How To Actually Vibe Code

You know what's just crazy though is, like, this space is moving so fast. Like, I remember we published one of, like, the first reports when generative AI came out, like, 2 days after ChatGPT came out.

Yeah.

I last— what, 2 and a half years ago now? Maybe 3 years ago.

Mm-hm.

Then last year's kind of been AI agents, now it's like vibe coding. This space is just moving so quickly.

Billy, welcome to the podcast. Great to have you here.

Hey, Ollie. Happy to be here.

Awesome. Um, so this is a slightly different episode to previous episodes we have done, as everybody is talking about vibe coding now. And I don't know about you, but when I see these posts on X where people are creating, like, these pretty simple products and generating, you know, a couple thousand dollars a month in revenue, we are sometimes like tens of thousands of dollars a month in revenue in the space of weeks and months of launching.

That really annoys me. So I'm like, these are basic ideas. And when I was looking around to see, like, who's the best person to get onto the podcast to talk about this, I found Billy and A, you're just super awesome, and I think it's just so cool what you're doing.

And, um, awesome to jam on kind of what's happening today.

Awesome. Well, thank you. I appreciate it. That was a glowing introduction.

Yeah, I truly get them all the time, but, you know, super excited too for today's episode. Um, so what I thought we'd do, so the first, um, part of the episode, we're going to go through how to build a directory.

Um, and for those of you who are not, uh, too technical, you know, we're going to make it super simple. I'll ask some Billy some, you know, questions throughout the episode, um, on just how to build a really cool directory and maybe, you know, how to sell it, as I know you've done that a few times now.

And then the second part of the episode, we'll talk about which tech trends, um, Billy is really excited about and the types of companies that could be created from those trends. So, um, let's begin with the first part, vibe coding.

What the hell is happening?

So much is happening. It's— and it's happening so fast that people who work full-time in AI and software are coming to me and being like, "Billy, like, I feel like I'm behind. Like, did you see XYZ model release?

Like, what do I do?" And it can be overwhelming, but I think it's just important to just try new stuff when it comes out. But like, don't sweat that you're missing stuff because, um, you know, I think it's a rising tide tech deal.

As OpenAI gets better, as Claude gets better, as Gemini gets better, the products that use them under the hood, like Replit, Lovable, they're going to naturally get better too. Um, so it— yeah, once you kind of pick your stack, um, check in like once a month and see if you can improve your stack, maybe, uh, by building an app and a couple different ones.

Uh, I tell people, actually, when they come to me and do like a strategy call, you know, pick a single feature app and, you know, use ChatGPT to make, uh, a product requirements document, you know, that lays it out as if you're going to hand it to an engineer.

We need these features. We're going to use these integrations. Go ahead and pop that plan into like every single coding agent and see what you get out. And I really think that's kind of the best way to learn the best tools for you and get a lay of the land on, "Okay, well, maybe like this one's better at UI, but this one's way better at the database stuff."

So I know if I'm going to make like a UI heavy landing page, I'll use this one. But if I need something that's really tight on the database, like I'll use this one. And so that's how I stay on top of stuff.

You know what's just crazy though is, like, this space is moving so fast. Like, I remember we published one of, like, the first reports when generative AI came out, like, 2 days after ChatGPT came out.

Yeah.

I last— what, 2 and a half years ago now? Maybe 3 years ago.

Mm-hm.

Then last year's kind of been AI agents. Now it's like vibe coding. This space is just moving so quickly. But I think, like, I don't know about you, what I think is, it's such an interesting and great opportunity to build.

And we think, like, at some point, are we going to see the first, um, single employee unicorn company at some stage? Because it's possible now, you know, vibe coding and, uh, such as that.

Yeah, I hadn't thought about that. That's totally going to happen.

But, um, we're super excited. Um, we're super excited to dive in. So, um, I'll let you do most of the talking and we'll jam for some questions, um, throughout.

Yeah, sure. So as a first kind of project, a really good one is a directory website. And a directory website just kind of, you know, aggregates, um, I don't know, like there's ones for AI products, uh, there's ones for restaurants.

Um, my buddy Ian has one, uh, breweries nearby. Actually, I might switch to screen share real quick. I'm going to have to do this eventually, so let's just get it going. Um, breweries nearby. I think this is a great example.

And it just lets people upload breweries. I don't even know if he's scraping breweries. Uh, and then they can add pictures and five-star reviews. And it sounds like kind of silly,right? Because there's Google reviews, there's all sorts of, you know, foodie apps and things like that.

But I think when you

kind of pick a specific niche to bite off and you kind of can tailor a product just to this,right? Like, um, you know, it's got all these different things that you care about for breweries. Like if they do flights, if they're dog friendly, if they have an outdoor patio.

Uh, that's not like standardized. It might be standardized in Google reviews, but I don't think people are going to give you that kind of rich information. Uh, so you're able to just build a way better experience for people who are looking for information.

And I think that's why Ian's site has done so well. Uh, and I think you could unbundle like anything, whether it be different kinds of softwares or a specific category of like restaurants. Uh, and so I was thinking we just do this for barbecue restaurants because I'm crazy about barbecue.

It's like my favorite food. Uh, and I definitely think it's a little niche that would be helpful for people because you could have, you know, this, but for, do they have brisket? Like, do they smoke their meats? Like, do they have this and that?

Uh, do they sell beer? You know, I think that could be cool.

