bro
422 posts








Tencent’s HY4 is currently in training and will be larger than HY3 (295B total parameters, 21B active) HY4 will also be multimodal. Hunyuan aims to enter the top tier of Chinese models by 2027 ByteDance’s Seed team is training an unprecedentedly large model Source: LatePost (晚点)



The Chang'e-7 spacecraft arrived at Wenchang spaceport on April 9. The robotic lunar south pole landing mission will be prepared for launch for the second half of 2026, according to CMSEO. mp.weixin.qq.com/s/B8Hh_lCU43YR…

Inner Mongolia boasting it has "world-class" DC cluster including DCs from HW, Tencent, Alibaba, China Unicom, Mobile & Telecom + banks. Moonshot & DeepSeek are both training models in DCs in Horinger, Inner Mongolia. Utilizing abundant wind/solar resource as well as cold + dry climate to full advantage in providing cooling & power for these DCs. Other big companies like XPeng & Sinopec are also training their models here.

Can the impossible be done? Yes. Professional VFX with Seedance 2.0. Workflow: @midjourney 8.2 (image creation) + Seedance 2.0 in @higgsfield Suite Prompt + Tip in the comment: (This is not an advertisement)






















