you really understand how it all fits together, you can then deconstruct it on your own. You guys are all smarter than me in your own domains. I promise, I promise you guys are smarter than me in your own domains. And I hope you can take what we learn over the next 2 days together and apply it to your domain without having to be like, hey, Jon, help me do this, right? I want you to really just use your power and then hopefully these skills that we use with how to use AI to unlock almost anything very quickly and be able to just move a lot faster in whatever it is you're doing. So excited again to have you guys here and this is going to be great. So let's do a quick intro. I want to hear your— and I know we kind of know this, we kind of did it at the beginning, but let's officially— I want to hear your first, you know, who are you? Where are you from? First and last name. And then where do you live? What do you do? And then I want to know whether or not you know how to code. Just curious. I always like to know if there's any developers in the room. And that's okay if you're not. Remember, I'm not. But it's always fun to just kind of lean on some developers if we have them. So we're just going to do it in the order that I see here. And so it's going to go Marta, Eli, Travis, Tyson, Bardhi, Brittany.

So Marta, go ahead. So Marta Miner. I live in Scottsdale, Arizona. I do a lot of different things and I've done a lot of things, but real estate for the most part for about 25 years, a real estate investor. I'm also a stay-at-home mom of a 7-year-old. And recently since this year, I decided I'm gonna be a master of AI because we're heading in that direction so quickly. I'm like, okay, forget real estate that, you know, that's gonna be small fish pretty soon. So, so I'm just here to learn. I'm learning from a variety of different sources, YouTube and different courses, but this kind of caught my eye because I'm like, this, you can make progress within a very short period of time. So I'm excited to learn from you.

Very cool, Marta. I'm really excited to have you here. Eli, go ahead.

Yeah, I'm Eli. I'm a freshman at BYU, and I recently got back from a mission for my church and started an AI business when I got back. And then I connected to Jon on LinkedIn, and he's kind of been like my AI mentor. And so I've been just doing like AI marketing and consulting with like video creation. And I do not know how to code at all.

Eli is amazing. Eli is somebody I met, what, maybe 3 or 4 months ago at this point. And we've just been kind of bouncing things back and forth off each other, learning from each other for sure. And, and we're basically going to hire him now. So that's pretty cool. Eli is very smart. And, you know, he's young, but, but that young mind learns quickly. Travis, go ahead.

I'm Travis Lamborn from Salt Lake City, Utah. I am part of the GenAI PI team. Zero experience with coding. I did about 10 years of door-to-door sales. I have a side project in mind that directly supports home services companies for being able to have an AI agent sales assistant that will also couple in through an Agentic dashboard that will, through an API connection, update the CRM of choice for the home service companies to be able to essentially be the field sales assistant for the reps, as the CRMs that are available for home service companies are not sales rep friendly. So being able to have that middle ground piece that is essentially missing as a simplicity for the salespeople that they can just update their end and it pushes and updates everything to their CRM of choice. For ops and production, which is really how the CRMs are geared heavily towards. So creating that middle piece that has as little friction as possible for the sales reps to always update everything for them and also be able to schedule everything appropriately in real time. So if a rep does not able— if they're not able to take a, you know, a new marketing lead, they can mark that accordingly. If they're going to be late to that appointment, it will auto-reassign to the next available rep., based on the addresses that are populated in that, in the entire sales org's, calendar system. So obviously, uh, address dependent, um, but it'll be able to just create such a flow that, uh, is very frictionless, um, which solves about 3 of the major pain points for all home service companies.

Is the frictionless part— is one of the frictionless parts just voice updates? Yes. Yep. Yeah, that's cool. I did door-to-door way back in the day for probably about 6 years. So I, you know, not as long as you, but long enough to know those pain points. Yeah. And honestly, back when I was doing it, I was just like on a pad. I got to have like a printout from Google Maps and I would just like mark stuff and that we didn't keep. We didn't keep, we didn't have sales rabbit back then or any of that stuff. Right. Yeah.

Like just having a sales rep update a CRM is already a lot to ask from like any vertical really. So if they're doing it, that it's benefiting them from an easy-to-use tool, it then helps ops and production be able to, you know, keep everything flowing accordingly. And so, yeah, I just, I've got a really cool idea and I'm pretty confident we're going to be able to build it in the next couple of days.

That's cool. Yeah, I built a— thank you, Travis. And yeah, Travis joined our team just what, a month ago? Yeah. Yeah. January somewhere. Yeah. I mean, just kind of just barely. And he was, I found him, he was off doing some AI consulting and things like that and doing some AI speaker circuits and things and just kind of working and doing some cool stuff. And we just connected and we said, let's join forces, right? So really excited Travis is with us. He's already made an impact and he's just getting started. There's a lot of stuff he's working on that's really great for us. So excited to have him.

Tyson, go ahead.

Hi guys. Hey Jon. I'm Tyson. I'm from California. CEO of a recruiting staffing company. I'm on the opposite end of the spectrum of Eli. I got the old brain, so I've got to— in order for our company to survive and exist into the future, we gotta embrace this stuff. I think all businesses will be in jeopardy if they don't embrace it, and we have not yet. And so I gotta look at myself and say it's because I haven't. So I think this is me pushing it, and I'm excited about it. I hope to dust off the cobwebs. Of that brain and, and see if I can keep up. But I brought Brittany with me just in case. She's with our company too. So, um, yeah, we're excited, Jon. And, and look, there's so many things we can use AI to make our business better, and I've got to be driving that, and I got to get our team behind it. So, uh, this is the first step in, uh, getting me up to speed.

So it's awesome, Tyson. I hope the thing that you get out of this most is realizing how easy it really is. Like, it just feels big and technical and AI and this just crazy, right? And obviously, you know, I know you already understand some about vibe coding and things like that and you guys will, we're gonna spend a lot of time in this course vibe coding together. But, but man, when you just break it down into these, these easy steps that I'm gonna use here, you're just gonna be like, okay, I can do anything fast, right? And that's, that's exciting. And I hope that we can get everybody at your company, like everybody at the company from CEO to secretary and everyone in between. Doing this kind of stuff. So that's great. All right, um, Vardip, go ahead.

Hi, my name is Vardip. I'm from Vancouver, BC. I'm an insurance agent selling home and auto insurance. So I'm just thinking about implementing into my business, learning more about AI. Um, been thinking about for the last year or so how to implement it and have some other business ideas as well. And, um, Yeah, that's about it, right?

You said something that I want to point out for everybody. You said, I've got a few other business ideas as well. I, I, I said in the email, kind of the prep email, I said, yeah, make sure you come with an idea. If you don't have a great idea, that's okay, I can help you. We can, we can come up with an idea in, in a minute, right? That's super easy. But most people have a few ideas. Don't stress too much about which idea you pick to work on today. You're gonna do this over the next 2 days and you may never come back to it after these next 2 days, but you're gonna learn this process where you can say, okay, that was cool. Now I know how to do it. Now I wanna go to the real thing that I'm gonna work on, or man, I really wanna try this. I've had people build 3 businesses in these 2 days during the course, right? And, and like launch them, like crazy stuff. Like people that are just, you know, working now, they, they finished the 4 hours today and then they worked like 10 more hours that same day and then came back the next day and they're like, yeah, I've built 3 things anyway. So, um, it's very, very fast once you learn this process, but don't stress too much about which idea it is you're doing right now. Okay, just kind of pick something and roll with it.

Brittany, last but not least. Hello everybody, Brittany Ray. I am the Director of Sales. I work with Tyson in the staffing and recruiting industry. I've been in the business for about 20 years, so everything from tracking applicant flow on Excel to all of the automations and technology progress that we've made over the years. I'm really hoping to be a driver within our company. Obviously, from a sales perspective, there's a lot of value proposition that comes with a company that can embrace this. So really hoping that if I can understand kind of those core concepts that you've mentioned, that will start to drive some creative juices and I can partner with Tyson to to really drive this home in our business.

Very cool. Thank you, Brittany. And, and to just kind of comment on what you said there and finish up here, guys, um, there are a lot of companies— I would say most companies have looked at AI, right? Some have even signed up for ChatGPT or Copilot, right? But honestly, that is where most companies have stopped. And I would say over half, maybe 70% of companies that have stopped at that point really aren't using even those tools well. They're using ChatGPT as a glorified search engine. Copilot helps them write an email. I mean, it's very, very high-level stuff. It's not fundamentally changing the way they do processes. How many of you guys have ever been to Zion National Park? Yeah, all right, a few of you guys here. There's this one spot, and you'll remember this, there's one spot as you enter the park, you drive it in from like the St. George side, you drive in, you drive past the gate, right? And it's just beautiful, whatever. And then you have to go up these kind of windy roads, and you go up then through a big tunnel. And once you get up through that, then you're kind of in this other section that's beautiful, and you can get out and hike around and stuff. But that that process is kind of slow, right? It's a slow road and there's usually a bunch of traffic depending on what time of year you go, obviously. But usually there's traffic and so it's just kind of slow, whatever. And so a lot of people, what they're doing is they're saying, okay, I'm going to use AI to drive up this road faster, right? And they, you know, they wind up the road and they just do it a lot faster and they get through the tunnel, whatever. And they're like, yay, I did it in 4 minutes instead of 40 minutes. Okay, that's good. That's progress. You reduced your time by 90%. Amazing. But I want to get up there in 10 seconds by flying to the top instead of going up the road faster. Most people, when they're looking at new technologies, are trying to figure out how to go up the road 10 times faster, right? And they might even accomplish that. But the people that really win say, okay, I'm here, I need to get here. This is how I normally get there. What if I just go like this? What if I just take this different route? And I think it— so I want you to just— hopefully one of the main outcomes of this is you just learn how to think a little bit differently and approach things a little bit differently. And this is an entrepreneurship skill as much as it is an AI skill or anything, right? And I'm not the best entrepreneur in the world, but again, not to toot my own horn, I'm a pretty good one. I've been successful. I've gone out, I've raised tens of millions of dollars, I've sold companies, I've received big offers that I've turned down from Snapchat and other huge companies, and then gone off and figured out how to do even bigger things. And anyway, I've had a lot of experience, and it's because I try to think differently. I try to say, what is the thing that no one has thought of yet, right? And you can even use AI to help you tackle problems this way, right? Lots of times you'll ask ChatGPT something and just give you an answer and you just take the answer. You're like, oh yeah, that's pretty good, that works cool, let's move on. But I hope in some of the exercises we do over the next couple days, you start to say, okay, actually, what would be something that no one has ever tried before? I want you to take 5 minutes, ChatGPT, and I want you to research the entire internet, and I want you to see if anyone has ever done it in this way before and what would be different and why would this work, right? And push the boundaries, push the boundaries, figure out how to do things differently so that you can come up with that next Uber or Airbnb or whatever it is. And you can even come up, guys, this is— I thoroughly believe this— you can come up with that next Uber or Airbnb idea inside of something like, like Balance for Tyson and Brittany, right? You can, you can take a new idea in the staffing industry and be like, no one's ever tried this before. We should. Let's go. Right? And we can— and what I also hope you learn in this course is that failing fast is very, very productive, right? If you fail slowly, it's expensive. It's very expensive, right? So, all right, that's just about right. We usually have about 25-30 minutes for this first intro part, and we're gonna dive in. We're gonna jump right into some some coding. And I don't think any of you guys said that you're developers. And so that's, and so that's, that's fun. How many of you guys, let's ask this a different way. Raise your hand if you've vibe coded something before. Yeah. Travis, Eli, Brittany. Okay. Yeah. So about half of you have. Tyson, you have, you've never vibe coded something. All right. This will be fun. This can be, this can be great. Okay. So we're going to start today. We're going to vibe code something. We're going to do something really easy. Hopefully you guys have signed up for Groq and ChatGPT and Replit. Those are the 3 main platforms. We'll use a bunch more. But you do not need to sign up for them. But at least for this, this course, you know, hopefully you've spent that, you know, $75 or whatever it is to have a month for each of those. Um, if you don't have Grok, it's not the end of the world, but, um, but anyway, okay. We are going to, let's see here. We're going to find and figure out what's the best place to hang out here. Yeah, we'll just, we'll just jump in right here. Okay. So let me share my screen. So I want you to just watch for a minute. Okay. Sometimes we're going to do things where, you know, we want you to— I want you to, you know, work with me right now. You're going to watch and then, and then you'll work with me in a second. Okay. Most of the time that's what I like to do. So I just went to x.com. I've signed up for Grok through X Premium, which you can do. You can also do it just by signing up for Super Grok directly. And again, for this specific task, you can certainly use ChatGPT. Up here, I'm going to switch from auto to fast because what we're going to do right now is not very hard. This is going to be very low-level coding. But, and so fast, SuperGrok can do it. We don't need it to think forever and do the perfect answer or whatever. But, but here's what we're going to do. I'm going to say using p5.js, I'm going to explain all this stuff, but using p5.js, create a mind blowing visualization for me. I want to run it on or at the p5.js editor website. So give me the code that will work over there and provide the link for me. Okay. All right. I'm going to give you guys this prompt here. Chat, paste. You're going to see a lot of prompts come through the chat today. We'll send those to you later. But anyway, here's a mind-blowing p5.js visualization, the hypnotic colorful— I don't know what it is. Okay, so I can come over here, I can just click this little button right here, copy, and that copies all the code. Okay. And I'll just head over to the editor like it's telling me to, and I'm just going to paste this and I'm going to push play. Hey, boom, I just coded something super easy. Okay, that's actually kind of cool. All right, what happened? Is it interactive? It's not interactive. That's kind of cool though. It's like a, like a really, a really neat, um, screensaver from like Windows 93 or something. Um, so, so that's pretty fun. Okay, super easy, right? Super easy to go and make that happen. All right, so What do you guys think of the quality of this prompt? A lot of you guys have used AI now. Do you think this is a good prompt or a bad prompt? What would you guys say? Brittany, I see you shaking your head.

Tell me why. I wouldn't. Yeah, I've been trained to add as much detail as possible, so probably I would elaborate on that a little bit more.

Very good. Why? It's because AI, they have— it has to fill in the gaps. If you don't fill in the gaps, it will do it for you. And so all I did was say mind-blowing visualization, right? I gave it a little bit of information. p5.js is just this easy kind of, kind of low graphic, you know, front-end language. And it's cool. And this is actually kind of neat. I like it. But I wouldn't say this is my creation, right? This is some random AI creation. Okay, so in order for us to give a better prompt, we need to make sure we do a better job at maybe communicating a little bit more info, right? If I were to hire a developer and I just said, hey, I want you to use p5.js, create a mind-blowing visualization, they could do it, but it's definitely going to be the developer's creation because they're going to have to think through, okay, what would this mean? What is mind-blowing? What do you think? What does he, what does he really mean here? Right? And so anyway, AI is extremely good at instantly just kind of figuring it out and doing that. Okay? So let's try a little bit better. Let's see if we can do a good prompt. I always start with bad prompts and go to, and go to good prompts later on. Okay? So let's try one that's just a little bit better. Okay. Now you, let's see here. Create an interactive 3D model of the solar system 3D, an animated model of the solar system. Is this a good prompt or a bad prompt?

What do you guys think? It's okay. It depends on what you want to do, but you could still add considerably more detail.

I could. Now, there's actually a significant amount of data here. For example, the solar system that comprises, you know what, 9 planets and a sun and a bunch of moons maybe, and, and, you know, space and all these different rotations and all kinds of calculations that if I were to give this again to a developer pre-AI, they'd have to go do a ton of research and figure out this stuff right here. Right. How big is it relative? What's the speed? What's this? What's the rotation? What are all these orbits like? Right. And so yes, I agree with Brittany again, right? This is better. I also said interactive, which means, but I didn't tell it how it's going to be interactive, right? I did say 3D. That's pretty specific. And I did say animated. So I definitely want it, you know, doing some sort of, you know, revolving around something. So let's copy this and let's go look at it, see what it does. So again, I'm just going to select all, delete that, paste it, and I'm going to push play. Okay. So hey, we've got, we've got a 3D version of the solar system here. Let's move these videos so we can see this. Okay. It's not super interactive though. If I try to click on it, we do have a cool kind of Saturn over here. That's kind of fun, but I can't, I can't move anything. I'm trying to move or like zoom in. I can't really do that. Okay. So let's go back to X here. Okay. It says simple drag to orbit. Yeah. See, that's not working. Right. And so what I can say now is, um, the drag to orbit isn't working. Also, please make the planets different colors that more closely resemble their actual colors. Add shooting stars, add intermittent shooting stars in the background, and have the starry background slowly rotate to give more life to the scene. Okay. All right. So now all I have to do is just give it a little bit of feedback. Right? Hey, this wasn't working as well as I wanted. And here's the new code. It did some research. Here's, you know, the colors of the planets. And now we've got ourselves a little interactive model, hopefully. Right? So we'll paste this and we'll play. Yeah, I still can't drag and it didn't change the colors, which is surprising to me. It says Venus cream yellowish white. So here's what I'm going to do just for fun. Take a picture of it. Looks like the stars are kind of moving in the background, which is kind of cool. There we go. Here is a screenshot of the solar system. There is no color showing. Please fix this and give me the updated code. Okay. All right. This is the last iteration I'm going to do. And then you guys get to figure this out. All right. We've got a WebGL problem here. No big deal. It can fix that. And again, I don't have to really understand this code. I don't have to understand it at all. I just say, hey, here's what it looks like. Here's what's happening. Everything's getting washed out, whatever. Anyway. So let's, let's try one more time and see if If this works, we'll copy back over to here, paste and play. Hey, we got some color now. So I figured it out. I still can't drag and drop. And so I can, again, I could keep pushing this and doing whatever. It is cool that there are some random intermittent shooting stars. The stars are kind of moving. And so it does kind of feel alive. It's not bad. I think it's okay. I think I did an okay thing here. I could certainly do better if I tried harder. But this is a lot of what vibe coding is about, right? You come up with an idea and then you just kind of massage it and iterate it until it's kind of doing exactly what you want. And we're going to do way more complex stuff than this. But for now, I want you guys to all do this and you're all going to share your creation. You're all going to share your screen in the next 5 minutes here.

Okay.

So I want you to take the prompt that I gave you in the chat using p5.js, create a instead of mind-blowing visualization, come up with something, a clone of Tetris. Or what is that, Galaga or whatever. Come up with something. Dinosaurs, you know, running away from dinosaurs, a little human, whatever. Make a little game, make a little model, make a little something. And, and then, you know, keep the— I want to run in the p5.js editor website so you can follow this same kind of thing. And I'll explain why in a minute. But anyway, go for it. I'm going to— I'm going to— if you have questions, ask me. Right. But everybody, yeah, go ahead and dive in and dive in and create something.

How much time will we have to do this?

You have about 180 seconds that started 10 seconds ago. It's not hard. The code can be created in about 10 seconds. You just got to think of an idea and it doesn't have to be perfect, guys. This is not— the idea here is not to create, you know, the next Call of Duty. It's just, you know, let's see if we can make something work.

Okay.

So once we have the code, where do we copy and paste it to?

Were you using Groq the same way I was? Yes. It should show you in the prompt that that p5.js editor should give you a link for where you can go to paste it. Then make sure you delete the code that's already there when you show up here. Just select all and just delete it and then paste it and push play.

I don't see that. I don't see any link that can give me—

That's okay. Marta, why don't you share your screen and we'll walk you through it.

Okay. So I signed up for the updated Groq, but it's on my phone and this one, for some reason, it's not recognizing my email on my computer. So I'm just in the basic one. Is that the reason?