I was just in LA, uh, 2 weeks ago and yeah, we went to a barbecue place. Wow. So many options. When I think about like just the whole feel is like, like you said, is it pet friendly? So many different like beers on tap.

How do you like your meat? Um, the size of it. Um, I mean, US loves to do things in big style,right? But making it just like so easy, you know, how much would you like? Like, I can just see a whole opportunity there itself.

Yeah. You buying it by the pound here.

Yeah. Slight difference from, uh, Asian cultures when based.

Yeah. Uh, so usually what I do to start is I just kind of talk to ChatGPT, like I said, and get that PRT. One of my big kind of pro tips is I use this app called Super Whisper to dictate.

And you just, you can install it, I think, on any OS and you just click command space, hello, and it'll put that in for you. And I think this is important for vibe coding because even like just when you're editing in like Replit or somewhere, uh, being more verbose than not, I think gets you better outputs.

And you're able to just be more verbose when you can type, when you can talk naturally and you don't feel like, you know, when you're typing, you don't want to type the extra a hundred characters. So that's my little pro tip.

So what's your thinking here? So, um, you have an idea and for listeners listening, um, what's going through your mindright now?

Oh, uh, I'm just thinking about how can I get this prompt into ChatGPT to get something usable that I can kick off in, uh, I'll probably use Replit for this. Um, so I'll probably just talk and just say, I'd like to build a directory style website for barbecue restaurants in the United States.

Uh, help me write a PRD that I can hand off to my AI engineer to develop this app. And so you just kind of explain the app. And then I would say like, we need to allow users to log in, create accounts with email and password.

They should be able to create new breweries if they don't exist, add in the information like the address, uh, the name of the brewery, maybe when it was created, if they know that. And then people should be able to check in and upload a review.

It's, you know, on a five-star scale. They can add photos and then they can tick little boxes for specific categories.

Uh, and so then we'll probably need to fill in like some of those categories. I'll probably give ChatGPT some examples and then ask it to suggest some to me. Um, so I might say, I think I start recording too much.

For example, for specific categories, I was thinking they could check a box if they have different kinds of meats, if they use a smoker, um, what style of barbecue, like mesquite, uh, if they use dry rub, uh, if they have like vinegar base, like North Carolina sauces.

So the specific types of barbecue, maybe if they serve beer, have a patio, are dog friendly, they have live music. Uh, please respond with some other ideas of categories that would be useful for people looking for barbecue nearby.

The website will be called barbecue nearby.

Okay. What do you think, Ollie? Is this ready?

This is pretty cool. Let's go for it.

What's cool about it?

How quickly and easy it is.

Yeah. Yeah. Yeah. So I'm going to go ahead and send this and see what we get back.

And I'm just using 4.0 here because it's the default.

Allright. Seems to be kind of nailing everything.

Okay.

Oh, this would be cool. Autocomplete with Google Maps API if possible.

That's cool.

Um, I would probably,

I'd probably leave that out of like a v1. So when I'm thinking about apps like this, I want to minimize the number of kind of integrations and features for like the v1, especially if you haven't done this before.

If you haven't worked with the Google Google Maps API before, I would get the kind of the core features of the app down first because you'll learn so much about like, oh, well, I shouldn't have prompted it in this way or actually this user flow is probably better.

And you can always try to add it on later or you're even just start again from scratch just with a better prompt based on what you learned. So that's probably something I would leave out of like a v1.

What do you think around, um, so if you can, if you can log in or have a Facebook plugin, um, maybe there's something you can do where to see if any of your friends have checked into the place where you're going.

Um, or maybe that's too complicated for like a v1, but there's probably like a whole community element there,right? Um, how do you connect or join like other, um, barbecue, you know, um, lovers to go and hang out with?

Yeah. Yeah. That's interesting. Um, to me, not to just be like, that's a dumb idea. To me, that kind of sounds like feature creep,right? I think the core, the core value of this is that it's UGC, um, for people who are interested in barbecue and are kind of, are kind of virality, like our growth model for this is SEO,right?

Um, so I don't know if I would goright to like Facebook and Facebook's kind of sticky to work with,right?

Yeah.

Um, but I think in our own app, we could add in like a friend feature, like a profile feature. I really like gamification. So I think having a profile where it just shows your stats, like, you know, a number of barbecue places hit or like shows on a map, like where like I've had barbecue in these states, blah, blah, blah.

Like the completionist thing is kind of more interesting to me.

I love that. There's a whole gamification you could do where, um, or even like challenges. The first one to go to a barbecue place in all, all states in America, like that could be quite cool.

I love that. I love that. Um, so maybe let's add that in. Um, let's add in a feature where there's a user profile page that shows on a map, uh, how many, which states they've had barbecue in, uh, and it'll show, you know, like a completion percentage.

And this profile page is shareable and it should look like really clean visually so they could screenshot it and share it on social media. Um, and also, also we should add a list, um, that shows maybe all the different styles of barbecue that they've tried.

Um, and if you have any ideas for other kinds of stats like meals eaten or, or, or anything else, uh, please let me know. So let's see what ChatGPT says to that.

And they share it to the end of the year. Like the Spotify wrapped.

Yeah. Yeah, yeah, yeah, yeah. And look, uh, so I have my ChatGPT like prompted to by default ask three follow-up questions whenever you send a message. So it actually kind of keyed in on gamification. So that's pretty cool.

Okay. Perfect.

Looking good.

