Qwen-Image-2.1 GGUF Files Enable Local Runs
Quantized Qwen-Image-2.1 files now run locally via GGUF image tooling. — huggingface.co
Zig Music Player Sheds AI-Generated Code
A developer's Zig music player shed nearly half its codebase in a manual rewrite after an AI-assistant ban and a code host's no-AI policy, and emerged more robust with the same features. — alichraghi.github.io
FLUX Quantized for Low-VRAM GPUs
The quantized FLUX image model runs on GPUs with as little as 4GB of VRAM. — huggingface.co
bhyve Gains Live Migration on FreeBSD 15.1
Live migration now moves a running bhyve guest between hosts, pausing it only for the final copy phase. — gyptazy.com
Antithesis Fuzzer Beats NetHack Reliably
The Antithesis fuzzer now beats NetHack on nearly 80% of game seeds. — antithesis.com
OpenMausBot Brings Local-First Agent Chats
OpenMausBot is an open-source, local-first chat app that runs a roster of AI agents on your own machine, each with its own model and computer. — github.com
Android Clipboard Toast Bypass Exposes Copied Credentials
Android's clipboard-read toast can be bypassed by a malicious app that quickly shows its own toasts, hiding the system warning when copied credentials are read. — toninoble.com
WebKit Promise Fix Ends Endless Awaits
A WebKit fix makes awaited promises settle correctly when a property lookup throws during the await operation, instead of leaving them pending forever. — alistair.sh
DeepSeek-V4-Pro Gets Uncensored GGUF Release
An uncensored DeepSeek-V4-Pro variant is now available as GGUF quantized files, ready for local inference with common runners. — huggingface.co
llama.cpp v1.0.0 Stabilizes the API
llama.cpp and its tensor library reach v1.0.0 with a stable public API and clear versioning for downstream developers. — huggingface.co