Okay. So no, this is perfect. This is perfect. So you see that little copy button right there? Just Cut to the right. Yeah, so push that. And then just above this, there's a little link. See that link right at the top there? This one, editor? Yep, click that. Perfect. And then you're on a Mac, so go ahead and click Command+A and yeah, just delete that. Select all, make sure you get rid of that little bracket there even. Perfect.

Now Command+V, paste.

That didn't work. V as in Victor, Command+V. Yeah, my bad. Sorry. All right. And then push play. Push that little play button up there on the top left. Boom. You have something. Now, what did you ask it to make?

Your first prompt that you did.

Just the mind-blowing visualization. Okay, perfect. Yeah. So now you can go back and here, I'll have you— I'll help you stop sharing your screen here. Now you can go back and you can give it some other idea instead of mind-blowing visualization. You can have it do I don't know, whatever you want. Okay, make something cool. But that's, but that's the process. Okay, thank you. Yeah, and as soon as any of you have something to show, again, doesn't need to be amazing, just something ready to, ready to show, just, you know, speak up and we'll have you share your screen.

I made something, so Share the screen.

Yeah, yeah.

Can you see my screen?

I guess I do not see it yet. You might have to enable some permissions or say Sometimes you have to like click share screen and then select the screen you want to share. This exercise, guys, is actually twofold. It's one, can you vibe code something simple? And two, can you share your screen? Because we'll do that more later. There you go. All right, it looks like it's coming through now. Or it's almost coming through. There we go. All right. Hey, Tetris. I love it. All right.

Does it work? Yeah, it's coming down. Yeah, those things, whatever.

But can you, can you like move the blocks? Probably with your arrow keys. Um, click on that. Click right there. And then now try using your arrow keys.

Oh, okay. Yeah, yeah, yeah. Oh, wow.

Okay. Yeah, you created Tetris. You're now a good enough coder to create Tetris. I think that's cool, guys. I think that's pretty— the first time I experienced this and I was like, oh my gosh, I made something that works as a non-developer, that was really fun for me. That was really exciting for me to be able to feel. And now I was like, man, what else can I create, right? I made something simple, probably. In fact, I know I've made multiple Tetris clones. And, uh, in fact, the other day I was just sitting here, I was just kind of like, I was listening to something and I was like, I'm kind of bored. And so I just, I, I in about 60 seconds, vibe coded myself a fun version of Snake. I just sat there and just tried to get a high score on my own little snake, and I never came back to it, right? But you can create throwaway software now, which is cool. All right, who else? Who hasn't gone yet?

I'm waiting for my code to finish.

You must have complicated code if you're waiting that long.

Well, I made a mistake and did not do the fast, and so I didn't want to stop it halfway through, so I'm just letting it finish.

That'll do it. That'll do it. Yeah. Push that fast button because AI is trying to please you as best it can.

I accidentally did the same thing. My code is still going.

Tyson, you got something?

I changed it up. I changed my original. Let me see. I'm not sure.

Oh yeah, maybe.

All right. I'll share my screen.

Let's see here. Let's see it. Nice. What do we got here?

It must be a baseball player hitting something. So let's see. Um, see, I put a batter, a pitcher, and supposed to be able to hit the ball.

So, right. So, oh, that's interesting. So it is actually hitting it. Um, It's not hitting it exactly in the direction you want it, but it, but it did create some physics to where it's throwing from the pitching mound and then it hits that thing and it either goes past, swing and a miss, or foul ball goes off to the side. So obviously there's significant improvements you could make to the, um, you know, the, the experience and the visuals. And but guys, this is actually how video games are made. I don't know if any of you guys have ever even you know, kind of seen or even watched a YouTube video about how a video game is made, it's always started with just like a circle and a block and a cube and whatever, right? And you know, you're running around and you're shooting these little things. And then later what you do is you go in and you make really nice 3D models of whatever it is, maybe a human or an alien or whatever it is in the video game, right? And then you just tell the computer, okay, whenever this block shows up, that alien is going to be here. And then, and then here's some animations and whatever. Anyway, so that's, that's how it is. So this is the very first step of creating an actual working video game. But P5.js, this editor, this language that we're using is not a visual editor. It's not a high-quality visual editor, let's say. Perfect, Tyson. That was awesome. You're now a developer. Boom. Love it. All right. Travis, you ready now? Yeah, mine's ready. Okay.

Let's see it. This one's for my kids. Spidey and friends swinging through Central Park.

It's pretty cool, actually.

Yeah, it turned out quite a bit different than what I was expecting, but using just a simple prompt without adding like context and output and everything.

And does it work when you click? Does it do the web yank?

Yeah, I mean, it'll just go faster and faster until they get a headache.

Obviously you can work through that. You can make it a moving thing to where you're actually going through the city. You could do different stuff, but that's, it's still, it's still cool. All right, awesome.

Awesome.

Nice work. Nice work. Um, all of you are passing the test, by the way. The primary thing here is, can you create code that plays, right? It doesn't have to be beautiful, doesn't have to be perfect. In fact, none of it is, except Badeep's. Badeep, like, made Tetris. That was like a finished product ready to go, right? Um, so that's cool. All right, Brittany, Eli, Marta, have you worked on something else, or have you just—

Yeah, I'm working, but again, I can't find this editor. So first mistake I made is I didn't put the p5.js, so it created, let me just show you. So it just created an image, but then I'm like, why is it not moving? So this is what it created, right? And I'm like, okay, I'm doing something wrong. Right. But then I, I added that. Oh, here it is. So let's copy it and see what.

Yeah, let's see what happens here.

Did you say it was Command+V, right?

Yeah, Command+V is better. Yeah, it's just the fast one.

You'll do that a lot.

Hey, there's your walking the dog.

There is a lady walking the dog, but she's not moving. The grass is moving instead.

Yeah, that's all right. That's one way that animations work. Perfect. Awesome. So you got something working. Again, what you're seeing here is a bunch of circles and ovals and squares and rounded edges and triangles for the mountains. You know, it uses these kind of rudimentary things and figures out how to how to give you a scene. So very cool.

And why is it that as a picture, original picture, it created like an image of a person, actual person, but in this editor it creates cartoons?

Yeah, it's because this editor isn't capable of doing what's called rendering. It can't create that realistic of a, of a thing. There, there are some video game engines now that are starting, just starting to be able to, to build something that's interactive like this with real pictures. In fact, we might look at one of them today, Project Genie by Google. It's pretty neat.

Okay, Brittany. All right. All right, I have a prompt of just a tortoise and a hare having a race here.

In the forest. I love it. Very cool. Tortoise and the Hare going. Now what happens if you restart it?

The tortoise always wins.

Yeah, tortoise always wins. Tortoise always wins. All right, well, this might be a little antithetical for the, you know, what we're talking about today, building a business in 2 days, you know. It used to take years, but anyway, no, that's— I love it. I love it. Fantastic.

All right, Eli, did you go yet?

Uh, not yet.

Yeah.

Okay.

Let me share my screen.

We're going fast, guys. This normally takes— I normally go all the way through 10 o'clock on this, on the first hour. And anyway, you're doing good. Um, great. No big deal.

We'll see you in a minute. Oh, mine might actually take a while because it says I have to allow, um, permission to share my screen.

Yeah. So go ahead and click that allow thing. What it's gonna do is it's gonna kick you out of the Zoom and then you're gonna have to join again. So go ahead and do that. Again, this is part of why we do this right now. So we get this, these little technical things out of the way. All right, guys.

Why are we using Groq instead of like, I've downloaded pretty much all the ChatGPT and, and, and other ones, but why are we using Groq? Is it something special about that?

We're gonna be using Groq a little bit more today for other things. This specific task doesn't necessarily need Grok, but I just want you to get comfortable using different interfaces. Grok is really good at market research. It's better than any other tool out there, and there's a big reason why. Grok is the only AI that has access to Twitter. It has access to billions of things that people from every country, every gender, every nationality, every political persuasion, every interest group there. Everyone's there. And so if you're looking for market research on some product or some idea or some something that you have, right, even about staffing, there are people that post on Twitter, on X, about staffing, right? You can find that information on there. You can do research on there. And you can also— another thing that's really important that we're going to talk about in a little bit here, a little bit in more depth, is you'd never want to trust a single AI. Trusting a single AI is about the same as trusting Fox News or MSNBC. Right. They are very biased. Right. And I don't care which one of those you like. I personally watch MSNBC and CNN and Fox News because I want to hear what people are saying on all sides.

Right.

Just so that I can then hopefully figure out what's actually happening. Because if I just listen to Fox News, it's all pro-Trump, it's all pro-whatever, it's all pro-ICE, it's all pro-this. And then if it's MSNBC, Trump's the devil and ICE is terrible and whatever, right? And, you know, Hitler, right? And the sky is falling, right? I think neither are true. I think it's— the truth is really somewhere in the middle. And so we have to be able to figure that out. The difference between asking something to Grok or asking something of ChatGPT is sometimes quite different, right? Ask a religious question to 3 different AIs, you get totally different answers, right? So anyway, that's why we're using multiple AIs. All right, Eli, can you share your screen now?

Yes, I can. Okay. All right, so this is mine. Oh, maybe, hang on.

Yeah, yeah, you're good. Yeah, just click over there.

There we go. So for mine, it's a knockoff brand of Mario. Oh, nice. So basically, you can do a double jump, and if he falls in one of these green tubes, then it takes him to the, like, to be upside down. And then if he falls in another tube, he goes back.

Wow. That's pretty cool. It looks hard because it's like an infinite runner. Yeah.

Hit it just right.

Exactly. Pretty cool. That's pretty cool. My kids love doing this exercise. For those of you that have kids, show them how to create a game with just saying a few words. It's amazing. My son, my 13-year-old, has created some pretty impressive stuff. Stuff that pushed the graphics engine farther than I even thought it could go. p5.js is a very lightweight engine. Very, very easy on your computer. As long as you're doing kind of basic stuff, you can, you can use any graphics engine and break your graphics card. But for the most part, it's really, it's really easy to build just all kinds of stuff. I was sitting in church a few months ago with my family, and my son leaned over to me, I was like, hey, could I, could I, um, you know, use your phone to play, play a game? And I was like, no. Then he just sat there and he's like, what if I make like a church image matching game and I vibe code it? I was like, you can do that. So I gave him the phone and let him just sit there. He didn't use p5.js. He used JavaScript. But again, you know, you're not going to use— you may never, all of you, you may never use p5.js again. Okay? It's just— this was just kind of for this exercise to introduce you to what we call a low-level frontend, frontend meaning what we can see, frontend language. You got frontend and backend. We'll talk about that more later too. Okay, so any questions on this, guys? We're a little bit ahead of schedule, so we can go ahead and dive into the next thing. But, but anyway, any questions before I continue on?

If we put this code in ChatGPT or Perplexity or any of them, it would work the same way? You just choose? Okay, so we can play around and see what it creates in different engines.

Okay. 100%. Again, if you take away this constraint that we artificially added of using p5.js, this language, your idea would be much prettier. It would be much better and it would do it differently. We just did that for fun. It's just a fun language and it's fun to see what you can create with those things.

If you want to create—

I have a quick question too.

Oh yeah, go ahead, Eli. You're talking about how all the different AIs have different perspectives and themes. Yes. For example, with Grok and ChatGPT, what are the biggest differences between those two that you found?

Depending on the time, like, like if you were to go back in time a little bit, I don't know if you guys remember this, and I'm going to answer, I'm going to answer this by talking about another one. But Gemini is another one, Google's, Google's AI. I don't know if you remember, maybe about a year ago there was kind of a, I wouldn't call it a scandal, but in the AI world it was a little bit, right?— which is kind of a big piece of news, but everyone was really excited about Google's ability to create awesome images, right? Quickly with AI, right? It's really, really awesome. But it was asked, it said, hey, create an image of the founding fathers of the United States, right? And you could not get Google's AI image generator to actually create something that looked like a bunch of, you know, George Washington, Jon Adams, like a bunch white dudes. Why? Because Google had this bias of DEI and inclusion, and so there were women and Indians and all this stuff, and this nice multicultural representation of the Founding Fathers of the United States. Well, regardless of your political leanings or views, that's just not true, right? And so, and so it became very, very clear in that moment that, holy cow, these AIs have their own political bias based on what, what they're trained with, right? And they were trained with all this stuff that came from Google, which is a more liberal organization, right? And so Grok, I bet you guys could guess which side it leans towards a little bit, right? It's probably going to be a little bit more pro this, but I can tell you that Elon has, has focused really hard on making sure that it's, that it's neutral. He wants it. He wants truth. He wants to seek for truth. That's really what he's going for. And so at times in the past, X has been more right-leaning and Google has been more left-leaning. But Google and X and Sam Altman sitting there somewhere in the middle, right, with OpenAI, all these guys have really tried to pull the political bias out of it. And to be able to display what's as close to the truth as they can get. There's also going to be some bias in terms of who it knows you are, right? We're going to talk about that next. We're going to jump in and talk about, you know, using ChatGPT and how to personalize it to yourself and how you can make sure you're getting answers that align to you, because it understands maybe you're a Jew or maybe you're from Albania, or maybe you like chicken instead of steak, right? Or maybe you're a vegetarian, right? If it knows those things about you and you ask it for, hey, what's, what's a really good meal for tonight? If it knows you're a vegetarian, it's gonna give you a different answer than if it knows Eli's a carnivore, right? It's just going to do things differently. So, so I hope that answers your question, Eli. The reality is they all do kind of have some of their biases, but they're getting flattened down a little bit more, but they're not completely gone. And I would still say that, some of that bias is there in the, in the ways that I said. I would say OpenAI is going to be a little bit more liberal, um, and as is Gemini, and X is going to be a little bit more conservative. Although X is going to be a little bit more liberal in what allows you to do. They have less safeguards and things like that, right? And so you might be able to generate a more explicit image on X as opposed to OpenAI or Gemini. You might not be able to. So just be aware, understanding kind of what these different systems are capable of and the dangers. Maybe you don't want to see that explicit image. So, you know, be careful. Now I can tell you this, I have never had an AI just give me an explicit image like, oh my gosh, I can't believe it did that. Right? I would have to very specifically say, hey, generate this image of, you know, whatever, to get that. So it's not going to just throw that at you. But if you ask that of like OpenAI and say, hey, show me a picture of, you know, something inappropriate, OpenAI will just say no. ChatGPT will just say no, I can't do that, sorry. Whereas X will say, hey, I want you to know I'm going to do this. It's going to be 18+ mature content, whatever. And then, and then go in there. So, so anyway, again, I don't want you guys going and exploring that side of AI. I'm just trying to help you understand that there are safeguards in place to protect you and to protect your children or whoever it is, right? And, And then if you want to venture off into the dark web, the dark web's there, you know, have fun. That's up to you guys. So now we're going to do— we're just going to have you just watch for a minute here. I'm going to open up Replit. We're going to dive right into the fun stuff and we're going to do it a little bit, a little bit slowly and we're going to kind of back— actually, I'll tell you right now, we're going to do it the wrong way. That's, that's really what I want to— what I want to say here. We're going to do it the wrong way first., and then, um, we'll go back and we'll, we'll really do this right and get you guys thinking the right way. First of all, I just wanna show off, here's my, here's my screen here. Let's get this sharing here. Um, I'm a level 4 advanced builder on Replit, which is super cool. This means that I'm a senior developer and I don't know how to write code. And a year ago, no, not a year ago, 2 years. Let's say 2 years ago, I had never done any Vibe coding at all. Right. And now I've used Replit enough. I've built over 200 apps and it's really cool. It's really fun. So anyway, I know what I'm talking about with Replit. I've built a huge amount. I know the CEO and their team. If you go onto my internal Slack, whatever I have, I have my external Replit. I can, I can speak to the Replit team anytime I want. I've got my direct lines in there anyway. And, and they're, I'm a great partner with them at this point. Okay. So, um, when you get to Replit, it's very, very easy. replit.com. You get to this thing, it says app or design, describe your idea. Okay. I'm gonna start with design for, for today. And, and I recommend that. And that's because the design agent is insanely powerful here. And then you can convert it to an app later. If you start with app, then it builds the functionality and you have to kind of fix it later. I like to start with a good design, at least something that looks good. So, um, one of the fun things I do with this course is every single time I, I, uh, teach the course, whether I'm teaching it to, you know, in this 25th cohort or any of the other 24 before it, um, or teach into a business, I always come up with a brand new business idea and I get to itch my entrepreneurial, you know, my entrepreneurial designs and desires. And I get to come up with a new idea. So, and I always do it on the spot. So I gotta come up with an idea. Right now, and I usually just kind of look around my room as I think and try to see something that I like. I got to come up with an idea that's actually useful. I see— I'm looking— this is what's just kind of, you know, is popping into my head. I've got this drink here, Bundaberg Peach, Australian, family-owned. I'm a drink guy. I really like drinks. This has a lot of sugar, and so it's been sitting here for a while because I'm kind of trying to avoid sugar. But this tastes so good. But anyway, what kind of business could I create? Could I just make a drink review business? How many of you guys know who Gary Vee is? Gary Vaynerchuk. Yeah, so Gary Vee is a super, super smart guy, extremely successful, 9-figure, probably headed towards being a billionaire, but he really started his career reviewing wine, a YouTube channel reviewing wine. And now he has this big, huge wine kind of thing on top of all the other stuff he does. But thinking it could be cool to have some sort of website that ranks the best drinks out there. I'm going to make money with this right now. We're going to figure out how to make money with this idea. So, all right, I want to create a website that allows me to review my favorite drinks. Others will be able to come in and do the same. There will be ranking systems, um, categories, um, and much more. I don't know, I can't think of— I can't think of how it's going to do this. Um, we'll likely monetize through ads or affiliate links for drinks. Yeah, okay, this is good. This will be for all non-alcoholic drinks, because I'm a member of the Church of Jesus Christ of Latter-day Saints and I don't drink. And so we're just going to make this for that audience, for me, for all non-alcoholic drinks. Okay. All right. Do you guys think it's a good prompt or a bad prompt? Pretty good. I think it's a bad one, but that's okay. I'll tell you that the good prompt we're going to use in a bit here will be more like 3 to 4 pages long. Okay, so that's where we're going to go. But for now, I just want to show you how powerful Replit is. I think Replit's so cool. Now, while this is thinking, we're going to— we're not going to sit here and, and wait for it. We're going to talk about what it's doing. One of the things that AI has been— has learned how to do, has been trained to do over the last year is multi-step thinking, or something called chain of thought reasoning. Okay. Tyson, we're gonna pick on you. Um, tell us how you get ready in the morning, right? You get up and you got to leave the door, right? How do you, how do you get ready?

What's your routine? Yeah, my kids wake me up about 5 AM, run into the room, um, jump in the bed, um, and then I get them. I, uh, I check my emails probably while they're laying in bed next to me, and I then get up and I get them going with their changing their clothes and getting their breakfast, and then I'll get myself going with, uh take a shower, get dressed, come back out, and then I see where they're at and they're usually off track and then I'll get them back on track, make their lunches, and sometimes I'll dive into a company meeting for a half hour and put them in the back. So this all starts early in the morning. And then I, as we get closer to 7:15, then I'm go time. I gotta make sure they're on schedule and ready to go by 7:40.

So that's very loosely my morning. That's cool. That's great. Let's zoom in on one thing that I don't think you mentioned, but I'm sure you do it.

I want you to tell us all step by step exactly how you brush your teeth. Okay. Step by step, I go to the sink. I pick the toothbrush out of the holder, get the toothbrush wet, put the toothpaste on it. I brush my teeth from left bottom to right bottom to left top to right top and brush the front and brush the tongue and then clean out the toothbrush and put the toothbrush back in the cup holder.

