Tag
A user on Hacker News asks for help identifying a mysterious legacy networking device with a CC- serial prefix found in a data center, which exhibits consistent 0.4 ms latency from multiple connections and appears in records from before the facility was built.
A developer successfully runs DeepSeek-V4-Flash (284B total, 13B active) locally on four RTX 2080 Ti GPUs with a $2,500 budget, achieving 255 prefill tokens/s using custom Turing CUDA kernels, W8A8 quantization, and heterogeneous inference. The implementation is open-sourced.