[00:00:00] Gary: If you punch your card 10 times, do you get a free sub?

[00:00:03] David: You do not get out.

[00:00:08] David: Hi everyone. Welcome to the Biz Dev Podcast podcast about developing your business. I'm David Baxter, your host. Joined today by Gary and only Gary, which makes me sad 'cause I have no one else to talk to and I don't want to talk to you.

[00:00:21] Gary: Thanks for the great introduction. Looking forward to it.

[00:00:24] David: That's what I'm here for, man. I'm here to make you feel like crap. 'cause that's all I You know, every Every week I look forward to this, and every week you disappoint me.

That's fair. That's fair.

[00:00:36] Gary: You've been dabbling with something new and posting about it on LinkedIn. I wanna know more about that.

[00:00:41] David: I have been playing around with a gizmo. We haven't

[00:00:45] Gary: business related gizmo.

[00:00:47] David: it's a Yes, that's true. That's fair. It is. We have officially not said the name. It's not a secret, but I have not posted the actual name. But the name of it is Teela which is just a fun name. We started calling it Teela in internally just 'cause that was the name of the ui and I was like, that's a cool name. And we just kept going with it and so now it's fun. We haven't done logos or design or anything.

The screenshot, I've shown one screenshot and it is painfully boringly ugly, and that's on purpose. It's just enough UI just to get the point across, and everything is really bland, it's brown.

[00:01:18] Gary: Yeah, like we talk to business owners and entrepreneurs and people who build products and apps, and we always talk about the stage before you make something fancy where you could actually just test your market fit with an Excel spreadsheet and a telephone.

[00:01:34] David: Yep.

[00:01:34] Gary: In that, that Excel spreadsheet phase.

[00:01:37] David: We are, but it's harder when it's new. For me, it's ironic because I've never launched a product before. It's like we've been doing this for 12 years. I've been doing this for 25 years, and I've always launched everyone else's product. But ultimately the launch is not me. The launch and getting out to the public and all of that.

I've never really been involved in. I build it, I give them lots of ideas because I've seen people work and whatnot, but I've never been the one in charge of it. So this is all very different and weird for me to think about what to post about and what to do and who to invite and who to talk to.

I have a nice network which helps. And so we're trying to get more pilots is what we're basically doing. But it's just been very interesting. But we are in that same stage of. Of what you would call the Excel spreadsheet, but we're not a. Two-sided market, like a two-sided market is a, Airbnb is a classic example.

You've got your, the app is a middleman connecting two types of people. That's two-sided market. We're selling a product, so it's a one-sided market. It's, I have customers and they buy product, so that is different than I'm used to. So I have to build the product, right? This isn't something I can do without tech.

The whole thing, point of it is the tech. Now the good news is I can build this myself so I don't have to invest in that. And we're showing it around. We've got some pilot customers and we're playing with it and they're giving feedback and it's very fun and interesting and fun to build.

And we're also using all the AI tools to build it.

[00:02:58] Gary: That's what I was gonna get at. Yeah.

[00:03:01] David: That are we building this on the side and how are we building it on the side? What's the purpose of it and how the process of putting it together with robot help?

so the purpose of it was. It. We have lots of clients and they have for many years. They ask us for a different slice of data. Hey, can I see this report but a little different? Can I get a download of these users? And this very specific window, as it were. And the answer was always, sure, let me go and build you something custom.

It's usually not very difficult. Some of 'em are because some of 'em are really in depth kind of things. And secretly, not secretly, but we're not a SQL guys. We are good at sql, but we're not great at sql. And so sometimes we'd run into a wall and some of those queries would be crazy and we wouldn't be able to do that very well.

'cause again we're coders and designers. SQL is secondary. And so sometimes we'd run into a wall, but. We would always get asked these things, Hey, can you just show me the data a little differently? And you really have two options, maybe three if you wanna stretch it. If when you have that, you can hire someone like me to build you a customer report, or you can hire someone internally or use your IT guide to write you queries, which is basically the same thing as a customer report.

Or if you're using a custom report builder, like a Crystal Reports, something like that, you can have someone who really knows how to do that. Power BI is another example. It's all, they all kind of fall in the same bucket. If you have

[00:04:26] Gary: it's a complicated, it's a complicated process to take the data you have and rearrange it in a way that you didn't think you would need it until that time.

[00:04:35] David: Yeah,

[00:04:36] Gary: I. So just random customized reports took a big lift just to get, a decent set of answers.

