The Forgotten AI of Video Games
summary
Shows concrete game AI work: Unity project for a non-player character walking a demonic hellhound, built navmeshes and A pathfinding with G/H/F cost tuning, then replaces a costly goal-oriented action planning system with a faster hierarchical task network after FPS drops. Covers classic techniques used in production—finite state machines*, behavior trees, blackboards—and specific optimizations (convex polygon navmeshes, tick rates, decorator nodes) and tradeoffs between predictability, emergent behavior and runtime performance.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:05AI in video games
- 3:58Symbolic AI history
- 12:09Symbolic AI examples
- 14:38Expert systems
- 17:49Finite state machines
- 23:56Behavior trees
- 30:48Navigation meshes
- 36:07A* algorithm
- 40:02Goal-oriented planning
- 47:49Hierarchical task networks vs GOAP
- 53:57Director and utility AI
- 57:55Game AI experiences
- 1:00:04Q&A
related talks
01h 06mThe Forgotten AI of Video Games
52mThe Journey of AI in Games: A Technical and Philosophical overview
59mGame AI 4N00bz
29mThe Present & Future of AI and Games
38mPractical AI in Unity
39mEmbarrassing Mistakes: a short story about AI
02h 02mAI & Video Games, from Origins to Today: a (neural) link to the past
01h 01mIntroduction to Practical AI