Tony Simons(@tonysimons_):DeepSeek just open-sourced DeepGEMM for Huawei Ascend. Same API. BF16, FP8 + FP4 kernels. The wild number: 99.8% of the Ascend 950’s theoretical compute limit. This isn’t another model release. They’re going after the AI infrastructure stack now. GitHub:

Tony Simons

@tonysimons_

I help people get @NousResearch Hermes Agent working. Setup, configs, troubleshooting, workflows & automation. | Founder: @AIowaLLC | Book a 1:1 with me 👇

Joined December 2025

909 Following 12K Followers

Tony Simons@tonysimons_

2026.10.01 04:51

DeepSeek just open-sourced DeepGEMM for Huawei Ascend. Same API. BF16, FP8 + FP4 kernels. The wild number: 99.8% of the Ascend 950’s theoretical compute limit. This isn’t another model release. They’re going after the AI infrastructure stack now. GitHub: