Billy Quit his Job to Start an AI App Agency: Here's How

Chatgpt rolled out. I started coding myself from self taught and then when Replit came out that was kind of like a 0 to 1. You could go from nothing to full time freelancer or agency in two weeks. Now I run an app agency and I've got a team of four guys.

You just stay in the game and you'll win. How do you know? Like this. I need to lean into this side project or this client. With all the opportunities out there, it's really easy. What business ideas have you seen or had lately that can be created with tools like these?

So there's a lot of stuff to build.

All right. I had a new guest on today by the name of Billy Howell and Billy is a vibe coder. Vibe coding is basically using natural language like I'm talking to you right now to talk to apps like Cursor, Replit, lovable, et cetera, to build stuff with code without knowing how to code. And this guy's doing a great job of it and I wanted to learn from him in public with you. So he talks about apps that he's building, cool app ideas that he has, how to make money on upwork, how to land jobs on upwork, how to how to hire people on upwork to do work for you, which AI coding apps to use for which type of app projects, whether it be a web app or mobile app, et cetera. You're going to love it. Share with a friend. We'll see you next time on the Kerner Office. Yeah. So tell us who you are and what you do.

I'm Billy. I make stupid simple apps for business owners and startups. Come from a defense consulting background, so not really. The two don't really make sense. But I think any good entrepreneurship path is kind of crazy like that. And kind of quit defense contracting during the pandemic. Started a marketing agency because I knew I wanted to work with startups. Learned a lot about kind of automations, helping clients automate marketing operations and things like that as ChatGPT and cursor and all that rolled out. I started coding myself more. I'm self taught and then when Replit came out that was kind of like a 0 to 1. Like hey, I can make really useful internal apps for my clients without taking too much of my day up. And it's just really fun for. I think it's fun to tinker around and like Cursor and repl it. Now I run an app agency because I posted like a video about selling an app and it went viral. I've got A team of four guys now and we're making stupid simple app.

What are some examples of a stupid simple app?

The original one that I made is just data entry for auto mechanic coaching business. So there's this big need.

Yeah, okay. Okay.

Yeah. Pretty sexy right there.

I've never heard those words strung together before.

He is a successful garage owner and he now coaches other garage owners on how to get their techs to see more cars and have a higher per car revenue. He thought he needed like a no code solution with like airtable and softer. I'm on upwork sitting at my desk and I'm looking at this. I'm like, okay, like blue collar guy. Like matter of fact, this sounds like my kind of guy. And I'm kind of in the process of writing a little proposal. Like I use Airtel for my marketing clients. I can do this for you. Here's some examples. And I, I'm thinking to myself, I'm like, hey, like this is really simple app. I could just make it in replit. And so I did and I kind of showed him, I made like a demo and I kind of took a loom and showed him both solutions side by side. His custom replit app or what it would be in like software and airtable. And the reason I did that is because the software cost of using that no code stack would be like 250 bucks a month for him. There's no reason you need to be paying that. So let me build you one and then I'll charge you 50 bucks a month to run it. And he said, I like yours better. And we went with that. And, and it's just, yeah, now he has car techs put in like how many cars they saw each week and he can kind of track their progress.

Now at what point during the the upwork application process did you build these MVPs for him? Or was that early just to get the job or was that after you had the job?

I want to say he said yes before I did a full demo. But I think I also may have just made something in V0 because it takes 30 seconds to visualize what he wants. So I would say like, if you're going to try to do something similar, you just have to find a way, whatever stack, if you're using V0, Repli, Cursor, Windsurf, Bolt, lovable, just get good at making like something that's like a clickable interface that brings their idea to life. And if you can do that in five minutes, I think it's definitely worth Doing it to make the sale. If it's your first job and you really want it and you want to prove yourself, go ahead.

All right, let's back up and educate the listeners a little more. How would you define the differences and similarities between V0, lovable, bolt, replit and Cursor? Or is there anything else I missed there?

Windsurf.

Windsurf.

