How Athena Crisis Built a Fast, Deterministic Game AI Without LLMs
cnakazawa · x · 2026-08-04
The author explains how Athena Crisis was powered by a conventional game AI rather than an LLM.
- The system combined the game’s existing pathfinding, k-means clustering, and greedy action selection.
- He modeled the heuristics on how he would personally play, and chose a stateless design so the AI recalculates from scratch after each deterministic move.
- Determinism made the AI easier to test and helped players optimize against it, though a few deliberately limited random choices were kept to avoid making matches too predictable.
- The first version took 2–3 seconds per turn, then was optimized to be 10x faster so the server could compute an entire turn while the turn-transition animation played.
- After months of iteration based on player feedback, the game launched to the surprise result that many players considered it the best AI in its genre, despite the author having never built a game AI before.
Related event: Y Combinator to Host Student Hiring Day on August 15(2 posts)→
More from Research
- AlphaFold’s real impact, the thread argues, is reshaping biology workflows — rishabh16_ · 2026-08-04
- Paper photo shows “Fundamental Limits of Caching” in IEEE Transactions on Information Theory — cneuralnetwork · 2026-08-04
- An ARC Prize team says its CUDA C stack is already 10x faster than the baseline — GregKamradt · 2026-08-04
- Wayve says its GAIA-4 world model is now useful for robotics simulation at scale — alexgkendall · 2026-08-04
- NVIDIA open-sources NOOA, a Python object-oriented framework for AI agents — Roger_M_Taylor · 2026-08-04
- Light Origins demo shows a humanoid robot switching between walking, climbing, and vaulting — Distinct-Question-16 · 2026-08-04