Cohort 23 - Session 1 - 11-18-2023

Okay.

Yeah. I'm not sure how I got this, the two meeting links.

I've never had that happen before.

Yeah. Andrew's trying to join and he says that there's another meeting in progress. And so I don't know if that's kind of why.

Yeah. He's— could you just send him the link for this one really quick? Oh, here he is.

Okay, I got it. Yeah, I'm sorry, I was on the different Zoom links.

Yeah, I know.

I have no idea where the second meeting came from or how the system did, but obviously I need to check it. So just FYI, tomorrow we will be on this same meeting. Okay.

So yeah, just use this link tomorrow and we will, we will be good.

Okay. All right, guys. Well, welcome. I'm Jon Cheney. I think Most of you know that. A couple of you I've known for a long time, like Morgan Maurer here. This guy, we go way back, back to high school, back to high school. I haven't— I think Morgan, I saw you maybe at like our 20-year reunion. That was the last time I saw you. High school reunion.

Yeah.

But yeah, good to see you here. And then some of you I met as recently as just a few days ago. Drew and Andrew. I'm excited to have you guys here. Met them over at an entrepreneurship conference that happened last Friday. So, and Jabulo, nice to finally meet you.

Hi, Jon.

Thank you. Thank you so much. Nice to meet you as well.

Yeah, for sure. We've been, we've been going back and forth a lot for a while. So I know all of you guys to some extent and excited that we have a relatively small group here today. You know, sometimes this does get up to like 20 people and You know, it's good. We have good meetings, right? But this group is— this group size is like primo because we can actually talk, we can have discussions, we can argue with each other and yell a little bit. And anyway, we can have a good time and get to know you. Now, there won't be any yelling. I don't think we've had anybody yell in 23 cohorts. There have been no yelling, right? Other than like passionate, yes, you know. So, well, guys, really excited to have you here. You know, this course is, is very different than any other AI course I've ever seen. And I've looked at a lot of them. I'm in the business and I've, I've looked around to see what people are doing. I've had many people take this course that have taken other courses and then they've just told me, they said, hey, you know, this is very, very different. It's different in a few ways, but two primary ones that I want to point out is one, I really actually view this more as an entrepreneurship course. With, with the AI as a superpower, right, that's going to help you, you know, push forward and actually, you know, get things done. And, and the second piece is we don't spend a lot of time on the fluff, on the, on the, you know, hey, let's dig into what an LLM is and how does an LLM work and what is AI really. And, you know, no, we're just going to get busy. We're just going to start using it. It's going to be very practical, very hands-on. In the email I sent you guys last night— and I'm sorry I sent it a little bit later, I normally send it, you know, 2 days before— but hopefully you guys had a chance to just jump in, sign up for a couple things. You should have ChatGPT, most of you probably already did. Groq is optional but very helpful, and we'll talk about why in a bit. And then Replit, right? And I gave you guys a link for that. Replit is definitely— out of all the things you're using today, that's going to be the most important because we're going to be diving into that and using it a lot and really understanding how it works. So, so anyway, guys, it's very practical, very entrepreneurial in focus. And, and I really want you to understand, I do, you know, we won't cover a whole lot of like the technical, like what is the LLM and stuff like that. We'll talk about it a little bit, but we will talk a lot about the why, right? The why behind it. Why does this matter? And think about it, right? And this is very much an interactive discussion. This isn't just me sitting up on screen. For those of you that tuned into the AI Advantage event that was last week, that was very much not interactive. That's just someone talking on a, you know, a talking head and you just kind of, you know, cheer when they want or clap like Tony Robbins does, you know. Anyway, he's a fun guy. But anyway, guys, this course was not— I did not set out to build a business that teaches courses, right? And at this point, this course is probably about 5% of the revenue that, that my company makes. Early on, I made a lot of money with this course just because I wasn't doing some of the other things that I've figured out as time has gone on. But I started this company, Jenny IPI, March 1st of this year, and I did it in a weekend, right? And, and, you know, many of you are here because you saw that story or you heard it. But, but just the brief, you know, kind of 2-minute version, or maybe even 60-second version, is, you know, I kind of gave myself a challenge. I, I post every single day on LinkedIn and on Facebook and Instagram. And some of you follow me and see that. And, you know, I'm sure you enjoy a few of those posts and other of them you're like, okay, this guy's just trying to post every day. But I really, I don't use AI. I don't use AI to post. I really actually think and I say, hey, you know what, what do I want to say today? Occasionally, if I'm saying something longer, if I'm writing an article, I'll use AI maybe to edit, right? To act as my editor. I'll write the whole thing. Like, I just write it, right? I just let my conscience stream and I write. And that's one of my talents. I enjoy the writing process. It's on, right? And but then I'll take it and I'll say, okay, hey, don't change any of the language. I want you to just tell me, hey, do I have any misspellings? Is there anything that could be a little clearer? Is there any point that I need to add a little bit more things to, right? But I'm really not having it write for me, right? It'll change maybe 2% or 3% and then I'm ready to publish. But anyway, I sat down and that's important because we'll talk about that a little bit more later. I sat down, I said, okay, what am I going to do today? And it was a Thursday morning. I said, hey, you know what? I've gotten pretty dang good with AI. I've gotten good at vibe coding. I've gotten good at this stuff and I'm an entrepreneur and I've I've got an idea. I wonder if I could build this business and start it today, build the software out, build everything that I need and get my first paying customer by the next day.

Right. That was the goal.

And ended up taking me 2 days or 3 days, really. It took me Thursday, Friday, Saturday, right, to finish it. Like 12-hour days, staying up till 3 a.m. type stuff. And, and, and then I started some ads and by Monday morning I woke up. So, the ads started basically Sunday morning at 3:00 AM, right? Saturday night, really late. And woke up Monday morning and I said, "Okay, you know what? I wonder if I got any sales?" And I didn't have any sales yet, but I'm an entrepreneur, I'm a sales guy. And I said, "Hey, you know what? I wonder if some businesses would be interested in buying this?" And I just jumped on LinkedIn. I started reaching out to a few CEOs in the area where I was like, "Yeah, I think their business probably has enough people where this could make sense." And by the end of the day, I had a I had a sale. I had my first $15,000 contract, and, and I was ready to rock, right? And so, you know, within, within a very short 4 to 5 day period, I was able to start coding a website as a non-developer, by the way. I don't know how to code at all. I was able to then build that thing out, launch it, have it work, and sell my first contract. And I realized in that very hyper, you know, condensed timeframe, I was able to pivot a little bit. My original idea was I'm going after consumers. I want consumers to just come and see how good they are at AI, take my tests, you know, get a certification, take my training courses, things like that. That was kind of the original idea. It was all going to be automated. I wasn't going to do any live teaching. This was something that just came after. But anyway, I pivoted over to B2B and, you know, because I sold one, right? Hey, that's a market signal, right? So, I listened to that market signal and said, okay, I wonder if other businesses would be willing to do this. And after I've sold it once, I, I just took that same pattern, went to a few more people, and within about 6 weeks I had almost $200,000 of, of revenue, right? And, uh, revenue booked, I guess, is the right way to say that.

How quickly do you think, knowing what you know now, yeah, could you have done this business idea? You said it took 3 days, and, and obviously you playing around with it, trying to figure out how to do it, and you still done 3 days, which is commendable. But I'm wondering what you now know and what you've learned How quickly could you do it?

Probably half the time. It's still a decent amount of work. It probably took me 35, 36 hours to get Replit to really put everything together, and I'll show why, right? You really have to— yeah, like the GenAI API is actually quite complex. The website itself is quite complex, um, and there's a lot of different moving pieces. Now you could obviously, if you're just building a website that just needs like a booking something, that, that's not gonna be very difficult, right? But, um,, but because it had, part of what I had to do is I actually had to build curriculum, Morgan. Like I had to build the training courses. I built 5 different courses out and I had, even though I was using AI to do it, right? I still had to be like, okay, these are the topics we wanna cover. I had to kind of look through it, make sure it was good. Right. And then go through it. So it just took time to do that, to go over it.

Sure.

Then I went back and forth with AI and I actually, I spent probably 3 hours building. I, I wanted, to build what I called the AIQ test, right? It was just a quick test to see how good are you at AI? How, how, how, you know, how good is your brain at understanding and using AI? And so I actually debated and learned for a lot, a long, for a long time, for hours with ChatGPT about like, okay, what is the right way to build a test that really is going to show, you know, an employer or themselves or me how good this person is at AI, right? And so I had to kind of learn instructional design. To build this test. Then I had to build out the functionality of the test itself. I had to build out the databases that, hey, what's all the logic on how we choose what type, what's the right balance? It was a lot of stuff like that. My business was granted probably a little more complex than, like if I was building a CRM, if I just said, hey, I'm going to build a CRM for painters or for people that build pools or something like that, right? I could probably do that in about 6 or 7 hours.

Right.

Right. So it just depends on the complexity of what you're doing.

Right.

Sure.

But yeah, there's no question I've gotten faster. Now that was not the first time that I had, I had been vibe coding, right? I had, I had been doing that for a while and I'd learned it about probably that skill really probably about 6 months earlier. And I'd been doing it a lot. Right. And so I was like, okay, I'm pretty good at this now. I wonder if I can now just build this whole business in a weekend.

Right.

So anyway, built it, launched it, pivoted, found the thing that stuck. And then I had bunch of people say, hey, you know, that was, that was cool. Cause I kind of, I talked about it on my socials. I did, I kind of lied, kind of built in public, you know, and people, people just said, hey, will you do a course on how you actually built the business? Like the business is cool. Right. But I want to learn how you did that thing. And so that's where this course came from. And, you know, I don't know how many of you guys have taken courses online to learn how to build courses online. I have for sure. I, you know, I, and I kind of, hate it. I kind of hate that business model because you're not really— you're just like teaching courses to build courses to build courses. Like, I don't know, it feels scammy, right? Um, and spammy really, because you don't really add any value. You're just teaching people how to— yeah, anyway, the things that you'll learn in here, I really, really hope you'll be able to use. And not— maybe none of you will do courses. Fantastic, right? Most people that have taken this thing have not built a business that teaches courses. They build some other cool thing. I had a guy build a a volleyball software that's kind of like Cameo, but for volleyball players. So you'd film yourself, you know, or someone films you doing a spike or whatever. You're a high school volleyball player, you submit it to the platform and Olympic-level volleyball athletes will look at your thing. They'll draw on it, they'll give you feedback and send it back to you. And you can, you can go back and forth with professional athletes on, on that. Right. And, you know, I had a, I had a 60-year-old mom that, you know, has, has like, I don't know, 80,000 followers on Instagram or something doing this. She does baking and she wanted to build a baking website. She'd never had a website for her business and she wanted to sell cookbooks and things like that. And she built that, right? I mean, you can build anything you want with the stuff that we're going to learn today. This is really going to be very focused on you guys, what you want. If you don't have an idea, a business idea, totally fine. For the purposes of today, Just pick an idea. Maybe you got 6 ideas. Okay, you're going to learn this skill on how to quickly stand it up, quickly test it, and then move forward, right? That, that's what you're going to learn today. And so Drew and Andrew, I know you guys have a business, right? Amazing. Maybe rebuild it today from scratch, right? Even though you've got, you've already got, you know, a good, a good starting point. That's awesome. But don't worry too much about like, I don't know if it is. I've had some people take this course and by the end of the course, they've built 3 things. They built 3 businesses because it's so fast. But when you learn this really nice kind of sequence of events, this process, right, you're gonna, you're gonna say, okay, I can, I can do this.

Okay.

All right, we're gonna take a 10-second break while I go grab a water bottle over there because I worked out hard this morning and I'm thirsty.

All right, Costco case of water bottles.

All right guys, um, so to start, we're going to dive right into some AI cool stuff, and we're going to start with Grok. Now, um, one of the cool things about Grok, and I think Grok 4.1 actually came out like last night. Didn't even announce it. They're just like, yep, we came out with it, right? And it's done, which is really cool. Um, Groq 5 will probably release later this year. Um, AI is not slowing down. Anyone that thinks AI is slowing down, uh, just isn't paying attention. The, you know, there's pundits out there like, yeah, Nvidia is overbought and all these things.

Okay, you know, whatever.

