I am using AI a lot in my work and a few weekends back I thought: maybe with how smart AI has become these days, it could help me get that game done I’ve been always postponing? Well, it does.
I refactored the whole game from Game Maker Studio 2 (where my account was deleted because of inactivity) to PixiJS, React and Electron. Or to be more correct: I didn’t really do anything - AI did refactor the entire code base. And it took half an hour to do so.
So let’s address this. You can see on this blog how over several years I clawed through game development - using my free time as much as possible to get the most basic things done in-game. All these small steps did feel like successes - seeing how I was able to program those fundamental things myself. But did I get closer to a playable game at a speed that was satisfying? Not really. I kinda felt like I would need several years of either not working a job or having to spend all my free time working on this project if I wanted to finish it. And both sadly weren’t really an option to me.
Maybe that is how it’s supposed to be. Meritocratic in a way. If you are ready to give in to a project like this fully, than you will get the results you want out of it. No half measures. It sounds fair. But the question becomes: what is more important? Having the game or making it?
There are so many ideas I want to include in this game, that only a few years ago I considered out-of-scope, as they would mean I had to probably spend months developing them on the side. And now, those take an afternoon of back-and-forth with Claude Code.
I am conflicted about it. Especially when it comes to assets. A few years ago when I made the first characters and animated them, I realized how much time it would take to create all of those assets in all the characters' different situations. Every jacket would need to be hand-drawn for every piece of animation, every holding of an instrument meant a new set of images for every configuration. Seeing that I didn’t want to do all that myself, I tried giving an animation task to some people on Upwork. The result was atrocious. I guess I got what I paid for it (maybe 20 bucks each, for an animation that maybe takes an hour or two to make). But it showed me, why you need - or needed, if we are fully honest about this - hundreds of well-trained people in a game studio for big games, that basically make no gameplay decisions and sometimes not even artistic ones, because they are just working on assets day-in day-out. But it also stopped me in my tracks: I would never have those resources to make the game look like I wanted (or I would have to spend 20 years on asset creation myself). Either I considered myself happy with the game looking way more simplistic than in my imagination, there being way less options and outfits, or I rob a bank and employ 20 people for a few years. Those were my “realistic” options. And now a new option presented itself. Do it with AI.
I totally understand if people don’t want to support games that use AI for design & production. AI is an industry killer and supporting it gives power to a few companies that now sell back intelligence that they stole from the internet back to its customers - including the people who produced this knowledge and put it on the internet for free. But it is here now. And it is not going away. It is a tool that is available to practically anyone and working without it is just putting yourself at a disadvantage. I personally feel like one has to differentiate between developers that use AI as a tool - as I think I do - and “developers” that use AI as a decision-maker and end-to-end-producer for their game. I hate seeing all the slop on the internet as well. But I don’t think that “artisanal code” is a thing. Or that the 12th animation frame of a book-reading animation of the left arm wearing the outfit “blue blazer, white shirt” carries the weight of human purpose and is a symbol of its creator’s creative spark.
I think as AI matures and our use of it, we will more and more be able to discern, what kind of work is really just slop and can be reproduced in one prompt and what kind of work still needs a human artistic vision. I hope people recognize my game as the second.