AI News Archives

Cloudflare today announced the open-sourcing of Cloudflare OS as an "open platform for agents, apps, and work." Cloudflare OS is an AI "operating system" but not in the traditional OS sense.

Last month the Linux Foundation launched the x402 Foundation to standardize Internet-native payments for AI agents as their latest AI play. That followed other "AI" initiatives from dealing with AI exploits to AI asset and data exchange. Today they have another AI foray being announced with the creation of the Tokenomics Foundation.

When AI/LLM agents are used in the creation of Linux kernel patches, the policy for a while now has been that it should be specified using an "Assisted-by" tag as part of the patches/commits. But Linux kernel developers this week have been discussing whether to revise that policy or to potentially eliminate it.

The Linux Foundation along with others like Amazon, Anthropic, OpenAI, NVIDIA, Microsoft, Red Hat, and others have joined forces to launch Akrites. The Akrites project is aiming to help defend critical open-source software from the brisk pace of new AI/LLM-discovered software bugs and vulnerabilities in ensuring that said issues are effectively addressed before they can be exploited by bad actors.

The Linux Foundation continues working to get more involved in new AI initiatives. Today the Linux Foundation announced the OpenSharing Project with an effort to standardize AI asset and data exchange.

The linux-firmware.git repository that serves as the de facto home of all the binary blobs used by the mainline Linux kernel open-source drivers has now introduced AGENTS.md documentation and other preparations for embracing AI coding agents.

Microsoft today announced their newest open-source creation... Under the MIT license it's the Intelligent Terminal.

Ahead of the 52nd G7 Summit being held in Evian, France next month, the recently conducted G7 Digital and Technology Ministers’ Meeting came to agreement on shared language around open-source AI and on the importance of open-source in AI.

For those curious about the growing use of AI and coding agents within the Linux kernel, this week there was another large batch of new patches fixed that were generated or co-authored by agents like Claude Code and GitHub Copilot.

The Lemonade SDK for "refreshingly fast local AI" that is largely developed by AMD engineers as an open-source project continues advancing quite rapidly for serving optimized LLMs on GPUs and NPUs.

As first reported on Phoronix in early April, Linux's second-in-command Greg Kroah-Hartman has been leveraging new AI fuzzing tools for uncovering Linux kernel bugs. Prominent due to his position within the Linux kernel community and also being the primary Linux stable maintainer. His AI-assistance for fixing Linux kernel bugs is based on a Framework Desktop powered by AMD Ryzen AI Max. The "gkh_clanker_t1000" continues assisting in Linux kernel development along with the less frequent "gkh_clanker_2000" references

Earlier this month on Phoronix we were the first to draw attention to a new fuzzing tool / AI bot uncovering kernel bugs by Greg Kroah-Hartman, the "second in command" for Linux kernel development and stable maintainer. Greg has now shared more light on the "gregkh_clanker_t1000" for this tool that has been uncovering more Linux kernel bugs the past few weeks.

Old network maintenance drivers are becoming a maintenance burden in the era of fuzzing and predominantly AI-driven bug detection causing an uptick in possible bug/security reports to upstream Linux kernel developers but with these drivers potentially having no actual users.

The Spack package manager is quite popular in the HPC / supercomputer space for scientific software. Even with the more selective niche than a typical general purpose OS package manager, large language models (LLMs) have already proven capable of being useful in generating new Spack packages. But there have also been some headaches involved too for Spack developers.

Since last year the Linux kernel already supported the Microsoft Copilot key appearing on recent laptops to trigger AI agent interactions. That keyboard key is becoming more common but now three additional new keys have been standardized for additional AI integration on future PCs. Merged today for Linux 7.0 is supporting those new standardized keycodes for AI use.

Announced today from the PyTorch Conference EU in Paris is word that Hugging Face has contributed their Safetensors project to the PyTorch Foundation, which is an umbrella organization under the Linux Foundation for hosting AI initiatives. Safetensors aims to help mitigate arbitrary code execution risks and more.

Open-source friendly company Tiny Corp that is behind the Tinygrad MIT-licensed neural network framework and developing a "sovereign" AMD GPU driver stack with their Tinybox hardware offerings has their sights on shipping the Exabox next year. The Tiny Corp's Exabox is expected to retail for around $10M USD but offer immense AI compute power.