This isn't investment advice, but I'm just going to tell you guys, we're just getting started. We are just getting started. The demand for those chips from Nvidia and AMD and Intel and all these different places, Qualcomm, there are so many, so many opportunities ahead of us, and the rate of improvement continues to increase.

It's almost vertical. It's almost literally vertical.

So we'll talk about that a little bit more later, but today we're going to start inside of Groq. Groq is, it's made by Elon Musk and his team, xAI, and it is, it has a few things, it has a few things that are really, really good and a few things that I don't care for so much. One of the things that Grok is that I don't care for so much is you can kind of turn off the NSFW filters and have it give you stuff that, you know, I probably wouldn't be looking for. ChatGPT, you can't turn that stuff off, right? It's like, hey, we're going to, we're going to stay super professional or whatever. Although, you know, Sam Altman has said that he is going to unlock some of that stuff, but it hasn't happened yet and people are pretty mad at him for it. He might. He might backpedal on that one, which I think he should. I think ChatGPT has a good place in the market to be a general use for everybody. They don't need to get into pornography to win this game. Like, that's just ridiculous, right? But Grok, anyway, so they do that. So if you leave either the standard settings on, you're gonna be safe, right? But you can go and you can turn off some settings and you can, you can let it do some things you probably shouldn't. But beyond that, the thing that I love about Grok, first of all, its answers are oftentimes better than some of the other platforms just because of the way it reasons. A lot of people think that every LLM, it's like going between Chrome and Safari, right? And you go to msn.com and it basically shows the same stuff, right? If you got it on Chrome or Safari, you're basically going to have the same experience. And that's because a browser, all it does is render code. It just shows you what the code is, right? That is not what is happening inside of Grok versus Anthropic, or Grok versus Claude versus ChatGPT versus whatever, right? They actually are trained completely differently. They are trained on their own unique data. Now, are they trained on the same Grapes of Wrath text? Of course, right? But the way that they're trained to act, to respond, to analyze, to think about what they say, to think about what you say, to challenge you. They're trained to do that differently, right? And that's from master prompts and a lot of other things that are under the hood. But one of the things I really love about Grok is its ability to get real-time insights into what is going on in the world, right? Because it has access to X, it has access to Twitter, right? ChatGPT does not. No other platform has access to real-time social posts. And so Grok becomes your market researcher, right? If you're like, hey, I wonder if people would buy this thing, you can probably find legitimate answers on Grok, right? Because people are talking about every subject all around the world, you know, male, female, Democrat, Republican, you know, whatever, whatever, everything you want to, you want to look at, however you want to segment your business out, you can pretty much find out what's going on on X and you can do it very, very quickly with AI. Okay. But today we're not gonna start actually with, with the, with, let's see here real quick here.

We're not gonna start with, sorry, I've got like 1,000 things open and, you know.

Jon, quick question. Would you say Grok is superior to ChatGPT in your opinion? Like if you had to go to one or the other, you only could choose one?

If I, if I had to choose one, I would choose ChatGPT.

You would?

Okay.

And I'll show you, I'll show you why. Okay. I'm gonna share my screen now. This is just a watch section. Sometimes I'm gonna say, all right, let's jump in. In a second, I'm gonna have you you guys do this stuff. There you go. 4.1 is actually a major release of Grok. Boom. That's awesome. Actually, I should have called 4.1 to 4.5 because the jump is huge. Let's look at this really quick just as we can. Creative writing is better. Personal tone control is better. Emotional intelligence is better. Big drop in hallucinations.

12% down to 4%.

That's amazing, guys.

That's really cool.

3x fewer hallucinations. That's bad math. It's not how you do that. But that's okay. Anyway, we're gonna go into SuperGrok here. So you can go to supergrok.com or whatever, grok.com, whatever it is, or you can go to x.com and just click on this. That's what I like to do. We'll be using, yeah, we'll be using this. I'm just gonna keep it on auto and actually, I'm actually gonna use fast today just for the purposes of what we're doing here. Fast is gonna be more than capable of what we're doing here. You know what, guys, before we do this, I want to pause for one second. And I meant to do this earlier. Let's take just 30 seconds per person. I want everyone to introduce themselves. I want— I want you to just go really quick. We'll start with Dave, Nate, Jabouilh, Morgan, Andrew, although it says Andre on yours. But anyway, Drew, and then Jon. That's just the order I'm seeing on my screen here. But I want to hear your name, where you're from, and what you do. And then whether or not you know how to code.

I want to know that.

Okay, uh, Dave, go for it.

Oh, hey, uh, my name is Dave Saller. I live in Meridian, Idaho, which is just west of Boise, so about 6 hours or so north of where you're at, Jon Cheney.

Yeah, I lived there for a summer. Oh, did you? Summer sales. And out— I was living— I was— I living in Meridian? I think I was in Meridian, and I'd be there in Nampa and just all around that area.

But anyway, okay. Yeah, beautiful area. Love it. Been here for about 19 years. I know nothing about code. In fact, I intentionally made it not a career choice when I was in college, and I opened up a software development book, I was like, no, no, I gave it the good old college try for about 5 minutes and said, no, not, not for me. And so what I do is right now, my wife and I are acquiring a business. So we acquire businesses. And then my career has been spent developing and growing small businesses, whether it's on my own, with other people, or inside the umbrella of a big corporation.

I love it. Very cool.

That's awesome. I go up to— I go up to Boise all the time to kayak up the Banks.

That's right.

That's where I'm hanging out any time I can.

Well, next time you swing up, you'll have to let me know.

We can do lunch or something.

Absolutely.

I will.

Nate.

Hey, nice to meet you guys. I'm originally from South Dakota, but I'm now living in Utah. I've been living in Utah for quite some time, just south of Provo, so Springville area. I'm finishing my degree at BYU right now in information systems, so I do know how to code. That's right up my alley. That's all I do every day. So I did summer sales. I sold solar for 5 years. And I kind of got the bug for business. I like business and sales and I love tech and coding. And so it's kind of just a nice overlap between the two. So I'm excited to learn some more.

How much AI have you used up to this point?

I use AI every single day.

Yeah.

Are you using Cursor or what are you using primarily?

Yeah.

So it's kind of interesting, especially in, in the core right now, because in previous years, obviously we didn't have AI. So learning how to code, learning all your JavaScript and so forth. You had to learn syntax. Now it's like you don't need to learn the syntax as much, though you should still understand it and know it, obviously.

Right.

I honestly— Anthropic with Claude is obviously the best for coding itself. I use ChatGPT for darn near most of my things, though, especially just in the learning atmosphere. So I can take blocks of like blocks of code, and if there's some syntax or something I don't understand, I'll just say, hey, explain this to me line by line. And that's kind of the, the learning side that I use AI for every day, like constantly.

I love it. I love it. So good. I love having a developer too. We, I usually have one or two in, in this and that's, that's awesome. I also love your career path with summer sales and coding and now business. Like you are primed to just take off on anything. You've got all the skills. You're very well-rounded. I love it.

Jabulo. Hi everyone, my name is Njabulo. I'm from Lineasia in South Africa, Johannesburg. What I'm currently doing right now, I'm in sales, vehicle sales. I work for Toyota under a big mobility group called Motus, and I don't know how to code, but I'm just very, very curious, and I've been playing around with it here and there. I like using ChatGPT. Especially for a lot of stuff with research, creating images, and, well, other many, many stuff on a regular basis.

Very cool. I love— I've never been to South Africa. My parents have been like 20 times because they go to Africa all the time and they just always fly into Joburg. But, but anyway, my Siri is set to the South African female voice. That's just— I just like the sound, right? So I like it when you talk to each other. But anyway, I'm—

you should come by.

I, I absolutely plan on it, and certainly you're the only person I know there. So, um, anyway, um, glad to have you from other side of the world.

Go ahead, I'll show you around.

I love it.

My name is Morgan Maurer. I live in Highland, Utah. So, and right now I own a background screening company, so I do credit criminal background checks for landlords and for pre-employment. And had that company for a while. And I've had a couple other companies, one that I sold and one that I winded down, but this has been my full-time gig of developing this company and growing it.

How big is that?

How many employees? Honestly, we don't— we got 2 in the Philippines. Yeah, me and my business partner. And then we have a developer that we have on 1099. We, we have a really good system. We don't need 11.

Good. I love it.

We have everything working together that we— I don't have to have employees. I can double it without getting employees still.

I love it.

That's awesome.

Very cool. All right, Andrew.

Sweet. Yeah, so my name is Andrew. I'm from Highland as well, currently at UVU. I won't speak fully for Drew, but Drew and I are in a program called Sandbox. Where you take 2 semesters and instead of taking like regular classes, you just get to work on a tech startup. And so Drew and I are both non-technical, but we are eager to learn and been working together on something so far. So we love it. And Drew met with Jon at, like he said, at Boost, the Boost conference last weekend.

So yeah, we're excited to be here. That's awesome. I'm so jealous. I wish I had the Boost or sorry, the Sandbox program when, when I was at BYU, that would have been really— I would have been in that. In a heartbeat. So anyway, that's awesome.

Drew? Yeah, like Andrew said, so I'm Drew. I'm from Bluffdale, Utah, obviously currently living in Provo and going to BYU. Um, and yeah, that's kind of all I do every day, all day, is just kind of work with Andrew. And we're, um, and because of our, um, because we both are like product managers and more the business side of everything, we are trying to learn how to become technical product managers. And so we are We do know a little bit about developing. We've dealt a lot with AI over the last couple of months, but definitely not that full-stack developer that we wanna be yet. And so excited to be here.

Very cool. And last but not least, Jon.

Hello, everyone. Jon Van Dyke. I'm from Long Island, New York. You probably hear it in my voice. I had a business right out of college, and that's the only time I ever did. Always worked for someone else after that. Primarily B2B sales and advertising, then did software and financial services. Services and looking to get into the AI space on a sales basis.

Yeah, Jon's actually looking to work for me, so really excited to have him here. And we'll say as part of his interview process, he's here. But anyway, really excited to have Jon here. You related to Dick in any way? Dick Van Dyke?

No, no, I'm, you know, it's fading out. I used to get that all the time, but he's an older guy, you know, so I don't get it.

Not as many people.

Yeah, relation. Henry Van Dyke, Van Dyke Brown, Van Dyke Hat.

You know, heard it all, but I love it.

I love it. No, it's awesome.

Hey, Jon, I got the Premium Grok. Do I need the Super Grok? I didn't know there was a—

Oh, Premium's fine. Premium's fine.

Yeah, yeah, yeah.

Good.

Yeah, yeah. As long as you can— Yeah. And Grok is— we're gonna use it a few times today, but, but I want you guys to just be aware of it and to use it probably more than you do right now. It's, it's quite an amazing tool. The thing that I really love, guys, is having more than one LLM. If you just live inside of ChatGPT, you put yourself at risk of bias, of hallucinations, of, you know, not, not being able to really get the right answer. Right. And then, by the way, Nate nailed it. Claude is the best at coding. And actually, underneath the hood for Replit is Claude. Right.

It is amazing.

Right. And that's why, that's why I love Replit. We'll talk a lot about that and how we, how we use that. But okay, guys, perfect. We're going to dive in now. So let's— let's— thank you. Thank you for the interest. Nice to meet you all. And I guess I'll do me. I'm Jon. I live in West Mountain, Utah, which is just a little bit further south from, from Nate, from Springville. I'm looking at Springville there. I can see the Y on the mountain there. I've got a little 5-acre ranch up here on the mountain and I have a tractor and, you know, that's cool. But anyway, so we'll talk about a little bit more about some of the things I've done in the past, but when they—

when it becomes necessary.

So what we're going to do right now to start, guys, and some of you may have done this before, this might be super easy and rudimentary and whatever, but, but we're going to, we're going to do this anyway. So we're going to use Groq to code. Okay. Even though some of you aren't coders, that's fine. We're going to just do something. We're going to build something really fun. So I'm just going to say using p5.js, build me a mind-blowing visualization.

Hey Jon, I'm just in Groq. Should I be in SuperGroq?

Is that a paid platform? Yeah, you're—

if you're on groq.com, if I need to do anything pay, I didn't do it yet, just let me know, I'll do it right away.

Yeah, just use it for free for now for this. If you haven't paid for that one, that's okay, you'll, you'll survive. Um, show me where I can deploy it.

