Posts grouped by Category

This post kicks off a series of experiments by focusing on the first attempt to use Large Language Models (LLMs) to generate NPC behavior policies in a gridworld environment centered on survival tasks. It explores whether LLMs can efficiently create adaptive and complex strategies that might outperform traditional reinforcement learning in certain cases. Through this initial test, the post evaluates the potential of LLMs to drive intelligent, responsive NPC behavior that could transform gameplay dynamics in future applications.

This post kicks off a series of experiments by focusing on the first attempt to use Large Language Models (LLMs) to generate NPC behavior policies in a gridworld environment centered on survival tasks. It explores whether LLMs can efficiently create adaptive and complex strategies that might outperform traditional reinforcement learning in certain cases. Through this initial test, the post evaluates the potential of LLMs to drive intelligent, responsive NPC behavior that could transform gameplay dynamics in future applications.

This post kicks off a series of experiments by focusing on the first attempt to use Large Language Models (LLMs) to generate NPC behavior policies in a gridworld environment centered on survival tasks. It explores whether LLMs can efficiently create adaptive and complex strategies that might outperform traditional reinforcement learning in certain cases. Through this initial test, the post evaluates the potential of LLMs to drive intelligent, responsive NPC behavior that could transform gameplay dynamics in future applications.

This article delves into the Berlin Interpretation, a framework attempting to define the essential characteristics of Roguelike games. It discusses the criteria set forth by this interpretation and reflects on its relevance and impact on the genre.

This development log provides an overview of a tile map management system, detailing its architecture and functionality. It serves as a continuation of previous discussions, offering insights into the system’s design and implementation.

This post explores the process of landscape generation using Rust, focusing on a seeded expansion algorithm to ensure consistent chunk generation. It provides a detailed look into the methods employed for procedural terrain creation.

This entry offers background information pertinent to upcoming development discussions, setting the stage for more in-depth technical explorations in future posts.

This article introduces the “Flexible Architecture for World-building and Narrative” engine, a project aimed at integrating artificial intelligence and machine learning into game design to create procedurally generated experiences that feel handcrafted. It discusses the motivation behind the project and its intended goals.

This inaugural guest post reflects on the evolution of gaming culture and the increasing integration of narrative elements into Roguelike games. It examines the balance between procedural generation and storytelling in game design.

This post kicks off a series of experiments by focusing on the first attempt to use Large Language Models (LLMs) to generate NPC behavior policies in a gridworld environment centered on survival tasks. It explores whether LLMs can efficiently create adaptive and complex strategies that might outperform traditional reinforcement learning in certain cases. Through this initial test, the post evaluates the potential of LLMs to drive intelligent, responsive NPC behavior that could transform gameplay dynamics in future applications.