Brief

Preferred Networks is exploring technical roles within the CG and visual effects industry, as detailed in a CGWORLD feature. Separately, Moonlake has released a beta version of its "3D Agent," an AI designed to operate Blender like a human user. This agent can perform tasks such as modeling, scene construction, and rigging using simulated mouse and keyboard inputs. AI

Researchers have developed Code-as-Room, a new framework that uses a multi-stage pipeline and a structured execution harness to generate 3D indoor rooms from top-down view images. This approach synthesizes executable Blender code for geometry, materials, and lighting, addressing limitations in precision and stability found in previous text-based or image-conditioned methods. The framework incorporates a cross-stage memory module to prevent context forgetting and includes a dedicated benchmark for evaluating code-based 3D room synthesis. AI

Researchers have developed SceneCode, a framework that generates editable indoor scenes by compiling natural language prompts into executable Python programs for Blender. This approach moves beyond static meshes, creating physically interactable assets with simulation-ready articulation metadata. The system uses an agentic backbone and a planner-designer-critic loop to generate and refine object-specific code, enabling traceable and locally editable scene construction. AI

A game developer is experimenting with Unreal Engine 5 to enhance gameplay. One modification involves adjusting the camera to provide a clearer view of the game environment, making the action more understandable and improving spatial awareness. Additionally, the developer is exploring interactive particle effects that respond dynamically to player actions. AI