highvramhardwarewatch.com

Wednesday 18 November

BIS targets remote access to advanced US chips through third-country data centers

AI-generated summary BIS says a new interim final rule requires licenses in certain cases where overseas data centers provide large amounts of computing power built on controlled United States technology to companies headquartered in Macau or countries in United States Country Group D:5. The Semiconductor Industry Association welcomed the removal of the AI Diffusion Rule while stating it is still evaluating the impact of the new regulations.

Also covered byFusion42
Earlier:

NanoI1 puts private extraction in 311 MB—but proof of usefulness is still missing

AI-generated summary KiteFishAI has launched NanoI1, a 155.3-million-parameter language model designed for structured extraction, classification, routing and short question answering over supplied documents. KiteFishAI says its multi-head latent attention design reduces the key-value cache sixfold compared with standard attention at the same width.

Monday 16 November

Cantwell says Commerce left AI-chip export loopholes open

AI-generated summary Senator Maria Cantwell accuses the Commerce Department of failing to stop advanced US AI chips from reaching Chinese companies through overseas data centers and affiliates. Cantwell argues that Alibaba and ByteDance used foreign data centers equipped with advanced Nvidia chips to train AI models while Commerce left export loopholes open.

Also covered byInnopark IT
Earlier:

Qwen4-Max is a cloud launch, not a home-AI model—yet

AI-generated summary Alibaba has released the Qwen4-Max native-vision mixture-of-experts model featuring 2.4 trillion total parameters and a one-million-token context window through its cloud services. Alibaba says Qwen4-Max activates 71 billion parameters per token and achieved higher scores on coding, agentic-workflow, and visual-reasoning evaluations.

Also covered byQwen (@Alibaba_Qwen)RSS.bzBest-AI.org
Earlier:

Cloudflare Kitesurf Is Now Open Source: Built for AI Agents

AI-generated summary Cloudflare open-sourced its Kitesurf browser engine on November 13 for AI agents to run inside Workers V8 isolates without containers. The Apache 2.0 release includes Blitz for HTML parsing, Stylo for CSS processing, and Boa for a WebAssembly-based JavaScript fallback.

Also covered byHelp Net SecurityPYMNTSHighVRAM Hardware Watch
Earlier:

Sunday 15 November

llama.cpp 1.0 gives local-AI integrations a compatibility signal, not a painless upgrade

AI-generated summary The llama.cpp project reached version 1.0 on November 13 with a new rule that a breaking change to its public C API should require a new major version. Homebrew listed llama.cpp 1.0.0 as its stable formula on November 15, while Debian unstable still listed version 0.9.0.

Also covered bybyteiota | From Bits to Byteselseif
Earlier:

Saturday 14 November