So there's local coding environments like Cursor and Windsurf. They're pretty much like if you were gonna use like a VS code, like that's a place where you can like store all your files, write code, do terminal commands. Purser and Windsurf are kind of like that, but with an AI chat assistant that'll write code for you or answer questions, debug a little more advanced. Like if you have coding experience or an engineer, you would go that way. Something like Bolt or Replit or lovable or V0 are browser based. So you just go on your browser and that's kind of more beginner friendly because you can just type in your app or like I've done before, like put in a sketch of an app and it makes it for you. You don't have to know how to install packages locally. What kind of stack am I going to use to build this app? Where am I going to put the directory? How do I demo this locally? You don't have to know any of that. So it's more beginner friendly and it's quicker to MVP with those.

I'm guessing most of our listeners are more on the beginner phase. Not a lot of engineers in this crowd. So if I want to build an app, like why would I choose lovable over V0, over bolt over Replit or vice versa?

It's a hard decision to make because as the underlying models like OpenAI or Claude, that kind of power, these things change. Then you know, Bolt might be better because of the, the most recent ChatGPT release. But then maybe Lovable is going to go down, right? So they all kind of are changing. So it's really personal preference and just putting reps in to one that, that works with your workflow. So what I would say if I'm setting out to kind of start trying to build apps either for my own business or to sell to people or just as a hobby, I would think of a really simple one feature app that you want to build and literally do it side by side, repl it, v0lovable, bolt, whatever else there is and try to make the same app and see what comes out like See what frustrations you run into. Maybe you really like the UI for one. Like, I really like the UI from B0. So a lot of times I'll have it make UI and then pull that into my main app in, like, repl it. So I would really. Rather than getting kind of pigeonholed and stuck in one, I try them all first. If I were to, like, start this.

Over, here's an example. Just because this is an idea that came to me today with this new 01 image generation. Right. Have you played with that at all on chat GPT?

I've done it. And yeah, chatGPT4.0, whatever the standard model. I don't know.

The new, like, image editor, they came out like a day or two ago.

Yeah, I've been seeing it. I didn't. Do I have to, like, select it?

I guess you just have to, like, prompt it. Right? Like, users don't really know if they have it on their Pro account until they start, like, playing with it. I was thinking it'd be interesting to have just like a stupid, simple web app where it's just one prompt and you upload your resume and it converts your resume into like an infographic. Right. For five bucks or something. And you just connect the API on the back end. You connect Stripe, Email, something like that. If you were going to build that, what would you use and why?

I would use Replit, probably.

Okay.

That's kind of my go to for starting point for a lot of things, because you can just get to MVP really quickly. I know it plays nice with OpenAI, which is what I'm going to use a lot of times if I'm building something with AI features, we could build an MVP literally in like 10 minutes. Probably. The hard part that it comes down to, I guess, is that consistent prompting. Right. So I might want to do more parsing of the resume upfront so there's less ways for the open AI model to kind of go crazy, right?

Yeah, yeah.

But, yeah, I think that's a great use case. I think you could do it with any of the ones that we mentioned earlier.

For sure. I'm glad you said replica, because that's the one I'm most familiar with. Do you find it to be difficult getting across the last mile, like, getting like, literally a sellable product that's published? What are your shortcuts to that, if any?

It's paying someone to do it and I actually haven't done this yet, but in Replit, there's like a little bounties tab where.

Yeah, I'VE used a bounty.

Or you could pay, you could pay like 150 bucks to get 3 hours of dev time from I think a pool of, like, agencies they work with. I need to partner with them on this. Actually, I haven't done that yet. I would go that route. If you're going to do it yourself. The Last kind of 10% is starting fresh with the agent or pulling that code into like a cursor or Windsurf and just talking back and forth with the agent. Okay, look at all my pages and tell me any concerns from a security or whatever. And doing that just like with literally each page, step by step with the AI, like a partner and being like, how can we refactor this? How can we make sure it's tied up, it's tight and ready to go, like production.

I hired a guy for a bounty once, but it didn't work out. Like, he had the same problem I did. I guess he wasn't that great. But it doesn't mean I can't find a better way.