We'll probably need some sort of plugin for this map, but we might just see what it does off rip.

Okay. Great. Give me the full PRD with the original features and then the social sharing features.

This is making me hungry. I don't think we should have done this before lunch.

Hey, you can go for one afterwards. Well, when you, when it comes to the US, let's go for one.

Yeah, yeah, yeah.

I think I'm going to have to go grab barbecue for lunch after talking about this.

You should. Man, this is so good.

But this is, I think this is a really interesting, I think it's a great kind of starter project because it's not going to be, it's not going to have a million users overnight, but over time, um, you could, it, it, it's like, because it's a niche,right?

You're not telling your friends like, oh, I made a restaurant review app. It's like barbecue or whatever. So your friends and people maybe who follow you on Twitter will think about you whenever they go to this specific niche you've picked and hopefully just kind of grow users over time.

And all it takes is just getting lucky and, or maybe you spend some time doing outreach and, and asking barbecue restaurants to like link to their profile,right? And then you can just kind of build up that SEO juice over time.

And it's a simple product that you can kind of stand up in a weekend, uh, and just kind of build over time as like kind of like, like a hobby,right?

And also like the great thing with directories is, um, for sure, like SEO, massive opportunity, but secondly, there are so many different types of directories you could build and how you can monetize that,right? Um, like if this becomes the go-to, um, directory for all things barbecue, like you can create a whole community out of it.

You can do like dinner clubs, brunch clubs, barbecue clubs. Um, and that's like, like it's such a huge thing. I mean, probably like this is probably a very US product as you guys are barbecue kings. Um, but like I can see that becoming like huge.

I, I mean, the more I think about it, the more I like this, uh, because I mean, like you said, you could do clubs, you could do like a newsletter, uh, and then there's all sorts of, you could sell your own sauces, you could sell other people's sauces, uh, sell gift cards.

I mean, I think there's a lot of great monetization on top of, you know, you could charge restaurants to, to be featured, uh, when people from their city look or, or to be featured like in a newsletter,right?

For sure. Maybe you don't want to do like, yeah, maybe like there's one idea like you probably don't want to do the whole operations, but if you can get like all the data on where your users are going to have their barbecue, that becomes really interesting.

Um, where they've checked in, where they've made a reservation, you could probably even get like restaurants to pay you per booking or, um, or per, you know, um, per, per bill spent based on the customer coming to the restaurant.

Yeah. You could probably then integrate, you know, if they're on resi or open table or something, you'll probably put in like that booking embed and then get attribution for that.

I'm liking that.

So we've got the, the, the PRD back from ChatGPT. I'm going to go ahead. I usually cut out when it recommends, uh, the tech stack. And the reason why is that all these, uh, cloud AI coding tools, they kind of have their preferred tech stack that they're good at.

Um, and, and so you just don't need to put stuff like this in like React plus Next.js. Um, I think if you have a preferred auth solution, that's something you could put in. Most of them have built-in kind of a preferred database solutions,right?

Um, Mapbox is not a bad idea. We might add that in actually. Where are we? Okay. So here's Replit. Um, this is kind of my go-to for kind of starting out new ideas. And I'm just going to paste it in here.

You see, it comes in as pasted text. And then the last thing is I'm just going to go ahead and I think Mapbox is, what's their pricing?

I'm going to just do a quick open source map software. I know there's a good one. This could be a whole rabbit hole in itself.

Let's go ahead and just try. What do you, uh, what do you think, Ollie?

I mean, I was actually going to ask what you, um, you've built and sold a few similar products and, uh, you've mentioned Replit a few times. Is there a, is there a reason why you use them in particular and any other tools that is your preferred tech stack?

Yeah. Um, I, Replit, I think just is a workhorse and gets to the core functionality, uh, most consistently. Um, I, I think like V0, for example, is one that I use sometimes when I want like really clean UI.

So I'll just have it like create like a dashboard page and then take a screenshot of that or take the code and put it back into Replit. Um, I think Lovable is really good, um, for like flashy landing pages.

Like I think they're, the UI that they output is, is usually pretty good, but I'm just most comfortable with Replit. I think it's quickest to idea, especially I think the database is really easy to spin up, whereas that slows you down a lot of times in other, other kind of cloud AI coding products.

Awesome. Yeah. Let's go for it. Whatever is easiest for, um, beginner vibe coders.

Let's just try, um,

build this app, use Mapbox for the map feature.

Not sure how it's going to go, but like I want to give it some direction and so that it doesn't just try to like build a janky map using JavaScript. I don't think that's going to go well. Um, Google, the Google API, uh, Google Maps is actually pretty robust and pretty easy to use.

Um, but for the purpose of this podcast, setting up like a new Google project can be kind of time intensive. Uh, if you're going to use, if you're going to rely on like Google Suite, like for any of their APIs or, or even login, I would use, um, Firebase.

I, I would try Firebase. Um, you can use it for authentication. Um, you could use it in tandem, um, with something like Replit or Cursor. I think it would probably be easier in like Cursor. And then to throw another wrinkle in, Firebase has its own like vibe coding product.

So you could build your own app in there. So that's, um, I guess the main takeaway is like if you're going to use Google Suite, uh, use Firebase.

And so as you, as we're talking about this, do you have a, um, do you already know in your mind exactly kind of step by step what's next? So we'll do like the integrations and then so on. Do you have like a blueprint on how you actually build these products?

