Side Projects And Building Games With AI Development Flows

Transcript is auto-generated and may contain errors.

Hey folks, I'm going to talk through some stuff I'm building like I was doing in one of the last videos. Um, so one of the last videos I was talking about a SQL workbench style application that I'm putting together, but um, in previous videos I've been talking about some other things I'm just kind of building on the side. Um, of course, if you've been around my channel for a while, you know that I'm building something called Brand Ghost that continues full speed ahead. Um it's been it's been pretty exciting because with Brand Ghost we've um a little bit earlier this year we pivoted over from just posting things onto social media. Um, you know, uh, if you're familiar with what Brand Ghost does, you might know know already, but if not, Brand Ghost is something that I build for helping me share my content out across, uh, different social media platforms so I can make it once and then effectively automate the posting of the content I've created to the different platforms.

And so, we were going down that path. And um it was it was a really interesting I guess like exploration for market fit and stuff like that because uh you know there's been products around for over a decade that do like you know crossosting and stuff like that. So we weren't attempting to like say hey look we we invented crossosting and that kind of stuff. Absolutely not. But um we were looking at some of the competitors that have been in the space for a long time and it's like they're just not they're not innovating and they're not really in my opinion like progressing what you can do in the space. Um so while they're you know incumbents they're also just like not pushing the envelope. It kind of it's a pretty common thing to happen right when you're when you have the sort of monopoly or the market share.

It's like it's easy to kind of sit back and and relax. So, you know, we were building Brand Ghost primarily around this and um it's been, you know, absolutely a lifesaver for me in content creation. Um, you know, I was just in Africa for 2 weeks and I could because I have all my content created, I could just keep it posting while I was gone, right? There are people that would not know I went on vacation for 2 weeks because what they see online is my social presence is still active. So to me, Brand Ghost, you know, has has and continues to solve that problem. But what we shifted over to earlier this year was um kind of solving the the problem that we have for our own business, which is like great. Yeah, social media is a huge part of it. And you want to make sure that as a business you have your social media presence active, that kind of thing.

Um but there was kind of like two factors. one is like we we need more like marketing support and so unless you have a team of marketing people it's really difficult to kind of ramp that up. You have to invest the time into it yourself and if you're investing the time into that yourself then you're not investing the time into other parts of your business that you're trying to grow. So that was kind of a tricky balance. Uh so you know how can we look into automating like um you know information collection for where we should be investing into uh marketing, branding, advertising, that kind of stuff. Uh lead generation, SEO, all this kind of thing. Like how can we get ahead of that? And then we said wait a second like this itself fits like perfectly into brand ghost if you were a business, right?

like why don't we solve the problem we're solving for ourselves for other people and so um sort of expanded what Brand Ghost is capable of to to make sure that we have effectively like um marketing automation for for businesses uh especially around their website and their social media. So Brandos has shifted that direction. The other motivating factor there was that when we would talk to content creators, right, like we would explain what Brand Ghost is able to do for them and literally we'd sit in these conversations with people and they're like, "That sounds great." Like, you know, uh definitely seems like it would help me. Definitely seems like it would take the load off. like um because if you're into content creation um one of the challenging things is like staying on top of constantly creating content is is a pain in the ass, right?

And if you have, you know, a period, it could be a week, could be a day, whatever, um where you're like, man, I just don't know what content to create, then like it's really easy to lose that momentum and kind of fall off. And so that happens to a lot of people, right? Like they're they're trying to create, they hit their stride, and then inevitably, like anything in life, you kind of hit like a bit of a a lull. And some people will push through and they're just like, I understand that's part of it. And for other people, it really just like decimates their ability to stay on top of it and they give up. And it sucks, right? Like it really sucks that people go through periods like that where they're doing something that they're really into and then it falls off for them and they can't get back to it.

Right? Not I don't blame them. I understand that it's challenging. So that was resonating with people and then we'd say cool like here's you know here's the price and it almost didn't matter what the price was. if it was more than $0, it was like, "Oh, well, you know, my my content isn't actually it's not generating money." And that's a tricky thing, right? Because with content creation, anyone can go create content, right? Like anyone watching this video right now, and by the way, I I'm not saying this in a way that's like condescending or facicious, like anyone who's watching this, if you wanted to, you absolutely could go start creating content there. It's such a great thing that nothing's actually stopping you from doing that. So, that's that's awesome. Monetizing content is challenging, right? There's lots of like, you know, things online like quick ways to like get monetized and create content and retire from your job.