For helping with the increase of AI tools scouring the Linux kernel source tree and sending security bug reports, a pull request sent today ahead of the Linux 7.0-rc7 improves the documentation to better guide AI agents -- and anyone reading the documentation -- how to send better quality bug reports.

KTransformers 0.5.3 released today for this framework for efficient inferencing and fine-tuning of large language models (LLMs) with a focus on CPU-GPU heterogeneous computing. With this release, KTransformers 0.5.3 is now more applicable for CPUs lacking Advanced Matrix Extensions (AMX) and AVX-512 in now providing some AVX2-only kernels too.

Last month NXP posted open-source Linux kernel driver patches for their Neutron NPU accelerator. The NXP Neutron NPU aims to help with edge AI applications and this neural processing unit is found in their different SoCs. Unfortunately, their GitHub repository for the user-space software ends up containing a binary-only blob that will end up delaying plans on getting this driver into the mainline Linux kernel.

Longtime Linux desktop users will likely remember the glorious days of the XMMS music player inspired by Winamp. It's been about two decades since the last official release but thanks to AI there is now a modern port of the codebase to GTK4 and GStreamer/PipeWire.

A few days ago Google engineers went public with Sashiko with their agentic AI code review for the Linux kernel. The Google Gemini Pro powered AI code review service is automatically monitoring the Linux kernel mailing list for new patch submissions and has proven useful already. Interest continues to build by upstream Linux kernel stakeholders around Sashiko and the latest addition is now covering the Rust-For-Linux mailing list submissions.

The last release of Llamafile was back in May and it's led me recently to wonder if Mozilla was slowly abandoning this AI project like they had done in the past to DeepSpeech and other software projects. Fortunately, that's not the case and out today is Llamafile 0.10 with some big updates.

Google engineers have been spending the past number of months developing Sashiko as an agentic AI code review system for the Linux kernel. It's now open-source and publicly available and will continue to do upstream Linux kernel code review thanks to funding from Google.

AMD, Broadcom, Meta, Microsoft, NVIDIA and OpenAI jointly announced today the formation of the Optical Compute Interconnect (OCI) Multi-Source Agreement (MSA) optical scale-up consortium.

The newest open-source concern around AI that is seeing a lot of interest this weekend is when large language models / AI code generators may rewrite large parts of a codebase and then the "developers" claiming an alternative license incompatible with the original source license. This became a real concern this week with a popular Python project experiencing an AI-driven code rewrite and now published under an alternative license that its original author does not agree with and incompatible with the original code.

The Linux kernel continues seeing more open-source kernel drivers emerge for supporting different AI accelerators / NPUs. The newest open-source driver breaking cover today is from NXP and is for enabling their Neutron neural processing unit.

Josef Bacik, of Btrfs notoriety before leaving Meta and stepping back from kernel development last year, announced the release of Systing 1.0. Systing is a newer eBPF-tracing tool for Linux complete with AI integration.

The open-source ollama project that makes it easy to get up and running with a variety of LLMs under Windows, macOS, and Linux is out with a new release. The ollama v0.17.0 release is driven by new functionality around enhancing the OpenClaw onboarding process.

The b4 tool used by Linux kernel developers to help manage their patch workflow around contributions to the Linux kernel has been seeing work on a text user interface to help with AI agent assisted code reviews. This weekend it successfully was dog feeding with b4 review TUI reviewing patches on the b4 tool itself.

PyTorch 2.10 is out today as the latest feature update to this widely-used deep learning library. The new PyTorch release continues improving support for Intel GPUs as well as for the AMD ROCm compute stack along with still driving more enhancements for NVIDIA CUDA.

A significant update to Burn was released today, the MIT and Apache 2.0 licensed tensor library and deep learning framework written in the Rust programming language. Burn 0.20 brings some low-level changes as it continues to strive to deliver high performance AI across the diverse hardware ecosystem.

Whisper.cpp as the open-source high performance inference project built around OpenAI's Whisper and from the same developers as Llama.cpp / GGML is out with a big new release. Whisper.cpp 1.8.3 is capable of delivering a 12x performance boost for systems with integrated AMD and Intel graphics.