Yeah. Uh, usually the first thing is just kind of testing authentication, um, poking around and seeing maybe what database fields we didn't think of. Um, it's just kind of like an exploration and kind of like debugging, but, uh, yeah, usually check authentication first and then go with integrations.

Uh, try to get like, you know, the user profile page down, kind of the core flow, uh, and the core UI down before I do anything with like branding or like the landing page. Um, yeah.

Cool.

So this is pretty cool. It's, it's kind of building out a, uh, visual preview here. Um, and then we can go ahead and approve it now based on this kind of plan that they've showed us. So you can see they've just given us this plan of things to build.

Uh, user authentication and profile management, barbecue spot directory with CRUD operations, interactive map view with Mapbox. Um, I don't, I like the check-in system, but I don't understand why we would cap it at one use per day. Um, if you're going to multiple barbecue spots in one day though, like you got to be a beast.

I'm sure that's not that.

Badge system and gamification features. That's sweet.

Uh, I'm going to let it actually build out this landing page, uh, because it's telling me up here, agent will use this design as a reference. So I think that'll be helpful to the agent that's actually writing the code.

This is cool. These achievements meet sweets, meet sweats rather. That makes sense.

So cool. And just how quickly it is as well. Yeah, we pretty much got it covered. Um.

It's bonkers,right? And just as comparison actually, I might just throw this into Lovable and V0 just to see what we get.

For those of you who haven't tried Lovable yet, um, check it out. You, I think you get like five credits a day and, um, you give it a prompt and they'll basically build a landing page for you. And I, I did this on my 16-hour flight, uh, last night and it's crazy.

Like just how much you can do in the space of like hours. I was looking to build like a community type of app and, um, it's insane. It did it exactly like I wanted it.

The agent thinks it's 2024.

That's funny.

Things didn't upgrade. Seven months out.

Okay. So I think we're ready to approve this. And I was going to go ahead and scaffold it out. Um, so it's already just kind of initialized like a basic project structure and we'll be able to see it kind of add in.

Look, I just added a bunch of stuff. Uh, but we'll see, we'll be able to see it build out the main pages. Um,

this can take anywhere from like, I don't know, pretty wide range, like three minutes to like 10 minutes. Um, just kind of depends on how complex your app is. And I've noticed another thing, which we just kind of take for granted, but all of these AI coding apps are really good at pulling like stock images.

Like, look at this. It's unbelievable.

Yeah. Actually, similarly to the, um, uh, product I built over the weekend, like, yeah, unbelievable. Yeah. Exactly. Like not perfect, but doable,right? Um, it's insane. And also just how accurate like the descriptions are. Like if we look at Discover America's best barbecue, like, okay, phone's quite nice.

Like just, it's slick,right? Um, as a first look.

Yeah. And, and I mean, look how solid this is with the little emoji or whatever this is that they picked.

It's pretty nice. Like it does look like a directory. Okay. It's not like the slickest like in the world,right? But, um, it's, it's quite cool.

Yeah.

Very barbecue-y.

And again, making me really hungry.

Same.

Oh, that's cool. So it'll show, yeah, the stats, like how many users have checked in here. This is a famous, probably one of the most famous in the country, Franklin barbecue.

I need to go. It looks good.

It is.

So maybe let's check in and just see where other little minions are. So it looks like they're both, uh, booting upright now. This is a very slick UI. This is V0.

Would you pick Lovable?

I really like it. And a lot of my friends are diehard Lovable guys. Um, I think it's super useful. I built some prototypes in it. I like it a lot. Honestly, I think I need to spend more time in it.

I just, there's, the thing that slows me down is the friction with Supabase. So just comparing it to like Replit and Replit, like they do the database all in here. And then with Lovable, you have to go into Supabase, initialize a new database, then come link it back into Lovable.

And then for some reason, I don't know if it's probably my fault, but Lovable always wants to charge me for each one of those databases I spin up like five or 10 bucks a month. And that's just, just friction that makes me not want to use this.

It seemed quite interesting. Um, so for, I think $25 a month, they give you, I think it's like 100 to 150 credits a month. You can very, very easily spend those credits pretty quickly,right? So there's going to probably be like a whole Lovable game here where, um, are they just going to make a ton of money just because people need more credits?

Um, or can they, can they, can they be slightly more user-friendly and give more credits, you know, stuff like that?

Yeah. And, and Replit switched over, I think, like to full usage-based pricing for the agent.

So that, that's pretty interesting.

Yeah. Which, I mean, I think is good because I would be frustrated sometimes when the agent would just like, I don't know, check something, like look at the database to, to see if the data is working properly. And then I would get charged the same for if it had written a thousand lines of code in a checkpoint.

So I, I mean, I always like when SaaS moves towards usage-based. So it's probably a good thing.

Yeah. We actually did a whole episode with Wade Forsyth, the founder of Zapier, um, Zapier on this. And, um, we said exactly the same thing, like is the future of software usage-based pricing? And it probably is.

Yeah. Yeah. And I think, I think those automations are moving more towards usage-based pricing too. A lot of those platforms.

It's great. It's great for the user and it's more, it's more ethical as well for the user.

Hey, definitely. Yeah.

Allright. Our friend Sarah is still being, still jamming us up.

Yeah. So the one, the other thing, uh, in Replit is, let's see, you can switch up. They have like this high power model now and, um, they've got the $3 signs here. Wow. 5X increase in cost. Uh, the first time I used it, it like immediately just like spent my, it got up to like my budget limit.