Amazing. Thank you. Well done, Tyson. Okay, so, um, it's really interesting. Everybody— I've asked this question of easily 50 people and 100% of those 50 people get their toothbrush wet before they put toothpaste on it. I don't know why. Why do we do that? It just like feels like wrong if you don't. Anyway, so thank you. And I find it fascinating just to hear the different ways people brush their teeth actually, because I've asked that question a lot. So anyway, long story short, hey, by the way, my battery's dying, so I should plug this in so we don't lose you guys. Okay. What you just did is you broke down, Tyson, a high-level task of get ready, right? And that get ready task is, you know, it's routine, so you kind of know, you don't have to think through it really well. But if you were to break it down, there's all these subtasks, and you listed them out for us, and that was great, right? And then you can get to each of these subtasks, like, like, you know, putting your kids in the bath, right? It's a lot more than just putting your kids in the bath. You're gonna have to help them shampoo their hair and do this. You have to turn the water on, get it right, fill it up, all this stuff. You're gonna have to— when you're brushing your teeth, you have all these different steps. And so even these subtasks tasks have subtasks underneath them, and some of those even have sub-subtasks, right? And so this chain of thought reasoning takes a big task and breaks it down and then reports back up. And so that's what this agent is doing now. You can see we're already starting to get a website that's starting to look— and look at this, guys. All I did was just tell it this. This is pretty cool. We've got a fully functioning website, or at least it looks pretty good. I think it's doing pretty decent. For, for giving it very, very little, um, information on what I want it. Now it's all— it's called Zero Proof. That's not exactly the idea I want to go after. I kind of want to go more after, um, drinks that aren't alcoholic. I didn't want it to focus on like mocktails, and that's kind of what it decided to do. Um, and what I really want is like a drink review site, like what's my favorite soda or what's my favorite hot chocolate. I don't know, you know, simple things like that. Or might the best flavor of Gatorade or sport drinks or whatever it is, right? And so, but it didn't know that, right? It didn't know that. Again, this is why this is a bad prompt. Okay. Ranking system categories much more, but we didn't really tell it exactly what's going on here, but hey, it's pretty cool. View details. I can click on this. It even created a product page here, right? I can buy now or add to cart, write a review. Hey, cool. This is nice. Whatever. Right. Category, functional, your rating, 5 stars. Yeah. See, the 5 stars doesn't work. So this is more of a visual preview than an actual working thing. But look at how dang fast that got to a point where, I mean, look guys, in probably an hour I could have this site operating to a point where I could put it in front of a customer and see what they think. And, and by the way, what did that cost me? 48 cents. 48 cents to get to, uh, an idea. Then it would cost me maybe another $20 to $30, maybe, of, of the credits, you know, that Replit, that Replit uses. All of you, if you've signed up for today, right, you're, you have, um, $25 or $30, I think you have $35 of credits that you'll use. And then after that, then it'll just charge you. But I doubt you'll even use that in this course. I think you'll, I think you'll be fine unless you spend a lot of time, um, on your own working on it. But, but bottom line is we're to a pretty good spot already. And, and I can come through. Okay. We got dynamic homepage review system. We've got categorization design system. All right. It's kind of cool. Refreshing mint citrus creams, deep navies to go crisp and non-alcoholic. I think it's, I think it's pretty cool. If I was wanting to make a zero-proof website, not bad. Not bad at all. Our mission. Yep. Didn't create that. This doesn't mean it did it wrong. It means that it hasn't created this yet. Okay. So what do you guys think? Isn't that cool? I, I don't know about you guys, but I think that's cool. When I have an idea, I love to be able to get it to a point where I can just kind of see it and then say, okay, what do I think? Is this good? Should I pursue this? And most of the time the answer is no, cuz I don't have time. That's really the real reason. But I'm like, pretty cool. I wonder if someone else could take this idea and run with it. So, Now we're going to back up. We're going to back up here and we're going to say, all right, let's do this the right way because this is the way that most people vibe code. They just jump right in. They give it 1 or 2 or 3 sentences here, just some very, very high-level stuff. And then from here I could say, all right, you know what? I actually don't want this. And then they try to just fix it. Right? The reality is I left way too much for the AI to guess, to figure out, to fill in those gaps, to say, okay, I think this is what he's wanting. I think this is what he's trying to do. And in this case, it got it wrong. This is not what I wanted. All right. But again, that's my fault. I'm the user. I'm the boss. This is my developer. I gave it poor instructions. So it's my fault. So let's do it the right way. Okay. We're going to use ChatGPT today for this. ChatGPT. Is, um, uh, why are we— oh, you know what? I know why. Because I'm in the wrong profile. I want to use this one. And yeah, we'll leave X open. We're going to use that later. All right. ChatGPT. Okay. So ChatGPT, guys, is for me, my project manager. This is where all of my different ideas live. I've got all kinds of, of, of chats, thousands of chats. And, and then projects. Projects are awesome. We're going to talk about them in just a second here. But this is where I, this is where I live. Okay, this is where I focus on, on pulling my ideas together. I probably spend more time inside of Replit than any other platform, but, uh, but ChatGPT helps me organize it and do it well. Okay, so let's talk a little bit about ChatGPT. Some things you may know, and some of this might be obvious, I hope, you know, some of it is, but we're going to go over a couple of things just to make sure we're very, very clear. One of the things that AI— a very important word in AI is this word called context. Okay? And it means basically the same thing that it would mean if I was having just a conversation. Brittany was over here and we were having one of our, you know, zero-proof drinks, right? And we're just talking. If I just jumped right into a conversation, I said, I said, what do you think about that? And that was my first sentence of our conversation. She'd be like, what do you, what do I think about what? What are you talking about? Right? I need some context. I need to know what's going on. Okay. So there's a couple ways that this word applies here, but let's go to, um, the, the, the, the, the ultrasonic liver cancer treatment. Okay. So I had, I have a, a, a close, uh, friend who just got diagnosed with liver cancer. Okay. And luckily they caught it really early. And I was asking about this ultrasonic liver cancer treatment that they were gonna be working on. So I, I started asking about this. Okay. So that's the, that's the context of this conversation for you guys. But if I were to come here and just say, help me understand this, help me explain this to a 5-year-old. I didn't have to come back and— well, please retry work. All right. Normally that works, and I think it is going to. I think it just had an error there. But I didn't have to go back and say, hey, remember, we're talking about liver cancer here and this ultrasonic treatment. I didn't have to do that. It remembered inside of the conversation what I was talking about. Now, that might seem very obvious. I'm glad if it does, right? But I can, I can come in here and I can just kind of start, I can start talking to it. Okay, so, um, yeah, help me explain this to a 5-year-old. Great, here's the 5-year-old version. Sometimes people get a bad lump. Okay, all right, great. So now I have this context caught up. So there's context and understanding at the conversation level. Okay, 2, there's conversation understanding at the System level. Okay. And you can click just on your thing and go to personalization. You can tell it how you want it to act. Default, professional, friendly, candid, quirky, efficient, nerdy, cynical. Okay. Awesome. You can say, hey, you know what? Do you want it to be more warm, more enthusiastic, more whatever? I just have it set on these defaults. Actually, I'm going to set less for emojis. That's a new setting. I've never seen that one before, and I'm so happy I can turn those down. I hate the emoji answers. I don't want emojis. Okay. Custom instructions. This I can, I can give it information about me. Now, I haven't come in here and told it, you know, that I, I do AI and I do this and I like kayaking and I like the piano and, you know, whatever. Right. I haven't told it that, but it has had enough conversations with me that its memory actually has a huge amount of information about me. It knows exactly what's going on. It knows, things about my business. It knows that I'm an entrepreneur and a composer and a consultant. And I live in West Mountain on a 5-acre farm, and I'm fluent in Spanish and Chinese. I'm a kayaker and a member of the church. Right? Again, these types of things are important for it to understand because it gives context so that when I ask it a question, I don't have to explain myself every single time. Okay? If I ask something about the nature of God and I'm a Jew, I don't want it answering it as if I'm a Catholic. I want it to understand that I'm a Jew, so I don't have to explain, this is what I believe and this is this, and, you know, I read the Torah and this is where really where I want you to pull answers from. If it just knows you're Jewish, awesome. Now it's going to give you that better answer for you, that better answer for you specifically. Okay. So being able to use this memory thing, and, and this just takes time. And by the way, you have the ability to, um, to tell it, hey, like, hey, once this text saved as a personal strategy guide, whatever, you can tell it, hey, I want you to save this, right? I want you to remember this, add this to your memory. You can't actually come in and add, you can delete all memories, you can automatically manage, you can do that stuff. You can't just come in and, and like tell it a bunch of things. Welcome back, Marty. Right now we're just going over ChatGPT and talking about context. Okay? So we have context at the conversation level. We have context at the system level. So it knows who we are and what we're about. Okay? It even knows, like, when I ask it to build, like, hey, I'm gonna, I'm gonna build something. It's like, okay, I know you're gonna use Replit. I know you're gonna do this quickly. And I know you're gonna use this pattern. It just knows my style. And so it works very, very quickly with me as a partner. Okay? And then the third place where you can get Awesome context is this project tool, and I love the project tool. So we're going to say drink review site or drink. Yeah, yeah, drink review site. We'll just call it that. That's fine. Okay. So I'm just going to call it that. I'm going to create a project. Okay. And you guys don't have to do this quite yet. I'm going to give you time to do this in a minute for yourself. Okay. But then here, if I have files, maybe I've got company files, right, for balanced staffing or something like that. Right. I want to add all this stuff and I can drag a bunch of things in. That it needs to know about the project. Or maybe I want to, or maybe I already have, maybe I've been working on, you know, my own drink reviews, you know, for, for years. And I've got all these, you know, maybe I have YouTube videos that I've done or I have these other things. Now I'm making a website. I could drop all that stuff into here and it's immediately going to have all that context.

Okay.

For, for now we're just going. And by the way, every, every time I come back to this, it's going to remember everything that's inside of this. I'm going to show you in a second, but I need to get over to here. I'm going to grab my little— to be clear, this is not a mocktail site. It will review all kinds of drinks, hot chocolate, workout drinks, or yeah, protein, etc. You know, sodas, mocktails can be a category, etc. Okay, so, all right, so I'm just going to tell it this. Don't do anything yet, just acknowledge this message. All right, perfect. So I've given it that just so I can kind of click away for a second here. Message received. Thank you. Got it. All right. But if I come over to General AI Proficiency Institute, this is my, my company. If you go back here, you can see when I actually started GenAIPI, you can go all the way back to February 27th of last year and where I came up with my first idea to actually turn what I was doing with all these AI consulting things into a bigger company and a bigger idea. But everything I do here, if I ask it something, I have hundreds of hours of conversations that I've had in here and I can just say something and it understands that it knows that it can pull from it. Okay. If I just go to new chat right here in the kind of base level, I'm not inside of a project and I just start asking it about GenAI API or ask it about my drink project, it actually won't know much, right? It doesn't have the ability to understand that much context. There's something called a context limit with AI. And so it has a limit of how much it can understand at once. And that's one of the big barriers to AI development. And it's kind of like the iPhone, right? Every time the new iPhone comes out, you go from 30 megapixels to 35 megapixels to 40 megapixels with the camera, and you go from 512 gigs of storage to 1 terabyte to 2 terabytes, and maybe we'll get 4 terabytes Here it just kind of improves as time goes on. That's kind of what's happening with context limits in AI, right? Anthropic, which is a major model, a major company like OpenAI or X, right? This is one of the big AI companies. They, they have Claude. Claude is, is like ChatGPT or Grok or whatever, right? But this 4.6 that just was announced literally 5 days ago has now a million, um, con— token, uh, context limit. And a token is about equal to a word. So you can— it can now analyze in one query a million words. That's, that's crazy. And taken and really take into context. This is extremely helpful for things like coding, stuff like that. Okay. All right. So now we're gonna come back to drink review site. Okay. So it's gonna help me make this thing. So what I want to do here, one of the things that again is very, very easy to do, and I'm sure all of you have done it. I know I've done it. Okay? So this isn't, this isn't bad, but it's just, you gotta be aware that you're doing it when you're using AI. Have you ever told it something? You're like, I have this idea and it comes back and it says, you are a genius. This is so good. Right? Let's build this out together. And you're like, great, let's go. Right? And then it says, okay, what do you wanna do here? And you tell it something and then it says, Okay, good. I did that. I built the marketing plan or I built the business plan. Would you like me to make a marketing plan now? Yes. And you read the marketing plan. You're like, that is amazing. It's perfect. Would you like me to build out some financials? Sure. Right. And okay, so good. I love this. This is really good. Okay. Would you like me to now, you know, help you figure out, you know, the right— the first blog for this? Sure. Yeah, do that. Right. And then you've been working for 20 minutes. On this cool idea, having this conversation, and you realize that all you did in the entire conversation was you gave it a one-sentence idea at the top, and then you said, yeah, sure, great, do that. This is great. Go. What have you actually contributed to the world at this point? Basically nothing, right? AI is just spitting back, and it has taken over. It's taken over your business. It's taken over the job, your job of creation, right? You're the one with the quantum computer in your, in your head, right? You're the one that has this ability to receive inspiration in a way that ChatGPT and AI cannot do. You need to be in charge. You gotta make sure you stay in charge, right? And so there are, there are several ways to do that, but my favorite way that I, I came up with this on my own, um, and I've seen other people do similar things., but I do it just a little bit differently. And, and this is really to have the AI interview you, right? And when you have the AI interview you, it, it completely flips the script. It puts you in charge, right? And so you say this, I want you to interview me one question at a time about this idea. Um, let's cover the, you know, name of the business. The USP, the unique selling proposition, is what that is. The monetization strategy, the intended audience, the marketing plan, the, you know, let's see, and anything else you think we need to go over in order to build a solid business plan. Okay. So I am going to copy this one over into the chats as well. Okay. So you'll, you'll use this in a minute. Okay. But, but this, there's an important thing here. One question at a time. If you say, I want you to interview me about the idea, it will just give you 10 questions in one prompt. And nobody likes to do a test or a quiz, right? That's boring, right? But we do like to have conversations. And so if you say one question at a time,, right? Then it will just start to break it down and say, okay. And what it's done is it's basically doing what Tyson was doing earlier with brushing his teeth and getting ready, right? It's breaking down this thing and it's saying, okay, step 1, here's what we're gonna do. And then I'm gonna answer and then it's gonna go to step 2. But it's kind of maintaining this context, this ongoing like flow, and you're just gonna kind of slowly feed it these things. Okay? So love it. We'll do this like a tight founder interview. One clear question at a time. What's the working name of this business right now? Or if you don't have one, what vibe should the name have? For example, playful versus serious, nerdy lifestyle, mass market, enthusiast, brand first. You can answer with a specific name, a few rough ideas. I don't have a specific idea, but I want to go playful and enthusiasts with the vibe. Enthusiast. Man, how did my brain do that? With the vibe. What ideas do you have? All right, so then it's gonna give me a few ideas and I'm not gonna spend a lot of time on this. If I was building my own business, I would spend more time on that, okay? All right, vibe, playful, here we go. What do we got? Sip score, drink rank, sip list. I don't like any of these so far. Drink index, sip review. Eh, sippable, that's okay. I don't like the word sip. Give me better names and avoid the word SIP, please. Go. Dropping SIP actually sharpens this. So AI is just gonna make you feel like good all the time. Even though you have to be able to recognize that some of your ideas are bad and you need to ask it to give you that feedback. But anyway, I actually kind of like Drinkable because I can actually use that. We're going to go with Drinkable for now. I think it's good. Now, if I was building a business and I'm like, hey, I like the— I like the name Drinkable, okay, I'm probably going to zip over here and I'm going to go to like Facebook really quick and I'm going to Search right here, drinkable. Just going to see what comes up. Is that a brand that someone already has? People, pages, here we go. Drinkable. There's a community with 12 followers, 63 followers. I don't care if someone else has used it. No one else has actually branded it. Okay, so that's cool. There's not really something out there. I'm then going to zip over to GoDaddy just for fun because I'm turning this into a business. Right? I'm just going to look for drinkable. I guarantee this is taken. Wow. It's not taken, but it's a premium domain. It's going to cost me a little bit of money to get that. So if I go with drinkable.com, now I can definitely do drinkable.inc or drinkable.me or drinkable.live or fun or whatever, right? There's other things I could go with, right? Or, you know, anyway, maybe I don't like that, but I'm going to go with drinkable for now. I think it's fine. I think it's a good, I think it's a good name actually. So, let's go with drinkable. Okay, done. So now I didn't tell it to continue the interview, but it knows, right? It's gonna gimme a little bit of feedback. Tell me why I'm super smart and visionary by picking drinkable. Okay, don't let that get to your head. But then it immediately goes to question 2, right? Why would someone use this instead of these other things? Okay. Anyway. All right guys, so I don't need you to sit here watching me do this. I am gonna do this. We're gonna spend the next, about 15 minutes, and you all are going to do this as well for your own idea. Okay, so you're going to start, open up ChatGPT, click new project, call it whatever you're going to call it, and then just explain to it in a few sentences what you're doing. Okay, and then say, um, don't do anything yet, just acknowledge this message right at the end. Okay, so that— because otherwise it'll just like go off and build a plan for you. You don't want that, right? Just have it acknowledge this and then copy paste this thing. I want you to interview me one question at a time about this idea. Now, if you're doing this, Tyson, for example, if you're saying, you know what, I'm gonna work on something for, you know, that's internal, then obviously you don't need to cover the name of the business, right? So just, just, you know, we don't need to do the name. Just be like, hey, let's cover the USP monetization strategy, intended audience, marketing plan, whatever. Give it the whatever you need. It's still very useful for you to think through that, even if it's for an internal tool. What is, what, what, what is the business plan? What is the business case, use case for this? Okay. So, um, everyone go ahead and dive in. I'm gonna just keep sharing my screen as I answer these questions.

And if you have questions, feel free to ask me. Jon, with the project, are you able to insert a chat that you maybe already had and can you put that into like the file of that in a project?

Yeah. Yeah. If you have, so create the project and then if you already have a file or something, You can just click this little three dots right here and just move to project.

Oh, perfect. Thanks.

Yep.

Jon, I've already got a project, uh, in Prompt and then also a PRD ready for Replit.

Um, can I just utilize that for the idea that I already have ready to go? You can certainly start there. Have you gone through this process of interviewing and diving deep into the monetization strategy and plans and all those things?

It, other than doing it one question at a time, yes, I did a, a variation of that, of ask me questions till you're 95% confident you can complete this task, and I gave it a task assignment, context output and that type of.

But do you have, you have a prd, but do you have a business plan?

Yes.

Okay.

No, no, no, no, no, no.

I do not have a go to market plan. Get a business plan. Okay, okay. And the business plan isn't just a go to market plan. It, it talks about why this exists and who is your target market. And then, you know, All of that, all of that's part of it. So get the business plan. That is the end goal that we're going for, guys, here. The reason for this, and go ahead and get started, right? You can just listen to me in the background. But the reason we go for the business plan is because the AI, even in considering whether or not to add a button to a screen, will think and do things differently if it knows you're trying to accomplish some business goal, like getting a signup. Or getting a subscription or getting someone to submit their name here or whatever it is. Right. And so the more it understands about the why, not just, hey, here's a product design, right? It actually understands why this matters from a business perspective.

It will do things differently.

Yeah.

So what should I ask? Can I ask ChatGPT on insurance?

