The WAV file is a valid audio file. It passes MIME-type checks. But the audio frame data contains a base64-encoded payload. Decode the frames, take the first 8 bytes as the XOR key, XOR the rest, and you have your executable or Python script.
Cohere's Transcribe model is designed for tasks like note-taking and speech analysis, supporting 14 languages and optimized for consumer-grade GPUs, making it accessible for self-hosting.
The Google Open Source Software Vulnerability Reward Program team is increasingly concerned about the low quality of some AI-generated bug submissions, with many including hallucinations about how a vulnerability can be triggered or reporting bugs with little security impact.
3,847 brand icons with multi-variant support (color, mono, light, dark, wordmark). Tree-shakeable npm package - import one icon, ship only that icon. TypeScript types for every icon module. Instant search with fuzzy matching and keyboard shortcut (Cmd+K / Ctrl+K). Filter by category - AI, Software, Framework, Language, Design, and more.
The distributed nature of the network is the greatest strength of Mastodon, but it also means that having a share button that takes you to the 'correct' Mastodon server for your account is a lot more involved than a simple hyperlink.
AI coding tools have caused as many problems as they have solved, according to industry experts. The easy-to-use and accessible nature of AI coding tools has enabled a flood of bad code that threatens to overwhelm projects. Building new features is easier than ever, but maintaining them is just as hard and threatens to further fragment software ecosystems. The result is a more complicated story than simple software abundance.
Qwen3.5 is available via Hugging Face and is released under an open-source license. With this, Alibaba is explicitly targeting developers and research institutions that want to work with the model themselves. The system can process very long prompts, up to 260,000 tokens, and can be scaled further with additional optimizations. This makes it suitable for complex applications such as extensive document analysis and code generation.
Entire's tech has three components. One is a git-compatible database to unify the AI-produced code. Git is a distributed version control system popular with enterprises and used by open source sites like GitHub and GitLab. Another component is what it calls "a universal semantic reasoning layer" intended to allow multiple AI agents to work together. The final piece is an AI-native user interface designed with agent-to-human collaboration in mind.
Alibaba has launched RynnBrain, an open source AI model that helps robots and smart devices perform complex tasks in the real world. The model combines spatial understanding with time awareness. Alibaba's DAMO Academy introduced the foundation model that enables interaction with the environment. RynnBrain can map objects, predict trajectories, and navigate in complex environments such as kitchens or factory halls. The system is trained on Alibaba's Qwen3-VL vision language model.
That mismatch worked, if uncomfortably, when contributing had friction. After all, you had to care enough to reproduce a bug, understand the codebase, and risk looking dumb. But AI agents are obliterating that friction (and have no problem with looking dumb). Even Mitchell Hashimoto, the founder of HashiCorp, is now considering closing external PRs to his open source projects, not because he's losing faith in open source, but because he's drowning in "slop PRs" generated by large language models and their AI agent henchmen.
Moca has open-sourced Agent Definition Language (ADL), a vendor-neutral specification intended to standardize how AI agents are defined, reviewed, and governed across frameworks and platforms. The project is released under the Apache 2.0 license and is positioned as a missing "definition layer" for AI agents, comparable to the role OpenAPI plays for APIs. ADL provides a declarative format for defining AI agents, including their identity, role, language model setup, tools, permissions, RAG data access, dependencies, and governance metadata like ownership and version history.
If you've worked in a technical role in news for long enough, you likely remember when the "show your work" spirit was everywhere. Newsroom nerds shared code on GitHub, swapped tips on social media and unfurled long blogs guiding others on how to get things done. You might also have a vague sense that - like reaction GIFs, demotivational posters, and that guy who sang "Chocolate Rain" - you're seeing less of it these days.
On Wednesday, the Paris-based AI lab released two new speech-to-text models: Voxtral Mini Transcribe V2 and Voxtral Realtime. The former is built to transcribe audio files in large batches and the latter for nearly real-time transcription, within 200 milliseconds; both can translate between 13 languages. Voxtral Realtime is freely available under an open source license.
Trying to write on a laptop means fighting a machine that is also a notification box, streaming portal, and social feed. Distraction-free apps help, but they still live inside the same browser-and-tab chaos, surrounded by everything else your computer knows how to do. Some writers just want a device that only knows how to produce plain text and does not care about anything else happening in the world.
Completely free and open source (view our licence here). data_object Supports export for integration with frameworks including React, Vue, and Angular. Fully configurable, featuring custom triggers and adjustable text to support multiple language locales. 60 languages supported by default (view the languages here). Includes multiple views, including Map, Line, Chart, Days, Months, and Color Ranges. export_notes Export data to multiple file formats (view the supported types here), with system clipboard setting support.
When I moved to VMware, I expected things to continue much as before, but COVID disrupted those plans. When Broadcom acquired VMware, the writing was on the wall and though it took a while, I eventually got made redundant. That was almost 18 months ago. In the time since, I've taken an extended break with overseas travel and thoughts of early retirement. It's been a while therefore since I've done any direct developer advocacy.
Poettering is best known for systemd. After a lengthy stint at Red Hat, he joined Microsoft in 2022. Kühl was a Microsoft employee until last year, and Brauner, who also joined Microsoft in 2022, left this month. The trio are leading lights in the Linux and open source world. Brauner posted on Mastodon: "My role in upstream maintenance for the Linux kernel will continue as it always has." Poettering will similarly remain deeply involved in the systemd ecosystem.
The ad industry is racing toward a not-too-distant future where AI agents negotiate programmatic deals on their own - and Prebid doesn't want publishers to get left behind. The group that turned header bidding software into an open-source standard announced on Thursday that it's taking ownership of code developed using Ad Context Protocol (AdCP) that will power publisher-side AI agents.
