Skip to content
TrackPodcasts
educationMar 17, 202635:53

My 13-Year-Old Built a 3D Game in a Week Using Vibe Coding (No Coding Skills Required)

About this episode

What happens when you hand a 13-year-old an AI coding tool and zero instructions? In this special episode, Paddy hosts his son Arjan for a hands-on look at vibe coding — the AI-powered approach to building software using nothing but natural language prompts.

Together they walk through Dolphin Quest, the fully functioning 3D game Arjan built using Replit Agent in roughly a week — complete with combat mechanics, an in-game economy, AI-generated dolphin artwork, multiple worlds, achievements, and boss battles. No coding experience required.

Whether you're a product manager, business analyst, developer, or just curious about AI, this episode shows what's possible when you combine human creativity with the latest AI tools.

What vibe coding is and why it's gone mainstream in 2026

Live demo of Dolphin Quest — a 3D game built entirely by a teenager using AI

How Replit Agent plans, builds, tests, and debugs its own code

The prompting techniques that worked (and the mistakes to avoid)

Why creativity and domain knowledge matter more than coding skills

Practical tips for anyone wanting to start vibe coding today

Timestamps

00:00 Introducing Arjun and Vibe Coding

02:35 What Makes a Great Game

03:31 Early Creativity and Dolphin Obsession

06:30 The Vibe Coding Experiment Setup

08:53 Tool Choice and Prompting Tips

14:35 Dolphin Quest Demo Begins

16:21 Worlds, Chests, and AI Dolphins

18:22 Submarine Speed and Cheats Testing

21:18 Achievements and Boss Combat

26:21 Build Time and No-Code Reality

30:06 Refinements and Vibe Coding Advice

31:47 Wrap-Up and Final Thoughts

⚡️ In each episode, Paddy Dhanda deep dives into a new human Superpower to help you thrive in the age of AI.

Host: Paddy Dhanda

Paddy works at the largest Tech training organisation in the UK and is passionate about helping tech professionals build human skills to thrive in the age of AI.

Contact Paddy: [email protected]

Subscribe to my newsletter:

👉 Newsletter: https://www.superpowers.school/subscribe



Get full access to Product Powers at www.superpowers.school/subscribe

Get every episode summarized

Each time Product Powers - Human Skills for Product Thinkers to Thrive in the Age of AI publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.

Email me new episodes

Free for 3 shows. No card needed.

Hosts & guests

Transcript ready

376 searchable segments. Every word is indexed and playable.

My 13-Year-Old Built a 3D Game in a Week Using Vibe Coding (No Coding Skills Required)

Product Powers - Human Skills for Product Thinkers to Thrive in the Age of AI

0:00
35:53

Full transcript

Product Powers - Human Skills for Product Thinkers to Thrive in the Age of AIMy 13-Year-Old Built a 3D Game in a Week Using Vibe Coding (No Coding Skills Required). Machine-transcribed; use the interactive transcript above to jump the player to any line.

So let me just give you a bit of a description about the game. So these are like pals, and this is your currency, and with pals in the shop, you can buy eggs when you ask me what makes a good game. I think achievements do, they give you a sense of purpose in the game. So every dolphin design is AI generated, they're made on Google Gemini and Chuck GT. From the city of Beaky Blinders, Birmingham, England, I would like to introduce you to Paddy Dandar. As the world becomes more automated and the robots take over, being an effective product thinker, still requires that human touch. Join me in a mission to champion human skills to help product teams boost creativity, enhance

collaboration and communicate with clarity. So pull up a chair, grab a small sore too, and make yourself very uncomfortable. So today we have a slightly different type of episode, I have a little human with me today. Say hello Arjun. Hello guys, my name's Arjun, and I'm 13 years old, nearly 14, and yeah, it's a bit about me. Very good, and you've got a lot more hair than me. You've gone bald. Thanks. So it happens when you get old. Anyway, so today we're going to be talking about a topic that's hitting all the headlines. It's going to be probably one of the big skills for the future, which is something called vibe coding. And we did a bit of an experiment.