Like I think I was near it, but it just spent like, I don't know, five or 10 bucks like pretty quickly. Um, so

I haven't played around with it since. Um, but I, I bet it's a lot more efficient. So I don't know, maybe it evens out.

Yeah.

The beautiful thing with this for listeners who are maybe not so technically savvy is it's just so easy to start. Like I, I'm not going to write to a developer, but I tried Lovable on the weekend and I was like, wow, if you were to pay a developer to do this, it'll probably cost you 2, 5K and it took me maybe a day on Lovable.

Yeah. Okay. So here's, here's what I'm talking about. This, this Lovable UI I think is a little more slick than the Replit one. What do you think, Ollie?

It's pretty nice. A little bit more updated, like 21st century tech. Um, beautiful landing page, you can filter and so on. Gives you some nice recommendations. Um, yeah, it's pretty nice. I mean, my preference actually is this one versus the other one so far.

It's like, it's slightly more user-friendly. Um, also like it'd be interesting to see what the comparison is on kind of mobile. Um, because if people are looking for barbecue, they're probably looking on their phones,right? Um, so you know, there we go.

Pretty nice.

I mean, it's too wordy, but you can probably fix it.

Yeah. I mean, well, we've done this in 36 minutes.

The cards look really good.

Yeah.

I mean, it's good enough for someone not to turn up straight away,right? Um, and if I kind of think if you're a barbecue lover, like we are, it sounds like, um, you're going to take some interest anyway and you're going to explore, uh, what's available.

Yeah. Honestly, I might, this might tell me that, uh, I definitely this UI, it's a lot cleaner. Um, but I think this is serviceable and

I, I, I think, I mean, I just know that this is going to, this is probably going to work better than this one. Um,

so yeah, the connect a project thing in Supabase, it's just a lot of clicking.

And it's just that on Lovable.

Yeah.

Yeah.

Barbecue.

And for, um, for listeners listening on YouTube, um, we've been recording for 38 minutes and this is what we've done.

Yeah. Wow.

It's insane.

Yeah. And

Replit is still chugging along, but, uh, do you see the friction here with where, okay, it's going to be $10 a month. Like just makes me not want to do it.

So we're kind of, we're kind of at an impasse here with both of them. Um.

It's insane.

So I think what I would do, you know, if we don't, if we're waiting, so this is spun up now. What I would do is

I would take screenshots after Replit is done spinning up like the prototype. I would just take screenshots and maybe even take bits of the code from the Lovable and be like, hey, like make it look like this, like make these cards, like make this hero,right?

That's a good idea. Um, almost like, um, try multiple products, see which one works and not one product's not going to have all of your kind of, um, uh, features,right? But if you can, if you can give us some ideas based on all the products you've tried, it's actually a pretty, pretty good idea.

And then like this, I mean, I think this is a pretty solid one too from V0 and we could probably take instruction from this somehow. Oh, they have a leaderboard. That's cool. That's not built out yet. I really like the simplicity of like this is, we do four things in our app and they can compete.

Like that's a really good way of framing it. They're all pulling in Franklin barbecue. That's funny. They have really good SEO evidently.

Yeah, clearly.

Look at that.

Damn.

It's still slogging.

Let's see how much we burn. Oh, it won't show meright now.

But I think the cool thing is like, like we've said a few times now, it's, um, like directories are so cool. Um, if you think of the, the legacy brands like, um, I mean, yes, you got Google, you have all the like yellow pages for like European markets.

There's probably such huge opportunity just to create like the best directories for the niche things, barbecue, running clubs. Um, there are so many digital nomads now, like globally, maybe create like a directory for them. What do you think?

Like where are the opportunities for people to create directories?

I, I think just exactly like this, like different niches and, and food or hobbies. Um, because I, I, I think what people maybe used to do is like go look in somewhere like Reddit because those subcommunities, you know, people like are experts and like really care about like whatever you're looking into.

So like you could search like Reddit, like best barbecue joints, Virginia. Um, but I think Reddit is just like flooded with like, I don't know, like normies and like spammers and stuff now, you know what I mean? So I don't think it's that useful.

And then on the other end, like the big aggregators like Google Reviews or like Facebook, like, um, I think there's lots of bots, um, people just doing it because like Google, like review, like they reward you with like points for rewarding.

So I also don't think, um, that's where people want to go, but like if there's a, a site like that's just people that are really passionate about your niche, like I think they're going to give the best feedback and I think people want that about, I mean, the same with like shopping,right?

Like figuring out what kind of like camera to buy or like water bottle or something. Uh, I, I feel like going just to like Amazon reviews isn't that useful anymore. Um, so I, I think you get by it off like any and every niche and, and just try it.

Um, and, and, you know, get it, get it rolling in like a weekend.

And so if you do have the, say, $500 to get your first customers, would you put all of that behind Reddit, um, SEO? How would you actually like first get this directory going?

If I had $500 to spend, um,

probably do Facebook ads. I would probably do, yeah, I'd probably do Facebook ads and, and, and just do like a simple ad that's like,

that's like appealing to like barbecue lovers,right? Or like finding like new spots, like, and, and then I think getting connected with, um, with, with like the actual like restaurants like pretty early on would be useful. Um, just getting them to input their information or ask people to like review, building out some sort of like incentive like program there, I think could be good.

