I’ve been Vibe Coding some simple games and want to build something more complex. Anyone else interested in this?
1 Like
What kind of games are you working on?
I’m interested in game development, but I can’t spend much time on it at the moment because I have to stay focused on work.
I’d be interested in seeing what you’ve got, sure. Game programming hasn’t been a thing for me in the past, but it might be in the future.
My many interests include Rust, WASM, RISC-V, Vibe Coding, “Personal Software,” gaming, machine learning, data visualization, CLI tools, web tools, software dev tools and process, etc.
So, I tend to create a CLI tool, than add a visual wrapper.
Latest toy project: a Sound Fonts plus synthesizer CLI tool specifically for AI agents to call to create intro/background/outro music for videos with preset moods.
This has morphed into a visual mood and melody editor. Hope to blog/vlog about it soon. To see what else I’ve been doing check out the Code tab in my Blog Index but I have not blogged about most of the game/demos I’ve worked on recently (yet)
1 Like
My favorite way to evaluate a new model/harness is to make a grid based multiplayer game on the browser and to see how far it can go before things go off the rails.
Here’s my game side project, working title of “Tactical Heroes”. Its a turn-based tactical combat game with a league and roster management metagame frame (like you might have in a sports game). It also supports multiplayer (but the ‘meta’ elements aren’t set up to translate to it, just the combat).
2 Likes
That looks cool; nicely implemented. I’d play it if there was a tutorial or AI-vs-player mode. I’m too distracted to play against people (they would get bored, annoyed, or kill my character). I tend to create games that are single player or AI-vs-player (where AI is used loosely here, but could use actual AIs).
Did you see my Rust/WASM homage to Vector Graphics Games (some playable) Vector Graphic Games Live Demo
I forgot to ask, are you using a game engine or graphics library (jellyfish-app in the URL suggests something, but not sure what). I’ve played with vibe coding Godot: a moving cylindrical matching game
1 Like
There’s working single player mode under the default “League Mode” settings (player vs player is only enabled if you click “Online PvP explicitly).
I do need to make a tutorial, you’re right. 
The url is just a default generated name for a digitalocean app platform deployment. I’m not using any frameworks or engines; the frontend is all vanilla js. The game board itself is html canvas.
I played the game (and lost, but it was close). It was pretty intuitive. I sent the link to my son to try. Are you vibe coding, or not using AI for this?
(Dave) So what are you coding lately and are you using AI, and if so which agent/LLM?
Entirely AI
Via IDE, or CLI? ChatGPT (codex), Claude Code, Gemini, opencode? BMAD, spec-driven, ad hoc? I’m curious since I’ve tried many of these approaches. Are you using playwright or similar tech so that the AI can “see” the UI? I’ve been using OBS to record videos of UI interactions and then have Claude Code use ffmpeg to pull out stills and analyze the video of the UI interactions.
I’ve been using Claude code with opus 4.6 (initially 4.5) and I experimented a bit with gpt5.3-codex.
I use them directly in the codex app and Claude Code (I tried Claude’s desktop app briefly but it’s still too unstable imo).
I don’t have them integrated into VS Code.
I used Claude Code opus mainly. I have had problems with codex. I sometimes use opencode with Z.ai coding plan (much cheaper than Claude) using GLM-4.7. I avoid IDEs (Emacs and terminal preferred) but I do have JetBrains with its AI agent. Do you use AI to help with sound effects or music? I’m working on tools to help with this.
I haven’t done any audio to date, no. If you’re finding success let me know, I’d be interested in adding sound as another layer of polish
During the pandemic, I started getting into synthesizers and tried making some game music and effects. It isn’t AI and I haven’t been able to produce anything that’s very good with it, but it’s a lot of fun.
This is a free one:
There are tutorials on YouTube.
AI might be able to help with some of the code-based ones like supercollider, sonic pi, or tone.js.
Some of these discussions might be related:
My sound effects project will require a ton of work (but it is an interesting “how to benefit from AI coding agents” problem). My music/mood generator is “good enough” for my current needs (it saves me time looking for and downloading YouTube-licensed music, and can be used for more than YouTube videos). It is more of a “here’s how I vibe code” project than a “here’s a good music generator” project.
Thanks Josh; there’s a lot there, but I will be looking through the resources for help improving both of my audio projects. I saw some graphics resources in your recommendations, too and those are always interesting with regards to how I can leverage AI to get better visual results. I already use D3.js, three.js, plotly.js, mermaid.js, and Blender via AI, but will check out others; thanks!
1 Like
I just thought of one more tool called Opusmodus. It isn’t free and I’ve only experimented with the trial version, but it allows composing generative music with Lisp. If I had more free time I’d probably get it.
There’s a game music demo here:
and an 80-second clip demonstrating one interesting feature:
I had an LLM review your links and offer some ideas for me (or Claude) to implement, and I came up with a new approach to generate music for my videos: music-pipe-rs: Unix Pipelines for MIDI Composition | Software Wrighter Lab Blog - has links to code, related posts, and a related video (working on a video for this new music pipe soon)
1 Like
Looks amazing.