Okay, let's just, let's just say that.

All right, here's a mind-blowing P5.js thing.

Okay.

It looks absolutely stunning. Full screen. How do you know? Anyway, it says copy paste this entire code into the P5.js web editor. Okay. So this is pretty cool. There's a couple things that happened here. Number one, I asked it to use P5.js, which is, um, what's called a front end language. Okay. There's no database, no backend, no APIs, none of that kind of crazy stuff. Right now, all we're doing is using something that's visual. Okay. Front end means it's what you can see. It's what you can see right here on the screen. Every button, everything. This is all the front end. Okay. So all I have to do is just click this little copy button and then I'll click right here on the editor.p5.js. And I'm gonna come in here. I'm gonna select all and I'm gonna click paste and then I'm gonna click play. Okay.

And all right.

Why does it want to use my microphone?

It's cool.

All right. Hey, look cool. We got an error. Okay. This happens all the time. That's totally fine. We got a couple errors. So what I'm gonna do here is I'm just going to copy this stuff.

Okay.

And I'm gonna come back to Groq. I'm just gonna paste this. So I don't even have to tell it, oh, I got an error, right? It's just gonna be like, okay, here's what's going on. I'm gonna rewrite this. Man, it has an audio reactive particle. So it, it's using audio.

This is kind of cool actually.

All right, so we're gonna let it go down.

It's writing code.

You actually have to scroll down and make sure it's done. Um, before you actually copy it.

So I don't know, maybe I did that wrong.

Let's grab this, we'll paste it again.

Play, play please. Well, my screen is like— we're just going to refresh this. It's like trying to do it but then it's struggling.

Reload.

I've never had that happen before.

What's it trying to do? I've also never had it try to do an audio thing. It's really struggling. Let's change the audio.

I don't know.

Let's do something that doesn't require microphone because I've never had it do that before. And this is actually a fantastic part of the lesson. This is part of what we're trying to do here. Let's open this up again here and get rid of that. All right, so we're just going to do something that's just awesome. Okay, looks like a $50,000 digital art installation.

Okay, okay, Grok.

Okay, we're just going to paste this in and click play.

Man, come on, man, give me something.

Give me something that's going to work here.

Copy, paste.

I usually get a one-shot, um, correct thing here. All right, what happened here?

It didn't tell me what actually happened.

Okay, copy and paste. Play.

Finally.

Okay, we got something working here.

Okay, I'm hearing something in the background. I'm not sure who that is.

Um, but anyway guys, so I was able to get something. This looks like a, I don't know, Windows 2000 screensaver and not like a $50,000 art installation, but you know, whatever. Um, it likes itself. I can't do anything if I click. It's not interactive at all. It's expecting something here. I don't know what's happening here, but basically it's just going to— I made something. Now here's why. There's a couple of things I want to go over. Number one is, first of all, look at how easy that is. It only took me, you know, 90 seconds, even with trying to, you know, troubleshoot this to get something up and working. Okay. Before this, and by the way, it's not super long, but it's 65 lines of code. Even though there's some, you know, just explanations in here and a couple things here, but still, there's no way I could do this without AI. I have no idea how to build all this stuff. I'd have to learn p5.js. I'd have to go and learn all that language with AI.

I don't.

Okay? Even when it goes wrong, I just give it the errors and it fixes it. And after a couple tries back and forth, we're good. But here's what we did wrong. And this is what a lot of people do with AI. Okay? I said, hey, build me a mind-blowing visualization. If I gave it that exact same instruction 100 times, I would get 100 different responses, 100 different things. Okay. And why? It's because I was a bad boss. I was not a good leader to the AI. You got to treat AI like a human.

Okay.

If I were to say, hey, Nate, build me a mind-blowing visualization. You'd be like, okay, like, you know, what, what, what do you want it to be like?

Right?

You know, and I'd be like, nope, just build me that. Right?

Just go.

Right? He, he has no direction. I'm being a bad boss. I'm, I'm asking him to do something that, that he just has almost no direction on it. So what he has to do is then fill in the gaps. And that's what AI has to do. They have to fill in the gaps. And so when you're talking to AI, you want to give something a little bit with a little bit more detail. Okay? So now let's go through, let's, Let's talk to it again here.

And now build me a 2D animated model of the solar system. Build me a 2D animated model of the solar system.

Let's just— let's start. Actually, let's add a couple of things.

Add a starry background that slowly rotates.

Throw a spaceship in there that I can control with the WASD keys.

If I hit spacebar, I can shoot out of the spaceship. Have waves of enemies come in slowly at first and get faster as I— or as time goes on, or, you know, sorry, as I progress to the next level.

Okay, let's just— this is, this is a good amount of work here, but now I basically told it you know, a little game, right? So it's going to build an animated model of the solar system, and then I'm going to have a little spaceship I can fly around and I can, I can shoot stuff, right? So let's see what happens here.

Hey Jon, is this also going to be in that p5.js? Is that kind of the standard, you know, interface that you use, or language?

No, it's not. It's actually just a fun language. We won't use it at all. I might never use it again. You may never use it again beyond today, although it's kind of just a fun one to know about. There's so many different front-end languages that you could play with. We could have used Three.js, we could use JavaScript, we could use all kinds of different things. p5.js is just kind of a fun little language that's easy, it's lightweight, um, and that's just what I chose to use today. Um, but yeah, uh, all right, I'm gonna paste this. We're gonna play the new thing here. Okay, wow, okay, we've got all kinds of crazy stuff going on here.

Um, where's my Okay, here's my spaceship. Okay, it looks like that might be a spaceship. Feel like I died. Yeah, game over. Oh, that's really hard to see. Refresh to restart.

I don't know how to refresh. This is good. Play again. Hey, I got it. It's really hard to see, like the colors that they chose are crazy. Yeah, and it did it more like a, like a, what do you call that, the one that comes down? Um, it's not Galaga, but anyway, so it made kind of a weird thing. I'm gonna, you know, so one of the things I could do is I could just take a screenshot of this and then come back over here to Grok. If I can get here fast enough, drag this into here. The, the colors are really hard to see, and it looks nothing like space.

Make it look better. Make space black and the stars twinkle. Also allow me to fly around more freely with the spaceship.

Okay, so I can keep doing this, by the way. I can keep just iterating on this now. Now, one of the cool things that I did is, um, And part of what you asked, Dave, I think, was, you know, how does it know now to build in p5.js? Well, it's because it understands the context of the conversation. At the very beginning, I said, hey, I want to use p5.js, and then I asked it to iterate on some more p5.js, and now I'm like, okay, now build me this, right? And so it just says, okay, I'm going to assume he still wants me to keep building in p5.js because that's what we've been doing in this conversation, right? And so no, it's not that, hey, that's the best language to use. It's that Hey, it understands what I already mean here. And so, all right, now we've got— no, no, that's not what I want. I didn't want— okay, great. Now you've designed it.

Actually looks better. Actually looks kind of cool. I like these better.

Now apply that to the code. All right, that's cool. It took this image and just assumed that I wanted it to adjust the image. But anyway, so now it's going to build it again. So at this point, guys, it's your turn. Okay, we're gonna spend the next few minutes and you're gonna open up X and you're going to build your own. You can use Groq, you can use ChatGPT too if you want. Like if you're already, if you're comfortable with that, I don't really care if you use X for this part. But I like using X here just for fun. But I'm gonna drop into the chat here and I'm gonna paste the prompt using p5.js, build me a mind-blowing visualization. So check the, check the little chat. Now I want you to replace mind-blowing visualization right? And I want you to come up with an idea, right? It could be, I don't know, dinosaurs chasing me that I have to run away from, or it could be a Tetris clone, or it could be whatever the heck you want, right? It could be something related to your business. It could be something— although, um, p5.js really is not meant to be used to build like things for your business. It's meant to build something visual and fun like this. Maybe you just want to build a mind-blowing visualization. You can do that too, right? Build a cool screensaver that you always wanted to build. Okay, but I want you to just jump in and talk to it, and then we'll have each of you share your screen and show what you build.

Okay, so get to work. And obviously ask me questions as you go if you need, if you have questions. Okay.

Actually.

Go ahead and speak up, guys, if you have a question, or if you're— when you're ready, as soon as you have something working. Should take you just maybe a few minutes here, 4, 4 minutes, maybe 3 minutes. You can even copy what I'm doing, guys, if you want to just build a little space shooter game or whatever you want to do.

Just want you to jump in and play with it.

Jon, do you ever use, um, one AI platform, for example, to just build upon your ideas and then allow one platform to prompt engineer another platform?

Sometimes. Yeah. Not as often as you might think though. I, like in Replit, for example, I, I want Anthropic, I want Claude to be doing it. And, and not only that, ChatGPT can certainly build me good prompts and things like that, but ChatGPT doesn't have the context that Replit does. Replit can see all of my files, right? It can see the whole project. And so because of that, it's going to give me better advice. What I will do, though, is get out of build mode and go into plan mode and really talk to it about what I'm doing. I'll have it make a plan, right? And then, and then do that. So we'll talk about that a little bit more later. But, but the reason I don't like to do what you're saying there is just that the, the, the AI doesn't understand the whole project. And so it might be able to get one little piece to work right, but then you—

it doesn't understand why it had to do that in the context of all of the code.

So, Jon, does it take a long time? As far as thinking, I get thinking about the user's request and just like bullet points coming.

Does that make sense? Yeah. Make sure you click, if you see my screen, make sure you change the model to fast instead of like letting it think hard about these.

Oh, okay.

Just change it to fast if you just want it to go faster. Yeah. You can even press stop and then change it and do it again.

It'll just, it'll, it'll zip up immediately.

I just— quick, quick answer.

Now it's going. Are you on ChatGPT, Jon?

I'm actually on Groq.

Oh, you are on Groq.

Okay.

Give me, gimme a, a first time try.

Awesome.

Hey, hey, Jon, for the p5.js, like, do you search it in the web editor or— yeah, it is.

Okay.

For some reason, the web editor isn't opening for me, so I'm trying to get in there, but But yeah, it should.

It's just, yeah, it's just editor.p5js.org.

Yeah, that's what I'm on.

Okay.

It's just like not loading. So I didn't know if the site was working, but no, it should work.

It should work just fine.

Although it did freeze on me earlier. I wonder if that was related to the site.

I mean, I don't know. I don't know what's going on there.

And we're just getting code, right, from this?

Yeah. Yeah, I got one done. It's a good one.

All right, let's see it, Morgan. Go ahead and share.

All right, this is— all right, share my screen.

Nice.

Utes dominating the Holy War. What was your prompt? I said, what did I tell it to do?

I got a lot of syntax errors and stuff, so it took me a little bit to get it. I just said, build me a screensaver showing Utah Utes football players move around playing a game of football. I want them to be going against another team like BYU football and show the Utah Utes beating them.

Yeah, you have to use AI to beat us these days, right?

We won 9 straight. Listen, okay, I'm still living there. I'm still in those glory days. But no, that was, uh, that was interesting because like I said, I, I kept putting it in, um, and it just kept giving me the syntax error that said, okay, now try this, okay, now try this. And finally I just cleared everything. I cleared the syntax, I cleared everything, and I just, I, I grabbed it myself instead of copying to the clipboard. Yeah. And put it in and then it worked. So I don't know.

Very nice. Try a few different things, but very cool. Well, you got it working. Awesome. Nice, nice work. Now, obviously, guys, what we're building right now, this is a 5-minute exercise. This is not supposed to build something that you're going to even ever share with anybody. It's just to show, hey, look, we can build some code, we can talk to AI, we can iterate a little bit, work through a problem or two, and boom, we've got something working.

Okay.

All right, who else has something?

Go for it, Nate.

Share my screen here.

What? Yeah, yeah. Let's see what you got. This is a—

can you see that one?

Yep.

This is just a little game.

Anyways, got it. If they hit the bottom, then you lose.

Now that's cool, awesome.

And you did it a little bit differently by throwing it into an HTML, correct?

Yeah, I, I think that website, the— I don't know, the P5, because I threw it in here and it just isn't wanting to render.

And so what you have to do sometimes is you have to just ask, you have to say, hey Grok, remember that this is being deployed in the p5.js web editor.

You're doing it wrong, right?

And then it fixes the code.

All right, cool.

All right, who else has something running? Or who's struggling? If you're struggling, we can help too, right? Share your screen.

