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
AWS describes a six-week programme in which four customer-facing professionals without engineering backgrounds built a multi-agent AI prototype, with about four hours of work each week. Its useful lesson is practical: give people time, mentors and a real problem to solve, rather than asking them to learn AI…
Read the report & discussion ↗AWS AI Watch
Cornerstone says its multi-agent AI system cut database diagnosis from 45 minutes to 10, while also reducing redundant alerts and replacing manual operational steps with a single interaction. The system, built with Amazon Bedrock and Strands Agents, offers a concrete example of AI being put to work inside…
Watch Desk
As agentic AI workloads grow, companies are trying to get more useful work from each unit of computing rather than simply buying more capacity. An Uber infrastructure leader said the company stabilised token costs while expanding agent use, using caching and a 24-million-node internal context graph to speed…
Salesforce Watch
When AI agents can work across company systems, a model’s own safety reasoning may not be enough to stop a harmful action. Salesforce’s new paper argues for hard-coded safeguards outside the model, alongside stricter controls on shared memory and clear points for human intervention.
AWS AI Watch
AWS’s AI business has passed a $25bn annual revenue run rate, while Amazon is spending heavily to expand its cloud and AI infrastructure. Yahoo Finance’s 7 October article also points to a sharp tension: the reported growth comes alongside a $7.6bn trailing-year free-cash-flow outflow, largely tied to higher…
Watch Desk
EqualAI president Miriam Vogel argues that public trust in AI depends on practical ways to check systems, correct mistakes and challenge decisions, not simply on assurances from companies. Her advice to organisations is concrete: map where AI is used, arrange independent evaluation, prepare to handle significant incidents, assign…
Meta AI Watch
Meta has brought its Muse AI assistant to iPad and added connectors for work and everyday apps, including Asana, Canva, Dropbox, GitHub and Notion. The expansion gives Muse a larger screen and a broader set of services to work with, while Meta’s new business features put the agent’s…
Watch Desk
Stacklok is building Mecatl, an open-source harness designed to run AI coding agents as cloud-managed infrastructure rather than as tightly coupled desktop tools. The pitch is aimed at organisations that want to govern agent sessions, execution and access centrally, instead of leaving them scattered across developers’ machines.
Watch Desk
AI agents may be setting up the next cloud-style cycle of vendor lock-in and expensive migrations, argues Guickly founder Prashant Jalan. His practical advice for companies adopting agents: measure costs independently, keep key assets in their own repositories and plan for switching before a platform makes it painful.
Microsoft AI Watch
Microsoft’s Agent Framework 1.24.0 brings a broad set of .NET changes, including fixes for session state and search memory, plus tighter checks around agent capabilities and MCP approvals. For developers building AI agents, the useful story is a release focused on making the machinery behave more predictably, not…
Watch Desk
AI rules that live in a handbook can be ignored by people and agents alike. Pipefy chief executive Alessio Alionço argues that governance must be built into AI systems themselves, with clear limits on what agents may do and a record of every consequential action.
Watch Desk
Bitdefender has opened a free macOS beta of AI Guardian, a security product that checks what supported AI agents try to do with tools, files and data. It is aimed at a growing weak spot in agent use: a system that can act on your computer needs more…
Watch Desk
Atlassian has announced a protocol designed to move AI use beyond isolated chat windows and into shared workspaces, where people and agents can collaborate. The central question is whether agents can take part in team workflows while working within clear permissions and governance, rather than acting as clever…
OpenRouter Watch
OpenRouter says its AI sales agent, Rasp, returns about 600 hours a month to the company’s five-person sales team, while handling much of the work around inbound leads, call preparation and CRM updates. The case study offers a detailed look at how the company says it built the…
Watch Desk
GitHub is rebuilding its Git infrastructure to handle the heavy, concurrent reads and writes associated with AI-driven software agents. Its internal benchmarks show up to 35 times higher write throughput, a potentially important change for teams running development work at agent scale.
Google DeepMind Watch
Google’s AI infrastructure chief Amin Vahdat says power is the most fundamental constraint on scaling AI, even as failures, networking and hardware efficiency all put pressure on large systems. His account also shows why faster chips alone will not settle the question: the useful measure is how much…
Watch Desk
Wajo is building an AI agent that can contact people and businesses, arrange services and pay on a user’s behalf. The startup’s pitch is that trust should be built into the product from the start, which matters when an assistant is being given access to your inbox, contacts…
Meta AI Watch
Meta has opened Muse Gadgets, a project for developers to connect its personal AI agent to home-built hardware. The useful shift is from an assistant that lives in an app to one that could reach buttons, screens and sensors around the home.
METR Watch
METR says a vulnerability in the Inspect transcript viewer could have let an AI agent change what a human reviewer saw, including earlier actions in the transcript. The flaw was patched within a day of being reported, and METR says it has not observed agents exploiting it.