It's going to be hard because even the best developer in the world, the guys I work with, say this, they're like, Billy, like, it'd be quicker. Rather it would take the same time to do that last 10% from your code than for me to just do it from scratch so I don't have to understand your code in your brain.

That's frustrating, but it is what it is. The last 10%, it doesn't even take as long as the first 90%. It takes longer. Yeah, 90%.

I'm on it right now with a client that I think I was supposed to wrap with like two weeks ago.

Yeah, okay.

And it's a little bit of perfectionism too, where sometimes, like, you'll get it to the client and they'll be like, this is great. And it's only going to be like two people using it. But then you're like, oh, but like, what about this, like, edge case? What about this edge case? And especially when you start thinking, like, how are we going to scale this? Then you get into those, I really need to fix this. But like, the approach of selling B2B for really niche things where one or two people are going to use it, I think it's really easy to get to that last 10%. Right. Because you're not thinking about crazy edge cases and redundancies and all that at scale just minimizes complexity.

What, like, business ideas have you seen or had lately that can be created with tools like these? Is there anything that you're excited about or that you've seen in the marketplace that excited you.

I'm excited about everything all day. That's just too much. I can't sleep because I want to be doing personal projects and projects for friends, projects for clients, projects for proposals, right? So there's a lot of stuff to build. I tell people all the time. I kind of just sound like a broken record here. But I don't start with direct to consumer. If it's your first app, like do not, it is a money pit, a pit of despair. You don't have enough confidence in yourself probably if you're just starting out to understand what to build without bunch of people on Twitter or your friends telling you, well, you should do this and you should do this. They don't know what you should build, right? Like they, they don't even know their own preferences. So you need to get someone who's going to pay you money to build out features and then like have them use it and get feedback, right? That's how you start building an app. And then maybe you could do, you could resell a similar implementation, but everyone wants little custom tweaks. So I think it's better to get good with like using the same stack of AI tools and maybe like your design components and then just sell kind of customized ones to other people. And so I've totally rolled over your question there, which is like, what's an idea you're excited about? But every kind of white collar business owner needs a smaller CRM. Like they're tired of HubSpot. I'm tired of HubSpot. It does too much. The pricing is opaque. As a CEO, like I shouldn't be spending all my time figuring out pricing models and passwords for three different integrations, right? So I'd rather pay upfront to have something that's mine that just does a CRM for real estate or car dealerships or something, right?

Like your CRM that has only the features that you want and use, right?

And then everyone wants custom document templating, right? It sounds like the easiest thing in the world, like, but everyone still just uses Google Docs and it's sloppy and messes up. I could automate anything. I could automate a horse, I don't know. And I still manually do the Google Docs thing because it's too fiddly to log in somewhere and then go to my templates, right? It's, it's just hard. It's too many clicks.

Docs new, right? What do you mean by a Document template. Like, what would that look like? What would be a use case for that?

Think about templates that are for a business. Like you're starting a new client, right? And you've got a bunch of documents you need to fill out. Some he needs to sign, some you need to sign. But they're overlapping fields with like, client name, my. My name, date, address, stuff like that. So like putting in, instead of pulling down a boilerplate document from like DocuSign or whatever, just uploading your own documents that you've been using in Cool Doc and then being able to like use those over and over. Right. It's like a custom DocuSign pretty much.

What are you working on right now as like a side project that you're excited about and why.

Okay, so I've got a good one, actually that I need to finish again. It's that last 10%. Right. I'm trying to do like stripe integration and I'm fighting with old code from when I wasn't as well. Practice. And the models were all different, so it might just nuke it and start from scratch. But it's chatgpt. It's a recipe generator. I've got recipe generator dot com.

What? How did you get that?

Well, I think there's a hyphen in there. Or generate. Generate dot com.

Okay. Okay.

Generate height. So I had been finding myself, like, you know, you don't like with your. With your fiance or your wife. Are you married?

Yeah.

