The MIT-licensed library, used in DeepSeek V3.2, V4 and V4.1, claims a 2 to 20 times gain over PyTorch’s stock topk. CUDA support arrived Sept. 10. Huawei Ascend kernels followed Sept. 30. DeepSelect is not a model. It is a…
Wednesday’s open-source update posts near-peak sparse attention on Huawei’s newest NPU, retunes a fused decoding kernel and breaks compatibility with earlier DeepSeek models. DeepSeek on Wednesday released sparse attention kernels for Huawei’s Ascend 950 processor inside FlashMLA, the company’s open…
Federal agencies have until Oct. 3 to remediate CVE-2026-76504 after Cisco confirmed unauthenticated admin access in active attacks. The Cybersecurity and Infrastructure Security Agency on Wednesday added one vulnerability to its Known Exploited Vulnerabilities catalog after finding evidence of active…
The TileLang kernel library for mixture-of-experts routing, quantization and Engram now runs the same Python APIs on Nvidia GPUs and Huawei NPUs. Hangzhou-based DeepSeek on Wednesday published Huawei Ascend support in TileKernels, an open-source library of training and inference kernels…
Treasury’s Office of Foreign Assets Control published SHA-256, SHA-384 and SHA-512 digests for human-readable sanctions PDFs, including the Specially Designated Nationals list dated Sept. 30, 2026. The Treasury Department’s Office of Foreign Assets Control posted fresh hash values Wednesday for…
The China internet registry and industry partners opened the platform in Beijing. Alibaba Cloud is the first vendor listed as having completed integration. China’s domain-name registry and a group of standards bodies, carriers, research institutes and cloud companies on Tuesday…
The open-source expert-parallel library aligns its public APIs with Nvidia DeepEP and reports 373–375 GB/s FP8 dispatch on eight Ascend 950DT ranks. DeepSeek on Wednesday published DeepEP-Ascend, an open-source communication library for mixture-of-experts training and inference on Huawei Technologies Ascend…
The MIT-licensed port keeps the DeepGEMM programming interface and, in company tests on Ascend 950DT hardware, reports dense matrix-multiply utilization as high as 99.8%. DeepSeek on Wednesday published DeepGEMM-Ascend, an open-source matrix-multiplication kernel library written for Huawei Technologies Ascend neural…
The Applied AI Index 2026, drawn from 1,330 executives, says firms that pair strategic clarity with applied AI extract five times the return — while most still lack controls for autonomous agents. Nearly half of companies now generate value from…
Andreessen Horowitz’s State of Markets II finds compute demand still ahead of supply, older GPUs holding value, and public software facing a prove-it re-rating rather than a collapse. Andreessen Horowitz released the second edition of its State of Markets report…