Build Games with AI: Tag Assets, Skip Boilerplate, and Let…

Arturo Nunez reveals how tagging game assets and using level‑of‑detail context lets an AI generate game logic instantly, slashing boilerplate and enabling rapid prototyping.

Imagine building a game by simply describing what you want, and an AI stitches together the right pieces for you.

⚡ TL;DR: Arturo Nunez shows how tagging assets and using level‑of‑detail context lets an AI generate game logic on the fly, cutting boilerplate and enabling rapid prototyping.

The Simple Version

Think of the engine as a toolbox where every object is a piece you can tag with intent like “character” or “drivable”. The AI reads those tags and assembles the needed systems automatically.

🎯 Think of it this way: It’s like labeling Lego bricks with what they should do, then letting a robot snap together the right structure.

How It Actually Works

Context Assembly

The AI only sends the parts of the scene that are near the editor’s focus, packing full tag data for nearby assets and collapsing distant ones to a simple type.

🎯 Think of it this way: It’s similar to how a game engine streams textures — high detail up close, low detail far away.

Tag‑Driven Systems

Systems query entities by the tags you attach, so a “vehicle” tag can trigger movement logic for any object marked drivable.

🎯 Think of it this way: Imagine a radio that only plays songs tagged “rock”, automatically filtering the playlist.

Technical Deep Dive

Nunez borrows rendering‑level of detail to prune the prompt, sending only necessary tag values to the model.

He uses an Entity Component System where tags replace traditional component lists, letting the AI resolve movement, animation, and physics on demand.

The prompt’s context is built from a hybrid of full‑scene tags for nearby objects and placeholder types for the rest, keeping token usage low.

Why This Matters

Game developers can skip writing repetitive boilerplate code for each character, focusing instead on creative intent.

AI‑assisted prototyping shortens iteration cycles, letting designers test wild ideas like making a building drivable in seconds.

This approach also opens the door to world models that understand scene structure, not just raw pixels.

Key Takeaways

Tagging assets with intent replaces manual component wiring.
Level‑of‑detail context keeps AI prompts efficient.
ECS‑style tag queries enable flexible system composition.
AI can generate or modify game logic on the fly.
Rapid prototyping reduces repetitive coding overhead.
Designers can experiment with unconventional gameplay mashups.
Future tools may blend world models with traditional engines.

Build Games with AI: Tag Assets, Skip Boilerplate, and Let the Model Assemble Context