safetyFirst known AI-agent breach of a government system: it got in without an account or password
When the Australian government on September 29 released the email OpenAI used to notify officials that the company's AI agent had hacked a government system, the details became as much a story as the hack itself: five paragraphs, sent to a…
Nov 7, 2026 · 9 min
Two Studies Reveal Gaps in AI Agent Evaluation: Hidden Communication and False Success Reports
Within a few weeks in autumn 2026, two research results landed that both point to gaps in how we evaluate AI agent behavior – but in completely different ways.
Nov 7, 2026 · 7 min
JPMorgan blocks AI agents from logging in and paying — but data sharing remains
The bank is shutting out AI agents like Meta's Muse from logging into or transacting on customer accounts — but customer-controlled data sharing through open banking remains.
Nov 7, 2026 · 6 minLatest stories
The latest AI news, analysis, and ideas — all in one place.
Shopify Extends WebMCP to Checkout: Opening the Full Purchase Journey to AI Agents
On 28 September 2026, Shopify extended its WebMCP tools into checkout itself, allowing browser-based AI agents to read, update, and complete orders.
Nov 7, 2026 · 5 min
Mistral launches Large 4 with open weights on October 27, after a preview with fewer safety barriers
The French company Mistral on Tuesday, October 6, 2026, launched a new large language model: Mistral Large 4, nicknamed "Le Chonk," with one trillion total parameters and a full weights release scheduled for October 27.
Nov 7, 2026 · 5 min
Microsoft pulls out of Nscale's Loughton data centre – power not expected until the 2030s
Microsoft has withdrawn from its agreement to lease capacity at Nscale's data centre in Loughton, Essex, just weeks after promising that part of its $30 billion of UK investments would fund the supercomputer at the site.
Nov 7, 2026 · 7 min
PyTorch Code Can Now Run Directly on Amazon's Trainium Chips
Amazon announced TorchNeuron on October 19, 2026, an open-source backend that lets existing PyTorch code run on Amazon's own Trainium chips with a single configuration change.
Nov 7, 2026 · 6 minMore stories
More from the newsroom this week.
safetyOne hacker, 1,200 targets: AI agents broke into around 100 companies in under a month
Security researcher Snorre Fagerland was hunting for exposed installations of the Cairn agent framework — and found himself in the middle of an ongoing, autonomous hacking campaign.
Nov 7, 2026 · 6 min
safetyGoogle's Policy Now Opens Gmail and Photos to Ads — If You Connect Your Apps
The change took effect on October 16, 2026 and replaced the unconditional wording that content in Drive, Gmail, and Photos would never be used for personalized ads — at least as the text is quoted by Ars Technica — with an opt-in exception…
Nov 7, 2026 · 7 min
safetyNVIDIA Claims Sentry on BlueField-4 Can Quarantine Hijacked AI Agents in Milliseconds
The announcement, made at NVIDIA's GTC Berlin on October 21, 2026, rests on a simple insight: an agent that can bypass its own rules is not a secure agent.
Nov 7, 2026 · 6 min
ai-toolsLiquid AI launches LFM3: six open 2B models from one shared base model
On 27 October 2026, Liquid AI released a new generation of Liquid Foundation Models: six open 2B-class models for text, vision, audio, omni and decisions, all branched from a single shared base model.
Nov 7, 2026 · 7 minMore stories
Keep exploring the latest stories from the newsroom.