[00:04:44] David: Yeah, if you want a custom report, it takes time. And so the idea behind this was, man, wouldn't it be cool if you had a tool that non-technical people could ask questions about their data and get answers? And so the answer is, yeah, that would be cool. And so the, that's how Teela was born. So we were building this for a client who had those specific things and we were like, Hey, can I turn this into a product?

And he was all for it. So we are, do running with it. And he's, he loves it 'cause it's way more advanced than the thing he would've gotten because it was just a small little development project that we were doing for him. And now he's got a pretty beefy little app that's way beyond what originally we specked out.

And so now you are not, now you have to train your data. That is the caveat. The data, your data is unique. It's your database, it's your whatever. It's structured data is what we use. And you have to, we have to understand that. And there's some big things we can do that are really easy.

And then the really cool part is you can use what we call colloquialisms, which is. You call your users cool people, right? For whatever reason in your industry, you call them cool people. And so you wanna be able to ask about cool people, not users, right? That's just a cheesy example. So we can tell your data that anybody asking about cool people that means go after this table.

And it, it's just very cool, becomes very natural.

[00:06:01] Gary: Yeah, and that helps if you have, A couple different people using the tool and they like you said, are asking it questions in natural language. You could have different associations for the terms they

used, To the same set of data.

[00:06:12] David: Yeah, so it's, it is been really fun. And so that's kinda where it came. And the reason we were able to do it now is because, a year ago it would've just been it, this really wasn't possible. And with now, maybe two years ago now with ai, we can turn natural language into queries.

That wasn't possible, not that long ago. And now that can be done, it just adds this whole new thing. That's really, when it comes down to it, if AI in general. The best thing it has given us right now is the first time in our history we can turn normal language into computer commands.

It's a new ui. If you boil down everything that AI is doing right now, it's essentially a new ui and that UI is turning normal speech into language and controls and tools and stuff like that. And that is something that is huge. That's where the vibe coding stuff comes from. I take normal language and turn it into code, for instance.

That's where, being able to get it to rewrite your stuff can be just natural language. That's a huge user interface change. Huge. 'cause before we had chatbots and whatnot, five, 10 years ago, we've had chatbots forever, and some of them were, quote, automated, right? But you had to use certain keywords and commands to make them work.

Google would be the biggest example 20 years ago, right? You wanted to search for something. If you were good at we used to call it Google fu, right? If you knew the right commands, you could get Google to be much more effective

[00:07:34] Gary: Yeah, you'd have to put in like a string of five different kind of question answer variables just in the search thing

[00:07:41] David: Yeah you put things in quotes and use, you could use variables and all sorts of stuff. Now that was early days, it gotten much better, but part of that getting much better is this concept of machine learning and language translation, which is what chat GPT originally was for. And that's, anyway, so that just allows us to have this whole new thing.

And so we're applying that. To your data, and that's what's we're trying to do differently. There's a million wrappers for different AI bots, but what's different here, what's unique about this is it's targeting your data. Not someone else's, not generically, it's yours. And that's what makes it unique.

And I think it's very fun. We've been using AI tools to build it. We use cursor.

[00:08:16] Gary: what I want to talk about more

[00:08:18] David: that's been an interesting ex,

I, will say flat out, I do not understand. And maybe I'm ignorant. But I've built several, we built a lot, a few internal tools and stuff like that. This isn't the first one. I have built quite a bit of code through these tools, and we are paying for all the top versions.

We're doing all the stuff, so we're not using the cheapy stuff. I don't understand how someone who doesn't know how to code can make a real app because it will make the most spaghetti weird code that's not, it's ama, sometimes it's magic. It'll just throw something up oh, that's really cool.

But then about, I'd say half the time, it builds stuff that's pretty, but does nothing, or does nothing or does something in completely useless from a UI perspective.

[00:09:03] Gary: yeah. Yeah. I saw a post recently, actually, I think it was this morning, from a developer, designer agency, where they were like, yeah, all these great new coding tools are awesome. I'm gonna start a showcase where you guys can post all the tools that you guys have built, all the products you have built, vibe, coding or otherwise with ai. And then, we'll help you promote 'em. And then right under that tweet two hours later was, okay. It's still waiting. So it's just a way to say, yeah, you see so much marketing around all these tools, how awesome they are, how they can code this, fix that, everything's great. I'm gonna build a product in two minutes and make a million dollars. Then when you actually get into building something that's beyond just showing off online and something you have to like test and sell and work with, it's a little bit harder than just telling, an MCP or some AI coding app. What to do.

You kind of need to know code and structure, backend, front end, all that

[00:10:02] David: There's a lot more to development and I think people are slowly realizing this. I've said this before, I don't know if I've ever said it on this podcast, but I'm old enough to remember the first wave of outsourcing. And that was, we know old friends of mine lost their jobs.