Like, figuring out what to eat is like the oldest problem. It's been a thing for 200 years. So we would just go in and literally just put, here's the ingredients we have in a chat GPT, make us a recipe. And it worked. And like, we would make sauces we would have never thought of before. I made an interface where you can go in and save those recipe cards and then share them with people. And it's super. It's sick. And I think I'm trying to charge just like 25 cents per recipe or something instead of having you pay per month, because that's annoying. Like, you want. I want less SaaS, right. I want less recurring SaaS in the world for people. So that's kind of what I'm really excited about that I'm trying to.

Is that out there yet?

Yeah. If you go to generate recipes dot com, if you go there, you can use it just for one recipe and then it'll say you need to pay. And I haven't set up the payments yet.

So now, Billy, you just told us to not do consumer apps. So what are we doing here?

This is one where I found myself. This is a solution for me, right, to save myself. Like, I want to have these recipes to reuse, right. Instead of just losing them in my ChatGPT app, I want to store them.

Yeah.

So it was for me. So. And you're right, I did tell you not to do consumer apps, but I'm not. The reason it's not finished working yet is because no one's paying me to do it. Right. So that's why it's time.

You did say that you're eating your own dog food.

And then the other one I found is that I like to do gardening and stuff. And so, like, I have all these different little plants, like succulents and little bonsai trees that I move inside and then move outside every spring. And I super, like just micromanage them and end up killing half of them. So what I do is like, every day I log, I take a picture in a chatgpt of like a plant I'm having trouble with and I'm like, does this look better or worse? Like, what should I do? And I really want to go back and see how it's changed over time. But then I'm scrolling through my photos app and there's plants, screenshots, pictures of my fiance, you know, random stuff. So I just want to be able to scroll and see how, like my bonsai or whatever plan.

Yeah, like a time lapse.

Time lapse. Right. And so I made like a demo of that, but it sucks, so I need to like remake it so it's good and I think it'd be super useful.

I would use it every day and other people would. I mean, that can be an idea.

Like, but I don't want to do marketing. You know, that's the thing with apps.

But like, if you're solving a problem for yourself, you're solving it for others. Someone could build a time lapse app for faces where you could just upload pictures, selfies of yourself over all the years and it could just show you like a time lapse video of yourself just because people are vain and they probably find it interesting.

Yeah. Yeah, totally.

Do you use like a specific API or LLM for images? Specifically a preference.

So no, that was something I was wanting to figure out. But now that like, kind of the general OpenAI model is so good at it, I would say probably if OpenAI has a machine vision, I would use that model. I don't know if that exists publicly, but right now like the General All Purpose ChatGPT, I think it does it. So that's what I would start with.

I think talk about your, your upwork process. Like how do you find alpha on upwork? How do you find customers? How could people that have no knowledge of upwork, no account, log in, create an account, get their first job, maybe it's free and then like get more jobs and then leverage those little, you know, 50 or $500 jobs into higher ticket jobs.

That's a good one. I'm going to look up this guy because Nick Sarev, he's got really good YouTube videos from. If you're an automations developer, how to set up your upwork profile, I would say watch that video. Basically what he's going to tell you is like do a professional headshot, shoulders up. If you're near a major metro area, put that as your location instead of some suburb no one's ever heard of. Right? So do Washington D.C. instead of Arlington for me. The next step. So once you kind of got your profile set up like the next step is to just randomly go on during the week because it's all about timing. That's like the main thing is you want to catch people when they're excited about posting their job listing. Like think about from their side. And as an exercise you should set up an upwork like client account and try posting a job just to go through the process and you'll learn. But think about it. So you want to catch someone right when they're excited about the project and that just means you need to check. You could try to do all these automations to alert you and automate stuff but I think really it's just timing and then get really good at doing a tight proposal. The thing that'll set you apart always is to record a loom so they see your face, see that you're real. A lot of people who are hiring online have been burned in the past so they want to see that you're trustworthy. A lot of people are also conscientious about am I hiring outside of the US A lot of people have been like tricked with like VPNs to hiring people they thought were in the US and then they weren't and they might have been a little lower quality. So just kind of reassuring people on that stuff goes a long way on upwork. Right? And literally just scroll through stuff that you've worked on and show them. A lot of people will just ask like can you show me some. Anything you've built just so I believe that you can do this because I guess a lot of people are hiring people who say they can do stuff and they can't.

