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
A new benchmark finds that even advanced computer-use agents can reach the wrong result by taking the wrong route. OSWorld-Pro evaluates more than 300 tasks and 2,800 subgoals, exposing the small, consequential mistakes hidden by a tidy final answer.
Read the report & discussion ↗Watch Desk
AI agents that rewrite their own instruction and memory files can quietly degrade their performance, according to Jamie Watters. The practical fix is decidedly unglamorous: keep a person in the editing loop, review the files regularly and delete the bits that no longer earn their keep.
DeepSeek Watch
DeepSeek researchers have introduced DSec, a production sandbox platform designed to run large-scale training and evaluation for AI agents. Its significance is less glamorous than a new model name, but arguably more useful: DSec targets the infrastructure bottlenecks that appear when agents need vast numbers of isolated environments…
Watch Desk
Coinbase has published an MCP server and command-line interface that let compatible AI agents manage portfolios, execute spot-crypto trades, trade US futures and equities, and pay for some data services with USDC. The important shift is not another chatbot learning to discuss markets. It is an agent being…
Watch Desk
Apache Ossie is weighing a hybrid query design for semantic layers used by BI tools and AI agents, aiming to make answers more consistent rather than letting every system improvise its own interpretation. That matters because reliable data access is less about producing another fluent paragraph and more…
Meta AI Watch
A reported zero-day in Meta’s Muse AI assistant could let locally installed apps or terminal commands take control of a user’s Muse account and its unusually broad privileges. The security report matters because Muse is designed to act across email, shopping, calendars and other services, rather than merely…
Watch Desk
Fireworks AI is arguing that the next performance leap in coding agents may come from routing each task to the right model, rather than asking one frontier model to do everything. Its new FireRouter approach keeps a pool of open and closed models available, while preserving context caches…
Watch Desk
A study of an AI coding agent found that describing the desired structure in plain language, without naming the scoring formula, improved results and reduced correctness losses. The trade-off was a roughly 20% increase in both agent costs and tool turns, a neat reminder that clever prompting can…
Watch Desk
Google has introduced GKE Pod snapshots, a new Kubernetes feature that saves a running workload’s state, including CPU and GPU memory, so it can be restored on demand. The practical prize is faster AI inference startup, with Google claiming latency reductions of up to 89% and a route…
Watch Desk
Xiaomi has released the MiMo-V2.6 family of omni-modal AI models, including Pro and Flash versions, with a research programme built around scaling reinforcement learning across different environments. The important bit is not simply another model joining the crowded buffet. Xiaomi says it is publishing the training machinery, environments…
Watch Desk
Large swarms of AI agents may finish work faster than one agent, but Toby Ord’s analysis suggests the speed comes with diminishing returns and substantially higher total compute. The useful takeaway is less “more agents equals smarter” and more “parallel work has a coordination tax”.
NVIDIA Watch
Nvidia is pushing deeper into server CPUs as AI agents create demand for the computing work around each model call. The important shift is that the agent boom may expand the market for processors, memory and virtual machines, while Nvidia tries to sell the whole rack rather than…
Watch Desk
Amazon Bio Discovery says it has published three computational studies on antibody design, including an agent-guided process that produced experimentally validated nanobody hits against a novel cancer target. The important point is not simply that AI suggested molecules, but that some of those suggestions were tested in the…
Meta AI Watch
Amazon has blocked Meta’s Muse AI agent from shopping on Amazon.com, according to Forbes, twelve days after Meta launched the tool. The immediate significance is bigger than one pop-up: the companies that control online shops can decide whether rival agents are allowed to browse, log in and buy…
Watch Desk
Bun, the JavaScript and TypeScript runtime, has rewritten hundreds of thousands of lines from Zig into Rust in an AI-assisted project that the company says fixed 128 longstanding memory-safety bugs. The result is a striking case study in using coding agents for a task usually associated with years…
Watch Desk
Cloudflare is proposing an Agent Development Lifecycle for software built and maintained by autonomous AI agents, putting orchestration, continuous testing, observability and tightly bounded credentials at the centre of the process. The bigger idea is not a new acronym for the filing cabinet. It is a challenge to…
Watch Desk
ASUS has unveiled the ProArt Air, a 1.28kg business laptop designed to run AI-agent tools with local Intel processing and Tencent’s WorkBuddy platform. The practical shift is modest but real: AI features are being bundled into the office computer rather than left as a cloud tab someone has…