Lots of layoffs. This is like 2003, 2002. There's also a.com bust. It all happened about the same time where they just started offshoring everything and. Every big company went that way and everybody was, this is gonna solve our problems. I don't need to pay you stupid American who's really expensive.

I'm gonna go get this guy over here and he's gonna be $5 for his doctorate degree. And that went on for a couple, few years. And then you realized the big companies IBM was probably one of the biggest, who realized, gosh. That didn't work the way we expected it to. We didn't save as much money as we thought we would.

And so they brought a lot of it back. And now they have this kind of mix where there's some offshore and some onshore and there's a mix and that a lot. Now, I think AI is following a similar thing. It's ai, I don't need you stupid expensive guy. I'm gonna just offshore you, or not offshore, incorrect word.

I'm gonna put you put a bot. I'm gonna replace you with a bot. I don't need, I don't need any junior developers. I don't need any new recent college grads. I don't need any of you fools. I'm gonna replace you with a bot. And you're seeing that, right? You're seeing a lot of this with I just read an article.

Gen Z right now is having a brutal time coming out of college and get hired because, especially computer science or any of those, because people are like, why do I need you? I got ai. I don't need you. I don't need junior people. I might not be ready to fire my senior people yet. But I don't need more juniors.

And that's brutal. And I think we're following that same path in a couple of years. We're gonna realize, gosh, these bots are cool, but man, they need a lot of babysitting, man. They need a lot of X. They need a lot of y. And in our case, if you just tell a code to just start building stuff, it'll build it all in one folder, or it'll build it according to a structure you're not used to.

Or it'll start doing things like, here's a cheesy example. Every website you ever go to has a header and a footer, and then contents in the middle. And that if you do not tell it how to do it, you might get a bot. That's dumb. Here's the funny part, when you open up a thread to do a new work with a bot, every time you open a thread, you get a new bot.

Some of those bots are dumb. I don't know why. I can't tell you why, but sometimes you get these like super genius bots who like do magic and then sometimes you're like, you're having to explain to it like it's a 3-year-old. I can't tell you why, but boy, that's common. Anyway.

[00:12:27] Gary: And that's all within the same chat

interface back and forth?

[00:12:31] David: Yeah.

You open a new thread and 'cause you can't keep a same thread open up forever. 'cause they start to get slow and it's a whole thing. But you always, you have to regularly make, every time you have a new topic, you need a new thread. And sometimes you get the smart ones, sometimes you get the dumb ones

[00:12:45] Gary: So it's not within the same conversation, but it

is within the same tool. Okay.

[00:12:49] David: Within the same tool. But let, just using my cheesy example, every page has a header and a footer and then your content that changes all the time is always different. If you just leave the bot to its own devices, most of the time each page will have its own copy of the header and footer.

[00:13:03] Gary: And they won't be the same

[00:13:05] David: Yeah. Yeah. They might be slightly off. And so when you click between five different pages, the, the little menu item will move three or four pixels, or the button will be a slightly off. And because they're not shared,

[00:13:17] Gary: Yeah.

[00:13:17] David: not, yeah. Different hover effects, different, slightly different colors because you're not directing it 'cause you don't know what you're doing.

You're a vibe, code or whatever. You have to tell it what to do, and I've learned this all the hard way. I've built all those ugly headers, and you're like, why does that look so bad? Oh my gosh. It's because each page has its own unique header as opposed to shared, which is not just for organization. Sure.

That's part of it, but a lot of it is so that it's consistent. Consistency is everything in ux and if you're not consistent, that hover button move or that little button moves a little bit to the left you're breaking your trust with your users. It, there's a whole thing on that, but. Anyway, I'm, I've gotten way off into the weeds there, but it's really interesting to me how everybody's talking about how great these tools are to build apps.

And I don't need a team. I've seen the LinkedIn post where, I built a 30 man team, but it's just me in 30 bucks. And you're like, what? No, what Now I will say I can move at, I call it hyper speed. I can build new features that would take me a week. I can build 'em in a couple of days.

I. It is crazy how fast you can be when you're doing this stuff. Now, caveat is it has to be a new code base. All my clients like, why am I moving faster? 'cause all these AI tools, because you have an older code base that wasn't built with AI in mind, and AI just gets confused and can't do it. It's that simple.

[00:14:47] AD: BigPixel builds world class custom software and amazing apps. Our team of pros puts passion into every one of our projects. Our design infused development leans heavily on delivering a great experience for our clients and their clients. From startups to enterprises, we can help craft your ideas into real world products that help your business do better business.