I want to create a business to monetize insurance or. Yeah, yeah. So whatever your business idea is, just start talking to ChatGPT about it and have it, have it help you work through that idea. Um, Marty, you just asked for the original prompt. Yeah, so my website, uh, that I was working on, yeah, this was what I, this is what I decided to do here. I'll share that with you guys here. Okay. Maybe another 6 or 7 minutes here, guys. Okay. And you want to get through probably somewhere between 6 to 10 questions at minimum. If you were really doing this, like really trying to build something amazing, you might do this for 2 hours, but we obviously don't have time to do that in this course. So just try to get through kind of the core. And if it's asking about too many things, be like, hey, let's skip to the monetization plan, or you can tell it to kind of jump ahead. Okay, take another about 2 minutes and then we'll move to the next step here. By the way, we will be taking a break here probably in about 10 minutes, maybe 15, for those of you that are a little restless. Okay guys, so you've gone through now, I've gone through, I don't know, 10 questions or so, and I now have a pretty decent idea of kind of what we're going after here. It's, you know, we've got, we've got, I've got some different access models and long-term visions and the interview is complete, right? For me, it completed, right? And that's cool. I have actually never seen it say interview complete, at least in this short of time, right? I've gone through 20 or 25 questions and it's finally like, okay, I get it now. Okay. So what we're going to do now is we're going to ask it to build us a business plan. Okay. And this is super easy. I want you to make a 2, 1,000-word business plan about, or yeah, for this, for Drinkable, for Drinkable, my business here.

Okay.

I'm actually super excited about this idea. I might do this on the side. I need to figure out how to, maybe I'll just put my Claude bot on it and let it just build it. But that's not really fun. Anyway, we'll talk about Claude bot later. So I want you to make a 2,000-word business plan for Drinkable. Use the Canvas tool. To do so. All right. So I'm just going to click go. I'll copy this little thing for you guys as well. Okay. The Canvas tool is a tool that's inside of ChatGPT and it allows you to basically make a document inside of ChatGPT that's editable. You like, as soon as it's done, you can click edit and go through and type and change it and make it your own, whatever you want to do, or download it or copy it or whatever it is you want to do. But it's now created an executive summary, right? Drink a Drinkable is a community-driven discovery platform dedicated entirely to non-alcoholic drinks, from sodas and sparkling waters to protein shakes, sport drinks, hot chocolate, mocktails, and everything in between. Drinkable provides a structured visual and social home for people who genuinely care about what they drink. Unlike scattered reviews on Google, Reddit threads that are hard to navigate, or influencer content that lacks depth and transparency, Drinkable organizes conversation around the drink itself, et cetera, et cetera. Right? But Travis, you can see how this is much different than just a PRD, right? It's talking about the why and who it's going after. What's the problem? What's the solution? What are the product page and what, what does it look like? Okay, we've got, we've got public drink pages, we have user accounts. So it's a little bit of the PRD, right? But very, very high level. We're going to let Replit take care of the technical stuff. Okay, you can certainly go technical if you want, if that's your style and you know what's going on. Um, community model target audience, whatever, long-term vision. Even here, let's see if it has monetization. I don't even know if it has monetization in here. Yeah, here it is. And I didn't do a, I didn't do like a subscription type model. This is one where it's like, I'm going to have a lot of people and then I'll be able to do affiliate links and stuff like that, right? Affiliate advertising, sponsored placements, and then eventually brand pages even, right? Like if you're Gatorade and there's millions of people on this site, Gatorade might want control over their own brand page. I could, I could charge for that, right? And then potentially premium AI feature features. Um, I talked about, hey, I'm going to go social first. I'm going to probably find some influencers to help me launch this thing. Reddit's probably going to be in there, and there may be some SEO stuff that's highly optimized, right? So now I have a pretty dang good business plan. If I just take this business plan right now and throw it into Replit, I'm— this is way better than my first prompt. Remember this prompt? I want to create a website that allows me to review my favorite drinks. Right now we're doing way, way better. Okay, so let's go to home again. Go to design. We'll come back here in a second, but let's make this even better. Let's make this an even stronger plan. Okay, so now what we're going to do is this. I'm going to copy my business plan by just pushing the copy button right here. Then I'm going to come back to X and I'm going to start a new conversation with Grok. I'm going to say, give me feedback on this business plan. I want you to focus more on the idea and the market. Check X for, for things people are talking about related to this. Do deep research to figure out if this is really a good idea.

Okay.

And then I'm going to hold Shift. Right? If I hold Shift and push Enter, Enter, or Return, it won't actually send the message. Then I can just paste my, my business plan here. Doesn't matter that it's all formatted weird. And then I just say go. Before I do that, I'm going to copy this for you guys. Okay. So I can come here to chat. You guys can have that. Give me feedback on the business plan. So now I'm coming from ChatGPT. And asking it to, you know, do some research on it, right? So, hey, the idea behind Drinkable is solid and timely. Non-alcoholic beverage space is experiencing significant growth and cultural momentum. Amazing. Here's the market opportunity: $1.2 to $1.4 trillion in 2025 with drinks. That's crazy, right? Growing to almost $3 trillion by 2030, 2035. That's crazy. Trends align well. Now, competition differentiation, right? No dominant Dedicated platform exists exactly what you're describing. A drinkable style, Untappd, Letterboxd, Goodreads for non-alcoholic drinks. Existing options include MyDryBar, Selecter Tool, apps like A Better Without. So that's all about kind of the mock, mocktail type stuff, but no one's really doing this kind of stuff. Fragmented or niche. None fully combined public drink pages. Strengths of the idea. Timely flywheel. Potential challenges. Cold start. Seeding with tastemakers. Curated content is key. Consider partnering early with Reddit communities. Yep, scope, engagement. But now I have some good feedback here on like, okay, this is actually a pretty dang good idea. Isn't that amazing, guys? I did not have this idea this morning when I was in the shower today. I did not have this idea until I looked over, sitting here with you guys, at my little Bundaberg drink. And I'm like, I like that drink. And I was like, I discovered that drink by standing in an aisle and looking at it and being like, I wonder what else is good here. This looks interesting. I'm going to try this. And I tried it and I was like, what if I could make a business out of that? Starting a business is not hard. It really isn't hard. Making the business really successful is hard, right? AI can eliminate a lot of the barriers, right? But I just, I'm just really excited about how cool of an idea I was able to come up with. By just looking at this drink, right? And I'm sure all of your ideas are amazing as well. Okay. So what we're going to do here now is I'm going to copy this feedback. Actually, what I'm going to do is I'm going to say this, um, give me, give me 5 concrete ways. Now I'm going to say 10, 10 concrete ways to improve the business plan that I shared with you. Based on your research. By the way, I can have it think harder. I can have it go and like really scour the internet and think harder and think more deeply and all this stuff. Okay. But for now, I'm just going to say, all right, give me 10 concrete ways to improve the business plan that I shared with you based on your research. All right. So narrow the initial scope, explicitly position against untapped recent expansion. Add a dedicated section on functional and wellness trends. Strengthen cold start strategy with specific tactics. Incorporate emerging apps. Enhance monetization. And I might like all these, I might not. Let's assume I like all of them. So here's what we're going to do. I'm going to copy its feedback, okay? And I'm going to come over here to ChatGPT and I'm going to say create a new and improved business plan using the following feedback. However, lean towards my answers and ideas that we've already discussed as you consider this information. Okay, I'm gonna copy this, give it to you guys. Okay, so you have that one. And now I'm just gonna click go. So now what I'm doing here, guys, I'm having these AIs fight against each other, right? I'm having them just kind of argue a little bit. And, and, and really what that does, arguing and debating, right? When Tyson and I get into a discussion on, you know, whatever about how we can use AI better at balance, That makes the idea stronger. And if we disagree, one of us is going to have to come up with a stronger argument as to why we should do it a certain way. And it doesn't matter really which one of us is right in the end. It just matters that the idea got stronger by working together on that, right? Same thing's going to apply here with just you're using AIs to have that conversation. And then obviously we're doing this in fast mode, hyper-fast mode, right? And so I'm not spending time to really read everything and really consider everything and really do that., which I would do if I was really building this business, which I might, which I might. But anyway, so now it's created Drinkable Business Plan Revised and Expanded. It's, it's a better version, right? And so now I can come in here and I can read my, you know, market opportunity. I can, um, I can focus on the problem a little bit more. And so it's just strengthened the entire idea. And, and now we have, um, anyway, risk mitigation, trust erosion, feature creep, control through phased rollout, long-term vision. Cool. So now I've got a better business plan. This is really, really cool. Um, man, I really like this idea. It's a fun idea and I actually really do love drinks like this. I'm not faking this at all. I really love this stuff. So anyway, um, now I've got a business plan. Now we can take this and we can put it into Replit. Almost. We're gonna do one more thing. Okay. Um, I'm going to pause for a second so you guys can catch up. Do you guys have any questions on this or are you just trying to keep up? Just catching up for myself. Okay, I'll give you guys like 1 minute to catch up. That's all you get, 60 seconds to just catch up for a second here. Any questions while we're waiting?

By the way, Eli had to step away for an hour and he'll be back. Hey, John and X, what was the prompt to pick the 10 things to break it apart?

On that, you went to X and you did that before you started going back and forth. Yeah, I did. And I did not share that.

Let's copy that right there. There you go. Perfect. Thank you. Mm-hmm. I'm catching up from behind. Just, I just had it do my 2,000-word business plan. Now I'm going to put that and my first prompt into X is going to be the give me feedback on this business Plan. Correct.

After that, is it going to be the create a new and improved business plan? No, after that it's going to be the give me 10. Yeah, yeah. So do that, give me feedback, then do the 10 concrete ways, then, then take that back over into ChatGPT, hand it the, the X, create a new and improved, right?

And then the create a new improved.

Yep, exactly. Got it. Thank you. You guys take another, take another minute or two here. Just, this is okay to get right. And I'll let you guys catch up. I'd rather wait 2 minutes now. And so you guys feel like you're a little bit more, more with me before we move into the next phase here. And we're again, probably about 5 or 6 minutes from break time. All right, guys, so the next thing we're going to do here is, um, I'm just enjoying reading, reading this business plan here. I think it's fun. The next thing we're going to do here is we're going to ask it for one other document. I want you to create another canvas entitled feature list. This will include all front-end and back-end features needed to make this work well. Don't specify— whoops— specify the, the actual technologies that we will use to build these. I just need to know the complete feature list. Okay. So go ahead and you can use that. I'm going to add one little more sentence on mine because I was thinking about something. Let's remember that the most logical way to do this would be to group drinks by brands. Drinks, individual.

Okay, Jon Cheney, copy that one in the—

oh, you got it in there. Perfect. Yeah, it's there. Yeah, I dropped that in. Yeah, this feature list is really cool, guys. I'll show you how we use it once we get into Replit, but it basically helps your AI keep track. Of where it is in a very long, complicated project. At least how well it's keeping on track as it relates to your vision, right, and plan. Pretty cool. So we'll take another just minute and let you guys let your feature list kind of populate here. It's good to just kind of, you know, scan it. Make sure there's nothing crazy here. And again, if you see something like, oh, that's kind of way off, just click Edit and just go in and edit it, right? Just change the text of that little thing that you want to do. You can also, by the way, ChatGPT is capable of editing that document. You could say, hey, edit that document you just made and change this and update it to this. You could certainly do that. So yeah, now that I've done this, it says, hey, this is a con— this is comprehensive, comprehensive enough to hand to Replit or an engineering team. I didn't tell it I was going to use Replit. It just knows because it's worked with me so much, right?

So that's cool. Jon, question: after finishing the, uh, give me 10 concrete ways to improve the business plan in Groq, are you downloading the file or copying and pasting it.

Just copy back in the chat. You just copy it into chat and just say, hey, make me a new and improved one with that prompt I gave you. Yeah. Okay. Okay. All right, guys. So, all right. So now we have basically 2 documents, right? Technically 3, but we have a second one. We're gonna take the second, the new and improved business plan, right? And then we're gonna take our feature list. And then we're going to ask it one more thing here. We're going to ask ChatGPT, we're going to say, give me a 3-sentence description of Drinkable or of your, your app, right? Or of this project, whatever it is. Okay. Give me a 3-sentence description of this. Okay. So I'm going to copy that. And now it's time to go to Replit. Okay, and I'm going to start in design. Remember, I want to start in design. By the way, I didn't really give it a lot of instructions about design. I might actually do that. Um, yeah, I don't know, maybe I'm just going to let it figure it out. But anyway, I give it some notes on design in a second here. But all I'm going to do is I'm going to paste that 3-sentence description, right? Drinkable's community-driven discovery platform. Right. So I got that. Then I'm going to come in and I'm going to grab my documents. Right. And I'm just going to do that by simply copying it. Copy. And I'm going to go over to Replit and I'm going to paste it. Now, when I paste a really long document like the feature list, it's going to look like this. It's going to look like an attachment. Okay. So you won't see it all, but I promise you it's there. Okay. And then I'm going to go up and I'm going to grab my business plan. I'm gonna copy it and I'm gonna paste it. Okay. Alright, so I gave it a couple thoughts just on the design, right? I was just like, hey, I want it to push the boundaries, right? I want a mix of retro and modern, whatever. But this is all I need for my, for my prompt for Replit. Okay. And so I'm just going to say, alright, let's do this, right? So I've got my 3-sentence description. I gave it a couple just notes on design. This is optional. You don't have to do that. And then I pasted my feature list and I pasted my business plan and then I click start and we are right on schedule. 11:00 AM, I push start. So let's let it go here for a second. It might ask me a question. Sometimes it will ask you questions for your project. And one of the things I like about starting in design is it won't start trying to dive into the different technologies, and we'll talk about some of that stuff in a bit here. But anyway. We're gonna, we're gonna make you guys low-level, low-level developers here in the next little segment. It'll be super easy. Okay. Um, all right. So it's 11:01. We're gonna take a break till 11:15. Okay. So what, what you should do is be to this point, let Replit just work while we take our break. Okay. So I'll turn off my screen or, uh, yeah, anyway, I'll just stop sharing and turn off audio and video and stuff like that.

And I'll see you guys in 15 minutes. Thanks, guys.

Jon, can I, uh, can I steal you for a couple minutes?

Yeah, what's up?

Okay.

Um, so on the— so right now it's just finishing the, uh, updated business plan. Um, then I'm going to put the, I want you to create another canvas entitled feature list. Um, and then I'm gonna download that file and put that into Replit, right? And then I'm gonna say, uh, as well in the chat, give me the 3-sentence description. Um, for my idea, and I'm going to use that and paste that into Replit with my attached file.

Yes. Yep. Okay, cool. That's all I need. Yeah, the feature list and the business plan. Yep. So 2 attached files, 3-sentence description, click go, and do it in design mode. Do it on design mode. Okay.

Okay, awesome.

All right guys, welcome back.

All right, how are your projects looking? Anybody have something cool? Good. It's really fun, actually, once you get the hang of it.

Like you said, you can really do anything. Yeah, let's see.

You want to share, Brittany? Sure. Yeah, don't seem so— me on the spot here. I love it. Okay, so mine was more, uh, value-add. So I was thinking in terms of, of balanced staffing, obviously from the sales team, I need a way to generate some leads while offering maybe a value proposition to some of our potential customers. And I know some of the, um, We've used like Market Insights before, but using some branded live intelligence to help our potential prospects and customers build hiring plans, essentially.

I love it. That's cool.

And I could see you're on the free version right now. I am. I need to upgrade. Tyson had invited me about 10 minutes beforehand, so I didn't have a subscription to this one.

But yeah, for sure, I'm definitely going to upgrade. Yes. Yeah, this is pretty cool. Yeah. I mean, everybody at our company obviously has, you know, we have, we have subscriptions to this and, and yeah, we encourage it.

It's nice. Yeah. Yeah. We don't have a strong internal like marketing department. And as you learn these things, I don't know if I know how to do this, that we would need any additional spend on marketing, having the ability to build some of this stuff.

So yeah, isn't that crazy? And it's crazy that you would say that, like, yeah, we don't need a marketing department now, right? Like, and it's the same way that I said when I, when I built, you know, GenAIPI, I don't need to hire developers, right?

It's everything from even Canva to building one-pagers and things like that. I do all of that in Canva with a couple prompts very easily. It looks like we can add landing pages and revamp our website with this tool. Obviously sales campaigns, drip campaigns, we use ChatGPT and everything to build very customized, not even just generic templates, but go find as much as you can about this customer.

So really, really cool stuff. Yeah, it's awesome. I just want to check. Looks like someone's coming into the room. iPhone is coming into the room.

That might be Tyson or someone that stepped away. Jon, so it's asking if I want to publish it, but I think it's still running in the background. No, don't publish. It's just—

don't publish it.

Yeah, but it's quite amazing. Can you share your screen, Marta?

Yeah, it's still working on it, but it's fun. Let me show Awesome.

Very cool. So tell us what your business is. I mean, it's just like a fun idea for to do a glow up for women in their 40s, like different trends, different methods, you know, meditations and mindset, nutrition, skincare, everything where people can— I kind of stole your idea with people coming in and being able to review products just like your mocktails or your drinks so they can add share your glow spark, like what's working for you. Yeah.

Um, so I mean, it just looks beautiful.

I'm like, I've hired before. It even looks like you.

Oh no, it's not me.

But like, I know it's not you, but like there's a very similar vibe there anyway.

But no, and somehow it knew, I guess. I know. No, I know, that's what I was impressed at, is that it somehow figured out who you were as you talked to it and gave it answers and stuff like that.

So I don't know, Yeah, that's cool. Yeah, I think that's pretty neat. And then will you teach us like about monetization of these things, or is that something that we talk to Chad about?

Yes and yes. Um, so yeah, we'll talk about that for sure. Um, we're going to talk— we'll talk more about sales and marketing a little bit tomorrow, but it really, it really is supposed to permeate through the entire process, right? That's why, you know, I had Travis say, hey, you know, you've got this product, you know, requirements document, whatever to push through, but does it really know the why in the business plan? Right? That's, that's where if you'd start from the beginning and have it understand some of that stuff, then it can, it can be really, really helpful. So, all right. For the next few minutes, guys, while you're, you know, maybe your Repwood's still going and maybe you're going to just poke around your site a little bit and things like that. You know, I've got, in fact, I'll share my screen just for a second, but, but we're going to talk, we're going to talk a little bit more about, So yeah, here's my review, what you drink, whatever. And I think it's okay. I think we've got a good start here. I don't think we should be drinking this lotion, you know. So again, this is just a mockup, so I'm not super worried about it. But we've got a couple of pages here. We can go from sodas or cocoa or herbal tea. See, I don't like the formatting of this, right? So we'll need to fix that. So we'll learn about how to kind of make changes and things like that as we continue to go through. And it looks like the tags aren't even working here. If I click on, on this classic hot cocoa by Ghirardelli, you know, it's, it's good, right? Check price on Amazon. That's cool. See, guys, I could, I could just put my AI to work and have it auto-populate thousands of drinks and find the Amazon links for me automatically and set up affiliate with Amazon, which I already do actually have my own. I never use it, but I have an affiliate account set up. But, but I could have it actually turn this into a working website and then I just have to start Driving traffic to it. And that's obviously some of that marketing stuff. But, but anyway, so maybe you'll see me, maybe you'll see me, uh, if you follow me on social media talking about Drinkable and being like, hey, you know, check out your favorite, um, your favorite drink there. So, um, okay. So anyway, I'm going to stop sharing right now. What I want to do right now actually is I want to take a moment to just kind of step back from the computer for a minute and talk, just just talk a little bit about, about what we're, what we're seeing here. And we kind of started the conversation, you know, Brittany kicked it off perfectly, like, I don't need a marketer. And I'm like, yeah, I don't need a developer. And, you know, I mean, it really, you know, people, a lot of people want to say, yeah, this isn't going to affect the job market. A lot of people are like, it's going to destroy the job market. And the reality is, I think both people are right. I think there's, I think there's a huge amount of impact this is going to have. I want to play a video for you guys really quick. Let me just pull this up here and we're going to share screen. It's only a 2-minute video. This is Eric Schmidt. Eric Schmidt. Anybody know who Eric Schmidt is off the top of your head? Anybody? Yeah. So chairman of Google for many years and, you know, very, very wealthy guy, very clued in, very connected in. This video was— this clip was actually from April 20th of last year, okay? So consider that as you listen and how much it's even developed since then.