I've been hiring on upwork for like 12 years, since it was odesk. And I even have like a creator account. What's the other side of the marketplace called? Yeah, yeah, yeah, I have that account too because I've, I've played around and like you, like you said, you can learn so much having an account on both sides of it. You can just learn what your competitors are doing. I like logging in as. What is it called? Like a client.

The clients client. And there's freelancer and then there's like agency is a third one that not many people do.

So like, if you log in as a freelancer and just start trolling jobs like you said, even if you have no desire to actually make money on upwork, you can learn so much. Like, you can learn competitive intelligence on what people are doing. Like this. I saw this one job posting was like, I need you to walk around the city and take pictures of cars in parking lots and, and upload them to an app. And I was just like, what are they getting at? Like, what are they doing? What is the play there? You know, Like, I'm just so curious.

It's.

You can learn a lot from doing that. Anyway, what I was saying is over the last three years, the bot from like the business side of upwork are out of control. Like, I'll. I'll post a job and I'll have like 15 applications within five minutes with like this much text. And I put at the very top of my application, don't use AI. Like, don't use AI. If it's ChatGPT, I'm gonna ignore you. Put this keyword, put the word orange in your application. All this, it doesn't matter. Like, how do you stand out from all those bots?

I have the same experience and I wish platforms again. I don't want to make fun of upwork too much, but whatever platform would build that in where not even like they are checking to see if someone's a bot. But you could just say, hey, if the response doesn't have this in it, filter it out. That would be nice.

Yeah. So video is cool. But you know what's better? Long form audio via podcast and my newsletter, tkopod.com go there to subscribe for free to my newsletter. It's one email a week, very tactical. And then go to my audio podcast, three episodes a week. Stuff like this. You're gonna Love it. All free. No sleazy sales pitch tkopod.com the thing.

For you if you're a freelancer trying to land a job is you can't be squeamish about using connects to boost your proposal. So in upwork you can pay. It's like $0.15 per connect. It's like a currency and whoever has the. You have to spend a minimum amount to submit a proposal. But then it's like you get bumped to the top if you spend the most and you can see what other people bid. So I just don't even think about it. I spend the money every time if it's something I want to go for because it's just worth it. If you're spending three to four bucks on a proposal, that's a great CPA for a potentially thousand dollar job and like lifetime relationship if you're doing that kind of work.

What type of like non obvious things do you look for in jobs that you apply for? Like I say, maybe the obvious ones are like you want to see someone that has some sort of a hiring history. You don't want this to be their first job because they might flake out, right?

Yeah. Look for how much money they've spent on upwork. Make sure they've spent a couple thousand dollars. And I think freelancers can rate them so they'll have like a star rating. Read enough proposals, you'll be like, is this person writing like kind of defensively? Or like, do they sound like a tire kicker? Right. You'll kind of get an idea of that tone. Someone who's like, has too much like weird requirements up front. Like I think those are the main tips really is is like think like, does this sound like something a good client would write? And have they hired people in the past? You can see what kind of rates they're hiring people for in the past. Or if they're only hiring people for $5 a day, probably they don't want to pay you for your time, assuming you're going to be charging more than $5 a day if you're doing this kind of work. The other tip I would say you talking about taking pictures of the cars kind of made me think of this thing you could do to find out what to build and who to build it for is think of companies or business owners that you have existing relationships for. Think about what technology keywords or just workflow keywords they have. Look those up in upwork, both on the client side and the freelancer side. So see what kind of jobs are Getting posted from their industry and then see what kind of freelancers have. Because on freelancer pages you can have a bio and then like off the shelf like like packages for like an existing job type so you can find out like what kind of services and then go pitch them to your buddy or your friend in the community or whatever. And then you've got that, you know, competitive advantage because you're sitting in front of them and you're not on upward.

