DeepSeek just open-sourced its inference engine

DeepSeek released DeepInfer, an Apache-2.0 inference framework powering its production services, supporting DeepSeek-V4.1-Flash on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs with optimized kernels, parallelism, and deterministic/structured outputs. A complete service still requires a KV cache, disaggregation, and replica load balancing, which are not included.

Things to know if you get cancer in 2026

After a May 2026 diagnosis of squamous cell carcinoma in the mouth, the author shares lessons for cancer patients: persistently pursue earlier appointments, push for comprehensive genetic testing and second opinions, consult NCCN guidelines, and document doctor visits.

Mistral Large 4 (Le Chonk) is now Open Weight

Mistral AI released the open weights of Mistral Large 4, a 1-trillion-parameter natively multimodal model improved by three weeks of additional reinforcement learning, claiming leading open-weight coding results—including 19.7% on FrontierCode 2.0—with substantially fewer tokens per task.