So before we go into the experiment, tell us a bit more about you. What does a 13-year-old do in their spare time? So my hobbies are things like sports and gaming, I have an Xbox Series S at home and a switch and times of sports, I play sports in my school, I have like football, cricket, basketball, you know, all the ones that are normal, 13-year-old boys like. Excellent. And you've got a passion for gaming, so what makes a really good game? What are some of the key ingredients that you need for a really good game? I feel like there has to be some sort of aim to the game. It depends what type of game it is, for example, if it's like a story game or if it's like a shooter game, obviously the goals will be different and what makes it good will also be different. But for example, in a game where you're like, it's a competitive game where you're versing someone else, you're working together. You want it to be like easy to, not too hard to understand, the controls aren't, they're

simple but there's some sort of skill to it. But whereas if it's something like a story game or like a collecting game, you might want like things where you'll just get interesting, good graphics, you want, maybe if it's collecting something you want to have some sort of aim to eventually complete, like achievements or something, but yeah. And I know you've been tinkering with games for a while. I remember you made a game with cards a few years ago. Oh yeah. You want to tell us a little bit about that? What was the thought process behind that? Because I thought that was pretty interesting. It was quite a while ago now, but I made some like game with cards that, I don't know, I was just bored one day and I got some playing cards and I just drew some things on and then made it like different stats a bit and yeah, I kept adding to it. I think I started when I was like eight years old and a few years, and after a few years I kind of stopped and built quite a few.

Excellent. And another passion you have, and I know this because it's an expensive passion, dolphins. What's the story with dolphins? So when I was like maybe nine, ten years old, I don't know what it was. I just like really liked them as animals and I want to see a dolphin. And then when we went on holiday, I think to Dubai one time, I don't think it was like three years ago now. We saw some dolphins there and I got to pet them and we saw a dolphin show, which was really awesome experience. And then I think a year after we went to Turkey and then we saw a little dolphin show and then we met some of the dolphins at our resort. But I think the obsession has died down a bit now as I'm getting older. But at one point it was just like I was really obsessed with dolphins. I just like, I don't know. I honestly don't really know why, but like I just really thought there were cool animals and really interesting.

And I know you've got a whole room full of dolphins, teddies from when you were younger. And I actually remember we were on the train once and I met a colleague and I think you spent like an hour and a half just talking about dolphins and geeking out on dolphins. Do you know any dolphin facts? You can share? Well, a bit rusty on them now, but all right. I think the most common species is the bottom of those dolphins. Some species can be found in like many different oceans, like even the Arctic I think might be wrong. I think one type of species can live to like 60 years old, I'm not sure what the name was again. I think maybe it was the Atlantic spotted dolphin, but I forgot now. They are quite fast and very intelligent creatures. They have been seen playing games with seaweed and stuff like that. And there's also like obviously dolphin shows and places.

They have been able to understand how to use like buttons, like a dog, a bit which is quite crazy for a sea animal. No, excellent, right. So I took my dolphins and games. We run this experiment and actually a bit of the backstory, so about a year ago because I'm a little bit obsessed with AI. I thought it would be a good idea to maybe try out some of the AI tools to build a game. And I think you're a bit disappointed at the time when you got what went wrong when we did that first experiment. First up, a bit is a bit of an understatement. You're very obsessed with AI. But yeah, like maybe a year ago when AI was still quite big, but it wasn't as big as all good as it is now. We tried making like games on like things like chat, TPT and like what things were available at the time. And I wouldn't say there were the best results.

Like they were quite simple, like 2D obviously, like you would never see a 3D AI game back then. But yeah, there was quite like just like a few button presses, there weren't really many controls. It was just really simple like games that you could literally create yourself with really basic coding skills. But yeah. So not impressed? Not that impressed, you know. So a few weeks ago when there was a lot of talk about some of the new vibe coding apps, I thought it would be a good idea. Maybe get you off your phone, off YouTube for a while and try and build your own game and we were looking to allow you to just unleash all of that creativity and see what you come up with. Now, if anyone doesn't know what vibe coding is, well, it was a term that was really popularized back in February of 2025 and it's very simply the act of using natural language and prompting

