14092026

Progress has been slow, but I’m still very excited about the project, and I feel my vision of it solidifies by the day. I recently have come upon a bit of a dilemma though, creatively and professionally. After some positive experiences using Gemini at work to handle some rote data analysis, I decided to see how it’d fare with GML. I wrote a fairly brief breakdown of something I’ve been working on — namely, the character generation system, and the various logics to randomise, save, and load characters effectively — and much to my surprise, the tool was able to very quickly generate some code based on my prompt. After carefully reviewing it, I found it to actually be more efficient in its implementation than what I had been planning on my own, and for the first time, reviewing my own logic reflected back to me in a slightly unfamiliar way, I felt I was finally able to grasp a number of concepts that I’d previously had little patience to learn.

My qualms are less so with the usage of AI in general in some grand moral sense, and more to do with my concern over the awkwardness of the workflow. It feels a bit backwards to me, describing the way something is to be built, then watching as someone — something, to be sure — rapidly builds it, and then reviewing it line by line, trying to gauge the consistency of a foreign logic against my own. There are times where it does things that I recognise are technically correct and probably even “better” as far as broader programming standards go, but because I was not formally trained as one and never cared much for others’ conventions, the results feel unintuitive and it takes me possibly greater effort (although certainly not time, to be fair) to slow down and understand it than if I’d just pieced together my own solution from the start. It’s like I’m not building the game on my own anymore, I’m just watching someone else build my idea off a spec sheet, and making sure they keep in line.

I’m willing to give it a bit more time to solidify my thoughts and come to a conclusion about how I will use it, if at all, moving forward; but thankfully, this project is highly unlikely to require any particularly inventive or demanding engineering to begin with, so scrapping the work that’s been put in heretofore and starting over with a more traditional approach would be no great loss for either the present or future. Strange times.