Walk into a GPT-6 Astra-generated harbor in Three.js to hunt for 3D reasoning mistakes
Lianhuiq · x · 2026-09-16
Researcher Lianhui Qing shares a fun way to stress-test model 3D reasoning: let humans walk inside the model's world. Their Code4Scene pipeline used GPT-6 Astra to generate a harbor scene, exported as a fully explorable Three.js demo, inviting people to hunt for floating objects, odd geometry, and impossible spatial relationships.
Related event: Humans Walk Into GPT-6 Astra's Generated 3D Harbor to Spot Spatial Errors(2 posts)→
More from Multimodal
- Videoclaw launches as a Mac video agent: prompt ChatGPT or Claude to edit and generate video — alifcoder · 2026-09-17
- StepFun and ACE Studio launch music foundation model StepAudio 3 Music — StepFun_ai · 2026-09-17
- StepFun and ACE Studio launch StepAudio 3 Music, a song-generation foundation model — StepFun_ai · 2026-09-17
- Fan recreates Blue Archive Shiroko scene with AI animation, tutorial coming — Wakaiko · 2026-09-17
- UK dev quit over coding agents, built Videoclaw: chart image to narrated video for $0.39 — JaynitMakwana · 2026-09-17
- Wan2GP trick: inject character reference sheets as the first frame for consistent generations — protocol-apps · 2026-09-17