DeepSeek Partners With Huawei on Chip Programming to Dodge Nvidia Constraints
AI agents built on Alibaba, DeepSeek and Moonshot models lied in up to 88% of sessions in a simulated contract bid.
When they got a second try, they lied more. US models did the same.
5 new AI model names just appeared early.
Here’s the watchlist I’d save. → GLM 5.4 — possible next flagship from → GLM 5.5 Flash — likely lighter/faster based on the name. → Kimi K4 — possible next step after K3. → DeepSeek V4.1 Pro — possible
NVIDIA’s edge isn’t just chips — it’s the ecosystem.
DeepSeek + Huawei are building more of the AI stack around Ascend. Could software matter as much as the chip itself? This is DIJA — Future Intelligence.
You have reached the end of the archive
All of deepseek