AI 导读
据彭博社报道，字节跳动正准备最快于下月推出一款基于Seedance的空间视频模型。创始人张一鸣亲自督战，但计划可能生变。 该模型将创建可交互的3D世界，能响应Pico用户的语音和动作。据报道，输出可达20fps，延迟0.05秒。 云端生成有望降低头显要求，将字节与Google、Meta和Apple的竞争转向AI模型、算力和内容。
正文
ByteDance is preparing a Seedance-based spatial-video model for launch as soon as next month, Bloomberg reports. Founder Zhang Yiming is personally overseeing it, though plans may change.
The model would create interactive 3D worlds that respond to Pico users’ voices and movements. Output reportedly reaches 20fps with 0.05-second latency.
Cloud generation could lower headset requirements, shifting the battle with Google, Meta and Apple toward AI models, compute and content.
来源：@thexpin · x.com