I definitely am. I put the prompt in and got the code and then, you know, that's okay.

Jon, share your screen. Let's see what you got.

Can you see that?

It's coming up. There we go. All right, we've got it. Okay, so you've got all this cool code, um, and did you go ahead and copy it?

I didn't do anything. A couple times I gave it answers.

I thought it was trying to get further through it, so yeah, that's all right. Okay, so go up, go up and just click that copy button right there. Actually, hold on, hold on one second. Sorry, I want you to say, um, right at the bottom it says, how can Grok help I can already see this is gonna fail. So I need to say, here, you know what I'm gonna do? I'm gonna type it to you, and then I'll just type it to the group.

I want you to format this so that I can just take it to and paste it and see it.

Okay. So grab what I just sent to the chat. Go ahead and copy what I just sent to the group there.

Right here. And then, and then paste it to Groq. Yeah, copy that and then paste it to Groq. Good.

All right, just click go.

All right, here we go.

Okay, so go, go ahead while we're waiting, click on that link editor.p5js right there. Here.

Sure. Yep, that'll work. Oh dang, it should have opened that.

Okay, that's dumb.

Click back.

It should have opened that in a new window. That's okay. All right, it's still writing. Okay, perfect. Right-click on that, open it in a new window. Yeah, open like a new tab. Yeah, right there is perfect. Do the, do the new tab. Right-click on it, do the new tab.

That'll be easier. Yeah, you're stuck on—

just close that new window that opened. There we go. All right, so go ahead and right-click on that again and just say open a new tab.

New tab.

Yep, that's perfect. Now, um, yeah, go click right there, kind of in the white next to all that text written there over on the left.

Yeah.

Yep. Right there. Go ahead and click there. And then are you on a, on a Mac or a PC? PC.

Okay. Left.

Go ahead and click Control+A and then just delete or backspace. Perfect. All right. Now bounce back over to Groq, that other tab above right there. Okay. Now let's scroll up and grab the new code. So just scroll up a little bit. There you go. All right, go ahead and click the copy button up in the top right there. Mm-hmm. No, well, maybe, I don't know.

No, let's use the one on the right.

This one? What that one is over there. Yeah, right there's perfect. And then go back to the next tab. Yep. And then just click Ctrl+V to paste it. Okay.

And click the play button over on the left.

On the bottom, I got, all right, I got, I got Play.

Yes. And yeah, the play button way over on the left above top. Yes.

No.

Oh, here.

Sorry.

Yep.

Yep.

There you go. Awesome. Okay.

You built something. There it is. It's working. Awesome. Okay. Nice. All you're doing is having Groq just kind of build this stuff out, and then, you know, you built some cool AI, you know, visual logo. With a logo in there.

Yeah, I love it.

Perfect. Okay, that's awesome. That's all, that's all it is, Jon.

Okay.

All right, very good. All right, who's, who's got something ready?

Um, I've got something ready on my side as well. Hey, let's see it. Um, let me see, let me see. Where do I share my screen again?

Oh my goodness.

This is actually secretly just a test to see if you guys know how to share your screen and know how to copy.

Okay, there we go. Share screen, um, entire screen.

Okay, let's see if this will allow me—

it wasn't not allowing me to share my screen. Um, okay, there we go. Can you guys see my screen?

Um, yeah, it looks like it just started, it's coming up. Okay, there we go.

Yeah, okay, perfect. Okay, let me go to my JS. Um, there we go. Can you guys see that?

Yeah, what do we got here? This thing looks great.

So, um, I went to Groq. Um, what I— my prompt was that using, uh, the p5.js 'Build me a mind-blowing photorealistic screensaver of a 2025 Yamaha YZ450 and show me where I can deploy it.' So that's, that's what it gave me. I'm a big dirt bike fanatic, by the way.

That's cool. Um, can you zoom out on it? Can you just like scroll back on it so that we can see? So we can see kind of what it did. It's cool because— so look, guys, p5.js— there you go, there you go. So p5.js— I see the tires, I see the tires, they're there. So p5.js is using WebGL and it's using some very basic shapes to pull this together. This is obviously not a language you're going to use to build Call of Duty. It's not a graph, it's not a high-powered graphics engine. It can do very, very basic stuff, but it's cool that it kind of thought through it. And there's probably some explanation inside of its brain where it's like, yeah, this is supposed to be a motorcycle, but But it's— yeah, anyway, that's cool. I like it. I like it. All right. Awesome. Nice job, Jabulom. Okay. All right. Next, who's up?

I can go next.

Go for it.

Okay. This is— I love— as a kid, I loved Power Rangers and the Mystery Gang. So I wanted a little show-off between the two groups.

And this is just an animated stare down.

Yeah, I was hoping to be more realistic, but I think, I think for this it does okay.

So guys, one of the cool things— um, Nate, have you ever tried building a game?

Um, I've never done a game, honestly. It's been more just, uh, web applications, not like—

got it. That's all right. Well, when you go in and you build something in like Unity or Unreal Engine or something like this, This is usually actually kind of the first level, right? It will build kind of a 3D environment and then you'll have a stick figure basically, or like this, this column or like a cylinder or something that just walks around, right? And then as you continue to build out, you build some art and you build some 3D models and all that stuff, and then you'll just replace, you'll tell the engine, hey, we're going to use this 3D model with these animations and all of these things to do that, right? And so that's okay, right? But even to build something dumb like this with the Power Rangers and Mystery Gang kind of floating here, right? Without AI, even if you knew exactly what you were doing and you were a good coder like Nate is, it would still take you hours to figure it out, taking a year to learn how to do it. Like, it's going to take a long time to figure this type of stuff out, and you can just whip it up. Now obviously, again, this is not supposed to be like super high-tech. This is actually really fun to do with your kids, by the way. If you have kids, show them that they can build cool little fun games like this. Like, they love Roblox, they love the low the low poly and Minecraft.

They love all that stuff.

So awesome, Andrew. Um, all right.

Um, yeah, I haven't gone. I can go. Yeah, my thing is kind of funny.

It's, uh, sorry, let me share the screen.

Um, it's not letting me change how the arrow keys work, so I keep, I keep trying to fix it so that the arrow keys work, but basically Restart this. Oh, maybe I think because it's screen sharing, it's not letting me.

One sec.

Sometimes you have to click. Oh, there you go.

So I tried to make this like Subway Surfers type game where you could move between 3 things and like avoid the obstacles. And I was just kept trying to get the arrow keys to work and it wasn't letting them to actually do anything. Yeah. Yeah. So I made it further and then it kind of like reversed it and undid some of the things that it made. And so that's just where I'm at right now.

But anyway, but you see, you got the power, you understood kind of like, okay, I can do this and kind of build this and play with it. Now, now, guys, the process that we're using, by the way, I'm purposefully kind of making it hard. There are easier ways to do this. And but I really, I think it's a very powerful way to teach is to teach something the wrong way and then teach you the right way and the better way. And then you can kind of feel the difference and understand why we're doing things in the way we are. Dave, last but not least. Okay.

Let me see here.

Oh, I need to turn screen sharing on.

That's why we do this. That's exactly why we're here.

All right, let's see here.

Yeah, you might have to, you might have to leave the meeting really quick and then come back. We'll just wait a second. It should only take you 30 seconds. You just exit, do that, click on the link, come back in. Yeah.

Okay, you guys seeing it now?

Yes, we are.

You're good.

Okay, awesome. So I like basketball, and so I told it to create an image of Kobe dunking over Michael Jordan.

Nice.

And this is actually, I think, the 5th or 6th iteration. The first image I got was, you know, some rectangles with, um, circle heads and, um, kind of dunking. And so I asked it to make it a little bit more realistic, but here's what it looks like when I play it.

So, okay.

Nice. So that's the best I got. I think it's telling me that the— I asked it to make it the images of Kobe and Michael Jordan more realistic, and I think it's telling me in those little text boxes that it's not going to allow me to do that.

Got it. Got it.

All right. Well, cool, guys. Thank you, everybody, for sharing. We now know that everybody can share their screen. Everybody knows they can code and That's cool. We are right on schedule. We're supposed to finish that by the first hour, introductions, all this stuff by the first hour. We are exactly on schedule. Really good stuff.

Sorry, let me turn off my notifications here because my team's going to be looking for me.

Guys, like I said a second ago, the way that we're doing that, p5.js is a low-level engine. It's frontend only. It's not super user-friendly. It's not going to give you exactly what you want. Although I have built some pretty dang cool stuff and I've seen a couple other people build things where I'm like, wow, that's actually really fun. Like you could make that an iPhone game and people would pay for that and download it. Like there you can actually build some cool stuff, right? If you just spend, spend 30 minutes on it, right? We're spending like 5 minutes on it. Right. And, and so to be able to go in and, you know, build something that fast, it's just incredible. Right. Just think about what I said earlier, right? You'd have to spend a year learning something. And then it would take you, you know, hours to pull that together. We're doing it in 5 minutes. And so for a completely untrained group, right, for the most part other than Nate, right, for a completely untrained group, you guys are saying, okay, we're going, we're skipping 99.9% of the work that it used to take to get here, right? That's what AI can do for virtually anything. Coding is obviously an amazing thing it can do. But, but if you use it right, if you point it in the right direction, it can just do, it can just do incredible, incredible things here. So, all right, let's, let's see here. So what I'm doing here, guys— sorry, I was actually just for fun, I, um, after Drew said his thing, I went over to Lovable really quick and I just said, hey, build me a Subway Surfer, right? And you know, in just a few seconds— now it doesn't work yet, the keys don't work yet. Oh, you know what, maybe they do. Maybe I was just using WASD.

Okay, why can I not Spaces jump, spaces jump.

I don't know how I'm supposed to play this. Why am I dying?

Also dying almost instantly anyway.

But yeah, but, but look at how much better, you know, a little bit better program can do, right? I don't know why it's not letting me do this, right? But I should be able to jump. I should be able to collect those coins. Maybe there's something invisible there. I don't know what it's making me do that, but in just a few minutes, I could really build this thing so that it's actually working and I can publish it to whatever. Yeah. See, why can I get past those?

Oh, hey, is, is Lovable like another— what is Lovable?

I guess Lovable is what's called a Vibe coding platform. Yeah. Anyway, somehow I died. I don't know why. I don't know why I died, but it's getting better. It's editing it right now as it, as I'm previewing it. Right. So It's getting better and improving, but in just a few minutes. Okay, slower starting speed.

Cool. All right, let's see what happens here.

Yeah, and I'm still dying for some reason. I don't know why. But anyway, I don't, I don't need to keep playing it, but pretty cool. Pretty cool that you can just, you know, whip something up really, really fast. And obviously I could share that then and and do that. So what we're going to use today, though, guys, is Replit. We're going to build a business today. One of the things that I do with every single course is I come up with a new idea kind of on the spot and build it. And, you know, I've built hundreds of apps now. This is my 23rd cohort, you know, with just this. But I've taught— I teach tons of businesses and other things, and I got to think of a new business. It's hard to think of a new business you know, every week or every, every month, every couple of weeks, especially when I've done a lot of the cool things that I like. Let's see here. My eyes just kind of went over to my bookshelf and I have lots of cool books.

I'm sure you've already built one on kayaking.

Yeah. Yeah, I've built a couple. I've built a couple. Actually, you know what? I just thought of one. So I built— I've built one that focuses on like river levels and stuff so that it can like automatically tell me when the rivers are in and when's a good time to go and all that kind of stuff and using APIs and different things like that. One thing I could do is I could build like an inventory management system. I've never built an inventory management system before. That's a fun one. I could build like, because I have, I have like, I don't know, 12 kayaks or whatever. I used to have like 30. I was running a kayak school at one point, literally out of my out of my apartment at BYU as a BYU student. Anyway, and I had just kayaks everywhere. It was awesome. But anyway, so I'm going to build, I'm going to build an inventory management system because I think that's actually a pretty useful thing for me to just kind of think through and learn how to do. And I'm going to do it just for my— I'm just going to be for all of my like equipment. What would just be kayaks, right? Okay.

I love Adventure sports, kayak— whoops, here we go—

kayaking, mountain biking, skiing. I don't know, we'll just even call it sledding. I love adventure sports and other sports. We're going to do this. We're going to make it a little bit, you know, golf, basketball, and tons more. I want—

I have tons of gear and I want to keep track of it. Let's build an inventory management system to help me with that.