But let's, let's listen to this. We believe as an industry that in the next 1 year, the vast majority of programmers will be replaced by AI programmers. We also believe that within 1 year, you will have graduate-level mathematicians that are at the tippy-top of graduate math programs. So that's 1 year. Okay, what happens in 2 years? Well, I've just told you about reasoning, and I 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 they're now somewhere around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement, is the technical term. So what happens when this thing starts to scale? Well, a lot. One way to say this is that within 3 to 5 years, we'll have what is called general intelligence, AGI, which can be defined as a system that is as smart as the smartest mathematician, physicist, you know, artist, writer, thinker, politician. I call this, by the way, the San Francisco consensus, because everyone who believes this is in San Francisco. It may be the water. What happens when every Every single one of us has the equivalent of the smartest human on every problem in our pocket. But the reason I want to— 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's much more interesting after that, because remember, the computers are now doing self-improvement. They're learning how a 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, so what do you guys think about that? You guys think about what he said there, and remember, this was a year ago, and a lot has happened in that And it's not even a year ago, but a lot has happened even in the, what, 8 or 9 months since he said that.

I think a lot of people are in denial so badly about AI. Like, when I have conversations with people in Europe— my family's from Europe— they don't believe that anything will be— my sister keeps saying people will never allow this to happen, that AI will replace humans. I'm like, I don't think we have a say in it at this point, you know, but they're, they're just in denial that this is not happening, that AI is not going to be as— be able to replace jobs and all that stuff to that extent. I think unless you're like always, um, on top of the trends and you see how fast it's happening, they think it's like some next generation type of deal, you know.

Yeah.

Where are you from in Europe?

Where's your family? Poland. Poland. Very cool. Um, yeah. I agree, I agree that— and you know what Eric Schmidt said at the very end? He's like, people just can't even get their minds around— we don't even have language to talk about what this actually means, right? What else?

What are the thoughts, guys? Thank you, Marta. Just in business, um, it's all, it's all available to you, Jon, if you don't get on board with this. But your, your differentiations of all this work you put in is anybody else can have it at their fingertips. Like you said, a brand new person coming in can, can, can build it better than what you established over 10 years. So if you don't— I mean, we always say, you know, Blockbuster, Netflix, but it's even bigger than that. It's a bigger gap than that is if you don't get going on this, everything you built for 25 years is gone. I mean, it doesn't matter. It's going to be irrelevant. A lot of that stuff, if you don't take this on because you can't compete.

You can't compete if somebody, if somebody at their fingertips has all that. Yeah. Yeah. I mean, I could, I could in a month, like a, a hardcore month, I bet I could build something that is as good as Balanced Staffing in terms of all the systems and things you have in place. Now, what I don't have is the experience and the existing customer base. Sure. So that's definitely where you have a benefit. But, but if you, if you stay there and you don't do the AI stuff and all that stuff and you let these guys catch up to you, well then that's your own own fault, right? But if you already have that established business and that name and you've already got momentum and all these things, and then you apply AI, well, yeah, now, now you're uncatchable, right? Because you can just move at the same speed as everyone else that's starting now, but you've got this massive multi-year, tens of millions of dollars of head start on them, right? And so, um, I think it's, I think it's crazy, right? Like, you know, by the way, when this video first came out back in April, And I'm posting, I'm talking about all this stuff. So many developers are just like, AI sucks. AI can't write any code. And I'm like, dude, I'm doing it right now without developers and making tons of money. Like, you could say what you want, but I'm doing it, so you're wrong. Right? Today, developers aren't saying that anymore. It's crazy. In just a year, developers are not fighting. Well, I mean, there's a few, there's, there's like the, We're on this big adoption scale or whatever, right? We're past the middle of the adoption level. I taught a group of student athletes at BYU last week, and I asked at the very beginning, there were 8, it was probably about 70 students, about 70 students in there. And I said, raise your hand if you have never used ChatGPT. No hand went up. 100% penetration. If I had asked that, I'm not kidding you, 12 months ago, I bet it would've been 20 to 30% of hands would've been raised. This year, the world changed. Everyone now uses AI, and that fundamentally changes. Now, the one thing I will say, so I asked them then, I said, okay, how many of you guys have used an AI platform, an AI system other than ChatGPT? And only a few hands went up, right? Out of the 70, maybe like 5 people, their hands went up. And so yes, this general like, hey, have we kind of replaced Google, or do we know that ChatGPT is available to help us with things? Yes, we kind of know that AI is there. But how many people have used things like Replit and some of the other systems we're going to use, some of the more agentic systems, Manus and ClaudeBot and things? We're going to talk about some of these other more advanced things. And, and a bunch of other ones in between. But the reality is, uh, it's amazing how much has changed even just in the past year. But what does this mean? What is— what do we do? How do we approach— and I'm not saying I have the answer to this, I'm legitimately asking you guys as my think tank today, okay? How does the world operate when we don't need developers anymore, or marketers, or paralegals, or accountants, or whatever? Because AI can do all of it. 100 times faster at a hundredth the cost.

And it's impossible as a human to compete with an AI.

They're gonna bring in universal income. Okay, that is an answer.

I personally believe that our government and systems that would supply that are not capable of doing so. I think that it allows you a little bit more time and flexibility to be creative and be more idea-focused on new things because you can be efficient in what you need to. I look at a lot of what we were paying consulting fees for and outsourcing a lot of these things. If you have one person on your team who's familiar and trained with all of these products, you now get to keep that in-house.

You improve your profitability and now you have more time for creative processes and ideas. But you know what's scary, I think, about this is that we're in this room. So obviously we're not as scared as, you know, nothing will happen to us. For the most part, we're progressing with, with the AI. But as we're laying off all kinds of people, especially people in their 50s, 60s, 70s, or whoever, whatever they were doing, they're not going to adapt as easily. And that's going to collapse the system because there'll be so many people who are not believing this is happening. I truly see they're not adopting it. They're hoping that somehow it will pass by or maybe it's not going to be as extreme. And they're not learning the basics. They may know ChatGPT, but that's not going to help them keep a job, right? And I'm just worrying about like, how do we deal with these masses that are not able, you know, because there'll be some jobs available with AI, you know, creating websites, whatever we're doing, programming, vibing, whatever. But a lot of people will be displaced. And that, to me, I've watched so many podcasts with some of these catastrophic guys talking about, like, you know, I mean, it's like end of the world type of scenario they're presenting because it's just too fast and we're not having time to adapt to it with the massive layoffs.

You know, that's what's scary to me. Yeah, you hit a very key, uh, element there, and I kind of mentioned it to Vardip a second ago. I said, I don't think our, our government systems can adapt fast enough, right, to provide UBI, right? Or in Elon's, you know, perfect world, UHI, universal high income, right? Um, Elon talks about a post-work world. I don't think that world exists. I don't think the world works without work, right? And it's because I don't think humans actually can survive without work. I think that we are meant to work, right? Not that we're meant to just do backbreaking work in the fields all day, every day, right? But I think we're meant to create, we're meant to contribute, we're meant to have stewardship over our families, our creations, our businesses, or whatever it is, right? I think that we need that as humans to actually survive. And so if we're in a post-work world and people are like, yeah, remember the Renaissance? Okay, you know how many problems there were during the Renaissance, societal problems and breakdowns and dark ages and things that happened that were a result of a lot of that stuff. And that's just not even close to what we would be talking about here, right? The reality is when you combine AI and robotics and then quantum, which we have, which we won't get into in this course, we don't have time to really dive into quantum.. But for the purposes of this discussion, quantum means the speed at which computers can calculate things is a billion times faster than it is today. And that's actually underestimating, understating the, the power of quantum. So when you add all that stuff in, it is impossible as a human to compete with them to do that. And then, so I did like what Brittany said about you know, having to be creative. And that's good. That is work, right? It doesn't necessarily mean we're all just going to become poets, right? But I think we're always going to have to push the boundary of what's next, right? Where are we going next in the universe? What is the next medical breakthrough? What is the next this and that and whatever, right? And yes, AI will be able to help and assist if you do it right and if you position yourself right and if you learn how to work with it. But if you just bury your head in the sand kind of like Marta said, maybe her family is doing and many people, right? Just using Marta's example here. You're not going to be in a position to win. You're just going to be sitting around waiting for UBI to show up and be like, why is no one showing up to help me? Right? I think that— That's my dad.

For sure.

And it's a lot of people. It's a lot of people. Well, it happens every generation, right? Like, look at how many new job titles we have. Like, you know, with the, with the output of something taking over another job, a new job is created in its place.

I feel like sometimes, sometimes, and historically that has been decently accurate. If you go back to like the Industrial Revolution, you look back at tractors and, and you look at the internet and you look at computers, you know, all these major revolutions. Mobile, social, all these things. Yes, it absolutely created new jobs. Here's the biggest difference this time. The internet couldn't surf itself and social couldn't post by itself and the tractor couldn't function by itself. So it shifted jobs, but now AI can, and I know I'm using this word incorrectly because it's not exactly true, but AI can think for itself. It can act, it can be an agent to go and take very complex tasks and finish it all on its own, right? To build the website that, that any of you built or the one that I just built right here, right? It had to think through all kinds of stuff. Now again, it's just a really just a mock at this point, but, but it, but it was able to think through, okay, how are we gonna really, how are we gonna make this happen? And it built this out, this level of a, a mockup. Up used to take at least days, usually weeks. And, and then you want to bring it to life, now you're talking months and, and all kinds of stuff. And I can literally, in a minute here, I can just say, hey, let's convert this to an app, let's make it work. I want you to go out and find the top 1,000 drinks and populate the site, and I'll come back in 30 minutes and it'll be done. And I will have spent $7.50 on it. I mean, that's breaking our system, right? That is breaking our existing system. And so the, the purpose of this discussion, of this part of our class here, is not to, not to scare you guys. And I know you're not scared. You're here because you're aware and you're trying to do that. But I'm just saying, those of you that, those that do take time to understand this, now you use your super intelligent brains to go figure out how you're going to navigate it, right? And how you're going to help your family navigate it, how you're gonna how you're going to get through this, what I think will be a painful transition for many. I think it's going to be a painful transition for society. I do believe, I do believe that AI and robotics and all these technologies are going to bring more food to the world, more housing to the world, better transportation, better medical breakthroughs, space exploration. I think it's all going to get better. Maybe our batteries will last longer than a day. You know, the world's going to get better, but the transition from— let's say that's 20 years from now where we're really feeling, you know, just we are in an advanced society. But this transition could be really, really, really, really rough. You get to 20% unemployment in some places in the world, right? A huge amount of the international population depends on the United States outsourcing those jobs to them and hiring them. And if all of a sudden AI can take all of those outsourced jobs, that's gonna wreak havoc in the Philippines and in India and in Bangladesh and in Poland and wherever, right? I mean, places where, you know, and it's not that the US is the center, I'm just saying the US and every other country that is producing at scale and outsourcing all these things, the number, the sheer number of jobs is just going to shrink, right? And so I do believe that it is ultimately not the government's responsibility to figure this out. I believe it is our responsibility to figure it out for ourselves and our family. And so, take that responsibility. Take that responsibility seriously. I think entrepreneurs like Tyson are going to be the type and many others, obviously. But I feel like entrepreneurs of the world are going to be the ones that are in the most control, that are in the driver's seats. And I also believe that we have this, I'm going to say, 3 to 5 year window where if you really understand what's possible now before everyone else catches on, you can just make so much money you're never going to need it again. Like tens of millions at minimum. I think any of us on here can do that, but you have to really apply it. Almost all, almost all money in this world comes from some sort of a knowledge gap.

Right?

You pay your lawyer because he knows the law. You pay your dentist because he knows how to fix that tooth. You pay your doctor for whatever. You pay your accountant because he knows the tax law. He knows whatever, right? But, but you pay my wife because she knows how to make a beautiful garden, and she'll go and, you know, set up a garden in your backyard that's better than anything else out there. You pay for that knowledge, right? There's also services. I get that too. But most of it is related to— is related to a knowledge gap. And so if you have this knowledge gap, you're up here and you're operating your staffing company company up here and other staffing companies are still working here, you're going to, you can monetize this knowledge gap by just moving 10 times faster than everybody else. So anyway, all right. Um, any other comments on that before we finish up this part of the discussion? Dive back into Replit here. Okay. All right. So I'm going to, and thank you all for participating, right? Um, we have one other discussion like that that we'll do tomorrow. On a concept called surf, don't swim. That's a really fun one. And anyway, these are some of my favorite parts of the meeting. Like, it's cool to look at cool tech, but anyway, I like to actually sit back and think about why. Okay, so I'm now going to— it's not perfect. There's lots that I need to change still, but I'm just going to go with this design. I think it's pretty good. We got ourselves a logo even, right? We just got a little, you know, cup up here, and that's fine. Review what you drink. I don't know if I love the message. I think it's okay. Maybe I could review what you drink. That's just all right.

I want it to be more like discover. Johnny, I do something with you sitting in the aisle, like you're sitting there thinking of what these drinks are.

