The Big Story
A major breakthrough in artificial intelligence research has been achieved with the open-sourcing of Agensh, a multi-agent harness developed by Microsoft Research. This innovative technology allows for the creation of large-scale systems that can scale from 1 to 1,024 agents, making it an invaluable tool for various industries and applications.
Agensh's decentralized architecture eliminates the need for a central orchestrator, enabling self-organizing workers to collaborate and share resources through a Git repository, chat, and findings board. This design allows Agensh to efficiently handle complex tasks, making it an attractive solution for organizations seeking to improve their operational efficiency.
Microsoft's decision to open-source Agensh will likely lead to widespread adoption and further innovation in the field of AI. As the technology continues to evolve, we can expect to see increased applications across various sectors, from healthcare and finance to education and customer service.
What Shipped
Cognition Open-Sources an Agent Plugin that Helps Secure Systems against Recent Cybersecurity Threats
According to a recent report from MarkTechPost, Cognition released the Devin Vulnerability Fixes plugin, which includes three MIT-licensed agent skills that inventory, patch, and test PHP, Node.js, and self-managed GitLab deployments hit by this week's critical RCE flaws. This new plugin aims to help secure systems against recent cybersecurity threats.
Microsoft Open-Sources Agensh: A Multi-Agent Harness for 1,024 Agents
Microsoft Research has open-sourced Agensh, a multi-agent harness with no central orchestrator. Self-organizing workers share a Git repo, chat, and findings board, scaling from 1 to 1,024 agents on Programmar here.
From the Labs
A new coding guide has been released for Underdog Saluki 27B, a tool that allows developers to explore quantization and memory usage in stock llama.cpp.
Cognition Open-Sources an Agent Plugin that Helps Secure Systems against Recent Cybersecurity Threats
According to MarkTechPost, Cognition released the Devin Vulnerability Fixes plugin, which includes three MIT-licensed agent skills that inventory, patch, and test PHP, Node.js, and self-managed GitLab deployments hit by this week's critical RCE flaws.
Block’s new voice model sounds eerily human
According to TechCrunch, Block said that the voice model can run locally on a MacBook M4 Max.
Other Notable News
A new coding guide has been released for Underdog Saluki 27B, a tool that allows developers to explore quantization and memory usage in stock llama.cpp.
A study found Grok's AI overwhelmingly recommends Republican candidates while OpenAI's GPT favors Democrats, with the two AI models disagreeing in more than 90% of nearly 500 election races.
We created a list of the most notable AI agents that can live in your text messages, from general assistants to agents designed for families, travel, and work.
A hands-on guide has been released on how to add guardrails to a local LLM agent using Gemma 4, Ollama, and the OpenAI Agents SDK.
The Take
The Take: AI's Multi-Agent Harness and Vulnerability Fixes
This week, Microsoft Research open-sourced Agensh, a multi-agent harness designed for self-organizing workers without a central orchestrator. This development is particularly significant in light of Cognition's recent release of the Devin Vulnerability Fixes plugin, which aims to secure systems against recent cybersecurity threats.
The Agensh framework allows for scaling from 1 to 1,024 agents on Proxima, providing a flexible solution for various use cases. The absence of a central orchestrator enables decentralized decision-making and increased resilience. By open-sourcing the project, Microsoft Research invites the global AI community to contribute to its development.
Meanwhile, Cognition's Devin Vulnerability Fixes plugin addresses critical RCE flaws in PHP, Node.js, and self-managed GitLab deployments. The MIT-licensed agent skills facilitate inventory, patching, and testing, empowering developers to fortify their systems against potential attacks.
In related news, researchers and developers continue to refine AI models for various applications. A recent study found that Grok's AI model overwhelmingly recommends Republican candidates, while OpenAI's GPT favors Democrats. This disparity highlights the importance of transparency in AI decision-making processes.