to be able to build software. And just like you may have built or actually just like you may have generated images using AI by describing exactly what you want, we use prompting to build software. As you can see, this episode is going to be a bit mischievous. So we put it to the test, we gave you a popular tool and the tool that we used is called Replic. Replic, there we go. And there are other fantastic tools that you can use such as lovable cursor and a whole bunch of others, but I really wanted to try out Replic because they just launched this new AI agent feature, which sounded really cool. So first of all, before we jump into demoing what you came up with, like what was your initial thoughts? So I think around the start of the academic year, I think about October, September

time, you presented me with the Replic app and at the time I was thinking like, well, I wanted to use something to do I guess dolphins because that's what I used to really like and I just thought like any other creature wouldn't really resonate as much with me, but I want to make it like a bit of like a collection thing with an aspect of gambling maybe. But yeah, to make, I want some sort of like aim and some sort of luck based thing in the game where you can collect new things and just keep expanding and defeat and some sort of combat system as well. Wow, that's a lot of different game mechanics in there, not too sure about the gambling side, but hey, that's, we won't talk about that. So in terms of Replic, there are, there's a free plan where you can literally just create yourself an account and give it a go and you get so many free tokens, but then we very

quickly use that up and ended up on the pay plan. So what we're going to do now, we're going to bring up the game that you created. And if you're listening to this on audio, you may have to switch over to the video version just to see some of the awesome screens and the 3D world that Arjun has created. So it's definitely worth checking out the video to see this in a bit more depth. So yeah, we're going to switch over to sharing the game. Okay, so we've got the Replic.com website that we're showing on the screen. As you can see, there's a whole bunch of different features that they claim, but the one I was mentioning earlier, which was the agent three feature. Now, this feature is super useful because it's just so much more intelligent than some of the other vibe coding tools that have used and it has some awesome aspects whereby you

can actually ask it to create your plan of the requirements that you're planning to build. And it can even give you advice on whether those are good features, bad features, and you can see inspiration from it. And it can also test your application as it's building it and it will go through step-by-step as if it was testing it as a human and that's super, super cool. But what we're going to do, we're going to just quickly log in. So if we were creating an app from scratch, we'd literally just start to enter a prompt here and start the design process, but we're not going to do that because we've already got an app. So I'm going to go on the left-hand side here. And we can see a few apps that we've been tinkering with, but the one we're interested in is DolphinHatch, although you change your name afterwards, I think, to DolphinQuest. All right, so just as the workspace loads, what we have on the left-hand side here is

the area for prompting and that's where we would enter all of our instructions, all the things that we want to build. And on the right-hand side we actually get to see the result of the build and this allows you to actually start to use the app as well and just really test out what you're building. But what I'm going to do is we're going to just, before we jump into demoing the actual game, we're going to get up some of your prompts from some of the functionality you've already built and then maybe you can talk us through that. So let me just find that so we don't want to roll back. We want to really view some of the conversations you've been having here we go, I can see some here in blue. So do you want to just talk a little bit about the prompting side of things? So it's actually quite simple, like if you just like chat with you, like to do something,

it's kind of similar way, but except like implementing something in the game, obviously you have to try and describe it like clearly so that AI doesn't like accidentally do something that you don't want it to do or messes up something that was already working fine. But yeah, I think it's not too hard, try a little bit of a clear description and as long as the AI knows exactly what to do, it should be fine. But also don't, I've learned that if you put too much description or like if you put too many features, it just gets a bit confused sometimes, so you have to kind of do it individually sometimes or like just a few in one rather than like 10 different prompts at once. It will just ignore some of them and or we'll just mess up the ones that's already done. Okay, shall we dive into it? Yep. Yeah, I'm going to give us a demo. Oh, Dolphin Quest. This 24 is false, by the way, that needs updating, but yeah, let's begin.