Okay, then we'll go. Now I will tell you right now, this is a bad prompt. But this is what most people do. This is most, most of the time what they— how they vibe code. They jump into AI and they just get to work. And that's cool. The cool thing about AI is it's actually probably going to build— and this is Replit, right? This is, this is, you know, what I believe is the best platform out there for people that know how to code, don't know how to code, whatever, but need kind of a complete solution, right? If you look at the you know, the market cap and the total number of users and total number of things, Replit sits heads and shoulders above the Lovable and Bolt and Base44 and some of these other platforms. They're, they're fine. Lovable is fine, it can do some cool stuff, but when you want to actually deploy, um, long term— let's just start with this design here, okay? Um, and by the way, I should have read all this, right? Here's all the stuff it's going to do. We'll look at that in a second. But when you want to actually deploy something and put it in front of people for customers that just, I'll say this, GenAIPI, this entire site, everything about it is completely run and hosted everything inside of Replit. I have never had a developer even look at it. Actually, that's not true. I've had one look at it just because he was my CTO and I was like, hey, take a look at all this stuff. And, you know, he looked at the code and he's like, yeah, it's actually pretty awesome. But, but I do everything myself. I am the only person who does all of this. I built everything with Replit.

Okay.

And I've changed it. It's different than when I first published it. If you were to go on the, you know, Wayback Machine or something and look back in time at, you know, what it looked like 6 months ago, it'd be a little bit different. But, but anyway, it's, it's pretty awesome. So what Replit is doing, guys, Replit is what's called an agent. And this is different than something like ChatGPT or Groq. Okay. An agent really is kind of defined as something that can do a more advanced version of what we would call chain of thought reasoning and then can act on that reasoning. Okay, so let's explain what that— let's, let's give an example of what chain of thought reasoning is that you guys will understand really well. Drew, tell us how you get ready for the morning. In the morning, right? You wake up, you got to get ready. What's your, what's your, what's your routine?

Yeah, I usually wake up. Try to— I mean, sometimes I get caught on my phone, but most of the time I try to like get up and shower and then go eat some food. And then I work in the morning. And yeah, that's kind of my— I usually work for till about like 9, 10 o'clock.

All right, well, let's zoom in on one thing I'm sure you do every day and brush your teeth, right? Yeah. So let's talk about brushing your teeth for a second. I want you to explain to me step by step, as detailed as possible, how you brush your teeth.

Okay. So I reach in my drawer, so I open my drawer and then I pull out my toothbrush and I also pull out my toothpaste. I unscrew the cap of my toothpaste and then I put just a little, like maybe 2 centimeters of toothpaste on my toothbrush. After that, I turn on my sink and I get my toothpaste a little wet. After that, I turn off my sink and I put my toothbrush in my mouth. And I usually start in the back of my mouth and I just start scrubbing my teeth. So I go back and forth motions and I just move through my entire mouth. After that, I spit out my toothpaste and I get a little bit of water in my mouth and I swish it around and then I spit out all of that and then I wipe off my toothbrush and I put it back in the drawer and close the drawer.

Awesome. Well done. Okay, so thank you for teaching us how to, how to brush our teeth. And, um, but, but, you know, it's kind of funny, you kind of laughed at yourself a little bit as you were talking through it because you're like, man, this is like weird to talk through, right? But, but the reality is our brains automatically do this chain of thought reasoning and thinking and will take a task, right? Like, hey, I need to get ready in the morning And then we break it down, right? I go to the bathroom, I got to get ready, I got to shower, I got— I got it. Yeah. Anyway, I got to put clothes on, got to shower, I got to go, I got to brush my teeth, I got to eat breakfast, I'm going to work out, I'm going to read my scriptures, I'm going to say my prayers, whatever it is that your routine is, right? You've got these tasks. And then within each of those tasks, even down to brushing your teeth, yeah, there were like 20 different steps there, right, that you had to kind of think through. You don't really have to think through, but you had to talk us through, right? And so you take a major task, you have smaller tasks and smaller tasks and smaller tasks. That all kind of work towards this goal of I'm getting ready. Okay. And so that's what it's doing right now, right? Agents can, can say, okay, I need to build a website, I need to build a dashboard, I need to build something, right? So look, it's, you know, it's, it's building it. Look how cool this is. In just a few minutes, I've got this thing. It totally works. I can click, I can add gear. Look at that. I built an inventory management system for my own farm in 30 seconds of work. Right now. It had to work for another few minutes to just kind of build it out. But that's crazy. That's crazy, guys. Like, to be able to build software that fast. Like, we'll talk to— we'll talk to Nate for a second. Nate, take AI away. How long does it take you to build what we have right here already?

Um, I mean, it would— maybe the front end, just like the, uh, to make it look all right, uh, couple days, you know, make functionality.

All right.

But then once you start mixing in the like backend and like start mixing in actual functionality, right? I could spend a week on that easy.

Absolutely. Absolutely. And it does it again. Think about this, guys. 30 seconds of work compared to, you know, a week, let's say. Right. A good developer would take a week to get this right and feeling nice. Right. And, you know, and he'd have to do the designs and he had to do all this stuff, right? And a lot of developers also aren't great at designing, right? And so they sometimes bring someone else in or, you know, anyway, they might go to AI to help have them help with that. But here you can just come in and say, hey, I need this. And if you're building something like for your personal farm management or your personal gear management, like maybe this is something I will use. Like, hey, I know what I have. I know how old it is. I know when I bought it. Here's the warranty information. You know, there's all kinds of other stuff that I can do here. And if I build it up and I'm like, man, I have this problem and I wanted to keep track of all my gear, I wonder if other people have that problem, right? And that's how you then can turn it into a business, right? So obviously right now it's not a business. Right now it's just, you know, some random, some random stuff. And one of the reasons why this is a bad prompt, guys, is because, hey, I love sports and I do this, right? I want to keep track of it. Okay, that's cool. So I give it a little bit of information and then I said build an inventory management system. I didn't tell it what to call it, what colors to use, what type of functionality it should have. I— it's having to think through and fill in all the gaps and say, okay, I think he probably wants this and this probably makes sense. Now, is it capable of reasoning and figuring out a plan that actually makes sense? Yeah, I'm gonna include these following features, right? I'm gonna add gear items, details, browse and search by sport type, edit and delete, visual dashboard, gear condition tracking, outdoor adventure themed with vibrant Like, hey, Jon, pretty cool. Go ahead, Morgan.

Is it, is it easiest to do, do all of that kind of chain of thought reasoning at the very beginning and making just, just long, big, thought-out thing for it to build, or is it easier to take something more high level and then as it— and then once it builds it, then start drilling down and add, take away things? Is there a difference between that?

Thank you for asking, because that's a perfect transition to where we're going. Absolutely, you want to plan, and that's what we're going to do next. We're going to go to ChatGPT to do that, right? The problem is, and those of you that have ever developed something before, even been around something that's been developed, right, Dave, I bet you've been around software that's been developed before and things like that in some of your small businesses and stuff like that. And even though you're not a developer, being able to understand, okay, we're doing this for this reason, we're building this backend piece for this reason, right? They all fit together. And some of you may or may not have ever participated in like a software architecture discussion, right? I bet Drew and Andrew, you guys have done a little bit of that, right, lately, where you basically get in front of a whiteboard and you say, okay, here's how it all fits together. Here's where the data flows and here's why we're doing this and this is what we need over here and here's why we need this because this is going to scale here. And, you know, I've had the privilege of being a tech CEO for about a decade and probably about 5 or 6 years in senior management in tech companies before that. So I've been around it. I've had amazing CTOs that were way smarter than me, where they would show me and I'd be sitting in one of those meetings and say, can you explain to me why this matters? And I would ask those questions a lot to the point where I feel like I know what's going on, right? And so I have a little bit of an advantage there, just kind of understanding a little bit of software. And how it and why it fits together. But the reason that's important, Morgan, is because if you just dive in and start, you know, we could, we could absolutely build something that works doing it this way. Okay. We're just going to keep iterating and iterating and iterating and iterating. And that's cool. But then all of a sudden we're going to get to a point where it's like, okay, now we're going to add some sort of monetization here. How are we going to make money with this? Well, we didn't actually think about that. And actually we're going to need to structure it a little bit differently and we're going to have to gate this and we're going to have to do these different things. And anyway, it's harder to kind of go back and figure out how the structure should be and how the architecture should be if you don't start from the beginning that way. So, well, I'm going to let this keep working and it can do it. It's now built a database, which is cool. And now that's why it says I have no gear, because there is nothing in the database yet. Right. Gear, zero rows. You know, here's all the category, brand, condition, location, all this different stuff.

Right.

So needs attention. Yeah, I definitely have some things that need to be repaired and And it's still—

this is still cool.

I like this. So it's building out all the, all the real stuff now. And guys, it's writing like real code. Like, like, look at this. You can click on these and see all the stuff it's writing. You can figure out what it's doing. Here's 27 more things it's done. It's working on, you know, it's changing the dashboard here and it's modifying this and it's modifying that. And it's really, really fun to go in and watch that. Even though, even though, again, I can't write this stuff, I do kind of know what's going on when I look at code. I know what's happening. And, and if you just kind of are patient with yourself and you say, okay, you know what, what is this doing? It looks like there's something to do with deleting gear, right? If we're deleting gear, then we're going to have to delete this mutation. Okay, that makes sense. If you don't know what a mutation is, that's fine. Be like Nate, talk to ChatGPT and say, what does this mean? Right? If you want to learn, if you want to learn, you can, you can learn just so that you kind of understand what it's talking about a little bit more. Ask the same questions that I would ask my CTO. You can ask the AI. Okay. So we're gonna build this for real now in the right way.

Okay.

And we're gonna use ChatGPT to do it. And I am going to, yeah, we have ChatGPT 5.1 now, which just came out this week or last week, I guess. And it is awesome. I'm a fan. Okay. Before we dive too deep in here, guys, I want to talk a little bit, look at how many projects I have. Goodness, I need to go through and delete some of these because I'm not doing most of them. They've all been built for things like this. But anyway, ChatGPT is, for me, my project manager.

Okay?

There's all kinds of stuff I do. For example, I can go into my General AI Proficiency Institute. And I was working on a new tagline for GAIPI, make AI work for your business.

Right?

And so I was working on that. That's cool because we're focusing a lot on B2B. But there are dozens and dozens and dozens, maybe hundreds of discussions in here about different different things. And I can go all the way back to the very beginning. This is the first conversation I ever had with ChatGPT back, you know, when I launched GenAIPI, when I built the business in a weekend, right? You can go back to my very, very first post or my very first prompt. It's like, hey, I want to create a website that checks the proficiency levels based on an intelligence test that shows the aptitude for an individual to succeed in a world with AI. This means they need to know how to talk to AI, how to be good prompt engineers, and they need to have a good brain to figure out how complex systems systems can quickly be put together. Can you help me build this? That was my idea. That's where GenAIPI started. And from there, I then built it out and we went deeper and I started figuring out the name and I did this and that and whatever. And you can, you can read the whole thing, the whole process, right? And started figuring out, okay, how am I going to do this now? First, we're going to design this and we're going to build a test platform. They're going to do some branding. We're going to do all this stuff. And, and I just dove in, right? And so anyway, we're gonna do that today. But, but one of the cool things about ChatGPT is it has this project feature and it's not the only platform that does it. I know that Claude has this as well. And I think Grok actually has a project feature now that's kind of brand new, but, but ChatGPT's had it for a long time. And so ChatGPT has become my project manager. And the cool thing about it is this. I can go down and I can look at chats. Like, let's say, like right here, I was asking ChatGPT about You know, tell me why ChatGPT 5.1 is better or different than ChatGPT 5. I wanted to ask it that. So I've got this conversation. I can go back in my history. I'm sure there's more history. Yeah, there it comes. But I have this. And by the way, guys, I'm operating right now inside of the app, the ChatGPT app, or, you know, desktop app. You can do this also on the browser as well, but it might just be a little bit different UI. But either way, I can come into a chat and I can, you know, I talk to it about it. Okay, give me a post for Facebook to share some of this. Okay. I don't— Jon? Yeah, go ahead, Jon.