You gotta, you gotta get tap into that person. Yeah, I'm curious. I want the messaging to tap more into the, you know, exploring the drink aisle vibe. Yes, there are reviews, but that's not the point. It's about discovery. Make it better. So I'm going to let the AI think about this for just a second, maybe just kind of tweak this a little bit, but It's as simple as that, right? All I have to do is, is just type, hey, I don't want to even mention when it first came up, you guys didn't see it, but it had alcohol everywhere. I was like, what the heck, man? That was not in my business plan. And so I had to like be really like direct about it and do that, right? And so it's formulating a plan here. You can look at all the different things that it's trying to do and you can kind of follow this while we're letting this kind of go for a second here. I actually want to talk just for a minute about, um, something we have. So let's see here, uh, where, where is this? Is this here? Yeah, okay. So as soon as you, uh, graduate from this course, right, you will get access to this folder right here that has a bunch of workflows and different things in here. Um, this is just one thing right here is overviews. You got a bunch of different overviews that kind of through this different software and tools. One thing that I want to go over here right now really quickly is just I want all of you to understand some very, very high-level, like this is a 3-minute crash course on coding, just to say a few words. Because when you're operating inside of Replit, there are some, the perfect SIP, there's that word SIP again. Why does it like the word SIP? Come on. Anyway, but when you're sitting here operating inside of Replit, you, you'll do a lot better if you know like how it works a little bit, right? And it's very, very simple. Most of you might even already understand this, but I think it's good to just say the words out loud. Okay, so you've got your front end versus back end, right? Front end, we used a front end language, p5.js, right? Other, other— here's Vue.js, React, there's JavaScript or JS, right? CSS is kind of like your formatting, HTML is web, web page Right? But this is the front end. This is what we're looking at right here, right? All these buttons and everything. This is all front end. Okay. The backend is going to be everything behind the scenes, right? It's where all the data is processed. It's where you check for authentication and, you know, permission and all these different things. And, you know, and talk to databases and all kinds of different things. But a good example of how those two things connect, right? You've got a login form where users enter their email and password. The frontend is like the little boxes where you type and the button you push to sign in. And then when you push sign in, it's going to send a little signal and say, hey, does this combination of username and password exist in the database? If so, give them a key to get in, right? Give them an authentication token, right? So you got frontend and backend, very high level. Then you have databases. Databases are essentially, I think of them like Tables, they're Excel documents, right? You've got lists of, of what it is and then all the data and those databases are everything, right? So if you, um, if you, uh, post something, in fact, I think I've got, um, yeah, in fact, I'll get back to it in a second, right? Basically the database is where everything happens, right? Where in terms of where all the data is stored. So if you don't have a database on a website, then, and you have a bunch of things you can interact with, as soon as you refresh the page or open it on a new device or a new browser, all of your progress will be lost. But if you store it in a database, then you have it retrieve that database from a login or an account, then your data would be the same everywhere. If you log in on Facebook on this MacBook and then come over to my PC right here and log in on Facebook, I'm gonna see the same friends. I'm gonna see the same comments. It's all connected to the same database. Okay. All right. APIs. APIs is where the world starts to get really, really fun with developing, right? An API basically is a connector. I'm connecting my button to some other system or my table or my display to some other system. For example, maybe you, maybe for your app, you need to be able to display the weather, right? And there's a couple of ways you could do that. You could set up a thermometer. Outside, and every hour, every day, you go outside and you check the weather. You check the thermometer, and you say, oh, it's at 60 degrees, and now it's at 62, and now it's at 64, and now it's at 69. Okay, and then you come in and you update your database every single time, and now you're keeping the weather of, you know, for your house, but you're not taking care of the rest of the world. Obviously, that's ridiculous, right? The second option would be you set up a thermometer that's maybe connected to the internet, and it sends a signal back to you. And so now you connect that via your own little private API. But the reality is there's already lots of weather APIs out there and you can just go find one and ask Replit, hey, use this API and connect it so that it shows the weather of the user's location. And it says great. And it goes and it grabs this database that's out there and it pulls all that stuff in. So it allows you to connect to other data. And this is really how the entire internet works. Every program, every system, every everything is connected via some sort of API to something. And it's really just a combination of how you connect these pieces of data and then present it to the world with an interface to use, right? That's, that's really how everything happens. Okay. Now APIs have really four, all they can do, the entire internet works through four different functions. You get data, you post it, put, and delete. That's it. So let's think about we're browsing Instagram. I turn it on, I click on Brittany Rae's profile, and she's got her latest pictures at the beach, and boom, I'm getting this data. She's got her post, she's got her pictures, and there's a whale in the background, and that's beautiful. She's got what she said. By the way, how did it get there? She posted it. That's actually a technical term, an API term. She posted that, she sent the picture, she sent the caption, whatever it is, and it then sits in the database. And then when I go and click on her profile, the API gets that data from the database and shows it to me. Okay, then she could say, you know what, I had a typo, and so I'm going to update that. I'm going to put that, right? Or how many of you guys have ever reset your password, right? You get a link. That link is actually a PUT request. You click on that, you update it, and it overwrites in the database. And puts the new password there. Okay. Hopefully hashing it so that it's secure. But anyway, and then, you know, maybe she's like, you know what, felt cute, might delete later, whatever. She decided to delete the, the beach, the beach photo. So that happened. Okay. But anyway, GET, POST, PUT, DELETE. That's how the whole internet works. It's how every social site, it's how Netflix works. It's how whatever, right? If I want to watch a movie on Netflix, how do I do it? I go and I click on my browser and it shows it's getting all of the different titles and then pulling up the thumbnails associated with those titles. I then click on it and push play because I liked it, and it then gets it, it pulls and starts streaming in a little bit different way, right? That video file. So the entire internet works this way. Our entire world works that way. So now that you understand, okay, I've got a frontend and a backend and they can communicate to each other, and then I can pull data from anywhere and have it display. Now you're a coder, now you're a developer, you know how it all works, and now you can start to kind of put this stuff together. I also really like this CRUD. CRUD is really good. Create, read, update, delete. Let's say you want to add a blog feature. Maybe I want to add a blog feature to my, to my website, and I want the ability to create a blog, review the blog just on the blog page, right? Update the blog. So you'll be able to edit it and then you'll delete the blog, right? Or maybe you create a user a user management system, right? I need to be able to create users, see the users, update the users, or delete the users, right? So you can say, yeah, build me a user management system with full CRUD functionality. It says great, and it'll do it, right? It'll build you a nice system. But because you know this phrase or this little word CRUD, you didn't have to say, okay, I need to be able to add users and I need to be able to remove them and I need to be able to update them. You know, you can just say I just need to CRUD users system and it knows what to do, right? So anyway, authentication security. Um, Replit has the ability to create something called Login with Replit. I hate it. It is the worst. Okay. I want you to, and you'll see me do it in a second here. I say, I want you to create the login system and we're going to use a JWT. Okay. That's all you need to know. All it is, is basically it's going to have a database with a username and a password., and it will check to see if it's right and it'll give a token to let you in. JWT is a very standard and secure way to set up all this stuff. You can also do OAuth, like log in with Google, log in with Facebook. It's a little bit harder. You're not gonna have to do that right now. Okay? I mean, you can later, but I wouldn't recommend starting there. Start with JWT and then you can expand later as needed, if needed. Okay? Um, hosting deployment. One of the cool things about Replit is that literally everything can be done inside of Replit. You don't have to go to another system to deploy or anything like that. We'll use this Replit document in a minute here. But anyway, that's, that's just a quick intro for that. So let's go back to Replit. So, okay, we've got this. I'm actually going to click convert to app now. And so what that's going to do is, yeah, we'll just, we'll just click convert to app. I know there's going to be more that I need to fix here. Whoops. There we go. All right. So we're out of design. We're going to app. Yep. Convert. Okay. Brittany, you won't be able to do this unless you upgrade probably. But, but yeah, I mean, hopefully, hopefully you can do that or everybody else. And if not, you can just sit and watch. But, but here what we're gonna do is we're just gonna start connecting it, right? The design is approved. Let's wire up everything. Now it did not ask me anything about, about authentication. And so right now I'm going to say, please use a standard JWT. Auth system. Do not use the Replit auth system. Copy. Go. And I will give this to you guys. And by the way, I sent it. I have to actually send the message, otherwise it will do it, finish its stuff and then read it. So I'm sending it to it and kind of interrupting its thought process and saying, hey, make sure you do this. Okay. So I'm going to go ahead and send you guys this chat. So you have that little message there. If you have a login system that you need, So got it. Standard JWT, no replit. Let me install the JWT and password hashing packages and build everything out. Okay, so guys, I have become a pretty advanced programmer, partially because I spend a lot of time reading what the AI says back to me as it works, right? I just read it. I'm like, oh, that's cool. Now I understand that. Or if I— sometimes I'll see something, I'll say, what does this mean? Just like I used to do to my CTO, right? Just be like, what are you doing here? What is— why does this matter? I'll say, oh, if I do this and then this this will work better. And now I'll catch it doing things sometimes and I'll say, that's not the best way to do this. You didn't think about this and you need to go do this. And it's like, oh, you're right. That's actually a good catch. And so I become a developer to a certain extent just by sitting here and reading. And I've spent, I don't know, man, I'm probably, I'm probably 5,000 hours live coding at this point. I'm a lot. I have spent a lot of time. I did it long before Replit was around. I was using it, you know, for years before, and I've spent just a huge amount of time teaching it. Like I said, I'm probably close to 1,000 hours just teaching it, right? It's like speaking out loud, talking to people while doing it, right? So now, you know, again, you can just follow along with it, right? You've got what is hashing. Okay, well, hashing is how it's going to keep your password so that I can't see what the actual password is. It's a is just kind of encrypted, right? What is the schema? Well, that's how everything talks to each other and kind of the layout of the app, storage layer, API routes, seed script, all at once. Somebody, hey, what is seed script? I don't know exactly what seed script is. I don't think I could define what seed script is. And so I might just for a moment while this is going, I might jump over to ChatGPT and say, hey, what is the seed script? Right? And learn that while that's, while this is going. When I was building my business in a weekend, um, I, uh, Um, and by the way, some of you may or may not know the story, but, but I, I had been doing— I, I had started another company back in 2016, raised much money, sold it in January of 2023. And then I had been doing some consulting and things like that and working on AI and stuff. But then I had this kind of just broader idea for an AI education training company, because I just saw so many people were not, were not following AI like I felt they needed to. And I was just kind scared, right? And so I said, all right, I'm going to start a business. And so, I gave myself a challenge and I publicly posted about it on LinkedIn. And this was last February 27th, last Thursday, or it was a Thursday morning in February last year. So almost exactly a year. And I posted about it and I just said, hey, I'm going to see if I can vibe code a business today and get my first customer by tomorrow. Right? That was the idea. And, but my idea actually initially was I wanted it to be completely hands-off. Like AI will deliver everything. It will be completely, you know, just, just a SaaS product that I don't have to touch. Okay. That was the idea. It didn't exactly work out that way and that's fine. Right. But, but I just kind of got to work and, and I just started, I just started building and it took me, it ended up taking me instead of, you know, I, so I Thursday, thought I could get my first customer by Friday. It ended up taking me till Saturday to finish it. Saturday at 3 AM, so really it's technically Sunday, but really 3 days of work, probably about 40 hours put in during those 3 days, maybe 35 hours. But while I was doing all of that, I, I wouldn't just sit here and read, you know, what Replit was saying. I would say, all right, work on this, and then I bounce over someone else, uh, somewhere else. All right, here's, uh, here's Eli. Try again. Let him back in. But anyway, I would, I would sit here and let Replit work, and then I would, I would say, okay, I need to create some images, and okay, I need an LLC set up, and I registered it with the state, and I need a bank account. I got that process going. I needed an EIN. I got that process going. And some of these things take a few days anyway, right? But I just kind of went through that process really fast and started all the kind of, you know, organizational things that were needed. I needed an operating agreement. I needed, you know, and obviously using AI to help me do a lot of that stuff. I needed a logo. I needed branding and all this different stuff. And so I'm going to show you guys kind of all these different platforms and things that I used to actually make that happen. Right. But bottom line is you don't have to just sit here and, you know, watch it work through its agentic, you know, chain of thought reasoning. And they're cool. It's cool. It's cool to watch and you can learn. But when you're actually really trying to get something done, done, let this go, get another agent to work. That agent's working great. Now let's come back over here and do this. Now let's do that. Right. And so I spent a lot of time. I had to build Facebook pages. I had to build Instagram pages, X page. I had to build— then what I wanted to do is build a bunch of posts to be on those pages so that the pages weren't just sitting there empty. If someone landed on it, I wanted the website to feel and the company to feel complete when I was doing it. So anyway, I, I ended up launching it after 3 days, ran some Facebook ads, and I'll talk about this more. I'll talk about this a little bit more tomorrow. But, but by Tuesday, I had my first customer. So it ended up taking me 5 days. I want— my goal was 2 days, ended up taking me a few days later, but I was just relentless in pursuing the goal of I'm getting my first customer. And I had the first one, it was a $15,000 customer by day 5. And it, you know, took off from there. But, you know, and I took took it to about $1 million of revenue within the first 6 months, right? From, from basically March 1st to about September 1st, it was about $1 million revenue. I did that completely solo with me and my AI team, no employees, right? Now I've got 9 and it's awesome and we're continuing to grow really fast and it's, it's, it's getting better as we go. But, but yeah, and we, and we now have dozens of customers doing incredible things and we're, we're really making an impact on a lot of them just really saving them. We got this one, a pool company. They do about $25 million a year and we're on their way to helping them get to $40 million a year because of the implement, the things we're implementing and the changes in the systems that are saving them so much time so they can really focus on really what they're, what they're able to spend the extra time on is finding more business and they are finding it and we're using AI to push that and find more business. It's really awesome. It's so amazing. It's so amazing what is happening now. But, but it all started from just an idea and then being relentless in pursuing that idea. And so that's something you can't really learn from AI. You just have to have that inside you and say, I'm going to go do it and I'm just going to make it happen. And that's my entrepreneurship bug that I have inside of me. So, all right. Yeah. Digital Grocery Isle. I'm still not loving the branding. That's okay.

Whatever.

We could still— I could work on that. Taste buds are bored. I like this branding better. This is cool. Let's make this the main message. I can tell it to do that later. Right now what it's doing is it's building a bunch of things behind the scenes. Okay. So right up here in Replit, you have tools and files. So I can click right here and I can see all kinds of different stuff. It's going to ask you like publishing right here. If you want to publish that and take it, you know, public and all I'll talk about that later. We're not gonna do publishing yet. Um, you also have this database. The database is important. Okay. Like I said, the database is where all your data is stored and it's what allows you to have, um, all these things. When you first start, you're gonna have a development database. When you publish, you're gonna have a production database. You'll have two. We'll talk about that more later too. But when I look at my, it's actually, I, I think very, um, smart to spend time looking at your database. You don't have to spend tons of time in here, right? But let's click on like brands and see what it has here.

Okay.

So we've got brands. We've got Olipop, SpinRift, Bundaberg. Hey, it found my Bundaberg. That's cool. I didn't even tell it that. That's so cool. So those are cool drinks. Those are nice. Swiss Miss, Langer's, Shirley's. But then I can look at this slug. Anybody know what slug means in this context? It's actually the little thing up here that will be like, you know, drinkable.com/olipop, right? It's the, it's the slug. It's, it's whatever it is. Right. And so it might be different, you know, in a, in a, in a URL, you're not gonna have any spaces. So traditional medicinals isn't gonna work. So traditional-medicinal. So it's created a slug. So now, you know, okay, that's a little bit of the structure of my site. And just by looking at that, you just kind of, you don't have to write it down. Or anything, you just kind of put that piece of information in your brain. So now you have just this, this picture and then we've got a description and then we've got a logo URL and look, they all say null, which means we haven't gotten any real logos yet. So obviously we're going to have to build something out there at some point. Okay. If we want to, right? You've got categories. We could obviously add more. Now you've got a slug. You can have like sodas slash, you know, Coca-Cola slash Sprite or whatever it is. Right., right? And then you've got some images, you've got image URLs for the categories, you've got images and brand IDs and different things. But coming in here and looking at users, I love looking at users, right? We've got ID, it'll be like ID number 1, right? And username, John, email, John, whatever. And you know, there's my password, display name, right? Okay. So it's, it's considering I'm going to have a handle like on Instagram or Facebook or X, right? My avatar URL. So I'm going to have some sort of a profile image.. I'm going to have a bio. I'm going to know when I was created and then what are my favorite, what are my favorites and what are my reviews, right? You can see all those things. Okay. So you can see all this different stuff here. And just by looking at the database, you're starting to understand how the system is set up, right? You're starting to just kind of build this puzzle in your head. At least that's how it is for me, right? So I like to always understand what the database is and I'll just kind of click through and I'll just, I usually will spend no more than, no more than 2 minutes just kind of looking at it, but just quickly. Okay. All right. I see that now. I kind of know how it's working. I kind of know how it's put together. Okay. And then I can come in here and there's a lot of other things. Security scanner is amazing. Once we get to the point where we're saying, okay, we're ready to publish this, we want to do that. Well, we want to make sure we don't have any ways people can just break in or we're not exposing API keys or giving away people's passwords or anything sensitive like that. It will automatically take care of it for you. Right? You'll just run the scan. And you'll say fix with agent. Boom. Right now you have, you can't do it while it's working, right? It says, it says an action is in progress, please wait. But, but eventually we'll do that. So it's just a lot of really neat tools in here that will help you do some really cool stuff. And then if you do have other developers that want to use your app, you can click on this Git thing. And GitHub is a way that you can actually store all of your code outside of Replit in case you ever decide to leave Replit or you want to use some other system. You're not stuck inside of Replit. Okay. All right. So now it's, it's done a bunch of things. It's actually kind of made it real, right? And so now it says, all right, publish now or configure myself. I don't want to do that yet. You don't always have to listen to this. Okay. You can, and if it's your first time, it might not be a bad idea to listen to it, but I don't want to do that yet. Right now what I want to do is I want to log in. I'm going to have to sign up. I'm going to be Jeannie Piano because that's who I am everywhere. Uh, actually, sorry, display name is this. Janie Piano, and my email will be this one right here, and we'll go password is my password. Password. Create account. Okay, and we'll save it. Nope, that's not my username. I accidentally saved the wrong username, but that's okay. But now I'm logged in. So this is important, guys. I know it seems like I just logged in and it worked, but that's a big moment. It worked, right? It's very possible that when you try to create a username or whatever, It doesn't work. And so what do you have to do? Well, you're going to have to tell, hey, I tried to create a username or register and it didn't work. Okay, so let's see if we can break it. Let's see if I can make it do something that doesn't quite work yet. Hey, there we go. I did it. Okay, so brands, brands doesn't exist yet. Okay, so I clicked on the brands page, whoops, on the brands page and it didn't work. Let's build it. Make it awesome. Okay. So obviously I can give better instructions than that, but it's really as simple as that. I go in, I find something that doesn't work, and then I go through. Now it used to be that kind of the, what I would call the QA side, quality assurance, or the testing side, you would, you would have an idea as a product designer or as a, let's say the CEO, right? And then you, you, you get some sort of a mockup in Figma or something like that, and then you give it to a developer, and then they create it, and then you send it over to someone to test it, and then they work with the developer and test it until it's finally there. Now you get to be all of these things, at least you get to direct all of these things. Um, Replit has some ways where it can actually help you do that for you, or you can, you can do it yourself, right? So I have right here, down here in the bottom, um, corner where you actually to the agent, there's a few different things that I want to just show. Okay. Number 1 is agent tools. Okay. This little piece right here, you have, whoops, sorry. You have the ability to, I would stay away from fast. I'm actually not a fan of their fast feature. Just don't do it. Low, medium, high, or max. I will tell you that this also correlates very well to how much money you spend on each task that you ask it to do. If you're on low, it's not going to cost as much. If you're on max, it's going cost more. I operate almost exclusively on high and max, and if I'm working all day on something, I'm spending $80 maybe, right? All day. That's not very much money, guys, when you're considering the types of things you're doing. And if I'm working all day, I'm building something that previously would have cost me $20,000 to $50,000, and I'm getting it done for $80. It's 99.99% 50% cheaper. It's unbelievable. I like working in high. The only difference as you click through them is you can see these little checkmarks here, right? Below, it will just generate and it will execute on a task list. Okay, medium is going to say, all right, let's, let's look at the code changes and see if I made any mistakes. High is going to say, all right, not only am I going to look at what you asked me to do, I'm actually going to consider also how this affects the rest of the app. And then max is going to say, okay, I'm going to do all that, and I'm gonna consider something you didn't even think of, and I'm gonna add this additional thing that's probably gonna help you. Right? And so it goes into a very autonomous thing. I like to kind of chill in high cuz it really does what I ask it to do, and then I can kind of stay in control. Sometimes when I'm asking it to build out an entire new set section or new thing, I'll go into max so that it just considers all the things that I maybe didn't consider as well. So there's the agentic tools.

That's, that's, that's That's part of what you need to understand there.

I also have— Can you show again where that tool is? Yeah, yeah. It's right here in the bottom left.

Right where you talk to it, right where you type to the agent right here. There's just a little settings button right here. Hey Jon, when I click on that, it doesn't— I have low as my only option.

Is there some setting or something like that that enables medium, high, and max to be clicked? It's possible. Oh, you're still in design mode. Okay. Yeah. You just have to say to the agent, Convert it to an app. Then a little thing will come up and you'll confirm it. Then as soon as it finishes making it into an app, that's when you'll have the ability to have these things here. Now, remember when you do that, if you have a login setting, remember after you tell it to make an app, to give it another message and send it and use this, please use a standard JWT auth chat system, right? This little thing, I just copied it over from the meeting chat. And remember, when you send something like this, if I send it right now, it will not— it enters the queue, right? So in order to send it, so you kind of interrupt its thought process, you just have to click this send message button. If you do that, if you have a login, any kind of login, trust me, just tell it that thing and it will save you a headache. Okay, I'm going to delete the queue because I don't need it to say that again. Okay. Um, all right. So my brands page is now made. Look how cool it followed all of the same stuff. Like I've got a brands page now and yeah, I can go to, go to Bundaberg, right? Here's the collection Bundaberg Ginger Beer. This is actually how I discovered Bundaberg was cause I loved their— I was looking for other ginger type products, ginger beer and ginger ale and whatever. And, um, anyway, so that's awesome. Right? So now the This is cool. So what I'm going to do, again, I'm just— I'm kind of the product guy now. That's my role, or your role. What is the experience you want to have? How many of you guys think you're good at understanding how a product should work?

Travis, yes.

Okay, everyone else is kind of like, I don't know. All right, anyone that said I don't know, um, whether you said it verbally or just with your cringy face um, cringing face. Not cringy face, those are different. You were scrunching up. How many apps do you guys have on your phone? 10, 20, 100, 200, 400? Yeah, I mean, I, I have like pages of folders and like just— I have so many apps. And then how many websites have you visited, right? Probably thousands.

Okay.

Do you guys all pretty much feel like what a sign-in experience should feel like? Or sign out? Yeah. Do you feel like you know what should be in a footer maybe at the bottom of a page? Or, you know, when I click on something, like if I click this all brands button, it should go back to the brands page. If I clicked on that all brands button and it went to the home, I'd be like, hey, that didn't work right. I mean, your brain will just tell you. You're all expert product managers. You really are because you're product users. And that's all we do. It's all we do in our life is you've got your bank app and your grocery app and your Costco app and your Amazon app. And you just kind of know how the world should work. And when you click on something or do something inside of your app that feels off, you're just like, whoa, hold on. And your brain just says, hey, let's fix that. Right. And I think this is really important too. When I was launching the GenAIPI website in a, in a weekend, right? What I was doing is I poked through that entire thing was I was trying to make sure that nobody felt like they were getting scammed, right? I tried to avoid what I like to call paper cuts, right? Because what would be a paper cut if I click on the about page and it doesn't work, or I click on start your AI journey today and somehow that didn't work, right? Or I go down to the footer and I see a link to an X page and I click on it and it doesn't work, right? But if I click here, oh hey, there's things and there's posts. Well, this is real business, right? And so I might not even spend time looking at it, right? But if I come here and I'm like, oh yeah, there's stuff here. This is a real page and I have, you know, posts and people and things that are going on here. And my privacy policy is there. Guess who wrote this privacy policy? ChatGPT, right? And the About Us page, right? I know I spent time writing a lot of this stuff, and I actually have a huge update coming to this. This is actually really ugly, and I'm really excited to push the new update. But the point is, it all worked, right? I got— I had a— I set up a blog. I actually didn't set this up the original, the original day, and I, I put some thoughts there, and I think that's great. But again, I need to be able to click here, and if I click on this, it needs to work. If I click here, it needs to work. I'm avoiding paper cuts. If I click share this article to LinkedIn, that should work, right? If it doesn't work, your brain just has this little paper cut and it's like, okay. And I don't think you're gonna die from a paper cut, but you're gonna die maybe from 1,000 of them, right? And so death by 1,000 paper cuts, right, is what you're trying to avoid. And so you wanna just make sure that your, your, your website works, right? It needs to go through and it needs to, Every button needs to do something that it's supposed to do.

Okay? If it doesn't, you're not done yet. Okay, Jon, can I ask, are you, are you, as you kind of put the final touches on, are you prompting very specifically, when you click this button, I need this to be linked to this video or this page or this social media?

