I hope to use this blog to document my experience making games with agentic engineering, the practice of developing software with the assistance of coding agents.

For a long time, I treated AI as a smarter but unreliable search engine with a chat window. But in the beginning of 2026, I stumbled upon a github page about building a game entirely with AI by "Vibe Coding". I had no idea what that meant, and 90% of the page was filled with technical lingo that went over my head, but when I was about to close the page, I saw the "get started now" section. It outlined just a few steps and looked simple enough to try immediately.So I followed the instructions, downloaded and fired up Cursor AI. By the end of the day, I had something that felt roughly like a full game loop.

It blew my mind how much I could accomplish in half a day, no, less than that, because my frugal self spent over an hour debating whether to pay $20 for Cursor after hitting the free limit. My desire to see things through won, thankfully.

Looking at the rough prototype, I thought I'd finally found the cure for my procrastination. After all, if I could finish a prototype in a day, I had no excuses left. Convinced I could then ship a game in a week, I jumped straight into another idea I'd always wanted to try.

Welp, a week was most definitely not enough, BUT!

Turned out unrealistic expectations made the best motivation.

My first not too cringy attempt worthy to be called a game — Second Shot Stardom: Prelude — took more than a few weeks and a few iterations. But It was complete. Procrastination didn't beat me this time.

I once thought, now that AI could write working code better than I ever could, all I had to do was give it an idea, then sit back and catch up on my TV drama backlog while it worked.

I was utterly wrong. There was still so much left to do.

The story behind developing the management sim Second Shot Stardom can easily fill a few blog posts on its own. Quoting Simon Willison, "getting great results out of coding agents is a deep subject in its own right"; besides, game design is never a field as easy as it sounds.

But now I know I can handle more ambitious projects and get closer to my dream game. It's bewildering and trying, yet I'm starting my second project carrying the hard-won lessons from the first. Stick around to see how it goes!