I have been working on a game on/off for over a year now, the idea itself came about probably two years back. I was thinking of fun things I could do as a developer and I've always had an interest in game development. I mean I tried it many years ago with Unity and a bunch of other software. However most of the stuff I tried then were platformer game ideas and I really wanted something that was easy to get into while travelling. I then remembered the sliding picture puzzle game (reference below, although this is the numbers one) I used to play when I was a kid, the one which had only one empty space so you had to move a piece into that space until you could get them in the correct order for the picture to show. I then thought what if this was combined with a Rubik's Cube, so instead of creating a picture lets try to order the colors according to their rings. Think of it as a 2D version of the Rubik's Cube with one tile missing.
The Prototype
The first ever version was attempted with pygame and Python, this has never seen the light of day as it was in its earliest form which I was pair programming with another colleague called Andrew Berkowitz. I don't even think we had a working version of this at all as I lost interest in working on the project after a month or two. The project was left for dead until around a year later when AI started to become more popular, I have always been a skeptic of AI but since this was just a fun project, I took all the plans we had written up and just dumped them into ChatGPT, it gave me back some results but the performance of the game wasn't great and it didn't fully interpret my game requirements well enough, so I decided to try Kimi instead and got a more performative game which was playable. I tried the same prompt in Claude and got an even better result. The performance had improved but it then seemed the bottleneck wasn't the AI writing the code but rather the game framework I had decided to use at the time. I then asked Kimi and Claude to change out the game framework and use PixiJS instead, this yielded the best results for performance and rendering. The final prototype was created using Claude, but moreso Claude Code as there was one bug that none of the other two AIs at the time could fix and only Claude Code was able to do so. You can try the first prototype on this link.
Version One
Now I decided to get a little more serious with the game development after having my kids test the prototype and realized they were enjoying the challenge. I wanted it to be more mobile friendly and be installed as a PWA on your mobile device. This way it feels more like an app without the need to put it through all the hoops of the App Store. For this I wanted to give the user interface a bit more polish and get the branding correct for the game too. This is where the explorations began, as I was now thinking how I could make this be a game where you could share your results with others and encourage them to try the game themselves. This was a huge leap from what the prototype was doing, but with the AI models improving it didn't take much time to get this correct and in the game. I now had a version where you could share your results via X and WhatsApp with those in your network. I was happy with where the game was at but didn't have the time to put into marketing the game, so it was left there for a while with no love and not many people playing it besides myself and those who I had shared it with. I didn't believe it was ready for public consumption as there were other mechanics missing from the game that would make it more competitive and fun. Even though this has been my thought for the past year and the reason why I haven't been promoting it, I think it's time to change that and start promoting the game now as I am working on a new version which I'm even more excited about.
Try Pattern Panic Now.