You probably want to switch off the music, which is great, but yeah, you've actually added a feature, which is great. All the music has also been made by the AI, didn't ask it to make it specifically. I just said, make music and I've kept the music since the beginning. As you can see, we've got this like diver guy, pretty cool character to be fair. All the modeling and design is by the AI and it's completely made itself. So let me just give you a bit of a description about the game. So these are like pals and this is your currency and with pals in the shop, you can buy eggs and these eggs will give you dolphins, as you can see, you start off with this one and look, I've got egg and look, I've got Spike Fury, which is a new dolphin. So if you can also sell dolphins and you can use potions as well, but I'll talk a bit

about that after. So the main objective is to get pals, which gets you stronger dolphins and then with stronger dolphins, you can beat these enemies and eventually the bosses. There's a connections menu, so obviously you can see there's two worlds, I'll show you that after, but all different types of rarities, every image and like dolphin is unique. By the way, these, whatever the descriptions are usually doesn't mean anything, but so just ignore that for now. But let's move on to the frozen world to just show you. So this is a portal and if I click on this portal, boom, we are in the frozen land. As you can see, I'll probably show you later, but some of the bosses designs have changed. It's like this guy, he's turned completely frosty, whereas before he looked more like a shark. There's also chests, as you can see, the designs have also changed of the chests and the

chests can give you, if you look at my pearl count, they can give you pals and they can give you eggs. I'd say when you're starting out the game, they're a very good way just to get three quick pearls and they're also good for achievements, which I will show you just a second. But look, now you've collected a decent amount of pearls. We can buy, we've got some eggs from the chests, we can buy some new dolphins. And those dolphin designs, can you just explain how you create those? So every dolphin design is AI generated and they're not replicate generated though. They were made on Google Gemini and Chad GPT, which again, everything is just AI now. So you prompted those apps to create those images? Yep, and then I uploaded them to here for the dolphin designs and down the connection.

Obviously, but the stats were all made by the game itself. However, some specific dolphins are like, purposely, fully given stronger stats or weaker stats. So there's different eggs and things in each land shop, world shop. As you can see, there's like these only 250, whereas this one goes all the way up to 5000. And the cool thing that I implemented recently is the submarine. Now, let's just see how fast our character moves. It might not seem too slow, like it seems alright, but it seems a bit slow to me. After playing the game for a while, you just get a bit annoyed, especially since the bosses are quite spread out now. But if I go in the submarine, just click on it, also keep an eye on this spinning road. Look how much faster that is now. Oh, that's like the submarine with the submarine.

Super fast. Three times faster. Oh, just run away from this enemy. But yeah, once you click off, obviously you'll get out the submarine. But I implemented that, honestly, just to help. But if it was a bit simpler, it was also quite a cool feature. How difficult was it to implement that feature? Actually, surprisingly very easy. Some features, even just really simple ones, feel really hard because the AI keeps messing up. But this one, which I thought was going to be quite hard to do, was actually really easy. I didn't even have to ask twice. It just did it the way it is now. Which is really impressive. Let's show you a bit more actually what you know. There's a different thing which I can show you. The cheats menu. Before you think I'm like a big fat cheat and just lazy. Well, let's say, I didn't think you were cheap, but lazy, I think that's definitely true.

Okay, well, I guess that is kind of true because the reason I implemented this is because for testing to get every dolphin, which, by the way, spoiler alert, once you've got all 23 of a section, notice how it says 48, but there's only 23 in each section. That is because in each section there's one hidden dolphin, which you get after getting every dolphin as section. So that I had trouble with that at one point and it just wasn't really working. But if I didn't have the cheats menu, it would have taken me a very long time to get every other dolphin. So yeah. There's a way to actually speed up your testing when you're testing as a user? Yeah, it can definitely speed up a lot, like massively. And if like, same with the submarine, if these two features didn't exist, there would just be a lot harder to test and also to check

our last feature, the achievements. Like I said at the start when you ask me what makes a good game, I think achievements do. They give you a sense of purpose in the game and like something to aim to do. And as you can see, I've already got to, right by playing right now, but obviously there's easy ones and there's hard ones like get every dolphin in the game. That's really hard. Get every achievement even harder, whereas get 10 chests really easy. B06 bosses, not too hard. And the frozen ones obviously ramp up in difficulty, where let's just get one more chest and more unlock achievement. Alright, let me switch. There we go. One more achievement done.

