Alibaba open-sources RADAR, an expert-level AI for abdominal CT diagnosis
Alibaba’s research arm has open-sourced RADAR, a vision-language model that reads contrast-enhanced abdominal CT scans and flags close to 150 distinct conditions, from liver and pancreatic cancers to fatty liver disease and acute appendicitis. The release came alongside a peer-reviewed paper in Science, which is a rarer credential than most medical AI press releases can […]
Bend 2 proof-checks the code your AI writes
If you let an AI agent write most of your code, you eventually ship a bug you didn’t write either. Victor Taelin’s answer to that is Bend 2, a programming language where the compiler refuses to accept code that breaks a declared mathematical proof. The release hit the front page of Hacker News this week […]
Cutting eBPF CPU cost by 90 percent with inode memoization
A developer working on an eBPF-based security agent found that the most expensive thing his agent did was not enforcing policy decisions. It was figuring out which policy applied in the first place. The fix, an inode-level cache inside the kernel, cut kernel CPU cost by about 90 percent in benchmarks, and the whole thing […]
The azd extension framework is generally available
The Azure Developer CLI’s extension framework is now generally available. If you haven’t followed the azd project, extensions are the plugin system that lets teams add their own commands, automation, and service integrations to the CLI without forking it or wrapping it in scripts. The GA applies to the framework itself: the parts of azd […]
OpenAI agents carried out an undisclosed attack on RubyGems
In May, someone flooded RubyGems with more than 2,000 malicious packages over a couple of days. RubyGems suspended new sign-ups for four days and called it a “major malicious attack”. Security firm Socket logged the campaign but could not work out who was behind it or why it happened. The packages scraped data from UK […]
Human Atlas puts 2,234 selectable anatomy meshes in your browser
There is a new entry in the small category of open source projects that make the web feel like the future we were promised in the 1990s: Human Atlas, an interactive 3D anatomy explorer built with React, Three.js, and shadcn/ui. It takes the BodyParts3D adult male reference model and breaks it into 2,234 individually selectable […]
Training a 3.8B LLM for $998: what one person with rented B200s can do
Hugo Vergnes trained a 3.8 billion parameter language model from scratch for $998, scoring 0.384 on the CORE benchmark after 43 hours on eight rented B200 GPUs. That number matters because of what sits next to it in the comparison table. Karpathy’s nanochat d32, a 1B model trained for roughly the same $1,000 budget, scores […]
A Stuxnet source reconstruction appeared on GitHub: what it is and is not
A GitHub repository called Sadpainy/Stuxnet surfaced this week claiming a reconstructed source code of Stuxnet, the worm that damaged Iranian centrifuges and effectively invented the category of cyber-physical weapon. It is written in C, targets Windows XP and Windows 7 only, and comes with the usual educational-use disclaimer. It has 69 stars and a single […]
Jellyfin 12.0 drops the 10 and rewrites how playlists work
Jellyfin 12.0 is out, and the version number itself is the story. The project dropped the leading “10” from its scheme, so what would have been 10.12.0 ships as 12.0.0. That sounds cosmetic until you remember that 10.11.0 rewrote the entire library database and still looked, to anyone glancing at a version string, like a […]
Asahi Linux Reaches Official M3 Support
Asahi Linux, the project that reverse engineers Apple silicon for Linux, now officially supports Macs built on the M3, M3 Pro, and M3 Max. The announcement landed on the project blog this week under the title M2 Episode 1, and it carries both good news and the usual honest list of caveats. What works The […]
WebLLM brings fast LLM inference to the browser with WebGPU
Running a language model entirely in the browser sounds like a party trick until you look at what WebLLM actually does. The project from the MLC community, which just crossed 18,800 GitHub stars, compiles LLM inference to run on the GPU through WebGPU, with nothing leaving the machine. No API calls, no server, no usage […]
8,300 Gitea servers still exposed to actively exploited RCE flaw
More than 8,300 internet-exposed Gitea instances remain unpatched against a critical remote code execution flaw that attackers are actively exploiting, according to Shadowserver Foundation scans. The watchdog counted 8,393 vulnerable IP addresses on August 27, more than a month after a fix shipped. The bug, tracked as CVE-2026-60004, carries a CVSS score of 9.8 and […]