Red Hat has introduced asago, an open-source project designed to automate the conversion of AI governance policies into deployment-ready code. This initiative, supported by NVIDIA, IBM, and other major contributors, aims to simplify compliance with regulations such as the EU AI Act by automating risk mapping and mitigation processes. While asago is still in its formation phase and lacks production-tested results, it promises to reduce deployment timelines from months to days. The project is open for collaboration on GitHub, inviting developers and researchers to contribute to its development.
Read originalStanford's Evo 2 AI model has made a significant leap in synthetic biology by generating phages that effectively target E. coli. This breakthrough demonstrates the potential of AI to design entire viral genomes, moving beyond simple DNA edits. The model produced thousands of candidate genomes, with 16 showing strong E. coli-killing activity in lab tests. By releasing Evo 2 as open-source software, Stanford is inviting further exploration and innovation in genome design, potentially paving the way for new treatments against resistant bacteria like MRSA.
Alibaba is testing a new business model for its upcoming Qwen open-weight AI model, introducing revenue-sharing terms for commercial users. This move targets larger companies that profit from offering the model as a service, requiring them to enter a commercial agreement with Alibaba. The approach mirrors the licensing model used by Moonshot for its Kimi K3 model, which includes revenue-sharing for companies exceeding certain revenue thresholds. This shift signifies Alibaba's strategy to monetize its open-weight models while maintaining their open-source nature, potentially setting a precedent for other AI developers.
MIT researchers have identified a crucial challenge in AI health tools: their effectiveness varies significantly with the user's expertise. Published in Nature Medicine, the study found that non-experts improved their diagnostic accuracy with AI assistance, mainly by relying on the model's predictions. Conversely, primary care providers achieved better results when they received AI predictions without accompanying explanations, indicating that detailed explanations might not always benefit trained professionals. This research highlights the importance of designing AI interfaces that consider the user's level of expertise to mitigate automation bias and enhance diagnostic accuracy.
The latest b10310 release of llama.cpp introduces significant improvements for aarch64 architecture, particularly with the addition of HWCAP fallbacks and refined fp16 variant detection. This update ensures better compatibility and performance for devices using aarch64, such as those running on macOS Apple Silicon and various Linux distributions. By requiring HWCAP_ASIMDHP for aarch64 fp16 CPU variants, the release enhances the handling of half-precision arithmetic. While no new models are introduced, these technical adjustments make llama.cpp more robust for developers working across diverse hardware configurations.
The latest b10312 release of llama.cpp continues its trend of broadening platform compatibility, now including support for a variety of systems such as Ubuntu with ROCm 7.2 and Windows with CUDA 13.3. This update ensures that developers working across different hardware configurations, from Apple Silicon to Windows x64, have access to optimized builds. While there are no groundbreaking new features, the release solidifies llama.cpp's position as a versatile tool for AI inference across multiple systems. This means developers can now more easily integrate llama.cpp into their workflows, regardless of their preferred platform.
The b10318 release of llama.cpp marks another step in its mission to be the go-to tool for developers across diverse hardware environments. With the addition of ROCm 7.2 support on Ubuntu x64, AMD GPU users now have a more robust alternative to NVIDIA's CUDA. This update also includes builds for Vulkan, OpenVINO, and SYCL, ensuring compatibility with a wide array of systems, from macOS and Windows to Linux. While no new model architectures are introduced, the release highlights llama.cpp's ongoing commitment to enhancing accessibility and flexibility for developers. By broadening its platform reach, llama.cpp continues to solidify its position as a versatile and adaptable AI development tool.