Now, if I just show you some of the better dolphins so I can display the combat factor, then that would be good. By the way, those are the best eggs in the game. I just open sugar rush. By the way, I'm not going to show you every dolphin. You're going to have to look at some of them yourself and I'll give you a bit of a spoiler. The dolphin that you unlock after getting every dolphin in the frozen section and tropical ones are very cool. You don't want to miss out on those. But yeah, let's go find some bosses. I've got quite a good dolphin. So this dolphin pre-ease, they take out these lower bosses, but let's show you the combat system. Oops, that's a different one. Alright, captain shell back. So that's the boss. Yeah, captain shell back and you're a captain.

Cactus bloom. Okay, so that's your active dolphin at the moment at all of those, right? So there's a few different things you can obviously run away and then just completely avoid it. A bit like Pokemon, I guess, whereas you can attack. I was going to show you the, oh wait, now I can. The reason why I'm barely taking any damage is because my defense status is really high, which is also a reason the frozen bosses are really difficult, but you have to see that after but there's different potions as a freeze potion. The health potion obviously heals you, whereas let me just buy it on the freeze potion so I can show you the different potions in the game. Alright, I think his health's going to be back up full again, but yeah, that's fine. The freeze potion stops him from attacking for two turns, so you get two three hits and then you can attack you after. Oh, look, I already beat him.

Alright, that was a really nice strategy. It's sort of almost blew up at the point in which you defeated the body. Let's see that effect again. Is that something you had to prompt it? Yeah, it wasn't too difficult to add. I just, rather than it just like disappearing, I thought I'd be a bit boring. So look, it does a bit of a small explosion, which kind of gives you a sense of happiness. Need to feed him. But yeah, you can see how much the submarine helps running around. Oh yeah, these are different enemies that are poorly designed. Reef sharks pretty cool, in my opinion. So it's giant octopus. I'm the poison jellyfish. But some of them, like, you'll see this cube one here. There's a few that look like this. That's something I want to add in the future, or update in the future, as they don't really look the most appealing. But obviously those bosses were really easy,

right? Whereas these frozen ones, let's just try have a fight against these ones. I only did 18 damage, whereas he's doing loads against me. And see, this is why the freeze potion can be really useful. It just gets you free attacks, but I've got the feated, and I cannot attack now. Katka's bloom is no more. Yep. Look, his fainted switched orphans. But yeah, that's pretty much the game. Obviously there's some things maybe in the future, thinking of adding, but you have to wait and see. And can I cut all of our dolphins or achievements menu, which should be updated. Oh, find an exotic dolphin. However, we didn't get to finish the boss layer cheaper. Nice.

Okay, well, look, that game, I mean, there's a lot to it. There seems to be lots of little game mechanics that you've incorporated. How long did all of that take you? You would think this would take me like at least a year for a solo person to do about themselves. Like, remember, I have no team, nothing, just me and the AI. But I would say in total of actually like at prompting and stuff like that, I would only see if like a week, maybe like didn't take as long as you would think. It wasn't obviously like just one idea and just one prompt, and that was it, but it still takes like some sort of brain power. But it was very, it was a lot quicker than you would expect, like definitely. Like, if I'm all time and enough credits, this game could look 10 tons better in like just a few weeks. Yeah. And I remember the early version of it was quite

for me, really amazing to see was this 3D world, right? It created that really quickly. And I think from there on you, you refined it, but actually some of the core graphics in the world, it's very much what the AI came up with at the start. Yeah. So this scuba guy, he's been there, not since the start, but a very long time. However, some like core things like these, we're in the middle of the world, the collections been there since maybe to start nearly. These are all like ornaments all around all these like random things that don't really do anything. Just look there. They've been there since the start. And all the common eggs, like all this shops eggs have been here. And all you think that's been added as a submarine, but all these things were there since the start. And yeah. And at the start, I think there was only like 12 dolphins, which now there's

four times that I'm out. It's very impressive. Excellent. And so you know a little bit of Python coding, right? I know a bit. A bit. So you know like the basics of an algorithm and how you construct ifs and I guess loops and variables and all that good stuff. Did you need any of that knowledge for others? Zero. Pretty much. Whereas before in AI coding, you might have needed like some sort of like idea of coding to fix some bugs that the AI would have done and made. But the thing with the good thing with this is that like you, it kind of tests it for you. So if it, if you have to test it every single time and like see all the bugs that I made and tell it all that, it would take very long time because it would obviously make quite a lot. But it tested itself when it's finished