The open-source ZLUDA project for bringing CUDA to non-NVIDIA hardware that can run unmodified is out with a new progress report. ZLUDA had a productive fourth quarter with now enjoying better Microsoft Windows support, full support for running Llama.cpp atop ZLUDA, AMD ROCm 7 support, and other enhancements.

Tinygrad 0.12 is out today for this deep learning stack led by George Hotz.

In addition to Linus Torvalds' recent comments around AI tooling documentation, it turns out in fact that Linus Torvalds has been using vibe coding himself. Over the holidays Linus Torvalds has been working on a new open-source project called AudioNoise that was started with the help of AI vibe coding.

The ollama 0.14-rc2 release is available today and it introduces new functionality with ollama run --experimental for in this experimental mode to run an agent loop so that LLMs can use tools like bash and web searching on your system. It's opt-in for letting ollama/LLMs make use of bash on your local system and there are at least some safeguards in place.

The current incarnation of ZLUDA continues moving along for running unmodified CUDA apps on non-NVIDIA GPUs. Released on Friday was ZLUDA 6-preview.48 with now boasting CUDA 13.1 compatibility.

The Linux kernel developers for months now have been debating proposed guidelines for tool-generated submissions to the Linux kernel. As part of the "tools", the main motivator for this documentation has been around the era of AI and large language models with coding assistants and more. Torvalds made some remarks on the Linux kernel mailing list around his belief in focusing the documentation on "tools" rather than explicitly focusing on AI, given the likelihood of AI-assisted contributions continuing regardless of documentation.

Last year AMD announced GAIA as short for "Generative AI Is Awesome". It started off as a Windows-only AI demo but over time added Linux support along with introducing different AI agents. For going along with AMD's AI announcements at CES 2026, AMD released GAIA 0.15 where they are now positioning this software as a framework/SDK for building AI PC agents.

Intel engineers as part of the OPEA Project today released the Generative AI Examples v1.5 update. This "GenAIExamples" open-source project is a collection of GenAI examples as part of showing the capabilities of the Open Platform for Enterprise AI (OPEA) and also highlighting Intel's hardware strengths for generative AI.

A few weeks ago it was mentioned by a Canonical engineer how trying to use AI to modernize the Ubuntu Error Tracker yielded some code that was "plain wrong" and other issues raised by that Microsoft GitHub Copilot code. The same Ubuntu developer shifted to trying Gemini AI to generate a helper script to assist in Ubuntu's monthly ISO snapshot releases. Google's Gemini AI also generated some sloppy code for a Python script to assist in those Ubuntu releases.

There's another setback to the open-source driver code around Intel's Gaudi accelerator support on Linux.

Qualcomm upstreamed new Cloud AI 100 "AIC100" firmware today to linux-firmware.git to fix a rather significant power/performance issue for these AI accelerators.

MLPerf Client as MLCommons' machine language inferencing benchmark for client form factors / PCs now has a Linux build. MLPerf Client 1.5 was released yesterday with an experimental Linux build but for now at least is not nearly as full-featured as this AI benchmark on Windows and macOS.

Posted to the mailing list on Friday were the latest proposed guidelines for tool-generated contributions to the Linux kernel. The coding tools in large part being focused on AI generated content.

ollama 0.12.11 released this week as the newest feature update to this easy-to-run method of deploying OpenAI GPT-OSS, DeepSeek-R1, Gemma 3, and other large language models. Exciting with ollama 0.12.11 is that it's now supporting the Vulkan API.

Open-source developer Tomeu Vizoso who developed the open-source and reverse-engineered "Rocket" accel driver for the Rockchip NPU and also worked on the Etnaviv/Vivante NPU support and other related code like Mesa's Teflon is teasing new NPU drivers coming in the new year.

Crashes on NPUs and AI accelerators are unfortunately a thing and yet another obstacle to worry about it with modern computing. Qualcomm developers have sent out patches for Sub-System Restart "SSR" functionality for their Qualcomm AI Accelerator (QAIC) driver for Linux to handle restarts when workload crashes occur on their AI accelerator hardware.

Announced today at the PyTorch Conference was word that the Ray AI compute engine is becoming a project hosted by the PyTorch Foundation.