Indie developer Chris Ryan explains how to build an AI harness for game development
Overview
Indie developer Chris Ryan argues that LLMs struggle with games because much of a game is not code, so they cannot easily read or modify it.
He says that structuring a game as code, or building a harness around it, lets an LLM reason about and change the game. He expands on this in a blog post adapted from a September 2026 Indie Rodeo talk.
A separate post by kkukshtel, published October 10, 2026, describes techniques for making games legible to AI agents, including fixed randomness, deterministic state, command-based architecture, and profiling. The report does not establish how either author's methods perform in practice.
Written by AI from the articles below · updated Oct 14, 6:32 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- Hacker News · AI (150+ points)BlogHow to engineer an AI harness for game development
AIIndie developer Chris Ryan argues that LLMs struggle with games because much of a game is not code, so they cannot easily read or modify it. He writes that structuring a game as code, or building a harness around it, lets an LLM reason about and change the game, and he expands on this in a blog post adapted from a September 2026 Indie Rodeo talk.
Heat trend
Not enough continuous observations to show a trend yet.