Right.

So it's how you figure out like what needs does that market have?

Yeah. What types of jobs are you seeing more demand and supply for? Like what would be the lowest hanging fruit for someone coming on upwork today?

I don't know that I look for just general jobs enough to answer that because I'm usually looking at job keywords for specific no code or integrations that I'm used to working with. It's totally different. Right. So if you're anywhere global like anything like copy or editing or kind of automations that need a human in the loop that's there's going to be infinite work for that. And then for if you're trying to do kind of higher cost, more developer things like building apps or building out automations. So if you're new and you're really good@likemake.com automations instead of building out like an entire automation system for you know like a sprinkler company, whatever, maybe like start searching for troubleshooting like they have an existing one that's broken. Right. Or look for existing like go high level broken. Right. Look for those kinds of jobs because they're going to be smaller scope and you're going to be able to build up. They're going to be I think easier to get but it's something that like a big agency might not bid on because it's probably not worth their time.

I feel like upwork is so overlooked from especially Americans like someone that wanting to quit their job or start an agency or you know, build these no code tools. Would you agree that like there's a world where if you made upwork your full time job like just applying to job making looms. Making looms. Making looms. Giving it enough credits to be at the top of the. Of the pile. I just have a hard time thinking that people wouldn't be successful like assuming successful, assuming they did a good job like trying to convert those small jobs to more recurring or higher ticket clients. Like if you, if you literally took it seriously like a 10 hour a day thing after a month you Would have a. You would have a solid business.

Oh yeah? Yeah. But here's what's going to happen. And like I was that guy, right? Starting my marketing agency, looking for clients like on upwork, all day, grinding. So there's two problems. One is that, you know, depending on how good you are at reading between the lines, maybe half of your clients could end up being like annoying clients. Not annoying, but not a good fit, I guess.

High maintenance.

High maintenance. And in like with things like that. That's also probably me not knowing my capabilities as well. So it takes time to kind of figure out what your capabilities are, not just say yes to anything or to everything rather. And then the second thing is you can definitely stand up an agency in a month, but you're going to run out of bandwidth. So you need to have the foresight and be good enough at hiring to hire early or you're going to be on like a deliverable treadmill. What you said is totally true that you could go from nothing to full time freelancer or full time agency in two weeks, three weeks. There's enough work.

I know several people that simply apply for jobs as freelancers as an American and then they get a job and then they just hire people on upwork, offshore and outsource all of it to them. Like they're just playing middleman.

That's really hard to do.

Well, I'm not saying it's the ideal, but it happens. Like assuming you know the subject matter, like say make.com or zapier, like you know it, but you're trying to offload some of the work. A lot of people would, you know, they check the box and say, I only want to hire from the US and so everyone else is excluded from those search results.

That can work. But it's hard. I mean, I've made a lot of bad hires overseas. It's just like a muscle. You have to train hiring.

Yeah, well, like with any business, you pick your headaches, right? You'll never avoid headaches. I see that all the time. I post about business ideas in the comments, like, what about this, what about that? I'm like, there's always going to be headaches.

Like, what do you want?

What's an example of a client you found on Upwork that you were able to leverage into like something higher ticket or more recurring?

It'll be just like back to kind of my marketing agency, starting with like an SEO audit or can you tell me why our pages are ranking like this? And that call will just turn into like, okay, now Billy's my go to. We'll have like a recurring call just to get consulting. Maybe he'll have some special projects or, or look at my analytics. So like any sort of one off audit or troubleshooting for like Facebook ads, Google Analytics, HubSpot, things like that can turn into regular consulting. And I think the, one of the keys to success there is into not getting stretched too thin or is to just keep that as like instead of saying yes to some project because it's a really good fit with the client. But if you don't have those capabilities, just stick to consulting and analytics and all, all like I'm going to tell you what you're doing wrong and how to fix it, but if you don't have the team to implement it, like I would just stick on that. Consulting, like consulting versus building out projects to me are two very different things. Does that make sense? Am I just.