Uh, and then, yeah, obviously like, I think like just posting like, hey, I built this and like barbecue subreddits, like barbecue, there's, I'm sure there's barbecue Facebook groups. Um, honestly, like I think a lot of the organic stuff you do might do better than $500 worth of ads.

I mean, $500 actually, okay, I take it back. I think you would be better served finding like a food influencer on TikTok and being like, hey, can you like show off this site for 200 bucks? I think that might go way further.

Yeah. I was just about to say that. Go and find all the barbecue creators and maybe you start with one state in the US because there are so many. Um, and also like, you know, meetup.com, there must be like barbecue lover meetup groups.

Um, um, yeah, that's pretty why I've started as well. Meet up, Reddit and, um, find a couple of barbecue creators you can team up with.

Okay. So again, like we Lovable got us something pretty like really quickly, but it's not, it's not functional. Um, so we're going to have to go ahead and be like, hey, like do the rest of the stuff I asked you to do.

This is a great start. Um, and while you're doing that, like, um, so we've covered the how to find users, we've semi covered like how to just like launch it. Um, if you were building a directory to sell, what should users be thinking about next?

Oh, which should the, the people building the directory be thinking about?

Yeah. If they want to sell it.

Yeah. I mean, sorry. You, you just got to know your numbers. Um, so you got to make sure you have like Google Analytics and like at the very minimum and like Google Search Console and all that installed so you know your numbers.

Um, I would build out a some sort of like admin panel or like admin report that generates that tells me, you know, user growth, like, uh, DAUs, things like that. Um, and then obviously like the monetization stuff, uh, you could test, you know, if you aren't able to sell ads to like a barbecue sauce.

Well, honestly, like I, I think you're going to have to monetize it somehow before you sell it. Um, I guess having the traffic alone could help you sell it. Um, but yeah, I think just being on top of your numbers is the most important thing and, and pulling all that data into like one place so you can show it to someone who might want to buy.

So we have a beautiful project there. Who do you think would buy this type of product? Um, is it going to be a booking platform, um, like OpenTable? Is it going to be one of the big barbecue chains?

Um, who can use this as like their own SEO content automation tool? Um, yeah, how would you go, who do you think about sending this to?

I mean, my, my first thought, I think like just someone who lives in Texas or like Austin, who's an entrepreneur might find this useful, who knows how to pour on like the, the money to, to make this like a revenue machine.

Um, I, so I might be the wrong person to ask because I usually, uh, build apps for people who come to me with problems, uh, and, and then sell them to them. Less so kind of just build, um, random D2C apps like this and sell them.

Uh, but I bet you get a lot of bites on micro acquire if you threw something like this up there and had, you know, I think if you got up to like 50, 100,000 like, uh, like views a month, you'd definitely get some bites.

Just depends on how you package it. I don't know that I, I, I, I would be surprised if someone like OpenTable is buying platforms like this. Unless you really do what we said and build in that like booking tool and it's really working.

Uh, I just don't know if, especially in something like this where I don't know that many barbecue spots even do bookings, so they probably wouldn't be as interested. Um, but honestly, yeah, I think you could just catch some like Austin entrepreneur's eye that's like really passionate about barbecue.

So many, like when I think about directories, so many opportunities you could create.

Or, or the, I, I think like sauce, like sauce company that could just throw ads. They've got the production in place, they've got the fulfillment, be like, allright, great. Like they could chart out the dollars. Allright, I'm going to get this many impressions per month versus how much would that cost me for like Facebook ads or YouTube ads to get the same impressions?

I think that could be an easier sell too.

That's a great idea. Sauce company.

Okay. So we've spent $2 on Replit to build this. It's nuts. This is, I don't know, probably would have cost you at least a thousand dollars to get someone to do this a couple of years ago.

That's good. What do you think?

I think it's functional. Um, well, I mean, uh, I should say utilitarian in, in how it looks. We don't know that it's functional.

Okay. So it's going to use Replit. It's using Replit auth, um, which is really good for testing,

but I kind of find annoying when it does it without me specifying because I thought we said use email and password.

Okay. So it hasn't done the map box integration yet. Let's see if we can add. Okay. No, that's a glitch.

It's got my profile though. Look at that.

Pretty nice. There you go.

Oh, and look, it's got a little, this is so funny.

That's pretty cool. I actually kind of like this.

Not bad. That's pretty cool. Ratings. Nice.

You can share it, download. Oh, that's not out yet. Uh, this is pretty sweet.

Pretty sweet. Yeah. For a first version in, uh, 50 minutes.

Yeah. Yeah. So what do you want to do next? Do you want to keep debugging or do you want to move on?

This is really cool. Well, I think the, I think users can, you know, get a pretty good idea on, uh, on how to build that directory. So, you know, if anyone, put a new Uber links in, in the, uh, show description and, um, you know, anyone who's looking to vibe code, just give a product to go.

We'll actually give some recommendations in the, um, in the show notes, like Replit, Lovable, VO, um, and a whole bunch of others like I can think of. So we'll put those in the show notes. Um, but as we want to the second trends, you know, uh, the, the second, uh, part of the show where we love to ask our guests like which technology trends are they most excited about?

So today we're going to cover, um, uh, a couple of trends, uh, that you're all most excited about and as selling custom software to your employers. Um, the second trend is every business now has access to custom software.

Uh, lots of nascent demand is being, um, unlocked. Totally agree with that. And the first trend, AI receptionists for local services and businesses. Cool. Okay. Well, let's start with the first one. Um, selling custom software to your employers.