[00:15:15] David: It's been a very eye-opening experience. I've not never built a product. Everyone talks about. Forever. Hey. 'cause every agency ends their run by building a product, and I don't want to end big Pixel's run. That's not what I'm doing to do this.

But I do think that the world is changing underneath us, and I think we need to change some of the stuff that we do. I think this helps us and also gives us a lot of really good experience when someone says, what have you done with ai, with these AI tools? I'm like here's a professional thing that we've done.

[00:15:46] Gary: And we know what we're talking about because we've used that and it's fun to watch, like behind the scenes, you get really excited about all the little things that you can make it do,

Is cool to see. 'cause you're like, you wouldn't believe this. I was able to add this in just a couple of hours.

Yeah, it was a lot of prompting and fixing and changing and

repeating the

But still it was way faster than doing it from scratch.

[00:16:06] David: It is a very different way of coding too, as someone told me. And this is very true when you use AI tools. Let me back up. Before AI tools, we wrote

[00:16:15] Gary: smartest kindergartner.

[00:16:16] David: That's true. We, but we wrote every line of code. And you married that code. You owned that code. It was yours.

It was personal, right? 'cause you wrote every line of code, you hand typed it, right? You might have some auto completes, but really my, you were typing and so every single line was burned in your brain over time when you're using these AI tools, you don't have that anymore. It is disconnected. 'cause you're not writing the code, you're guiding the code.

You're changing the way, the trajectory of the code and how it's being built, but you're no longer writing it. So if someone says, Hey, can you walk me through this code? The answer is, I've never seen it before. Literally. Now you do go back and look at it later, to make sure it's following good standards and all of that stuff.

But I'm saying while you're building, it's shoveling out mountains of code and it's not all good. And you gotta go and clean that up. But it is weird, you get this disassociation with code that we are as old guys, as old, none of our team is are new developers. It's just a rule of ours. And so they're all dealing with that.

It's this doesn't feel like my code, which is a huge paradigm shift.

[00:17:21] Gary: Or it's, it can be, like you said, the code is personal because you're writing every line of it when you're a developer without using ai, it's very similar. When you're designing something like you, every pixel within the

group of things that you're putting together, you know exactly why you laid things out certain way.

You know exactly how much space needs to be here. Exactly. The difference in. Opacity between the grays and stuff like that. So on that end, when AI's writing code for you and then you're going to look at the code and you don't see it the way that you would've done

it, Or it's a little off, it's almost like, how dare you?

Like

how dare you try to do my work poorly?

[00:18:00] David: To poorly. I guess in your world it would be this equivalent as if I started from a stock image, and maybe you tweak it a little bit here and there, but that's not your image, right? You're starting from

[00:18:10] Gary: yeah.

[00:18:11] David: and now you're adding your flare to it, but ultimately the core of it wasn't yours.

And that's for a developer. That's just weird. We've never had that concept before. It's only ours. Now, maybe we've grabbed a template or a library or something like that, but those libraries usually state black box. We never went into their code. They just stay over there and we just use them. And so it is very different in our old, fogey developers that I prefer to hire.

They're not old. I'm just giving them grief. But

[00:18:38] Gary: experienced

[00:18:39] David: Senior developers I,

[00:18:40] Gary: not senior citizen senior experience.

[00:18:44] David: but

[00:18:44] Gary: shouldn't talk about Scott like that.

[00:18:46] David: oh, rude, he still has no gray hair, so he's not even old.

[00:18:50] Gary: Oh, you were talking about Jeff

earlier.

[00:18:51] David: talking about Jeff. Jeff is ancient man. I don't, but it, what I tell them though is this is a paradigm shift. What the thing is, I want you to take ownership now of is the final product, not the path it took to get there.

Fall in love with the things that you're building and you're guiding because those are being made with your skill. AI is writing code, but it's doing it based on your. Prompts. That's the thing. You marry, that's the thing you love. That's the thing you take ownership of and make sure that's right.

Now what we do is we stop at some point and we go back into the code ourselves to make sure the code is good, right? That's a difference between a vibe coder and what we're doing because we can do that. But that is a huge shift for us and it's just very interesting to watch all of our guys trying to wrestle with this.

'cause this is the first advance. Ever we've had changed what we code a hundred times. That's just nature of what we do, but how we do it, that hasn't changed ever. And I guess no, seriously, if you wanted to go way back, you're talking punch cards, that was the last time how code was written has changed.

'cause it's been on a computer with a compiler since the seventies. That's. We're going way back, but you'll punch cards. You got some of the old guys, when I first started the guys who were in their fifties and maybe early sixties back then, they're doing punch cards where they would literally write code, print it out, and then run it through the computer.