Yeah, it does. So stick, stick to consulting. If you had to choose between the two.

Yeah, I would say stick. I think that's if you. If that scale is better for sure.

Yeah. Okay. What is your framework for managing your shiny object syndrome? How do you know? What are your signals of like this I need to lean into this side project or this client with all the opportunities out there. How do you look at it?

That's really easy money. Are people paying me to do this? Yes or no. Are people trying to pay me to do this? Yes or no? Otherwise I probably shouldn't be spending time on it. How can I help answer that. Less like a cave.

No, it's just like sometimes this is the most like lame romantic thing ever, but you just have an idea and it just resonates with you. It keeps you up at night and it's like, no one's paying me for this, but this makes so much logical sense. Like when do you make exceptions to that?

Okay, so I think if it's something like me where I'm using chat GPT and I'm noticing myself doing this frequently or like every day. But I think just like this AI helps you iterate and get that shiny object out of your body so that it's like way less of a problem for me. So like think about like a year ago, right? I would have an idea for something for like a game or like an SEO tool and I would spend all night on cursor and OpenAI trying to code this prototype. I'm not like a full stack developer by trade, but I can kind of get to it. But then as soon as something breaks it's another eight hour night of me trying to troubleshoot it. And so then over two months, like I'm losing sleep over this stupid little shiny object. But now I can like realize an idea in 45 minutes in replit or bolt or whatever and it kind of gets that shiny object I can look.

At scratches the itch and I'm like.

Oh, that's not as cool as I thought it was. Or oh, like, dude, I should show this to people or keep going. So I think, yeah, I think maybe that's more of the answer, is that finding how to scratch that itch, it's easier than it ever has been.

But sometimes you just, you, you have to scratch that itch. Like you just have to check the box to know whether or not for yourself, you know, like me personally, I hate being years down the road and thinking like, I should have tried that. I just should have tried that. What would have happened if I were to try that? You know, that's just the opposite argument to focus on one thing. And a lot of times most of the time when you try that thing, you scratch that itch, you realize it's not for you.

But now you know, yeah, but I mean, I say all that and my entire kind of journey in entrepreneurship is the opposite of what I just said, which is only doing stuff that people pay you to do. I'm able to kind of do the projects that I do because I spent a weekend with some random idea I had and I learned about this library or because like I worked for random people like in college doing gigs like selling real estate or doing like a waiter at the country. All these random jobs, you know, kind of coalesce into who you are. So you're right, you do have to scratch that itch. And before I ever sold my first like AI coded app, I look back and I had 40 replicates I created for fun and 20 directories that I had made with cursor locally. So that's 60 itches I tried to scratch before I was able to sell one. So it is important to get your skills up. You just have to be smart and kind of direct with what you're doing. I'm doing this. Give yourself constraints. Say, I'm going to try to make this in two hours. I'm going to try to make this thing that does these two features instead of setting out to make the greatest hundred feature app.

It's a lot harder to take credit for the compounding of experience that we have because sometimes we don't have anything physical to show for it. It's like, oh, I just kept pivoting. I had shiny objects, so I kept doing. It's like all of those things help teach you what you know today to get you to this one thing that actually did work. A lot of times I hear people just talk about the being wasted time. Like, you keep resetting, compounding. Quit resetting. Quit starting over. It's like you're really not like, you just stay in the game and you'll win.

You know what the funny thing about that is that I never realized till you said that a lot of what consulting is and freelancing people pay you to tell them what shiny objects, not to chase what rabbit holes, not to go down. I could say, oh, I've used those SaaS that SaaS that SaaS. Don't use it. Waste of time, not right for you.

I scratch that itch.

So they're paying me. They're paying me to. Yeah. Save them on analysis. Paralysis. At the end of the day, that's all that business is, is like you're not selling like a product or service. You're. You're selling decisions made for someone who has limited time and resources. Right.

