AI Agents - Page 16

Two AI infrastructure launches from 8 October 2026: Goodfire's "inside-out" monitors, which read a model's internal signals to catch harmful agent sessions, reach Baseten customers (all performance figures are Goodfire's own), and Upscale AI unveils Token Fabric, a networking platform for AI data centres with general availability planned for early 2027.

A preprint from EPFL researchers, listed on Apple's research site, finds a single minimal coding-agent session matched or beat elaborate machine-learning agent systems in its tests. Pasqal releases skills that let coding agents prepare quantum experiments, and Microsoft makes 30 Dynamics 365 CRM skills generally available, with Autopilot still in private preview.

Anthropic says a team of Claude Science agents helped build the first complete UV map of the sky, about a third of it predicted rather than observed. A preprint finds agent benchmark scores can move because of the grader, not the agent, and Exa previews a search benchmark it designed and runs itself. What each one shows, and what it doesn't.

Two device stories from 8–9 October 2026. Nikkei Asia reports that Apple told some suppliers to cut production of iPhone 18 Pro components after price rises dampened demand; Apple has not commented. And start-up Natura has shown Interface, a $99 smart ring for sending voice requests to AI agents, which cannot be ordered yet.

CrowdStrike says an unnamed, likely Chinese-speaking attacker used ARTEX, an open-source AI penetration-testing agent, alongside several AI models to steal data from South Korean financial firms. What CrowdStrike found, what Korean investigators say, and what it does not prove.

Anthropic has published a new Usage Policy, taking effect on 12 November, with a section on deceptive campaigns, tighter wording on weapons and surveillance, and a ban on extreme abuse of its models. The UK's ICO says ten AI developers have made or committed to data protection changes, and has opened a call for evidence on AI agents. What changes, and for whom.

Three AI-agent start-ups announced funding on 8 October: Hone ($60M seed), Vesta ($30M, reported) and Catalyst ($30M seed, reported). Crossmint launched a toolkit that lets AI agents pay by card. What each company says, who reported what, and which figures are company claims.

Google Cloud's Gemini agent, in private preview, takes on whole tasks and can get its own Workspace email and calendar. Coworker pricing is not yet set.

Three agent-security stories from 8 October. Zenity says one prompt to an exposed AWS AgentCore agent could lead to control of every agent in the same account and region, and AWS disputes it. Claude Code adds an option that blocks an action when a safety hook fails. Cisco Talos finds malware that writes to AI analysts. What each shows, and what it does not.

An unreleased "Additional sandbox options" setting found in Google's Gemini desktop app for Mac would let the AI reach files, apps and network connections beyond the folders users connect today. It is hidden, unconfirmed by Google and arrives just as Apple is tightening Full Disk Access.