Is there a difference between chats and projects in ChatGPT?

Yes. Yep. That's what I'm talking about right now. So right now we're inside of a chat. Right now I'm inside of a chat and I can come back to an old chat and without having to catch it back up, it remembers what I was talking about. Okay, so I can just say, hey, give me a post for Facebook to share some of this. Okay, great, here's a Facebook post on this. Okay, I'm not going to share this, but whatever. I actually hate the way that AI writes a lot of the time. Um, and you know, you guys have seen all the posts on that, you know. Um, but anyway, look at this, there's no— oh yeah, there's one, there's one. Now I actually read that, uh, that Sam Altman pushed an update just in the last week that's getting rid of em dashes right now. There is one here right now. This is actually a better use of one, but it's always this, you know, it's not this, it's that, you know, I mean, it's the format. The style is very, very transparent. You can kind of see through it. But anyway, you can come into a chat at any time and it knows what you're talking about. It remembers what you were talking about. Okay, now if I go into Um, refund policy for PDFs. Okay, I had somebody— I don't know, I had somebody order some sheet music from me. I'm a pianist, right? Got my piano behind me here and two other pianos in this room and a grand piano in the other room. But, but anyway, um, that's one of the main things I do. But I was asking like, hey, give me something I can send to somebody because they wanted a refund on a PDF, um, because they thought it was a physical book. And I was like, look, I can't give you a refund because you already have the product anyway. So I just wanted them to help me write that. But if I were to come in here and then ask it about ChatGPT 5.1 versus 5 in this chat, could it answer the question?

Yes.

Is it going to be a little bit confused about why you're mixing these topics? Yes. Also, right. Can I ask it about something else I had in a conversation, you know, 2 months ago? No, not really. Right. It really can't do that very well. It— so this is talking about context. You've got this chat context, then you have your system context. Okay, we're gonna go bigger and then we'll come back to projects. Okay, your system context is right here in your settings. Okay, your settings are going to be right here in personalization. I actually have not filled this out and I have a reason for that, but a lot of people do love to do that. They love to come in here You can choose how do you want this to operate. Is it gonna be professional, friendly, candid, quirky, efficient, whatever? How is ChatGPT gonna respond? You can give it custom, you know, things like, hey, don't use emojis ever. You know, you can say stuff like that so that it gives you things in ways that you want it. You can tell it your nickname. What are you? Tell it more about you. And then one of the things that I do manage decently well is its memories. Now, its memories are things that it learns and knows about you, right? So for example, it knows that I'm fluent in Spanish and Chinese. It knows that I live on West Mountain, you know, in West Mountain on a 5-acre farm. I know it's kind of small text here, right? But as a member of the Church of Jesus Christ of Latter-day Saints, knows I'm a whitewater kayaker. The fact that it knows these things about me is actually incredibly helpful, right? If I was Jewish, I would want it to know that I'm Jewish, because if I ask it a question about God, I'd rather know what the Jews think instead of what the Church of Jesus Christ of Latter-day Saints thinks. Right. And that's— we're not here to have a religious debate, but what we're here to talk about is to make sure that ChatGPT serves you. It's going to amplify you. Right. And so you want it to— you want it to know as much about you as possible. And it's kind of a weird thing, too. It's kind of weird because we've been trained for our you know, for the last, I don't know, 30 years or so about privacy, privacy, privacy, privacy, right? You know, some of you that have lived through these changes as the internet came out and then as, you know, things started to go more online and people are like, there's no way I'm ever going to bank online. I'm never going to put that kind of data online. And hey, then all of a sudden there was encryption and security and things that allowed that to happen. And, and hey, we're never going to watch movies online. And then all of a sudden it was was fast and, you know, Netflix was fast enough, YouTube, or the internet was fast enough to handle YouTube, to handle Netflix, to handle those things. All right, knocked off my charger here. But anyway, bottom line is, um, going through this, you know, technology evolution, we've really just more and more focused on, I want to be private. And then AI comes along and says, I need to know everything. Tell me everything. What's your credit card? What's your login to Delta? Right? Tell me everything. Why? Because if you want an agent, an AI agent to be able to book a flight for you, which I have my agents trained to be able to do that, it's got to know your login to Delta and it's got to know your credit card. It's got to know your CVV code. Right? And so, you know, a lot of people worry about what can I put into AI? What can I tell AI is Sam Altman going to know? Are employees at ChatGPT or OpenAI going to know what's going on with my life? The answer is no. These are highly encrypted conversations. If you— there is actually one thing here. I'll click on settings and show you. There actually is a data control thing here where you can have improve the model for everyone on, or you can have it off. And the way that it does that is actually just looks for patterns. And it cuts out anything sensitive. It knows to not show anything like that. And so the data that's fed back into the system, it doesn't help it to know your credit card number. It doesn't help feed the system to know your credit card number. But if you're really worried about it, you can certainly turn this off and your data will just— you won't be, you won't be helping, you know, ChatGPT train, train its model. Okay, so you can certainly turn that off and stay, stay even more private, but it's very highly encrypted. And I really do trust the way that the data is flowing. I trust it enough that I can put my Social Security number in here and I can put my whatever in here and just say, hey, figure out my taxes for me. You know, I actually still pay an accountant, but you could do that theoretically. Okay, so you've got project or sorry, you've got chat context and then you've got system context and then you have project context. And the cool thing about project context is it actually can think back and, you know, I could say, hey, go back to the first conversation. What was the original idea? I want to do a post about, you know, how I started this company, right? And hey, here was the beginning, right? In the beginning, there was this, right? That's how, that's how you can use these projects. And so when, when I'm building something new, so this is going to be my, um, uh, or gear, my gear manager. We're just going to call it gear manager. Okay. And then maybe I have a subtopic in here and it might be, I don't know why they have things. There's a few more options you can have in here. But anyway, I'm just gonna call it Gear Manager. Okay? So I start this project, and then inside of here, I can do a couple things. One, I can just start having chats inside of it. Right? If I just start talking to it, it'll create the first chat, and they'll just start to show up. And eventually, I'll have a big project of lots of different chats and lots of things we're talking about.

If we already had a chat about something or an idea, can we move it then into the project?

Absolutely. Yeah. So you would actually just come over here and you'd say ChatGPT 5.1, add to project. Yep. And then just select your, select your project. Okay. Yep. No problem. You can also then add files too. Right? So you could, so maybe you're working on something for work or maybe again, Morgan, you already have a bunch of things for your existing business, SOPs, standard, whatever, emails back and forth, some proposals that you've sent to clients, whatever it is, right? Drag all that stuff in, right? Tell it as much information as possible. And then all of a sudden you have the ability to, to reference all of that and AI will just be kind of up to speed. You can also go in and connect different types of, different types of data sources, right? So you can connect actually, and you're not going to do that here. You're actually going to go in to your settings. I know how to do it on, yeah, it's right here. So you're just going to come in. So I have my Google Calendar connected. Right now. My Google Calendar is connected to every single project. Whenever— this is my main GenAIPI calendar, it's the one that I run my life off of, right? But it knows, you know, if I were to ask it a question, hey, tell me about, you know, what I have today, right? Or hey, when's the best time to do this? And I can connect all kinds of stuff. I can connect my Gmail and my Google Drive and HubSpot and Outlook and Slack and all kinds of stuff that you can connect here. And they all do different things in different ways, but For now, I really just have that connected. But bottom line is, guys, today we're going to build, we're going to build the, you know, company doing, doing this process. And I want you guys to just watch, and then I'm going to give you guys some free time to work on it here in a minute. So, Jon? Yeah, go ahead.

Sorry, real quick. Did you say you can add files from my machine to this project, to a project?

100%. Yeah.

Would you mind? I must have missed that. Would you mind showing me how to do that again?

Totally. It's just right here. So you've created your project and it's just right here. There's a little button, add files.

Oh, okay.

You just click that and then you click add files and pulls up your system and you just add what you wanna add. Yeah.

Yeah. And can I then include that? Like, I can have it read that and then say, hey, give me a summary of like whatever I just—

Anything you put in there, it instantly knows it all. Okay. It's now part of the context of your project. And that's why I love the project, right? Because you can do this. You can actually drag those files into a chat, but then you start a new chat and you would have to drag them in again, right? And so like, hey, I've got a project. I'm going to be working on this. Or I just have like my company project, or maybe I have my personal life project, right? Like, you know, or maybe it's my— I have a taxes project. And so it knows what happened last year. And so it can quickly bring me up to speed for this year. You know, there's all kinds of reasons and ways to use that project structure to be really, really awesome. Or maybe you're starting your taxes this year and this is the first time you're going to use ChatGPT to do it. And so you start a new project, but you drag in your last 5 years of returns and it instantly is up to speed on exactly everything that you have and need. Right? So really, really powerful. Okay. So I'm going to just start here. I'm going to tell it I'm going to paste the exact same thing that I Yeah, anyway, we'll just, we'll just tell it, right? This is what I copied, or sorry, what I pasted to Replit. I said, hey, I love adventure sports, another sports guy. I keep mountain bikes, skiing, sledding, golf, tons of gear. I want to keep track of it. Let's build an inventory management system to help me with that. Okay, so that's going to think longer for a better answer. I'm going to actually just skip that for now. I just want to tell it what's going on here. Oh, come on, what did you do? We'll paste it again.

I'm going to try cooking skip again here.

I wonder if I can't do that on the first answer.

I'm just curious. Yeah, that's interesting. That's okay.

All right, um, we'll let it think for a second here. Um, but really I haven't told it anything, right? Let's build an inventory management system. That phrase, let's build, could be interpreted a few ways. Like maybe it's— yeah, see, now it's planning high-level level design, all this stuff. This, guys, is a very important lesson. This is one of the most important lessons I want you guys to learn today. Um, you have to be the boss. You cannot let AI take over. It's so easy to come in here and I tell it what I want to do, and then it says, great, let's build this. Okay, core jobs: know what you own, know where it is, right? Keep ready to use, pack for trip, see history and value, optional gamification, XP for using gear. Like, it's coming up with its own ideas. I have all of a sudden lost control of my project, right? If I just— because what it's going to do, here's the interesting thing, right? Here's the data model, here's all this stuff it's building. It's kind of architecting a little bit here. Here's the dashboard, here's all this stuff, here's the gamification layer. I didn't ask for gamification layer, right? Maybe we want one, but I didn't ask for that. Now here's how you're going to build all this stuff and it's going to do this and it's going to do that. But watch what happens here. If you want, we can make this a Replit project next. Now it knows that I like Replit, so that's cool. It's building it in the right way. Here's the data model, here's the API, here's the mobile-first UI phase 2, we'll do this. Now if you'd like, next step, I can turn this into a concrete schema, start with Replit file structure for the API code, and I could say, sure. Then it would build all this awesome stuff and then it would say, next I'm going to build the marketing plan. Yeah, do that. Hey, next I'm going to do this. Great. Right. And so now what have I actually contributed to the world at this point? A description of the high-level description of what I want. And then, yes, sure, do that. Great. Thank you. Perfect. I don't like that. Try again. Yes, that's great. Perfect. I've contributed almost nothing. I have lost my position as thought leader. Okay, what you got to do is you got to turn the AI into your thought partner, and you have to be the one that's going to bring something unique and awesome to the world. Okay, so let's do that.

Okay, I want you to interview me one question at a time about this idea.

I'd like to turn this into a business because I believe other people have lots of gear and will have a similar problem.

Interview me—

whoops— interview me about the name, the colors and feel, the USP— the unique selling proposition is what that is— and why is this unique, why are people going to buy this. The monetization strategies, target audience, and anything else you think we need to discuss in order for you to help me build a powerful and effective No, I'm actually going to change effective to actionable business plan. Okay. So I'm going to copy this. I'll drop into the chat for you guys who use that in a minute. You build it with your own thing.

Okay.

But here's the cool thing, guys. When you turn it into this one question at a time thing. Okay, so I'm kind of just ignoring what it said above here. I'm just like, whatever, I don't need to talk about that. Okay, so perfect. I'll guide you through the process. Here's how we're going to do it. Question 1. Vision. When you imagine this product existing out in the world, not just for you, what's the core promise it delivers to customers? What is the single sentence that describes why this app should exist? I just need to know what, what I have, what its current— not its—

what its current condition is and where it is. I'm sure there are more use cases, but this is what I need.