Absolutely. 100%. So for example, if someone clicks now on my AI leadership thing right here, they click schedule strategy call. That is a specific link to a Calendly, right? And so, so I said, here's my Calendly link, right? And this Calendly link goes to, you know, it doesn't even go to my calendar. It used to go to my calendar. Now it goes to my team and my sales team that will take these, these things, right? But, and it's a shared calendar system, but I, I had to go in and grab, you know, the actual Calendly link. And then I go into Replit and I say, okay, and what I can do is really cool. Watch this. If I, if I, okay, if I click browse the shelves here, okay, that does, that's cool. What's new? What's new? Okay, perfect. So this, this, this, this doesn't work yet. This what's new doesn't work yet. So I'm going to click here again. I'm going to make this, sorry, I'm going to make this bigger so that you guys can see. I'm going to try to zoom in on the page here. So right here you've got this little, it's called element selector, right? Or select element right here. Okay. So I'm going to click on that and I can actually select the very specific button, right? And then I can actually change the text, right? I could just say, hey, let's change this to whatever, right? And I can literally see those updates live, right? What's the latest, right? I actually like what's new, so we'll keep that, right? We're not going to change that. But I can then tell it to do that. So I can actually like change the— I have control over individual elements. But I can say, let's get this button working. Um, I could give it instructions. I'm just going to say, let's get this button where I can give it instructions and say, all right, let's have an, let's have another tab that's like, you know, latest, you know, or trending. Right. Um, in fact, I like trending. Um, oh, hey, look at that. Scroll smoothly down to trending on the shelf section on the homepage. Oh, right there. That's where it would go. Perfect. So that's what it's going to do. So now I've got a what's new and here's the latest and greatest. We're going to drink this lotion. So anyway, you've got all these, these, these little tools down here. You can attach things here if you want to attach or upload a file. You can either drag it in here or click attach. You can even attach a Figma design for those of you that are familiar with Figma. If you design something out somewhere else in Figma, you can drag that in. Now here's a very, very important thing that I do in almost everything I'm working on these days. Okay. Down here in the bottom left, you also have this this build button and a little arrow right here. I can click here and I can switch it to plan mode. Build mode means it's going to take what I say and then get to work. Plan mode means I can say, all right, I need to build this new system, right? So let's, let's, let's actually do it, right? So I'll say, um, here, you know what, we'll keep it big so you guys can kind of see here. Um, I need to populate the brands with at least 100 popular drink brands. This will take some time to do manually. Um, will you help me do this? It's the best way to make this happen. Okay, so we'll just see what it says. Absolutely, I can help with that. Here's my approach. I'll expand the seed data in server/seed to include 100+ real popular or non-alcoholic drink brands with accurate descriptions. I'll research and compile brands across the existing categories. So does hot cocoa. By the way, the agent has access to just go surf the internet, right? So we can go do that research for me. You can do all that stuff. Super cool. Okay. And so then it gives me these planned tasks. Add many new drinks, brands, and, and sorry, drink brands and drinks. Update data storage with new brands. Check brands and drinks appear correctly. So it's going to check its own work. And so now it says, all right, can I go Go out of plan mode into build mode. Start building? Sure. Or hey, let's revise the plan, right? I like this plan, so I'm just gonna let it go, right? So now we're gonna go from 10 brands to 100 brands, and all I did was ask it to do it. And in probably 10 minutes or less, we're gonna have a lot of content, right? And I can keep doing that. I can keep expanding that. I can keep, I can keep making that happen. I can say, okay, great. Now let's add another. Let's, let's focus on. Herbal teas, right? Go find 250 popular herbal teas, right? Also go find the Amazon link to them. And I want you to use my affiliate, you know, thing. And so here's the structure of what that will look like. It will figure out the structure, right? I can say, hey, here's my affiliate link, here's what I normally do. And then it will go and it will say, okay, I see the structure for the Amazon link is your affiliate.amazon.com/affiliate/ Product page, whatever. Right. It'll understand that. And I can have it install affiliate links on 1,000 products for me and do all the research for me instantly. Really, really cool. And then I could say, hey, um, gimme a report on all of the drinks that did not have, or that weren't on Amazon. Right. Oh, okay. Here's that. Hey, you know what? I also found this other affiliate program for drinks and it's through Walmart, right? Or it's through Target, or it's through whatever, right? Um, can we build a system that has the ability to have multiple places to buy it? So you now, instead of saying buy on Amazon, it says buy now. You click on it, says, okay, would you like to shop on Amazon, Target, Costco, Walmart? Set up all these affiliate links, right? And then have it automatically go and do all that. And so I can set up thousands of links and change thousands of things in my website by just asking my dang agent to do it instead of taking what used to be days and weeks of just tedious doing stuff. You know, back to the conversation with Eric Schmidt, I think that what the world is moving towards and the reason I'm excited about it is because we're going to go away from doing mind-numbing repetitive robotic tasks., and we get to then move to those more human-focused tasks where we get to actually, you know, create, create in a way that, again, AI can't quite do. AI can create amazing things, but it still can't keep up with humans when you really get down to it. We're gonna do a lot more—

tomorrow's gonna be a lot more creation-type stuff. I'm gonna do some agendas. Can I ask just a couple of more questions? Sorry, I'm so curious about this. Okay. You have a couple of data points on the specific website. It says 12K tasted. You've got scores on here. Yep. Are you setting up prompts in the background to tie that data to specific as it's being input into the system? Is the system already recognizing that? Like, if you look at fine-tuning the prompts, on average, how many additional prompts are you really evaluating and honing in on before you go live with something like this?

This website— Are you testing everything? Yeah. Yeah, so this specific website, if I actually said, you know what, I like Drinkable, this is a freaking cool idea, I'm going to dedicate my life to helping people discover and talk about cool drinks, right?

Yeah.

I would probably spend easily 10 to 15 more hours working on this.

Okay. And clicking on everything.

And this 12K tasted, this is fake. This is— yeah, nothing is real. Dummy data. It's just, it's just the idea right now is, okay, here's what it could look like, right? And so obviously at some point I have to go in and I have to be smart hard enough to say, okay, this is— I don't even want that on there. Or yeah, maybe I don't want this yet because I have zero tasted so far, right? Or maybe I say, hey, you know what, if we want to do a metric, there's, there's 7,455 drinks, right? Yeah, listed here. Search, find your favorite, talk about it, add your favorite to your profile, whatever, right? You know, and so I'm gonna— I have to think through what What is this experience? Why am I here? What is the reason I'm doing this? What is the— what's the point for me as a user? And I am going to be my user. I hope you're not building products that are for other people, you know, to be the best creators are creating stuff that they want in the way that they want, because odds are you're pretty similar to everyone else, right? At least on those general things, right? Yeah. We might have different drink tastes, but we're going to have the the same idea of like, oh, it'd be really cool to be able to see what Brittany's favorite drinks are or what Travis's favorite drinks are. And so, oh, hey, let's, how does that work? Right? Like community. We haven't built the community element yet. Right? Where's, where's, what does my profile look like? Right. That doesn't even exist yet. Right. And so I have to build all this stuff out until I'm like, okay, this is starting to actually be something useful where I can say, okay, I want to, I want to go see what Travis likes and what his favorite Gatorade is. Their Gatorade flavor is, right?

So that I can give that to them when we go do a team workout or whatever, right? And I'm assuming after you publish, this is where you also will go in if you want to— let's say you wanted to add an additional landing page where people could upload something, like you're just coming back in here and you're building and then republishing?

Correct. Very quickly. Yep. So yeah, back to GenAI API, right? I've got my latest and greatest right here. This, this is my project. The same project I started a year ago. And it has a huge amount of files and code. And I mean, crazy amounts of data, right? And so I'm building this new, this new page, this new mission page right now. I'm working on GenAI API doctrine. Look how cool that is. Nice little like flowy lines and stuff. Why GenAI API exists, everything's like that. And this is actually, this is actually still a draft, but, but this page right now does not exist. If I go down on the published website, there is no mission page. There's an about us, but there's no mission page. And so I'm still working on this and I'm, until I'm ready for this to be out there, I'm not gonna push the publish, but when I'm ready, I'll come over here, republish, boom. So the last time I published was 12 days ago and I published a bunch of changes, right? But, but yeah, I come in here and this is where I change. This is where I manage. I've got my database with, look, look at how crazy my database is. Is in here, right? I've got, you know, 1,735, you know, tests that have been taken just inside of here. And those are like the free tests that are taken on the site. How many users do I have on here? Uh, 783 users, right? Those are people like you guys that have signed up for an account to be able to take this course or have taken other things like that. And, and again, this isn't even my main focus on our business at this point. This is, this is kind of like, like I, almost all of our revenue and everything is focused on AI Leadership, but all those users really are mostly in the live courses, courses and certifications, the kind of automated products that I started with. And so I've still gotten 783 customers over the last year with this kind of what I would call side part of our business. The main, the main driver right now is, is our AI Leadership and the way that we work with like Balance Staffing and many other companies. And we actually built an entire other platform platform for that, right? I built, uh, jennaiapi.ai, and this platform is, you know, where I can see our customers and I can log in and customers can see what they're doing. And this is actually where a lot of our, our work happens now. But I made that a completely different domain. Yes, it's connected with the same brand, but it's a different website. And quite frankly, I haven't even spent a lot of time on the— I actually hate the, the, the website itself. Like, it's not really— no one ever actually comes here. No one is clicking this link to buy our product. They're still buying it here. They're still coming here. And then this is really more of an internal tool. So anyway, but yeah, it's just this ongoing iterative process, Brittany, right? You're, you're constantly adjusting and adapting and figuring out what's next and what do I need to do and, and how can I, how can I really make this happen? This is cool. We're big data entry task. We use a sub-agent to build out the expanded seed file. I love it.

Um, okay, good. Other, other questions? Uh, hey Jon, I just had a couple questions. Some of it you may have already gone over. Obviously I apologize I missed the 1 hour. You're good. Um, but first off, I was wondering, just to clarify, Replit can create like integrations, things that you would use within a company interface, right?

Yeah, so yeah, we— you missed us talking about APIs, or maybe you heard— actually, maybe you were here for that. But the APIs, you can You can create your own APIs. So I could create an API between, for example, my, you know, genaiapi.org and genaiapi.ai. I could connect the user bases, right? The databases and things like that. I could connect stuff like that if I wanted to, no problem. Awesome. So yeah, you can connect to your own APIs, you can build your own APIs. And again, you don't have to know, you don't even have to know those words. You don't even have to know connect database accesses and APIs. All you have to do is know, okay, I have users on genaipi.org and I have users over on genaipi.ai. I want anybody who has a username on genaipi.ai on our, on our client dashboard, I want them to have unlimited access to log in here and have access to all of the different, um, you know, courses, like free access. They get access to all of that, right? Without having to pay for anything. Like right now I have access to AI in Business Continuum. You can go through and start modules and take quizzes and all these things. So if I wanted to add that functionality, I could very easily say, yeah, let's build that. And it will just, it'll just build it. It'll be like, great, I'm going to make an API and a connector that allows you to log in, kind of our own login with Google, right? It's going to be login with this other thing, right? And just sync the databases and the usernames. We could certainly do that.

Yeah, this is still working.

Big data entry task. I'm excited. Um, my next question.

Oh, go ahead. Go ahead. If you want to— yeah, I was just gonna say, um, can it trouble— what's the best way to troubleshoot? Is it with the AI, like, kind of going through the website itself and looking for gaps, or just you testing it out in terms of that?

So you can actually give it a task. You put it in high— in max mode. Sorry. Um, what you need to do is you actually have to to turn on this little thing, this little toggle right here called App Testing, right? And then App Testing, you can say, all right, go test every button, every feature, anything that doesn't work, build it out, make it work. I could do that. I could just go, Max, this, go build everything. And then it will go through, it will click every button. If it doesn't work, it will fix it and then come back and give you a report, right? And, and then you can go through and I, I personally, I don't always trust my AIs. It's good to be skeptical. And so I still am going to go through as a user and click on things and be like, yeah, it works, it works, it works, it works, it works, right? Or, oh hey, I found something, right? But you can actually have it. And when it's actually pretty fun to use the app testing thing. In fact, I'll have it do it in a minute here where it can actually do an app testing thing so that you can see. It'll actually like take control of the screen and go through and actually like act like it's a human or click through and do things.

So you can do it manually or you can have it automatically do it. Both, both work. Yeah, I had a question. Go ahead, buddy. Um, to make the website, can you use—

what's the difference? Like, can you use WordPress? Like, what's the difference? So WordPress is, um, it's a very powerful system It's insanely complex to set up properly, and once you get it set up, it, it's okay, but it's still very, very confusing, and it's all manual. So no, you cannot— you can't really use AI to help you do your WordPress other than like going back and forth with ChatGPT and saying, what do I do here, and help me write this and do that really quick.

And so this completely replaces the need for WordPress or or anything like that.

You're doing it all pretty much more simpler and easier to do, right? Yeah, because, you know, look, it just found 138 brands and 147 drinks across 11 categories.

Even with WordPress totally set up, like someone else just made it perfect, do you know how long it would take me to manually enter all of that crap?

Okay, I never use WordPress. I thought WordPress is simpler, so this is simpler. Okay. No way. This is 100 times easier because it can do all of the complex stuff, connect to APIs and do all that stuff. WordPress, you would be— if we just went in a competition and it's like, all right, build a beautiful website that can review drinks, right? And we just had to build this basic thing here, right? You're using WordPress. I'm using this. I'm gonna, I'm gonna build probably let, let's let's say we had a 24-hour competition, right? We start at 8 AM and we, you know, or 10 AM, and 10 AM the next day you got to show up ready to demo your product. I bet you in Replit I build 40,000 lines of code with all of the data populated nicely, and for you with, with WordPress, you would build 2,000 lines of code, maybe, and that's using AI to help you, and you would have none of the data populated.

Okay, right. And so next quick question I had was, you're—

so pretty much if you want to create a website and I spend all day, it'll cost about $80 to $100 to get set up? Possibly, possibly. This one right now, so far, I think we've probably spent— I don't know, that last task where I just had go work a bunch. Um, I don't want the shape, I want this right here where I had to go do all that research and find hundreds of things, things that would have taken me easily at least days, if not maybe a week or two to figure out. Cost me $4. Oh, well, okay. And it took 13 minutes. So now I can come into Brands here and I've got tons of stuff. Honest Kids. Honest Kids apple juice. With a mug of beer.

Yeah, exactly.

You're right. So, so obviously there's things that you're gonna have to go in and that's why, that's why Brittany, like even if it goes and you tell it to, to, you know, test it, it might not realize that, hey, you know what, maybe we don't want a mug of beer to represent the apple juice. Right. Um, yeah. Free Spirits, Spirit of Gin. But look at this, it's created all of these product pages already for me. It, it would take so long. There's hundreds of pages already created just sitting here in an hour's work. Unbelievable. Unbelievable. And then I can go through and I can actually have it find better pictures and I can make sure that the tags are right and I can figure out how to, how to then, how do I review it or whatever? Um, link it to shops, all that kind of stuff. Stuff. Here, here's a little thing. Let's say I come, I come here to Honest Kids again, right? And first of all, I clicked on that and it brought me to like scroll down. I wasn't scrolled up. That felt off.

Now if I click All Brands, where would you expect this to go?

Back to that page with all of the different—

correct. Where in the page would it would you expect it to go? I— where you left off. Exactly.

There are some times that take you to the top, and that's one of the most frustrating things. It's super frustrating. And so if I click all brands, hey, it took me back to the top. Okay, so I don't like that. I don't like that. So let's, let's, let's tell it, right? So let's say Honest Kids, and we'll click on this link here and say when I click this, um, wanting to go back to the list, I want it to take me back to where I was in the list, not to the top of the page. But when I click into a page like Honest Kids, I want it to automatically— whoops, I don't think I spelled that right. Oh, come on. Automatically. I know it understands me, but I'm just, you know, I want it to automatically bring me to the top of that brand page. Okay. So here's what I'm going to do. I'm going to turn on app testing for this. Yeah, that is on now and we're on high. And so we'll let it go.

So we'll let it, we'll let it work for a second here. Um, now sometimes— just a couple questions. Can you do multiple prompts at one time?

Like, if you find a couple of issues, can you do all of those fixes at once? Okay, so you can put it in a paragraph like this, especially if they're kind of connected. And this one, it was kind of two similar things, right? And so, all right, so yeah, two behaviors needed: going to a brand page, scroll to top; going back to all brands, restore scroll position, right? And so it will remember that, so it will save the scroll position on the brands page when navigating away. And whatever, right? And so you'll see actually it will be a more complex— sorry, it'll be a more complex URL now instead of just /honest-kids, it will say /source? This is, you know, have all these different little variables that it now saves in the URL. But again, you don't have to think through that. You don't even have to understand that. I've learned that as I've built stuff like that.

Right.

But it just— I experienced that just now. I clicked on this thing and I was like, Hey, when I click on here, hey, look, now it scrolls to the top. That's cool. I'm going to click all brands. Oh, it reset the server, so I got interrupted in my test. But what it's doing is it's— I'm starting to test the website by first setting up the browser and navigating to the home page. And so you can watch this. We can just sit here and watch it be an agent unto itself. It's a little bit painful to watch an AI agent work at it because it has to kind of think through it. It does a little bit differently, but it doesn't take that long. It doesn't take that long for it to figure this out. So it's just going to randomly click on something here and it's going to say, okay, hey, I'm at the top. Nice job. It worked. Then it's going to go up and it'll click all brands, probably. Wait for it. Come on, come on. And then it went right back to there, went right back to where I wanted it to. So we've solved the problem and now it's testing itself to verify that it did it. And you can actually see what's going on. It says, okay, there's some LSP errors. It found some errors, but all tests passed. Here's what I implemented. Selected. Clicking into a brand page, that works. Clicking all brands does this. Applied the same scroll to top behavior on individual drink pages for consistency. So it did even something beyond what I asked it to do, right? Not just the brand pages, but the drink pages also. So if I now come into brands and then click on my, you know, Fiji and then click here, it should pull up this page, scroll to the top instead of, you know, kind of down like this. A little bit, right? And then from here, I'm actually noticing there is no back to brands page or back to the brand page. So that's a problem, right? I can't get outta here. I'd have to go back. I don't like that. We need a little button right here. So on the drink page, um, like slash, whoops, slash drink slash 974. By the way, I'm seeing that right here. Right, this is the drink. There is no button to go back to the brand page. And there should be. Also, it should go back to wherever— whoops— we were scrolled to. Like we just did. Okay, so that's—

you guys could keep watching me do this all day, but, but you've got it now. You've got the basics at least, right? Sorry, Jon, one more quick question about the APIs. Are there— are you using ChatGPT? Let's say I do want a weather app. Number one, are most APIs— can you find free versions?