First question that just comes to mind is, well, first of all, like, what are you thinking here?

Uh, well, so I have a couple examples of this. After I did some kind of podcast interviews like this, I'm getting people DMing me that are saying like, hey, like I watched your, your Replit tutorials and I went and built something for my employer and I sold it to them.

And it's, it's insane. And, and so the, the, the employers want this. Like I, I, they want to invest in their employees and a lot of times companies are trying to, you know, haggle with like NetSuite or Oracle or someone to build up something that's like custom, but using outdated technology and it's going to cost, you know, like hundreds of thousands, tens of thousands of dollars.

Um, and so, yeah, I've had people DM me like, hey, we were going to, my employer was going to pay someone, you know, 120K. I've a, I have another guy who sunk a quarter million dollars into like custom software solution.

Um, and it just doesn't work like he wants. And so, um, he wants me to help him build a custom software. Um, so I, I guess like the trend one and two are kind of similar. Um, but yeah, I, I, I think you don't need, if you've got, if you work at a company that like has any footprint in, in software, like uses like a CRM or like inventory management, like I think you could build something useful and, and sell it to them, um, and keep working for them.

Uh, and I think they could help you depending on where you work. I think they would help you champion it. Um, so I think that's a really cool trend.

Isn't that off to the side? And the question was going to be, um, if your developer team is pretty large and okay, we both need each other, we both know each other, Billy has this cool idea. He's going to help X company, um, where we both work, but he gets paid a quarter million dollars.

Ollie's just as good, but, uh, he didn't have the idea, so he didn't get it. Uh, didn't get any compensation. Is there not, is there not like a flip side where people are going to get a lot more compensation for the ideas versus those who don't work on it?

Or is it going to be a team joint effort?

Well, I don't know. I mean, it's not that. I, I don't think it's the idea. It's the people actually building this out using like AI tools,right? So, yeah, I mean, I guess I could see something where someone in the break room is like, it'd be really cool if we had this, like Billy, we should go try to build this in Replit and then you do it.

And I don't know. I, I don't think ideas are worth anything. I guess is my answer. Like there's a dime a dozen.

True. Do you mean there are any, um, what types of businesses do you think are still yet to be created kind of in this space? Like, um, and this is more to help developers, vibe code,right? Or, or build products.

You know, we've seen a lot for like developers now. We're seeing a few more products and much more design, um, and product teams. Is anything missing do you think that could be created there?

Uh, any kind of like AI products missing?

Yeah.

Yeah. Well, so, I mean, I'll just kind of tell you that the hot niches that I'm hearing about a lot, um, real estate. I mean, I got real estate agents coming to me all the time, um, and, and wanting specific tools for helping them sell more listings, like pricing tools, like comparison tools, uh, just even managing like documents.

Um, and that's, I think that's a really interesting segment, um, that if you're starting out, you should look into because they're a lot of times happy to spend on software because if it helps them make one more, close one more deal, like then you can justify almost any price,right?

Because their deals are so big. Um, so I, I think there's a ton of open space there to play in. Um, like, uh, like just virtual staging. So generating different custom images of what the home would look like for a newlywed couple versus a bachelor versus retirees.

Um, and then the other one recently is, um, e-com and anything with like supply chain. Um, so you can bet, I mean, I would take this bet a hundred days out of a hundred that if you go to an e-commerce company, um, even if they've got every Shopify plugin, every WooCommerce plugin, everything dialed in, there's still going to be gaps with how they do business, whether it's, you know, how many components do I have left that make up this product?

Like where are my shipping containers? Like there's just, supply chain is just still so tough. Um, and that data is, um, all over the place. So that's another niche with pretty much any business is unifying and aggregating data.

Um, and I don't mean this as like a SaaS product because there's a million SaaS products like Databox, uh, even Looker Studio and Google that purportedly do this, but they all have feature gaps and limitations and it's just easier to do a code at a certain point.

So even just like if you wanted to start a business that's, hey, I'm going to make you a custom app that just pulls in all your business data and then maybe sprinkle on some AI insights, I think that's immensely valuable and can even save business owners money if you're cutting out seats on other, um, SaaS products just for, you know, just for like a manager that needs to go in and like monitor a dashboard, uh, that you could cut out that $50 per month seat for them if you roll it up into their own app.

Love the idea. That kind of, uh, you know, insights into the second trend, which is, um, every business now has access to custom software. Uh, lots of nascent demand is being unlocked. Um, again, like what are you kind of thinking here?

And it just feels like this whole, you know, AI space is moving so quickly. Um, are we kind of just in a bubble? Is this just going to keep going how it is? And there's just so, you know, much opportunity to unlock.

Yeah. No, I mean, I, I don't think it's a bubble. I think there's bubbly elements to it,right? Like I think business owners are shrewd and when they see that like, you know, Ollie, you made like an app in a weekend, that cheapens software to them in their minds.

So they may be like, well, these guys are ripping me off. And I, I think to your, to what we were talking about earlier, a lot of companies are ripping you off, um, by doing like monthly subscription versus usage.

Uh, but also those companies have been working and polishing on this product for 30 years and they got 24/7 support. So they're, it's not to say that legacy software isn't valuable anymore. They're going to have to recalibrate their models.

And I think there's, you can get your foot in a door by building these simple one business solutions for, you know, savvy business owners that want to save money and want to solve custom problems because they're, they're, they're aware of it now.

