Tag
CommonGrid is an open-source, community-maintained registry of U.S. power grid data, offering a connected graph of utilities, power plants, and infrastructure under an open license.
Release of SmolDataEnvs, a collection of 5,000 verifiable RL environment tasks for training small models in code and data science, fully open source.
ThinkingCap is a finetuned model based on Qwen 3.6 27B that reduces reasoning tokens by about 50% while maintaining performance, leading to significant efficiency gains in inference.
Analysis of 246 open-source repositories and 57 papers on agent harnesses highlights best practices like human-written context, limiting tools, and incremental testing to improve AI agent performance.
Jeff Geerling announces HomelabFest, a new in-person event for homelab enthusiasts in St. Louis, September 2027, focusing on self-hosted software, hardware, and community building.
Valentino Stoll announces the release of ruby-laya, a Ruby port of the open-weight System 1 decision engine Laya, making AI decision capabilities accessible to Ruby developers.
NVIDIA and Stanford developed a Contrastive Language Model (CLM) that frames decision-making in AI as a retrieval problem, achieving up to 9x lower latency compared to existing systems like Jev by caching action embeddings.
The article discusses a local and free version of the Muse AI model, making it accessible for personal use without cost.
A tweet sharing a GitHub repository that compiles explanations for key machine learning papers, such as Transformer, BERT, and GPT, for educational purposes.
Citation-Network v1.2.0 is an open-source app that builds interactive citation networks from Zotero libraries, with integration for Obsidian notes and a GUI for exploration.
An open-source coding agent harness was stress-tested by using at least 100 AI agents to update its own documentation, demonstrating efficient parallel task handling and identifying real issues in the docs.
BeeLlama v0.4.7 is released, offering gigabytes of VRAM savings through independent drafter ubatch sizing for MTP and DFlash, along with major ROCm and CUDA performance enhancements for AI inference.
The article discusses the emergence of LLM policies in GNOME and KDE, highlighting the philosophical divide between collectivist and completionist views in FLOSS communities and how LLMs are disrupting the collaborative balance.
Git-bug is a distributed, offline-first bug tracker embedded directly in Git repositories, allowing users to manage bugs without external services and ensuring portability, speed, and no vendor lock-in.
This blog post argues that open source software is fundamentally about the users, challenging Rich Hickey's view that maintainers are solely gift-givers, and emphasizes the ethical and social obligations in the maintainer-user relationship.
Graph engineering has become a standard method for building AI agents, with frameworks like LangGraph, CrewAI, and AutoGen highlighted for their use in production by companies such as Uber and LinkedIn.
Paranoia Sans is a self-censoring typeface that automatically redacts terms popular in conspiracy myths using ligatures, created as an open-source project.
The Dutch government is developing an open-source alternative to Microsoft products through the DAWO community, aiming to enhance digital autonomy, security, and verifiability in public sector IT systems.
Jev-Code-Reviewer is a GitHub extension and CLI tool that helps developers review pull requests from AI agents by classifying changes into priority levels and providing natural language explanations of code diffs.
Token Forecaster predicts the length of LLM responses before execution and monitors them in real-time, offering accuracy based on local history as an open-source tool.