Like literally put the card in and it would process it. This is different. You would write all the code beforehand 'cause you had one time a day to compile it and run it through the machine. That was totally different than what we did. And now we're like, we're the punch card guys. Because how code gets written now and in the future completely different.

And that's very strange.

[00:20:39] Gary: If you punch your card 10 times, do you get a free sub?

[00:20:43] David: You do not get out.

[00:20:45] Gary: I tried

[00:20:47] David: That's just bad. It is funny to talk to those guys though. I've, there's, I'm guessing there's a few of them still around, but they'd be pretty old if you did

[00:20:54] Gary: Yeah. That kind of hurts when you're like, yeah, we're the back in the day guys. Now the old timers.

[00:20:58] David: Oh my gosh. I can't imagine what this world looks like in five years right now in a, from a development perspective, I still believe you will need developers. I don't see, again, this is kinda like self-driving cars, that last 10% to replace us. Gonna be real hard. Like really hard. We're in that fun look.

I made a car drive a highway down a highway. That's fun and exciting,

[00:21:23] Gary: Yeah, the way it's talked about in design when it comes to AI and stuff is like you can have the tools create images all day long, even videos, even animations or whatever. But in the end you still need to have an eye and you still need to have taste. And you still, to be able, you

need to be able to go in and be like, this is good.

This is not good. This doesn't work 'cause of this needs to change because of this. So that extra little layer, I think is still gonna take. A human touch for sure.

[00:21:48] David: Yeah. But I do think. I don't know the bad coders. The fivers the outsource cheap stuff, the stock imagery, the I'm, that level of coder slash designer, i'm a low level website builder. Just like Squarespace came out 15 years ago. And made some really pretty websites.

They still do. I recommend Squarespace to anybody who just needs a beautiful little website. They don't need it custom. They don't need a special, they just need it out there. Squarespace is great. Fiverr not Fiverr. Stock imagery, stock videos. It's a good place

[00:22:22] Gary: just thought of a really good metaphor. I remember back when I used to work at the media company a friend of mine who worked right next to me, the other designer in our group, I. When Photoshop or Adobe would release new features, like every year we would, you know, wait to see what the latest, greatest thing is that are gonna

make our jobs easier and faster. And I remember every year we're just like, oh, let's automatic pen tool to make selections better than the magic wand. And we're just like, that's never gonna happen. You need someone to go in and use the pen tool, make bezier curves around these tiny little points. It's gonna take forever, but it's the only way you're gonna get a clean selection. And now the way

Photoshop uses ai, oh my gosh.

It is literally if it makes selections super good,

and if this was around 10 or so years ago when I needed this, oh my gosh, I would've been flying through work would've been

so

[00:23:16] David: Just some of the monotony is, and that's where I think it's exciting. It's an exciting time to be a developer. 'cause a lot of times it's I have this idea and I know to get from A to F to get this idea done. There's a lot of grunt work and a lot of stuff, a lot of typing, a lot of tr annoying stuff that really has nothing to do with my idea.

Now I can get to that idea so much faster and see, was that a good idea? How is this, was this crap? But maybe it is. But I found out in two hours, not two weeks, that's

exaggeration. Two days, maybe. Two days versus two weeks. And it, that's really awesome. That's where I think we have new this is not a sales pitch, but we have two new ways we're selling ourselves and this is all relative.

We have, I call 'em the launch and land, which I just love that. The launch is people who have ideas just like we've always done, but we can build a V1 using these codes this AI code. We can build you a V1 in two weeks and get that out the door. It's gonna be limited in features, but it's gonna be polished, it's gonna be beautiful of

[00:24:22] Gary: You to show and gain

some sort of reaction from the market to see if it's,

you know, persuable And that's a fixed price. And then we have another one I call the land. Which starts fixed, but could grow after that. The land is, I have this code that I vibe coded and made. Can you make this real?

Yeah.

[00:24:39] David: That, that, that's what I call the land. And so we're like, stick the landing because there's a lot of that going on too.

And so we're, we're changing too. And it's exciting and it's fun. I haven't been this excited. I love working with clients and stuff like that, but actually coding, it's been a while. It's been a, it's been a while.

[00:24:58] Gary: If you guys want to see under the hood a little bit about it, follow David on LinkedIn, check his posts.

[00:25:03] David: Yeah, man. Follow me on LinkedIn. I'm posting there a few times a week about the app as we're building it. And if you have any questions, you can always comment. Anyway, we will see you next week guys. Thanks so much.

[00:25:14] Gary: See you guys.

‍