Like, look at this channel. If you're watching this on YouTube, look at this channel. Look how many videos I've published on this channel alone. Right? I think this channel just hit its 2-year anniversary or something like that. There's over 500 videos on this channel. Do you know how much money I make per week or sorry per per month on this channel? Roughly $14, right? Like it doesn't it doesn't make money. So like it's not just make content. If I wanted to make more money from content on this channel, I would have to go seek out sponsorships. I'd have to go, you know, push products to all of you while I'm making videos. So, when we would talk to people about Brand Ghost that were content creators, a lot of them would be like, "Well, if it's more than $0, it just doesn't make sense for

me because my content doesn't make money." And it's a really difficult cell to tell people, hey, like if you want to make money with your content, if you invest into your content, you know, publishing, then it will help you scale to that point. And it's like you you're kind of asking them to like just like, you know, trust me, bro. Trust me, bro. like once you're able to scale your content then you can focus on parts that will make you money. So, it was a really difficult sell. And conversely, when we started approaching businesses who they're already selling products and we said, "Hey, look, we can help you with this." Instantly, they were like, almost like, "We don't care what the price is. Like, we're interested." So, it's almost like the conversation flipped. Now, I'm not saying that like businesses are just like blowing tons of money on us because they they'll pay anything.

That's not what I'm saying. What I'm saying is that like given the same I don't know like same technology and the same thing that we're trying to do the different audience is much more motivated to pay money for it because they know or they it's easier for them to trust if I put money in that should help me get more money back out. So those are big driving forces for Brand Ghost. Um, I actually didn't mean to make this video about Bran Go, so I apologize. But to wrap it up, Brando is still going full speed ahead. And I just wanted to say like it's been it's been really interesting to kind kind of like navigate um, you know, our target audience and adapt the technology. Now, if you're you are a content creator, you want to post stuff on social media, still perfect fit for that.

I still use it for all my stuff, which is the really cool thing. like if as a business it doesn't work, I will always run it because I need it for my content. So, it's it's not like a thing that's just going to die, which is uh which is very nice. Okay, so beyond that, um I had this thing called Squeal that I was putting together SQL Workbench. Um, I have I'm not going to make a full video on this now because I want to play more with it, but I have um this secure agent to agent communication platform that I'm building. And the purpose of this one is to be able to have you can use it beyond co-pilot, but for me, having co-pilot sessions, be able to interact with each other. Uh, kind of like chat for agents, but secure. Um and there's some things around the infrastructure and uh security encryption uh that make that sort of like the the feature set.

So um like I said, I want to exercise that more for myself before I uh talk a lot more about it, but something I'm building. And then it's been like a couple other things, but I'm trying to do I'm just trying to do more with games. So, if if folks have seen other videos of mine, um you know, I said this in a video the other day, right? Like I got into software development because of I wanted to make games. And so, um I talked a little bit about two of these games that I'm putting together. Um one is a role playing game that I've just kind of been building on and off for 20 years. And um the other is this sort of like creature uh breeding and caretaking game that's modeled after a game that I used to play when I was a kid.

I wanted to talk about this one a little bit more just from the perspective of using AI and building it out. So currently like obviously if folks are you know into the AI scene you know that models and stuff are coming out like seemingly every other day there's like the new world's best model. Um you know 10 10x better than the last one and and so right now I'm using Astra for this. And so I I had mentioned in the previous videos that when I was in Africa on my travels but with downtime I sort of did a highlevel game design I don't know talking about the mechanics with chat GPT just because it was convenient recorded all of that. Uh made a couple of technology decisions like I want to build it in Unity.

um you know some some of the highlevel architectural concepts um again not to do a deep spec in chat GPT but just to talk about some highlevel decisions and so when I got back from Africa I took all of those markdown files handed them to Astra and I said I want you to you know high level do a decomposition of these into some GitHub issues right you're not specking all of them out in detail right now but high level like these are the areas the game mechanics and then I want you to like order an implementation plan where we can do some vertical slices so we can basically get you know one vertical slice would be in Unity can we see a creature okay in Unity can we see a creature animated and then like uh given that the creatures have some AI mechanism can

we see them like walk around and like sort of change their animation based on on their their state and so just like delivering in vertical slices but I had it take the design docs make some highle GitHub issues and then have an order that they were going to work through these vertical slices and I said explicitly um don't spec out all the details now make it very apparent in the issue that when it's picked up to be worked on it probably needs to be decomposed and what's been super cool is that I've left it on autopilot and it's working through this entire game. And I don't get me wrong, I'm not going to like let it like air quotes finish and then I'm like I'm done. Like it's perfect to ship. Like absolutely not. But when I check in on it uh every day and sometimes it's a couple of times per day.

um when I check in what I'm doing is it's still you know still running in autopilot and I'll say hey like open up Unity uh you know to a a scene that demonstrates like some of the latest functionality and it launches Unity it's still working away like on the next you know vertical slice but it launches Unity and you know every time there's some other new like sets of game mechanics that are that are present so It's been really really cool to see that come together. The thing that I noticed it was doing that's kind of awkward um and I didn't I had to like ask it but it was instead of building like a cohesive game, it was building like each of the vertical slices like the functionality it was delivering almost like mini demos if that makes sense. So instead of being like here's one game, let me play it.