Okay, so that's not exactly answering its question, but it's giving it enough information. See, it gives us an anchor. Okay, that's good. Why is this existing? Okay, who's my target audience? Who will benefit the most?

You're trying to prevent it to go down a rabbit hole that you don't want it to go down, right?

Yeah, I'm directing. I'm being a good leader. Yeah, what I'm trying to do is I'm just saying, okay, but the cool thing is, guys, it's actually really amazing how good it is at setting up an interview like this. You can have it interview you. I have it interview me about stuff all the time. I'll say, I've got this idea that I need to do for my business. I need to write an article. Interview me about it so that I can kind of open my mind a little bit. Or, hey, maybe I'm writing a church talk, right? And hey, this is what I've been— this is what I've been asked to speak about. Ask me some questions and get me thinking a little bit here. Okay, great. Thank you. Right? And then, you know, you can go about and do it. And don't, don't have AI write your talk because that's just cheating. And there's, you know, without getting too religious, religious, the leaders really don't want you to do that. They want you to think. They want you to use your brain. And that's what I'm trying to teach here too, right? I want you to use your brain, your brain to figure out this business plan. Now it's going to guide your thoughts. It's going to guide your, your, you know, plan. It's going to help you work towards a business plan, especially if you don't know how to build a business plan. Right? Even if you do, this, this is going to help you take a nice thing there. So I don't need you guys to sit here and watch me answer all these questions. It is your turn. We're going to take the next, um, we're going to take 17 minutes. So it's 10:38 right now. We're gonna go to 10:55. Okay, we'll finish 5 minutes before the hour. I want you to take this. I want you to interview me. One question about this idea. So open, open ChatGPT, start a new project. Tell it your idea. Say, hey, I wanna build this, right? And it could be 1 or 2 sentences or 3 sentences, whatever. And then copy and paste this thing as part of that prompt and then go and then answer questions. You should get through probably, I don't know, 6 to 10 questions, maybe 15 questions if you're really fast. You've got 16 minutes now. Okay? Obviously feel free to ask questions. I'll keep sharing my screen. I'm gonna go through this. You guys will have a recording of this at the end so you can go back and watch what I did. If you want. But, but that's the plan. Okay, build an idea, come up with something. And if you don't have an idea, guys, just talk about a problem. Like maybe you, maybe you, you know, I don't know, something related to, I know, Jon Van Dyke. Maybe, maybe you're like, hey, you know what, I've been a salesperson my whole life and I think I could write a book and I want to create a website that helps people become better salespeople.

Awesome. The crazy thing is, like, I— because I've used ChatGPT in my own business, it already knew a lot about me. And so when I said, hey, give me a recurring business idea that can make me recurring income based off the knowledge that I kind of had, and it went and said, here's a bunch of ideas that you could use with what you know how to do. It's the coolest thing I've ever seen.

I love that.

I love that.

Pick one and build it now.

I love it. So cool.

But just a small detail, guys. I actually like to come up with a name for what I'm doing, and I like to do it pretty, pretty early because it helps kind of just guide my thoughts a little bit. And so oftentimes I will literally go to GoDaddy at this stage and see if my name is available as a.com or something.

Takes an extra minute, but I do that.

Hey, Jon, just a quick question. Some of this work I've done in Claude already. Is there a way to like provide ChatGPT the context, like to drag a chat over to ChatGPT or something and say—

No, but you can say you can run over Claude really quick and say, hey, I need you to give me a 2,000-word summary of exactly what we've done here and discussed, and I need to give it to ChatGPT so it understands. Just tell it what you're doing. And then it'll just give you something and then just paste it over. Say, hey, here's a good starting point. Ask me any questions that you need to clarify.

Oh, okay.

Awesome.

Thanks.

Yeah.

Hey Jon, I have a question for you. Yeah. This might be like a little bit of a side note, but so when Drew and I are doing a lot of discovery, would it be good if we like ask Chat specific, like, hey, give me 10 questions that I could ask to 5 different civil engineers to get a better understanding of what the product could be? Because have you done things like that where not only you are guiding the project, but you also can use your discovery to help guide your prompting as you're going, going about it?

So what's the question exactly?

Sorry. So obviously, like, when we're going through it, but do you think we should also, like, for Drew and I, take time to ask chat like specific questions that the users think is important and then add it in? Or as we're going, should we focus on just getting out what we've discovered so far?

No, yeah, I would just focus on what you have right now, right? I mean, you can always iterate later, right? Remember, you're the boss, right? Yes, you want to pivot and iterate and things like that, right? But that customer and consumer feedback will come as you, you know, at the right time after you get it in front of people.

Okay, perfect.

Thank you.

Yeah, Jon, um, with Replit, I've never really used it much. Do you know how their, um, I guess their billing model works. I know it gives you $25 with the credit, but what is considered like, how do you eat into that credit exactly? Is it just the designing and for every design is a dollar or you know how that works exactly?

With Replit?

Yeah.

Yeah.

So every time you ask it to do something, it's gonna cost, you know, a credit or something like that, right? So you're, you know, you guys have, I think if you use my link, you probably get $10 or $15 of free credits in there as well on top of that. But But yeah, in order to build— let me just put it this way— in order to build all of GenAIPI, which took me, you know, 35, 36 hours, 3 full days, cost me about $120, $130. So I paid about $100 beyond the, you know, the whatever. But obviously I'll pay that 100 times out of 100 times because that's amazing. Today with all the stuff we're doing, there's no way you'll go over your credits, right?

You'll probably use, I don't know, $7 to $10. My guess.

So you'll just kind of get a feel for kind of what, you know, how much credit is used. It's hard to, it's hard to be like, hey, I know exactly what this is going to cost.

You kind of just get a feel for it after, you know, about 1,000 prompts.

And is it only because, like, for example, I know you can go in and it gives you a kind of an architecture of like, hey, this is what we're going to do. You can edit that a little bit. Is it only when you actually go into the, like, creating the I guess templates like, like you did that it starts costing you money, or does every conversation even cost you money?

Well, everything costs you money, but if you're just doing planning, right, if you're just talking to it about things, it's going to be like low, like 0.01. Like it's very, very minimal, right? It really, it really gets into more credits being used when you're having to do long, complex, you know, development.

That's really where it's gonna, you know, and there's different levels.

I'll take you through kind of Replit, how you can manage that really well. You guys, another about 2 minutes here. Hey guys, we've kind of gotten there. Now this is something that, um, that you could spend 3 hours doing, and I recommend that And you don't even— it doesn't have to be 3 hours, but you know, you could spend an hour doing it, right? You could spend 3, you could spend 10. Okay. But it's amaz— isn't it, isn't it kind of fun though? Isn't it fun to talk about this and just like, oh man, this is really cool. And you just start getting excited as it starts to help you kind of weave through this. So I started out with this idea of this gear, you know, app, right? And I figured out I'm calling it the Gear App cuz I looked on GoDaddy and the gear, the gearapp.ai. Is available. And I was like, that's interesting. I wonder what AI could do here. And then I was like, okay, what if AI could actually, you know, help here? It can help me find shops to fix my stuff. It can actually email a shop for me and ask me about something. It can, you know, AI can do maintenance predictions. It can know, hey, you're going kayaking, you know, on June 14th. And by the way, you still have to get this thing fixed. You know, AI recommendations, proactive alerts, knowing what needs to be done. Hey, you're going skiing on December 26th, and so you need to make sure you get your skis sharpened, and here's a local shop that can do it, right? You know, being able to recommend stuff like that based on the types of things you have coming up. So I've built a free version. I've built a— so it's a freemium model. You could pay for it. And so I've got, you know, this exceptionally strong MVP, right? You've built a useful free tier, a clearly valuable paid tier, clean path from free to premium, product that solves real problems for household with lots of gear. And now I'm keeping track of all of my gear from from beach supplies to my adventure gear to, you know, whether or not my basketball is pumped up even. Okay, so we'll build a mobile-first web app first on Replit and all that stuff. So I've now pretty much got a plan. This is pretty cool. I've, you know, I've turned a cool idea that I have— I have this problem of I've got tons of gear, and I actually am totally going to build this and use this. I don't know if I'll actually launch it, um, for now because I just do too much stuff. I'm too busy. So So busy. But anyway, this is a cool app. This is a cool app. So anyway, I'm going to— at this point I can kind of just stop. It will keep asking me questions for quite a while. What I'm going to do here is I'm going to say, I want you to build me a 1,500-word business plan on this with everything.

And so taking into account everything we've discussed. Use the canvas tool for this.

Okay. So I'm just going to have it build me a plan. I'll drop this chat or this little prompt in here in the chat for you guys. You can just copy my thing. And so canvas, the canvas tool, some of you may have used it before, maybe not. But ChatGPT has the ability to kind of create a document inside of the chat and inside of the project. And so I can then go in and I can actually edit this document. Just like I can a Google Doc or something like that, right? So, but it's built it out. Hey, here's my executive summary. Here's my problem statement. Here's my solution overview, my product vision, my target audience, my brand identity, unique selling proposition. Here's the MVP for the free. Here's the MVP to add in the paid features. Here's the pricing strategy. Here's the revenue model, competitive landscape, technology stack, go-to-market strategy. Anyway, all kinds of good stuff. Okay, so that's cool. I built a full business plan. I love it. Okay. You can also at any time click on this and I can come in here, I can type, I can change that plan if I want. This becomes a fully editable document. Okay, so go ahead and do that, guys. Go ahead and get your, get your business plan created by ChatGPT. Hey, Jon. Yeah, go ahead.

I just went through the whole thing and then I reached our max. I had to pay, so I did that and I feel like it just just, disappeared everything I just did.

Does that make sense? Yeah. You might have to click back into the project or back into the conversation to pick up where you left off.

Like, hit the back button kind of thing?

Okay. Well, no, just over here on the left, you should see your conversation somewhere about what you're building.

I didn't name it. I don't know if that makes a difference.

Yeah. No, you shouldn't, you shouldn't have had to name it. It should be okay. If you wanna share your screen for a second, we can look really quick.

Yeah, so if we can find it for you, we don't want you to have to do that again.

Would it be better, Jon, for me to like do this in a separate chat so I can come back to these questions because I didn't make it through all my questions?

Yeah, you need to make it right now. What I would do— we're about, by the way, guys, we're about to take a 15-minute break so you can keep working on it, but you want to finish these questions before we go to the next step. And so because of that, we'll do that. So right here, Jon, you should, you should see your chat somewhere.

Yeah, I don't see it. This was, this was the last one prior to that.

That was one before that. And you tried refreshing the page?

No, I did not.

Let's try that. Yeah, let's just see if, see if it shows up.

Business plan interview. There it is.

Is that it?

No, because I didn't name it.

Well, click that.

Click that. Well, that's it.

Yeah, that's it, right?

Oh, you the man.

Yeah. All right. You're good.

I guess because of that order, I got mixed up on that.

Yeah, I know. That's weird. I, I have no idea why I did that. Okay guys, we are gonna take a 15-minute break till 11:15. Okay. I'm doing this on purpose right now because you guys sometimes get a little bit behind here and this is important. By the way, guys, remember you don't have to be perfect right now. So by the end of 15 minutes, you're done. Just say, hey, I need my 1,500-word business plan. Okay, so I'll leave this open. I'm going to turn off the audio and video for a minute and, you know, everybody step away, grab a drink, grab whatever, and we'll, we'll come back here. I can go for like 4 hours straight, but most people are like, we should stand up for a minute. So we'll do that. See you at 11:15. All right, guys, welcome back. Maybe some of you never left.

Yeah, I never left. Been having a good time.

This is fun.

Awesome. All right, it's about to get even more fun. This is where we get to start to bring it to life, build the business plan, then take it into Replit, right? But you can see what we're doing here, right, guys? And we're not ready to go into Replit quite yet, but already we have a business plan. Now, if you hand that business plan to Replit, right, imagine, you know, again, back to Nate as a real developer. Nate, if I gave you just my, hey, I want to build a gear management thing, right? You'd start to build this cool thing. It would probably all be all right, right? But it might not be exactly what I'm looking for. But if I hand you this fully vetted business plan now that has all these things and things I've thought through, now you're like, okay, I know what I'm doing, right? And everyone's going to be happier, right? We're building what we really want instead of just some kind of, yeah, this kind of gets it done, right? But when you're thinking about, hey, what features are going to be there and which ones are going to be paid, and you're thinking about that from the very beginning, well, you're going to structure that, that thing right, right? The system is going to be structured in a way can handle and is ready for that growth, right? If you go into it and you haven't thought about payment, you know, authorization or things like that in the beginning, then, you know, it's going to build something a little bit weird. So anyway, um, okay, it just keeps asking questions like how long—

just stop.