Yeah. Interesting. Talk to me about the prospect of building out a custom app for someone like your. Your garage consulting friend and then using that app and selling it to other people.

You mean reselling kind of the sim, the same app?

Yeah, yeah, yeah. Have you done that or is that kind of harder than advertised on the surface?

I have not done that with that example. I'm doing one where I have a better example where it's like a custom data visualization for HubSpot for like a specific industry niche. And it's like this revenue attribution. It's just like a feature gap in HubSpot and they want me to build it out and they're okay with me selling it to, you know, other clients or companies. Right. Because they just needed that. It's not like a full app. It's like a kind of like your.

Clients okay with it.

Yeah, it's kind of like, okay, it's like a custom dashboard. Right, Right. But like it's on a server and everything because I did the research and you can't make that custom Dashboard and the 20 SaaS apps that are supposed to be able to do it. Right. So we just had to do it custom. And like the garage one, I mean, that's data viz too. So I guess what I'm saying is that like basic, like CRUD apps and data visualization, like they're not like proprietary. Right. And you can kind of reuse the same visual components and architecture without, you know, hurting anyone's competitive advantage.

I think there's a world where you could G2 reviews for those listening review site for SaaS products. You could scrape all those for HubSpot and find all the keywords like the negative sentiment keywords about certain features that are missing and build tools around that.

Yeah, totally. This democratization of software development is going to change like how SaaS works because all these big platforms have APIs where you can send as many requests as you want like in a year. Why would Anyone pay for 20 seats on whatever SaaS when they could pay for one and then have a custom app built that just pulls all that data in and out for them. Right. Like Headless, it might push, you know, more software to be more open source and more, more headless solutions. I don't know what's going to happen.

Yeah, yeah. What about the, the hard part here which is distribution. Have you find, found any hacks outside of Upwork to distribute these ideas or SaaS tools or, or apps?

No. Like that's why I'm kind of optimizing for selling B2B relationship based because that's what I'm good at and I think there's just so many burning needs there and I think they are all kind of interrelated and you'll learn and you'll be able to compound easily. So personally, right, so me like my, my Twitter and YouTube have been popping off because I just one day decided to record myself doing replit for a client, right. Or for like a personal project. I post videos like that and then someone from like replit. I think like the CEO of Replit like retweeted me with the video and the link so that just massively boosted me there. I'm definitely like a replit shill and I like, I use it a lot and I'm not sponsored. I really like them. That's one of the things about kind of picking a stack and moving forward with it. Get really good at, you know, bolt apps for enterprise or lovable apps for teachers and try to put it out there and share with people. I'm sharing like a lot of sauce on like the upwork stuff and even probably too much on some of these like app ideas. So being open always, you always get like 10x value from sharing value with people. That's just really helped me grow and get lots of inbound. And then the last thing is just being ready to capitalize on any sudden, like, influx. Like, I've been posting stuff on YouTube and Twitter for, you know, four or five years for different business ideas and my personal brand and just like, nothing sticks, right? But finally this kind of app building thing really resonates. And I was smart enough to kind of set up like an agency landing page and a paid calendly. And all of that has kind of helped me take advantage of the extra eyeballs.

Well, it's like if you're building cool things, then you might as well hit record. You might as well publish what you're building, right?

Yeah, yeah. There's some industries, like if you're in finance, right, you can't do that as well. Or like healthcare.

Healthcare, yeah.

And that's why those personal projects are important too, because you can always record your personal projects. Maybe not like a client.

You have this agency, you have all this stuff. Where can people find you?

I'm on Twitter, Billy J. Howell, and my agency is stupid simple apps.com in between the words. I'm like a hyphen guy now. Yeah, those are my main two things. My YouTube channel is the same Billy J. Howell. I have lots of tutorials on kind of replit. And I play pickup basketball outdoors in Northern Virginia. So you might just find me there sometime.

Okay. Well, Billy, thank you. This was super enlightening and I appreciate your time.

Thank you.

All right, what'd you think? Please share it with a friend and we'll see you next time on the Kerner Office.