It was like here's a little game that demonstrates this functionality and if you want to see this other functionality, I'm going to go run this other little demo game. It's all the same game. I don't have a good way to explain it, but it's almost like running playable test cases. And so I noticed it was continuing to do this. And so yesterday I said, "By the way, just curious, did you have, you know, work scoped out to like to kind of move away from this so it's actually a the game that we've talked about instead of these, you know, individual playable little demo scenes?" And it was like, "Yep, absolutely." Like, you know, building out it explained like throughout, you know, this part of the development cycle, here's what we're focusing on. And then beyond that, here's this work item where we spec out like how to go uh merge all this functionality into a you know the cohesive gameplay experience.

So it's cool like it has it on its road map. Um I haven't checked today. I'm coming back from the gym right now before work. So I'll see if it's gotten there. Um and then the other thing that has been not working well is around the art direction. Oh my goodness. Uh this was this is definitely my fault. This is one of those things like you know uh we we see it all the time where it's like just because you can use AI like talk to AI doesn't mean that you can do something properly and like so very guilty here. Um, can I as a as someone who's been programming for 20 plus years, been in software engineering for over a decade, like can I have AI build software for me? Hell yeah. Do I feel comfortable doing it? Hell yeah. Am I someone who does 3D modeling?

No. Do I know anything about materials and 3D games? No. Am I any good at uh, you know, artwork for textures and graphics? No. So, what do I do? Hey, hey, Astra, go go go make me professional looking game assets. Make no mistakes. Um, and as you might imagine, it's been horrific. It's been absolutely terrible. I wish that in this moment I could show you pictures of what it's done because it is so terrible and it's like it's hilarious. It's hilariously terrible. It has MCP connections to Blender. So, what it did do is it was like making uh like a space station uh interior model and stuff like it did it, but it looks like it's it's not like air quotes professional. It looks like it looks like a child was able to get Blender and and make something. The textures are all like, you know, super smooth, like cartoony, not they're not textured.

Like, it just looks like it's a a joke. Um, but the the creature models are funny. And I will tell you that last night I spun up a second session and so I could let the the primary one work on the actual game. And then I spun up a second one and I said, "Look, you're in charge of art and like we're going to talk through some of the things that have not been working well and basically I'm leaning on you to be the professional uh artwork." And so it was like, "Yeah, we're on the same page. You know, we're going to have higher resolution models, do like materials that aren't glossy and shiny when it should be like texture for fur." Like, "Okay, this all sounds like really good. We're We're going to nail it this time. I woke up. So, I checked on the artwork this morning before the gym.

I I woke up and I looked at this thing and the creatures normally they uh what it created the first time was kind of like a fox looking creature. So, it's got four legs. It's kind of it's shaped like a fox, but it looks kind of stupid, right? It's not it's not a good fox. And I don't know if any of you have heard of like this raccoon Jimothy uh that's in Seattle area in Washington. Um that's where I am right now. Right. If you haven't heard of this Jimothy raccoon, it's got I think it's called like short spine syndrome. And so it's like compressed. Kind of looks like a a ball instead of a a raccoon. And so what what Astra made for me overnight and it was all like proud of itself like oh we now have like professional model with like more you know my god I can't even speak uh like more polygons like everything is it's all like everything's good it's checking all these boxes and like the screenshots available.

I opened this thing up and it took my stupid looking fox thing and made it look like a Jimothy version of it. It is so outrageous. Um, if I would have had coffee in my mouth, I would have spat it out on the spot. Like, it was unreal. So, one of the things I'm struggling with right now is like I know that I've seen people having really good success creating games and stuff with Astra. And that includes like the graphics and the artwork and stuff like that. But I did a quick check on chat GPT when I was like eating uh eating breakfast this morning and it was pointing out to me that a lot of those are because people started with some model or some texture to point it in the right direction. I didn't. I was just like make this and it was like oh yeah we got this.

So I think I'm going to try that. I'm going to go see if I can research um you know good free textures that are like high resolution for different materials. Um maybe there's some free models that I can use as a base. I don't want to rely on assets from like cobbled together free packs of resources because it's in the end it's going to look like it's cobbled together. But maybe I need to do that to give it a base because right now oh boy. Um, like I said, I wish I had a screenshot for you, but um, that game so far is called Evo Habitat Zero. I'll probably put it on, uh, on my website when I have some screenshots that don't look uh, completely outrageous. Uh, but my my sort of business website is nexus software laboratories.com. So, you'll see more of the stuff I'm talking about building on the side showing up there.

So, anyway, thanks for watching. um bit of a journey and yeah uh if you got questions about software engineering uh career development that kind of stuff leave them below in the comments or you can go to codecommute.com you can submit stuff anonymously that way and I will make a video for you and hopefully that helps other people as well. So thanks for watching and I will see you in the next one. Take care.