Yeah, you just have to stop. I feel like this would go— this could literally go on forever if you wanted it to.

It could do that. I, you know, it will come to an end But, but really at this point, just say— so what I want you all guys to do, even if you're not done yet, I want you to just say, give me a 1,500-word business plan, right? Use that thing that I put in the chat. I want you to build me a 1,500-word business plan. Get to where you have— I'm going to share my screen again here and pull this back up so we can resume what we were doing here.

Yeah.

Anyway, so I've got my 1,500-word business plan. So here's, guys, where, where we're going to, we're going to start kind of bouncing between a couple platforms now. Okay. So this is cool that we got a business plan. So I'm going to go ahead and copy this and then I'm going to come over here to, to Groq. Okay. And I'm going to start a new conversation with Groq. Okay. And I'm going to ask it for some feedback. Give me some feedback on this business plan.

I want you to focus more on the content and idea as opposed to the structure of the plan itself. I really just want to know what the market thinks about this. Focus on potential competitors, what people are— whoops— people are saying on X and whatever other research need to do to give good feedback.

Okay, then I'm going to hold Shift while I push Enter or Return, and I'm going to paste that business plan. Okay, I'm also going to get off of fast here, and I'm going to choose—

yeah, I'm just going to choose auto. I'll just choose auto. Okay, and, um, I'm going to give you guys this prompt here so you guys can copy and paste that. Okay, and I'm gonna click go.

All right, so now it's gonna think and it's gonna search.

It's gonna do a month's worth of research for me, and yeah, we'll see what happens here. It's running searches.

Boom, done. 15 seconds. You can obviously have it think longer too, right? You can say, hey, I really want you to dig deep on this, but Grok 4.1 is actually just really fast from what I've read over the last few hours. So now I can say, all right, overall impression of the idea, validation for the market, this problem is universal and growing. You know, here's potential competitors. So here's a gear inventory app. Looks like they've got $5 to $10 one-time in-app purchases, weight tracking, versatility. Okay, so here's gaps versus mine. There's no AI predictions or outreach or anything. So I'll read through mine, guys. You take yours, do this process, okay? Drop it into Grok, get your feedback. If you don't have Grok, if you don't have paid Grok, then, you know, that's okay. You can do this inside of ChatGPT and say, I want you to now do research on this plan and I want you to see what you can find, but you won't get any of the cool stuff you get from Grok in terms of, you know, what people are saying on X, right? So, and go ahead and take— yeah, yeah, go ahead.

And Jon, for the Grok part, do you just copy and paste it into it? Is that what you're saying?

Yeah, yeah.

Well, you'll— yeah, so you'll, you'll copy my prompt first that I just gave you in the chat, then go down a couple of lines, just hold Shift+Enter, Enter, and then paste your business plan into Grok and then click go.

Yeah.

Thank you.

Go ahead, Jabulo.

Jon, so I did exactly the same. I'm on a freemium model on Grok and it gave me almost a similar response that it gave you on your, on your paid subscription model.

Are you doing something that's in the same space?

No, I'm doing more of something that's going to be for unemployment in Africa.

Got it. Well, yeah, I mean, it might be similar. Yeah. I mean, it's pulling from the same data. And if it's just looking at a freemium model, then that's, that's fine. But anyway, guys, what I like to do here, when I, when I get all of this, you know, idea, I'm going to then actually copy the response from Groq. And I'm going to go back to ChatGPT. And I'm going to say, here's some feedback from Groq on this idea.

Implement or take into account what you need to, but create a new Canvas document with the—

and call it new—

let's see, updated business plan. We'll just say that business plan. Remember my thoughts and all our discussion. To make it more complete with this Groq feedback.

Okay, I'm gonna paste Groq stuff and I'm gonna send it. Okay, so I'll copy this for you guys as well. So I'm going back and forth, I'm having these AIs kind of just argue with each other a little bit, right?

All right, this is not, this is not what I wanted. I want just, no, I did not copy that properly.

Copy. Okay, there you go. So grab that, um, prompt for me, and then it will create a brand new updated business plan.

Yeah, your gear always ready.

Good. Awesome. Okay, so I'll wait for you guys to get that. So take your business plan, put it into Grok, get the feedback, take that feedback, put it back into ChatGPT, ask it to build the new plan.

That's where we are.

I'll give you guys 60 seconds to get—

So, Jon, just another one.

You what?

Sorry. Okay. Just another one. So with us getting Grok and ChatGPT arguing amongst each other, so could you potentially say that ChatGPT is the project manager and then Grok is the researcher in this process?

Yes, that's exactly right.

Okay. I'm actually going to ask it to do one more thing.

Create another canvas called competitive advantages. So it's better than the competitors already.

Cool. You guys could do this as well. So you can— I'll drop this prompt into the chat just in case you want it. For me, I saw that there are some competitors out there. There are some, you know, gear tracking apps, but they look pretty weak. And I really want to focus. I really want to make sure that we focus on why this is different, right? That's going to make a good business. Okay. So the last thing we're going to do, guys, here before we get into Replit is we're going to ask it one more document. We're going to say, please build—

or yeah, please create one last canvas document.

That focuses on—

hold on, I actually need to look at my— here, business plan, updated business plan. Oops, I accidentally said that wrong.

I accidentally said updates business plan.

I meant updated business plan. But anyway, um, please create one last Canvas document that focuses on, uh, a complete feature list everything that needs to be created, front end and back end. Don't specify the actual technologies that will be used, just, yeah, don't specify, just give me the feature list.

Copy that.

Um, create or call them. Yeah, call the Canvas feature list. Copy, go chat, paste.

Okay, there it is in the chat right there for you to grab. Okay, so you should finish here, guys, in a second with 2 documents at minimum, right? You could use the 3rd one that I threw in there, which was kind of the competitive advantages one. But, but really, you're going to have the business plan. The updated business plan and the feature list. Those are your kind of two primary things. Okay. At this point, while you're, you know, if you're caught up and you're kind of waiting for a second, let's bounce over to Replit. Go ahead and open Replit.

And, um, where is my Replit? I'm going to use this one here and I'm going to go home.

Okay. So on Replit, guys, it says describe your app idea. Okay, so we're going to give it a bunch of information in order to make that happen properly. So we'll come in and what we're going to do is we're going to ask ChatGPT now, now that we've built our feature list, we've got our business plan and maybe our competitive advantages kind of optional document that we had there. Okay, we're going to ask this. We're going to say give me a 3-sentence description of this app.

Okay.

And I'll, you know, I mean, I can copy and paste that prompt for you guys. Could maybe yours isn't an app, maybe it's of this business, of this whatever. Okay. But anyway, now I've got the Gear app is a proactive gear readiness platform that does this, this, and that. Okay, awesome. Copy. And now I'm going to go back over to Replit. I'm going to paste this. Don't click start chat for a while. We're going to paste a couple of things in here, but now I've given it just kind of a high-level 3-sentence description. Here's what we're doing. Okay, then I'm going to push Shift+Enter, Enter so I don't accidentally send the thing. I need to hold Shift down so that I can just get a couple more lines in here. Then I'm going to copy and paste my different things that I created. Okay, so I created the feature list. So I'm going to just come up here, I'm going to click the copy button and I'm going to come back to Replit. And I'm going to paste it and it's going to show up as like an attachment looking thing. Okay, then I'm going to go back to ChatGPT. I'm going to grab my next document.

Maybe.

Come on. And I'm going to copy it.

Okay, here's competitive advantages.

I'm going to copy that. Okay, I think I got it. Paste.

Yep, I got my competitive advantages.

And then I'm going to drop in lastly my business plan, my updated business plan.

Except my updated business plan has just failed me.

It's not showing up. At all. There it is.

Okay. And I'm going to copy this and I'm going to paste it in.

Then I'm going to say, please be sure to pay close attention to the attached documents, all of the—

whoops—

and all of the guidance given you. Um, as you create the auth system, do not use Replit Auth, just set up a standard JWT.

Actually, you know what, maybe I will use Replit for this. Yeah, I am. I'm not— I'm going to take that away. Okay, um, so here we go, guys. I'm gonna help you all catch up now. So you should have on Replit, if you look at my screen, you can see I've got a 3-sentence description. I then have one more sentence that I just pasted into the chat for you. Please be sure to pay close attention to the attached documents and all the I've given you. And then make sure you have pasted your business plan, your feature list, and your competitive advantages if you have that document. And you could obviously add more. You could have a branding document. You could have, you know, more details about pricing or more details about whatever, right? You can add all, all that you want here, but really, this is usually probably enough. You probably have want to have, have built this all out inside of ChatGPT. And if you had other supporting documents, you, integrated it into these things. Okay, so that's what it should look like. Anybody feel lost at this point?

Okay, and then click Start Chat, right?

So we're going to just let it start building, and while it's building, we're going to have a fun little discussion here. Okay, um, so everyone go ahead and click Plan or click Build. When it comes up, it's going to ask you— it might ask you a question or 2 as follow-up. And then it's gonna say, do you want me to just do the design or do you want me to build the whole app? Quick, build the whole app.

Right?

Quick. Let, let's, let's let it build the whole thing there.

Okay.

We're gonna approve. Also, it might ask if you have any AI-powered something.

I'm gonna approve it to use that.

Um, let's create this later.

Asking robot agent.

I'm going to connect it. It's asking me to connect to Gmail.

It's interesting. Okay, well, didn't like that. Now I'm stuck. Now I need to press open authorization. It's not cool. Connect to Gmail.

This thing's incredible, man.

What is going on here? There we go. All right, we're gonna do this one later.

So now it's going to propose a comprehensive plan. Hey guys, just read through it, click on a couple things.

I'm going to pull up Okay, build the entire app, start building. All right, mine is getting to work. Pretty cool, pretty cool stuff here.

All right, um, Has anybody not pushed Start Building yet?

Okay.

Are you close, John, or do you need some guidance? You good? You good?

Okay. All right, perfect.

Yeah, we'll give you another—

Hey, Jon?

Yeah, go ahead and jump a little.

With my end, because I think it was a couple of days back I tried to build some products already on Replit. So I think I used up all my credits for the month so far. So what I'll just do, I'll just, I'll just have a look at what the guys are seeing. And yeah, I think that's, that's what I did. I tried to build 3, 3 businesses. So yeah, I used up all my credits.

That's okay. You'll get more credit soon and you can obviously pay for more if you want. You just have to authorize it to keep going. But trying to save money, then just watch. Just pay attention to watch. You'll get plenty out of it.

So I do, uh, it's asked me to put in a Stripe account, I guess.

Yeah, so guys, this is actually building a product for you, right? And so if you need to connect a Stripe account and a Stripe key, um, then maybe you need to set that up. You could also just say, you know, I'll do that later, right? Um, you could just tell it if it asks you for a Stripe API key, say, um, ask me for this later, finish building the rest of it first.

You know, okay.

But yeah, we're going to get into APIs. We're going to talk about that a little bit here. You know, the next, the next hour, hour and a half is going to be talking about that. But while this is building, guys, in the background, we're going to watch a quick video that's like 2 minutes long and then we're going to have just— this is kind of an open discussion, fun part of the day. So yeah, let's watch this really quick.

I'm going to go ahead and make this full screen.

And here we go. So this is, um, a short video, again, 2 minutes 28 seconds, of Eric Schmidt, who is the former chairman of Google. He's a billionaire many times over and very connected into everything that's happening. Also, this video was recorded about April 19th or so of this year, and we are now in November, and AI moves at a billion miles an hour. And so a lot of what he says has already happened and faster than he even predicted, but we're going to listen to it and then we're going to talk about it. Okay.

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 there are now somewhere around 10 or 20% of the code that they're developing in their research programs is being generated by the computer. That's called recursive self-improvement, is the technical term. So what happens when this thing starts to scale?

Well, a lot.

One way to say this is that within 3 to