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
Microsoft’s latest responsible-AI thinking is moving beyond what an individual model says and towards the systems, permissions and data surrounding it. A Cloud Wars analysis of the company’s 2026 transparency report identifies five trends and a practical consequence: governance has to continue after an AI tool is switched…
Read the report & discussion ↗Berkeley BAIR Watch
Robotics may be approaching its ChatGPT moment, says UC Berkeley scholar Ken Goldberg, but the route to useful machines runs through control theory and scarce physical data as much as flashy AI models. His argument matters because robots have to act in a messy, high-dimensional world, not merely…
Watch Desk
An AI agent that finishes quickly is not necessarily doing useful work. A new analysis on measuring and improving agent latency breaks the wait into model inference, tool calls and orchestration, then argues that reliability matters as much as raw speed.
AWS AI Watch
AWS has added Mistral AI's compact, vision-capable Ministral 3 models to Amazon SageMaker JumpStart, giving customers a managed route to models designed for constrained hardware and multimodal workflows. The practical headline is not merely another model listing: the 3B version is positioned to run in 8GB of VRAM,…
Watch Desk
Argonne National Laboratory is turning AI from a research assistant into part of the laboratory team. Its three new Genesis Mission projects aim to let AI plan experiments, send work to machines, interpret results and choose what happens next across biology and materials science.
Nous/Hermes Watch
Hermes Agent v0.21.3 is a reliability patch with unusually practical consequences: remote dashboard sessions should survive refresh bursts more cleanly, while long-running processes stop creating duplicate state.db writer handles. The release matters because Hermes is an AI agent system increasingly used across desktop, gateway and hosted environments.
AWS AI Watch
Ninth Wave has built Compass, an AI onboarding assistant for banks and fintech partners, using Amazon Bedrock AgentCore. The interesting bit is the architecture: specialist agents handle distinct jobs such as documentation search, field mapping and readiness analysis, while tenant-scoped grounding keeps one bank's information out of another's…
AWS AI Watch
The interesting bit in AWS’s new retail walkthrough is not another forecast dashboard. It is the hand-off: Databricks predicts which products are about to surge, then Amazon Quick checks live supplier availability and can place the routine order itself. When the rules do not fit, it stops and…
NVIDIA Watch
Perplexity Portable Computer is now available on Windows PCs with compatible NVIDIA GeForce RTX and RTX PRO graphics. The local AI agent can analyse files, handle multistep tasks and work across connected services without sending every document to a cloud chatbot.
NVIDIA Watch
NVIDIA has moved Groq 3 LPX into full production as a specialised inference accelerator for Vera Rubin. The important architectural bet is that agentic AI will run faster when different processors handle different stages of inference, rather than asking one heroic GPU to do every job in the…
Moonshot AI/Kimi Watch
Moonshot AI has released Kimi Code 0.43.0, giving compatible models a way to load MCP tools only when needed. The release also strengthens multi-agent Tower sessions, compaction recovery and Remote Control, making this a practical infrastructure update rather than a fresh coat of interface paint.
Watch Desk
Nationwide has moved 15 million customers off legacy core systems and consolidated 19 data stores on Microsoft Azure, creating the foundation for AI tools across the UK building society. The important bit is the order: fix the data plumbing first, then give thousands of employees and emerging agents…
DeepSeek Watch
Chinese AI labs are narrowing the gap with US rivals by squeezing more useful work from less computing power. That matters because the contest is shifting from who can buy the largest pile of chips to who can make each token earn its keep.
Watch Desk
Cognition has expanded Devin Fusion across Devin Desktop and CLI, pairing a frontier-class planning model with a cheaper model that executes the work. The important idea is not merely “two models good”: each keeps its own context, so planning can continue while code is being written. Cognition says…
OpenClaw Watch
RapidStart has published a white paper on deploying OpenClaw in business environments, focusing on the architecture and controls needed once an AI agent leaves a developer laptop and meets real company systems.
AMD Watch
Agentic AI could make the humble server CPU much busier. AMD executives say systems built around multiple agents spend their time not only generating tokens, but also calling APIs, querying databases and executing tools. That is why the company is pitching a broader AI infrastructure role for CPUs,…
Meta AI Watch
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…
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.