And, and as more and more people become aware of it, uh, they're, they're, they're going to want their own software again. Um, their own custom solutions, I think. I mean, and, and you know how like small business owners are, a lot of people, they want this solution not only just for their business, but then they want to go white label it and sell it to other people after they've tested it, which I can't think of a better software development like, uh, like path, I guess,right?

Which is like the actual user using it for a couple of months, owning it and then making it better and selling it to other people like them. Like I think that's just a great way to do it.

I think like what's so amazing now, A, there are tons of opportunities. And when we think about this third trend, which is all around like helping local small businesses, so many like opportunities you could unlock. Like if I think of here in Asia, so much of everything's done on paper still.

None of it's digital. So like when we think about, you know, supporting, I think especially like AI receptionists for local businesses, um, what are you thinking here in terms of ideas and types of businesses which could be created?

Yeah, I, I just, I called like my local HVAC company the other day and like the, the receptionist passed the, the Turing test like totally. Like I, I, I talked to her for like 20 seconds before I realized it was AI, uh, because it was just so natural.

Like they had like, I don't know, programmed in little, like her conce, her speech pattern was consistent and like, like believable. I was totally fooled, but I figured it out because I would say stuff like, this is my address and she'd be like, oh, I see you live here.

And it would be just so much faster than someone typing,right? Uh, so I was like, wow. Um, so I was impressed by that. And I don't even think you need to build out your own custom AI receptionist product.

Um, I think you could just sell the installation and, and set up to, to local companies, just call them and be like, hey, like, do you want to give this a go? And then once you've done it for a couple, uh, companies, then you can, you can gather feedback from them and maybe build out your own solution, um, using, you know, like maybe like Replit and like something like LiveKit or VAPI.

Those are really powerful, uh, voice AI models. Um, and I think like it's just like every company needs it because you're losing money by not picking up the phone. And so even if you only turn it on like at night and you have a real human during the day, like you're just capturing more leads and more business.

I think it's a no-brainer.

What type of, um, platform or, you know, type of tool would you start? Um, and what type of receptionist would you, yeah, would you actually build?

I'm totally blanking on, there's something called like, is it like Ruby? I'm trying to pull it upright now. Yeah, Ruby is one that's, um, a virtual receptionist. Uh, Avoca AI is actually the one that, um, responded to my HVAC call the other day.

I, I would, I would look into one of those. I would look into, um, Go High Level, uh, and see what solutions they sell out of the, out of the box. Go High Level is kind of like a CRM that like helps you sell and like create accounts for different local businesses and like manage like their local SEO.

Like it's a big suite of things. Um, I think it's a pretty solid tool. A lot of people think it's pyramid schemey because you can make money for every new user you put in. Uh, but I think it just works well.

So I imagine they probably have an AI receptionist. And then as far as building, yeah, I would, I would go with like a, a Replit and/or Cursor and like LiveKit and/or VAPI. Um, and I think, I think maybe I would do the, I would, I would focus on what niche to start and just be like, this is an AI receptionist for HVACs or this is an AI receptionist for gyms,right?

So that you can sell like, I have this custom solution just for you. You're not going to run out of gyms and HVAC people to call,right? And you're just going to get better at, at servicing that specific niche.

True. I kind of feel like, um, what do people really care about? They care about the health. They like to travel. So helping with anything around that. How would you help people with money, fitness, and health? There's this company called Carlet, uh, AI.

There's like these 18, 19-year-olds. There's such a simple app,right? I don't know if you've tried it, but, um, you take a photo of what you're about to eat or drink. It, um, captures the, the, the amount of calories.

And these guys are generating like $15 million a year in revenue.

Yeah.

They launched like a year ago.

Yeah.

Just such simple ideas. But even if anyone listening, like if you want to do some of the travel, uh, had a 16-hour flight yesterday,right? Of course, the airline lost my bags and everything is done on paper still. None.

You can't track your bags. You can't do anything. You can't get a hold of anyone. You're just waiting there for like the bag to turn up. So I think there is so much opportunity to like, you know, fix these old school legacy problems.

And that, so that idea that you just talked about, Ollie, like bag tracking, like it might sound like a big scary problem and there's big multi-million companies involved, but with AI, I think it's going to take just a little bit more effort than something like a directory to fix.

Like it's a fixable problem. And I think if you just show people a demo and say, I've been thinking about this problem. I know you have this problem. Like, can I help you solve it? I think people will be a lot more receptive than you think.

Exactly. Well, hey, we've covered so much. Um, I know you, uh, have so much going on and super cool. You know, you've, uh, building and selling products. So hopefully our users and, uh, you know, subscribers learn so much.

So thank you so much. Um, where can our, uh, listeners find you and learn more about you?

Yeah, I think my YouTube channel is probably going to be the most valuable. I'll get tutorials about building apps, more importantly about selling apps,right? Like that's the most important part in my mind. Uh, so yeah, you can hit up my YouTube, uh, and I've got a playlist in there of like all my interviews too.

And then I have Twitter, Billy J Howell, and then stupid simple apps.com is my agency. So if you'd like an app, uh, built, hit me up. You can, you can fill out a form on there and, uh, we'll, we'll get to work.

Make sure you subscribe. Billy's content is awesome. Um, Billy, thank you so much. And can't wait to see everyone next time.

Thank you, Ollie.