Do you need any special permissions, um, to integrate some of those APIs? You know, yes, yes and no. I'm actually— I was just thinking about it. I'm going to integrate one right now. We'll integrate a weather API into this, right? And I'm going to tell it, okay, for the discover page, right, we want it to— or for the homepage, whatever that we've talked about, we want it to, if it's cold outside, focus more on hot drinks. And if it's hot outside, focus more on cold drinks. Drinks, right? Kind of cool idea, right? So that could be a cool way to use my weather API into this thing. Now, one thing I don't love right now, this is kind of weird for me, okay? When I'm logged in, I'm logged in right now, but I see the homepage, like, like the marketing page. I don't feel like I'm logged in. So I think that this page, when logged in, should be my like discover page should be my like, hey, here's the cool stuff that's happening. Right. So let's talk, let's say that for a second here. Um, I wonder if I can get this to work. Hello? Hello? No. Yeah, it's not working. My dictation isn't working. I was going to talk to it, but I can just, I can type fast. So, um, um, one weird thing, um, when I'm logged in at the root page, right, which is just this— that's called the root page, right? Um, it's still just showing the marketing landing page. I want it to show a personalized page for me. Um, part of this means that we'll need to have a better onboarding. So that— to do— this is interesting. We need to have better onboarding so that it understands my preferences. This also probably means we need to fully build out the profile profile page. Okay, now there's something that I forgot, and this is important. Okay, I'm actually going to copy this for now, and I'm going to just put it somewhere else for a moment. We're just going to go like this and go like this and go like this, and I'm going to delete this. I'm going to come back to it. But yes, you can say something, you can say like, do this in a minute, right? Whatever. And it's in a queue, and then you say, then do this, right? And see, it'll just kind of that. You can send it now if you need to tell it something about what it's doing. But, but you can always give it like lots of tasks to do, right? And so I do that all the time when I'm about to get up and like go to lunch or something. I'll say, oh, let's do this, work on this, work on that, and also do this. Also write a new blog here for me on this. Do this, do that. I'll get like 8 tasks. So it's working for like an hour while I'm gone, right? I think that's really a really cool way to, to do some things. Okay. This is important and this is actually kind of a critical thing that I do differently than many other developers using Replit. So over here in the top right on Replit, you have this folder looking, I don't know, whatever this is. Okay. But this is actually your file structure. Okay. This is the actual code that has been written. Right. And so you can click on your client and look at your index.html. And this is the code that has been written. This is what's running your website. This is what makes time travel possible. It's the flux capacitor. Okay. This is what's actually making everything work. All right. So, um, we need to, we need to update it. I feel like it's not doing a great job at remembering a lot of the things I already told it I want to do, like having a profile and a good onboarding experience and understanding the social aspect. Aspect of it and these things. It got— we told it that a long time ago, right? Like way up top. And it's forgotten. It is forgotten that I gave it these, this business plan and this feature list. And we didn't even really use it, right? We have the feature list and all this stuff, but here's what we're going to do. So I have this document here. This is like tricks for mastering Replit. Okay. What we're going to do is we're going to say, we're going to copy this this one right here. I'm going to tell you what it says as we kind of do it here. Please use the attached business plan that's currently in attached assets. Okay, so we attach the assets at the beginning and it's sitting right here. Those, that business plan and the feature list are sitting here in this folder. 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 Without any code, 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. So there's another file in here called replit.md, and this is like instructions. And Replit knows to check this file for what types of things should be happening. So like JWT system design, modern nostalgia, fonts, these things. Here's the project architecture. And there's all these kind of just technical things, right? But we have this overview. This is kind of all it is, right? But we need, we need to have a lot more information. So I'm going to go ahead. I'm going to give it this. I'm going to give it this information here or this task. And feel like this other task is basically done. Skip test. We're going to skip the test so we can be done with it. Okay, done. Here's what I added. Good. So yeah, now I should be able to check my— go to the preview here and we'll just test this while it's now— it's, it's in a second, it's going to be done. And now it just received my next message. Okay. Now I should be able to click on Dad's Root Beer, click on Dad's Root Beer, and then click back to Dad's Root Beer. And it remembers where I was. Okay. I did not like this. Actually, that was fine. That was fine. All brands. Capri Sun. Yeah. When it's landing on the brand page, it's not scrolling to the top. Anyway, I don't need to really worry about it right now. Also, the average rating is fake. We only have one drink. I have no way to add drinks. For example, I don't have a user management. I need a, I need like an admin tool or something like that. So anyway, right now it's going to find this business plan. It's going to figure out what's going on. I have everything I need. Let me create both documents and update the Replit.md. So right now it should be in a moment. You'll see a couple documents. Show up here. And all it's doing is copying from the attached assets. It's not a hard task. It's just going to copy it, format it, write it all out. It takes a second because it's doing it in the way an LLM like ChatGPT does. It types it all out actually, like, like it actually does that instead of just copy paste, which would be much faster, but that's okay. So it's going to create both documents, update repl.md, and we should be in much better shape now because the system has forgotten. It's forgotten all of the things that we started to do, right? And we started with this amazing business plan and the whys behind everything, and we're trying to keep it on task and make sure we build all the different things here in the feature list and everything we needed to be able to do or whatever, right? And it's forgotten that. It's just kind of doing stuff now, and we're kind be even drifting like I was drifting, as you know, just because I was trying to get some of these things to work. So this will keep us on task. So now it's created the README.md. I can click on that. And now here's my beautiful business plan. This is exactly what it was back there. I can edit this MD file if I want. I could just click open in editor and it looks kind of weird. It looks like code, but it's not hard. It's just text. You can just type You're not going to die, right? Just type it, adjust it, whatever you need. These little hashtags, these little pound signs, this is actually how MD files are formatted, right? And so if you, if you look at it, you can see the, the README file. This makes it— the one hashtag makes it a title, the two hashtags makes it a heading 2, right? And then you have just kind of standard standard text here or whatever, right? That's, that's all it is. Okay. So I've got that. I've then got my progress file and look how cool this is. It's gone through and analyzed the entire database and it's got checkmarks for all the different things that are, um, that are finished and all the things that are not finished. So now we have a good kind of status, right? Like I don't have password reset set up or profile pages and bios and all these different things. Right. And so now we have actually have a good, a good idea of what's going on. And then in replit.md, it added project documents, full business plan, market opportunity, competitive landscape, progress, complete feature checklist. So the Replit file that gives it its instructions knows to look at my instructions also, right? Really cool. Really, really cool. This structure adds what I like to call bumper lanes, right? So that you are staying in control. You're doing all that stuff. Stuff. Okay. All right, guys, we got about 15 minutes left here. Um, obviously, you know, we can keep working on all this stuff here. Um, yes, Marta, um, the, uh, that prompt is right here. I will copy this and drop it into— oops, I clicked the wrong thing. Yep, there's that. And I will also do this, and this is a little bit ahead of schedule, but I'm going to give you guys access to this folder so that you can, you know, already start to use it as desired. Let's see here. Share, copy link. Okay, so we'll then go back to— wait, sorry, it is right— go back to here, the chat, and give you this. Okay. So there's the link for, you know, you have access to all those different things. This one is in Software and Tools and then the Replit document. There's some great tips and tricks in there. Then I'm going to go back to my Word document where I said one weird thing. I want to actually get this.

Whoops.

I'm going to go copy. And then what I'm going to do is I'm going to paste this here. Whoops. I did not do that right.

Whoops.

Copy. Paste. That did not paste very well, but whatever. Um, please refer to the README.md file as you do this so we get it right. Okay, so now I've already spent all this time with ChatGPT figuring out what is my system supposed to look like, what's it supposed to do, how is it supposed to work, and tell you, tell Replit exactly how this is supposed to go. All right, all right, so very dark, couldn't see it clearly. Okay, I don't know. Okay, here's what we're gonna do. We're gonna go like this. Okay. This is what it said. Somehow when I copied from a Word document into here, it didn't like it. And so I did— this is just a trick— Shift Option Command V. So it's just like Command V, but you add Shift and Option to it. And then what that does is it it pastes without formatting. So it'll just paste the text. So sometimes you'll copy from like one system into another one and it pastes really weird into a Word document or a Google Doc. If you want it to just paste the text with no formatting or anything, Shift+Option+Command+V. Anyway, another thing I learned from my super smart CTO. Okay. Um, all right. So yeah, when you're logged in, the homepage should feel personal, not like a marketing marketing page. This means building out onboarding flow, personalized home feed, profile page. Let me, let me plan this out and get building.

Okay.

So one of the interesting things here is it's gonna build out a new onboarding flow, but I'm already signed up as a user, so that's not gonna work. I've missed the onboarding flow. So I need to make sure that I remember as the CEO of the company that I have to go back in. And try this again. I have to start a brand new user, experience the onboarding. Oh, that didn't work well. Okay, great. I'm going to fix it. Delete the user. Start again. Start over. Start again. Start over. Start again. Now, if I need to create a new user, here's how I would do it. If I want to use the same information, I don't have to keep creating new random users. You know, whatever. I can log out and come to my database and I go to users and I can just click this and delete. Delete it.

Boom.

Now I have nobody in the database. And when I come back in here, if I try to log in, it won't work. If I try to sign up with that same information, it'll work because there's no conflicting record. Okay? So super easy. So now it's getting to work. It's building all kinds of crazy stuff. Super fun. So, all right, guys. Questions? Day 1. Day 1 and Day 2 are very different.

By the way, but before we dive into day 2, any questions on this? So this is amazing, but I'm just curious if you can build in here also, let's say your automation for social media to promote the page.

Like, is it done here or separate app or something else? Depends on how you want to control it. Right. Um, the reality is yes, you can totally build it in here. You can build amazing automations in here. You could also use things that you guys may have heard of. Like N8N or Make or other agentic systems or whatever, right? Or maybe something like Hootsuite, right? That automates social media stuff, right? But how do you manage all of that? How do you, how do you fill it with data? How do you get the AI to fill stuff in and do all these things, right? You can do all that. You can create that entire system here inside of Replit so that you have this kind of management system. So you could say, hey, I'm I want to create automated social media management. Help me build that out. I'm not going to build the tool that actually publishes it to Facebook and Instagram because that's actually really difficult and requires months of permissions and getting approved by Facebook and Instagram would be forever. You want to use something like an in-between, like Hootsuite, right? So you'll use an API, right? So you go to hootsuite.com, okay? And you'll sign up for it., but it's not, it's not super cheap, right? This is actually one of the more expensive ones, but it's really good. And if I'm paying monthly, it's even more expensive. We pay for this as a company, but then we have access to APIs and things, things like this. There are cheaper ones, by the way, guys, you could use something like Buffer, buffer.com. It's awesome, right? Pricing $5 a month. Okay. So if you're looking for something like really, really cheap and whatever, and there's still good ways to automate this stuff, but this allows you to automate automatically post. Both of these things allow you to automatically post to multiple platforms with one, one place. But again, the, the amount of work they've done to make those connections to Facebook and LinkedIn and Instagram and TikTok and Pinterest and X and YouTube and Threads, all of these different systems, it is crazy. You do not want to do that. It is not worth your time to try and manage that. So you buy something like Hootsuite. Suite or Buffer, and then you create something inside of Replit where you'll say, hey, Replit, I need to create a management system. I'm gonna use Buffer. I want you to create content automatically inside here and I'll, you know, build me an interface to be able to do that. And then I want you to just push it into my Buffer account and we'll say, great, let's connect your Buffer API and get it all working. Okay. And so you can, you can connect these other powerful systems so that you don't have to, again, rebuilt that thing.

Okay.

Um, I wish this would finish quickly. I think it's probably gonna take another, I mean, might be another 10 minutes.

Very quick question. Yeah, go ahead, Marty.

So you can use like Buffer, Hootsuite to make your social media posts and they'll automatically create the ad or the post and you just post it yourself or they'll post it. No, Buffer is not going to create the content for you. Buffer is going to be the automated scheduler, right? So you basically use— you, when you're in Buffer, you'll type, um, had a great day today. There's your post, right? Here's my picture of it, right? And then you'll click post to X, post to Instagram, post to Facebook, post to this, and schedule. And then you could schedule it for right now or for tomorrow morning at 9 AM or whatever it is, right? Or at the end of the day, 7 PM, whatever, whatever it is you want to do. Okay? And then you post another one. And so you could have, you know, hundreds of posts just auto-scheduled out for the next month. All your social media is done, but that's a lot of work to go and schedule 100 posts. So what if instead you connect an AI on the front end to say, help me generate 100 posts on this topic for the next month. Okay. It generates all those and then you say, send these 100 posts to my Buffer account and automatically schedule them. So, like, Replit, will you do that? Or ChatGPT? Replit will help you build the interface for that. Okay. You'll connect via an API to ChatGPT, and we're gonna do more of that tomorrow. We're gonna do some more API stuff tomorrow. I wanted to do it right now. I wanted, I wanted to say, for my personalized homepage, I also want a weather widget. The weather, the current weather, whoops, weather at my location would influence the drinks that it suggests to me.

Okay.

And so then what I'm gonna do is I'm gonna go over and I'm gonna grab a weather API and I can, you can Google it and find one. There's one that's called weatherapi.com. Okay. This might look complicated and you're seeing JSON and XML and all this stuff. It's so easy. Okay. First, let's look at pricing. You'll understand this. Okay. Look at this. 1 million calls per month to this API is free. Once you get up to 3 million or you need to get, yeah. Once you get to 1 million and 1, then you're going to have to start subscribing to the starter package, right? $7. But even at 5 million a month, you're only paying $25. I promise you, if you have 3 million calls per month or 5 million calls per month, you can afford $25. I promise you, you can afford that. Okay, but anyway, you'll come in here and what you'll do is you'll grab your API key. It's free, right? I'm going to log in here because I know I have one already. Okay, you'll literally come in and you'll grab your API key. Here it is, it's right here at the top. You'll copy it you'll take it back over to Replit and you'll say, here's my API key, set it up. Right. But anyways, here's how we'll do it. Actually, I'll show you. I also want a weather widget. The current weather, my location would influence the drinks that suggest to me, right? Hot drinks when it's cold, cold drinks when it's hot. Not exclusively, but maybe there's a section on the on the personalized homepage for today's drinks based on weather. I have— and then what I would say is this, actually. I have an API key for you for weatherapi.com. Ask it from me when you're ready. Okay, so it would ask you automatically, but I'm going to just go ahead and send this message right now. It's working on something else, but I'm just, I'm going to interrupt its thought. And then when I interrupt its thought, it knows it's at this step right here. It was working on this. It will do this new thing for me. And then it will, it will come back and finish what it was working on. Okay. So here it is. Here's my weather API key. Now here's the cool thing. It already has it in here because it was already in my Secrets Manager. This is my Secrets Manager. I have in my account, link account secrets. Yeah, here are all of my account secrets. I have all these different APIs that I've used for different things. Okay. And so I don't have to go back and get it. So I already have it. I just say continue, done. And so now it will go ahead and set up a widget that will show up on the homepage that will display the weather at my current location once I'm logged in. Okay. API key is ready. Let me now build the weather-based drink suggestion feature. I'll add a weather API point on the backend and then add a weather widget section to the personalized homepage. Awesome. So now it's got the weather. It knows it's kind of snowy today. And it's 20 whatever degrees. It's 43 degrees. It's warm. Well, warm considering. But anyway, so it'll do that and then it'll set all that stuff up for me. And then Hey, maybe I don't like the way the widget is. And hey, move the widget over here, make it prettier, make it less dramatic, or add a little image so that it shows a cloud on a cloudy day instead of just showing the number. That's all just whatever you feel like your experience is or should be like. Hey, you know, it'd be cool if it had this. It'd be nice if it had that. You don't even have to be a designer. I am most certainly not a designer. I'm an artist. You put me in front of a piano. I'll blow your socks off. I'll knock your socks out. But you give me a pen and paper and ask me to draw something, or you ask me to design something in a system, I am— it's not my strength. It's not— I'm not good at it. But I know kind of what I would like to see, and I can at least describe it.

And if I can describe it, AI can build it for me, and then I just iterate from there. Okay. Um, if your business plan is to do— get money from ads on the site. Yeah. Um, obviously just same kind of prompts.

Is that considered like an API if you have ad integration? Okay, yeah, so there's Google Ads, there's all kinds of ad networks, and they have exactly what you're talking about. It's an API-driven thing. You're going to go and you're going to set up an account. You could even jump on ChatGPT and say, hey, I'm— I need to build an ad network for this. What's the best one that you recommend considering what I'm going doing here. I want it to be real. I don't want it to be scammy. I want it to be really kind of on point with just drinks, for example, or whatever it is, right? And then it will say, yeah, go over to AdMob, set up an account, grab your account key, take it over to Replit. Replit will ask you for your account key and it will connect it up for you. And then just start serving ads, right? Yeah. Yeah, pretty, pretty straightforward. Pretty straightforward stuff. Um, anyway, guys, Great job on day one. We'll see you guys same time, same place, same link tomorrow. Same spot, 9:00 AM to 1:00 PM, at least Utah time where I am. Tomorrow will be much more focused on bouncing through a bunch of different platforms. We may still keep building some things in this project, but maybe we'll add like a voice agent or something like that. And we'll add— so we're gonna do some voice tomorrow. Tomorrow. We'll do some music, we'll do some images and video, we'll do agentic use. So we'll do a lot of like, like really more— this is an agent that's a developer. We're gonna do more broad agent stuff where it can do anything for you, right? So we'll do some fun agentic stuff. Um, and, um, yeah, we'll bounce through a lot of platforms tomorrow. It'll be fun. Um, you won't have to sign up for all the platforms that I'm showing tomorrow. You just kind of watch in a lot of and please ask questions, of course. But we'll just kind of show how you can use AI to just fill in all the gaps. And we will, we will focus a little more on, again, marketing and sales tomorrow, which is, you know, it's so easy now to create. But still, even if you can create, not everybody is great at selling and turning it into actual money. That's where a lot of people run into problems. And so hopefully some of the things I share will, will help you. Man, this card is in a bad spot. Look at that. This what's new thing, it's just kind of in the way of everything anyway.

All right guys, thanks so much. Any questions before we finish up? Uh, thank you.

I, I do have just some quick last questions, but it might be slightly redundant for everyone else cuz the hour right now, I don't, you probably already went over it. By the way, um, you will have access, I will send out the, um, the link to this, uh, pretty, pretty quickly here, like this afternoon.

So you'll be able to go back and watch this so you can catch up. So I've been consulting with CEOs and for them, would you recommend to build an app like on Replit or a, like something simple, like a custom GPT in a project for them to use as an executive assistant, like having an AI that's grounded in their data and to make decisions?

A custom GPT, if you're building that on chatgpt.com, which we will do some of that tomorrow, guys. Is a terrible long-term tool. If you need something just quick and dirty, it's, it's fine. But if you need to actually be able to pull in live data and have things working, you need to build your own custom thing in Replit. Otherwise the data becomes stale very quickly and it's really not very powerful. Um, you need to build something that's much more intelligent than that, hopefully tied into like their Google Drive or something.

You can't do that that with ChatGPT, with custom GPT in the same way you can if you build something totally custom. So, and the next is, would you recommend for me with my marketing workflow, like all the, different AI platforms I use to make a video, would you recommend I just get an NAN subscription or that I build my own agent kind of interface with Replit?

So I personally build everything inside of Replit and Replit does have the ability to actually do a lot of those NAN end type workflows. You can actually create automations inside of Replit. We'll, we'll show some of that tomorrow. So thank you. No problem. All right guys, hold on. I wanted to just see if I could get in here and, and like, I don't know, just answer a couple things here. Um, it's giving me my, um, good comfort. All right, start exploring. All right, so I've got today's pick. 42 degrees in West Mountain. It knows exactly where I live. Feels like 42, 57% humanity I've got, or humidity, not humanity. And then, hey, some warm drinks. This is cool. Apple cider. It doesn't have the images right. And I'm not worried about that yet. I haven't gotten there. But I've got these different blends here and I've got hot cocoa and stuff. It's working. That's cool. That's actually doing what I want it to do. And then, hey, based on your taste, here's some other things you might like. Here's some hot off the shelf. Something, discover something new, keep exploring. But now I have a much better homepage. This is really cool. I go to my brands. We still haven't built out community. Lots of work to do on this thing, but, but it's turning into something real.

So on that note, we'll let you guys go.

Thanks guys.

Thank you.

It was amazing. Thanks, Jon.

See ya.

Thank you. Bye, Jon.