Dev Radar
Support
LiveUpdated 2026-10-30 20:27 UTC

Two open problems in world models: steering and physics. Tencent open-sourced a repo for…

Two open problems in world models: steering and physics. Tencent open-sourced a repo for each in the past week. 📌…

This is a dev post classified by AGK Intelligence as Other (a list of tools), kept by the Dev Radar because it carries real work, not commentary.

Two open problems in world models: steering and physics. Tencent open-sourced a repo for each in the past week. 📌 Hy-Nomas Multi-turn text and camera input. The world is generated segment by segment, with a KV cache that keeps its history. Fits on one RTX 4090 with the INT8 low-mem config. Diffusers support landed. The community already wired it into ComfyUI, SGLang, and WanGP. Repo: http://github.com/Tencent-Hunyuan/Hy-Nomas 📌 Looking-Ahead Neighboring video chunks look alike, so autoregressive models lean on local continuity instead of learning how the scene evolves. The fix: during train

Posted by Tencent AI (25.6k followers) 2 days ago · 41 likes · 3.1k views · view the original post on X. Kept by the Dev Radar as Other. Tools mentioned: hy-nomas, looking-ahead.

More dev work like this

Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 20:27 UTC. Full method.