Orkestr has launched custom sandbox templates for AI agents that preinstall tools like ripgrep and ffmpeg, eliminating the need for repeated installations at runtime. The templates are built once via …
Modal and Z Lab released DFlash, a speculative decoding model for Qwen 3.5 397B-A17B, achieving over 4.3x throughput versus baseline and 1.5x versus MTP on HumanEval at concurrency 1. The model uses a…
Microsoft, Hugging Face, Meta's PyTorch team, NVIDIA, and others launched OpenEnv, an open protocol for agent learning environments that standardizes how agents practice and improve. The protocol aims…
Hexo Labs released SIA (Self Improving AI), the first open-source framework that co-evolves both an agent's scaffold and its model weights in a single iterative loop. On the LawBench task, SIA achieve…
Anthropic launched self-hosted sandboxes in public beta and MCP tunnels in research preview for Claude managed agents, enabling regulated industries to execute tools and access private services within…
Anthropic CEO Dario Amodei said reinforcement learning shows the same log-linear scaling as pre-training, but RL system efficiency is critical to afford enough training. Experiments on open models sho…
A developer built Perri Comic Generator, a lightweight single-panel comic creator that combines an LLM (Meta-Llama-3-8B-Instruct) with a diffusion model (SDXL-Turbo) to generate comics from story seed…
Jian Chen, Yesheng Liang, and Zhijian Liu integrated Z Lab's DFlash block diffusion speculative decoding method into SGLang, collaborating with Modal and LMSYS. The team reports up to 4.31x higher thr…
A developer built a multi-agent AI system called 'The Price Is Right' that autonomously scans the internet for bargains, estimates product value using three pricing techniques, and sends notifications…
Omnigent launched an open-source platform that provides a unified layer over AI coding agents like Claude Code, Codex, and Pi, enabling users to swap or combine harnesses, enforce policies and sandbox…
SkyPilot released Sandboxes, a bring-your-own-cloud code execution layer that runs untrusted agent code on a user's existing Kubernetes infrastructure instead of a hosted third-party vendor. The tool …
A blog post titled "Failure numbers every programmer should know" provides reliability napkin math for hardware, cloud services, and software defects. The post is part of a collection that also covers…
Neuronpedia launched HeadVis, a tool for exploring attention heads in collaboration with Anthropic, now available for 37 models with over 36,000 attention head dashboards. The platform also added new …
A team of developers built Thousand Token Wood, a multi-agent economic simulation where five AI-powered woodland creatures trade goods using a 3-billion-parameter Qwen2.5-3B model. The simulation, cre…
Nous Research has released Hermes Agent, an open-source AI agent designed to run continuously on infrastructure users control, including a $5 VPS. The agent supports six terminal backends with serverl…
A developer's team has been running production AI agents on Bun for three months, finding the runtime's 15-25ms cold startup significantly improves interactive agent loops compared to Node's 80-120ms.…
A developer benchmarked the Qwen3.6 27B model on Modal using llama.cpp, deploying a serverless pipeline that downloads GGUF shards from Hugging Face and runs perplexity evaluation on an A100-80GB GPU.…
Hexo Labs released SIA (Self-Improving AI) as an open-source framework under an MIT license this week, enabling an AI agent to edit both its scaffold and model weights within a single self-improving l…
Cognition raised $1 billion in a Series D funding round that values the company at $26 billion, marking a 2.5x valuation increase from its September Series C. The AI agent lab now projects over $1 bil…
A team of Bayesian statisticians has developed a method to fit million-parameter models using serverless GPU computing for pennies, dramatically reducing the cost and complexity of Markov chain Monte …