AI Briefing#computer-use
The latest AI and developer news about #computer-use, with the original source and a short summary.
The latest AI and developer news about #computer-use, with the original source and a short summary.
LangChain unveiled new features for LangSmith and open-source tools for Deep Agents.
Computer Use capability is now built into Gemini 3.5 Flash, enabling the construction of agents that perform tasks directly in browser and desktop environments.
Gemini 3.5 Flash has built-in computer control features, improving agent development efficiency.
A new feature has launched that records a user's workflow on Mac and turns it into a reusable AI skill.
The next-generation computer use agent series Holo3.1, delivering strong performance in local and mobile environments, has been released.
Cognition is building an agent-based development environment where Devin can test and verify its own code on cloud VMs.
Anthropic has unveiled a new framework for managing a wide range of AI harms, including physical, psychological, and economic impacts.
Anthropic has released Claude Sonnet 4.5, a model with greatly improved capabilities in coding, agent building, and computer use.
A self-healing AI agent platform for legacy Windows desktop automation
Microsoft has unveiled MagenticLite and Fara1.5, agentic systems designed for small models.
It summarizes how to design agent evaluation harnesses that go beyond static benchmarks.
OpenAI is expanding Codex's Computer Use to remote desktop control.
opendesk, an open-source tool that lets AI agents control other computers over WiFi, has been released.
Open Computer Use has released local desktop automation as an MCP service.
Opendesk, an open-source tool that helps AI agents directly control the desktop, has been released.
For the same task, the computer use approach was 45 times more expensive than the structured API.
AI agents that directly manipulate UIs via vision are far less efficient in both cost and reliability than those using structured APIs.
Generates synthetic computers per user at scale to train long-horizon task agents.
Cua Driver has been released, enabling AI agents to control macOS apps in the background.
OpenAI has announced GPT-5.5, an agent-enhanced model.
Behind the glamorous success story, security incidents and scalability problems have come to light.
Codex received a major update spanning computer use, browser control, and even image generation.
Claude Sonnet 4.6 significantly improves coding, computer use, and long-context reasoning across the board.
Sandboxes for AI agents were compared and ranked based on their features.