the coding. And then if it makes some mistake, it almost like chats to you and it's like, oh, this name was this and it should have been that. And then that meant this. So I will quickly change that. And that's very useful and it makes vibe coding a lot more fun and easier. And so if you're going to now enhance the game, have you got any ideas that you're dying to add to it? Honestly, I'm not. There's nothing huge that I'm thinking of adding but just kind of refining it and yeah, I mean, I don't want to spoil too much away. But obviously like I said, those cube enemies don't really look the best. And these bosses are really, not these ones, but the frozen bosses are really hard and they should be a lot easier to beat. But yeah, nothing, well, not any huge features yet.

Got it. And so if you were giving some advice to anyone who wants to start vibe coding, like what are some tips that they could start off with, like some things you've learned, you already mentioned like keep your prompts quite simple. Don't put too much information in a single prompt. Yeah. Anything else? Do you have to be really clear with your instructions or can you be a bit vague? You have to be quite clear. Otherwise, if you're quite vague, then it won't do the thing that you're hoping it to do. But also, I mean, if a 13-year-old can make this with just one simple AI app, then I'm sure you can make an even better app than me. Oh, that's an interesting one. Yeah. Well, we'll have to put that to the test. So if you do go often create your own masterpiece with some vibe coding, then please do get in touch and let me know how it goes and I'd love to check it out. Well, there's been an interesting episode. I think

it really shows the power of not just AI, but I think just in terms of when you bring the human element and creativity together with the AI, it's pretty amazing what you can create. I mean, there's still a lot of creativity that's gone into this that obviously replete itself hasn't been responsible for, but it's allowed you to kind of get your vision out of your head. Yeah. And get it built really quickly. That's something that was quite difficult to do before, which is very easy now. And so thank you, folks. I just want to thank you for listening and you young one. Thank you very much for taking us through that. I, by the way, folks, was also building my own app and it's not quite as exciting as Dolphin Quest. It was an app to build agendas for workshops. And I am very pleased with it. I'll stick the link in the show notes if you're

interested. If you're a facilitator or a trainer and you're looking to build dynamic agendas, then I'll give it a go because it's actually deployed now onto a URL. But yeah, it's not quite sophisticated as this particular app. It's fairly simple, but very functional. And I think, for me, that was the key. And I recently ran a training course where I used that particular app to build the agenda. And what was really amazing was during the training itself, I was able to show delegates in the room the fact that I used an app I had vibe coded and they were pretty blown away. And so I think up until now, a lot of the AI tools that I've been looking at, there you could say quite hyped and don't always have practical application, but I can certainly see vibe coding being a huge advancement in software development. It's still not at the stage where you could build a whole enterprise app. I mean, there's a lot of other considerations to consider,

such as the architecture and making sure we're building the code in the right way. So for those hardcore developers, I'm sure that you're probably looking at this and saying, well, it's going to be messy code, all of that good stuff, which is probably true for now. But things are only going to get better. And I think the more accessible we make software for the everyday person, young kids, it's only a good thing because once we remove that friction of how to code and the whole friction around understanding different coding languages, we can get on with the good stuff. And that is really unleashing your creativity. Any final words from you, Ajahn? Use AI. Use AI, there we go. And play the dolphin game. What's it called again? Dolphin hatch. Not quest? Well, that's your name's Dolphin hatch, but like it says Dolphin

question titles. Oh, you've got to change it. Yep. Okay, well maybe we could put that to the vote as which name is better Dolphin hatch or Dolphin quest or a new name or a new name altogether. Yep. Okay, let's do that. Well, thank you once again and thank you to everyone. That's listening. It's the end of another episode. Thank you so much for listening. Please do connect with me via LinkedIn and drop me a message and let me know your favorite takeaways from the episode. Also, don't forget to subscribe to the superpower school newsletter so that you can be notified of all future episodes. Simply visit the website www.superpowers.school. Thank you once again.

More episodes

More from Product Powers - Human Skills for Product Thinkers to Thrive in the Age of AI

View all episodes →