AI and Games
Game AI, NPC behavior, and pathfinding with a focus on technical breakdowns and developer insights.
Nutrition Label
This channel produces high-quality educational deep dives into the algorithms and logic driving video game characters. The content bridges the gap between academic research and player experience, often synthesizing developer presentations or interviewing creators directly to explain complex systems.
Strengths
- +
- +
- +
Notes
- !Content often synthesizes developer talks or papers rather than showcasing original code implementation.
- !Explanations focus on logic and architecture concepts rather than copy-paste tutorials.
Rating Breakdown
Breakdown across the key dimensions we rate. Methodology →
Recent Videos
A Lazy Saturday Playing Resident Evil: Requiem
COTA Explained: GameBot’s Low-Latency FPS Bots
Snowcap: GPU Profiling Using Machine Learning | Al and Games #83
Ninja Gaiden: Ragebound's Self-Serve Difficulty is Great | Design Dive
Behind the Scenes at the AI and Games Conference 2025
MLMove: The LLM Trained to Play Counter-Strike | AI and Games #81
Building AI Speeders with Machine Learning in Star Wars Outlaws | AI and Games #80
Why this rating
Evidence receipts showing why each dimension is rated the way it is.
“Tommy sits down with David Durst: a Stanford PhD graduate to discuss his research”[00:00] →
The video is a primary-source interview with the creator of the technology, demonstrating the system directly rather than speculating on third-party news.
“The network takes a local height map of the terrain, the nav mesh... and the target path... and outputs acceleration and steering.”[04:15] →
Explains the specific inputs and outputs of the neural network architecture rather than using generic 'AI magic' buzzwords.
“Standard navigation meshes... are binary. You can either walk there or you can't... Speeders are physics-based... they drift.”[02:30] →
Clearly articulates the technical friction point (navmesh limitations) that necessitated the machine learning solution.