From the archive
AI agents use language models and software tools to plan tasks, write code and carry out multi-step work. Follow the latest AI agent news, coding assistants and automation frameworks, with attention to reliability, permissions and human oversight. Read the source-linked reports to understand what each system can actually do, where it still fails and what changes for developers and businesses.
Lead story
Meta’s attempt to collect employees’ keystrokes and mouse movements for AI training was suspended after staff resistance and a reported internal data exposure. The episode shows the awkward bit of building systems that learn from human computer use: the data may be valuable, but the people generating it…
Read the report & discussion ↗OpenAI Watch
OpenAI’s Codex 0.154.0 is a sizeable step for its AI coding agent: GPT-6-Astra is now available in the model picker and Amazon Bedrock catalogues. The release also adds experimental managed worktrees, inline questions during active runs and a shared background server for Windows sessions.
AWS AI Watch
Heurist Finance is letting its AI agent buy premium market data one query at a time, with a spending cap attached to every run. Built on Amazon Bedrock AgentCore, the system shows how agentic payments can move beyond the demo stage without handing a bot an open wallet…
Meta AI Watch
Meta has announced Muse, a personal AI agent rolling out in the US on iOS, Android and Meta’s web experience. Its pitch is less “chatbot with opinions” and more “digital assistant that carries the task to the finish line”, including while you are away.
Anthropic Watch
Anthropic has released Claude Code v2.1.265, a broad maintenance update for its AI coding agent. The practical headline is reliability: plugins should be easier to discover, large tool outputs are now bounded, and several failure modes around agents, sessions and MCP connections have been repaired.
Watch Desk
ToolGrad flips the usual recipe for training AI agents. Instead of starting with a user question and making an agent hunt for a solution, Google Research’s framework builds a successful tool-use chain first, then writes the prompt and response around it. That matters because agents are only as…
Alibaba Qwen Watch
Qwen Code v0.23.3 adds an OpenAI Responses API content generator and lets a subagent hand a turn to an external agent over ACP, with Claude Code named as the first hand-off target. The practical point is a coding agent that is easier to connect, inspect and recover when…
Alibaba Qwen Watch
Qwen Code Desktop v0.3.0-preview.0 landed on 10 September with a notably connective upgrade: it can generate content through the OpenAI Responses API and delegate a subagent turn to an external agent over ACP, with Claude Code first in line. For developers, Qwen’s desktop workbench is widening beyond a…
Moonshot AI/Kimi Watch
Kimi Code 0.42.0 adds a clearer control room for AI coding work: users can inspect progress from the main agent and subagents, give the /btw side agent read-only tools, and recover more reliably after automatic context compaction.
Alibaba Qwen Watch
Qwen Code v0.23.2 is a substantial control-panel update for an AI coding agent: it adds default web search for eligible ModelStudio plans, one-command remote serving and separate permissions for planning and execution. The release also improves how long and multi-agent sessions expose their state, although it does not…
Watch Desk
Meta has introduced Muse, a US personal AI agent that connects to everyday services and completes tasks on a user’s behalf. This is Meta’s biggest move yet from chatbots that answer questions towards software trusted to send, book, buy and keep working after you close the app.
Alibaba Qwen Watch
Qwen Code v0.23.1 is a substantial upgrade to the machinery around AI coding agents, adding live subagent status, isolated named tasks and much better ways to navigate, inspect and govern long-running sessions. The release also retires its web UI package, making that a breaking change for existing users.
Alibaba Qwen Watch
Qwen Code has published SDK TypeScript v0.1.9, bundling CLI 0.23.0 and a substantial set of changes for people building with or around its AI coding agent. The headline item is a command-injection fix in shell-command validation, but the release is much broader than a routine patch.
