The Forgotten AI of Video Games
summary
Survey of concrete game AI: shows classic, non-ML systems actually used, implemented in Unity for a commercial indie (The Third Tourist) and battle‑tested in titles like FEAR, RimWorld and Far Cry. Covers finite state machines for animation, behavior trees for modular decisions, A pathfinding and navmeshes for movement, then compares goal‑oriented action planning* (GOAP) — expensive and unpredictable — to HTNs, which replace GOAP in practice because they’re cheaper, more controllable and easier to debug. Performance, memory and ticking rates are measured and tuned.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 2:18What is AI in games
- 8:39History of symbolic AI
- 15:50Core NPC AI techniques
- 21:34Finite state machines
- 23:02Behavior trees
- 31:16Advanced decision making
- 36:57Nav mesh and pathfinding
- 43:00Goal oriented action planning
- 56:18Hierarchical task networks vs GOAP
- 1:00:59Complementary AI systems
- 1:04:06Why learn game AI
related talks
01h 01mThe Forgotten AI of Video Games
51mWhat AI Can, Can’t, and Shouldn’t Do for Games
20mStudio Speaks: The Reality of AI Adoption in Games
48mMachine Takeover - Improving Your Game With Machine Learning & AI
18mThink You Can Build a Game with AI? Think Again!
01h 06mDreams of Electric Sheep - Aspirations for the Future of Game AI
21mWhat AI Can, Can’t, and Shouldn’t Do for Games
18mSession: AI Effortless Game Testing With Planning Agents