Tag
A user showcases the capabilities of Opus 5.5 medium by integrating blender, image-gen, and three.js into a browser-based application, highlighting Anthropic's achievements and predicting its impact on future game development.
MengTo used the AI tool Astra to create a Catan board game in three.js, featuring animations, instructions, AI players, and trading, and is now planning to develop more games based on this foundation.
A developer showcases a game built using Fable 5.1, Three.js, and Cursor, featuring PS5 DualSense controller with haptic feedback, and notes high prompt rejection rates in Cursor.
A tweet indicates that more AI video generation demos will be shared this week, highlighting the complexity and time required for high-quality outputs using prompts and subagents.
The author argues that using multiple agents without critics loops, as demonstrated in a Three.JS project built with 18 agents and Claude 5.0 Opus, may be more efficient than critics loops, while noting token budget concerns.
An experiment comparing the qwen3.8:27b model with Claude Opus 5 on a complex code porting task from C to HTML/three.js, highlighting that local models struggle with long contexts and require effective prompting to avoid poor results.
Matt Shumer highlights a cheap and efficient controlled video generation workflow from GMI Cloud that uses code to create a low-fidelity skeleton scene, then passes it through a video model to make it realistic, costing just $1.97.
A developer showcased how Fable, an AI agent, generated a fully procedural Mars city in Three.js after six hours and dozens of subagents, calling the result beyond expectations.
A cyberpunk FPS tech demo that runs entirely in the browser, built in 3 days via vibe coding with multiple LLMs. It features procedural generation, WebGPU rendering, and fully synthesized audio.
Discusses how to quickly launch on the Douyin AI track using vibe coding, introduces the open-source project Rainform, which visualizes rainfall data as an interactive 3D particle rain scene.
Matt Shumer shares the GitHub repo for 'Claude of Duty', a first-person shooter built entirely in the browser using Three.js and WebGL2, with all assets procedurally generated by AI agents at load time.
img2threejs v1.3 is a tool that converts a single photo into procedural Three.js code without meshes or manual modeling, with major improvements to geometry reconstruction and material generation.
A 3D indoor map of Shinjuku Station built with Three.js using open government data from Japan's Ministry of Land.
Introduces a new open-source toolkit, img2threejs, that converts photos into Three.js models and automatically compares the rendering result with the original image to calibrate reconstruction quality, especially suitable for hard-surface objects.
User @mate_mattt gave their article to GPT 5.6 Sol to create a playable demo learning animation using Three.js, with excellent results.
Vanguard-Map is an open-source real-time 3D map displaying live ship traffic, aircraft, satellites, and more using Three.js, with optional local proxy for flights.
Proposes a blind visual paradigm using Three.js to test if procedural scaffolds extracted from large models can improve small model outputs without fine-tuning, validated by a blind judge model.
User tests Ornith-1.0-35B on an RTX 3090, finding fast inference speeds (1560 tok/s prompt, 78 tok/s generation) but consistently worse coding performance on Three.js tasks compared to Qwen 3.6, even after multiple attempts.
A tweet thread comparing recent AI models' ability to generate endless procedural terrain using Three.js, all in a single shot, with a mention of Fugu Ultra as a candidate.
A comparison of recent AI models, including Sakana AI's Fugu Ultra, demonstrates their ability to generate endless procedural terrain in Three.js with one-shot generation, highlighting frontier capabilities without export control risks.