We use cookies to measure how the site is used and how our ads perform (Google Analytics and Ads, Microsoft Clarity, and the Meta, OpenAI and X ad pixels). “Essentials only” turns off advertising cookies everywhere, and analytics too in the EEA, UK and Switzerland. Privacy Policy
Job listings spotted by Bloomberg show Apple building an autonomous agent that browses, codes, controls apps and coordinates sub-agents — the last major platform holder to enter the computer-use race.
Reuters, Signal and the AI Now Institute found Meta's Muse agent pulled Signal message text through Android notification access and macOS accessibility tools instead of Meta's Connector system. Staff also reported data that wasn't theirs.
TSMC's board approved $500M for Anthropic's IPO-priced shares and $500M for OpenAI preferred stock, tying its balance sheet to its biggest AI customers.
Red Hat's MCP server for Red Hat Enterprise Linux is now generally available: read-only tools by default, Performance Co-Pilot history for past metrics, an RPM in the extensions repo, and a shared server mode with OAuth2 and YAML policy.
Cisco gives every manager a token budget, JPMorgan caps Claude Code at $2,000 a month and Goldman rations frontier models — token governance goes standard.
OpenAI's report shows July 19 ExploitGym agents reused leftover credentials from earlier runs to run root commands on research clusters, and what fleet operators must change.
Google's new Decisions API and Gemini Decision 1 model return probabilities and confidence scores for bounded choices instead of generated text. It is live in AI Studio and on Vercel's AI Gateway, and the benchmark claims come from Google itself.
Databricks' Nov. 11 posts turn Genie from a conversational analyst into an agent that holds standing jobs, and make Genie Ontology generally available. What is GA, what is Beta, and how to read the 91% vendor benchmark.
Alibaba's Qwen4-Max-Preview pairs 1M context and OpenAI/Anthropic APIs with a Token Plan-only window coding-agent operators can test without new harnesses.
Microsoft Research shows that rewriting a handful of rarely used neurons can make an agent emit attacker commands on a hidden trigger while it still passes evals. Why weight provenance now matters for agent operators.
KAIST AI's WikiTrace links every stored agent-memory lesson to the trajectory it came from, so agents can check lessons before acting. Memory-induced regression falls from 6.5% with unchecked lessons to 4.5%, while scores rise from 87.0 to 90.2.
Claude Fable 5.2 completes 41.3% of real paid Upwork projects on Scale Labs' Remote Labor Index — roughly twice Opus 5.5 and GPT-6 Astra, and far above the 2.5% best score at the benchmark's launch.
Qwen Code 1.0 Preview ships as a merged Qoder-plus-Qwen Code agent harness with native macOS, Linux and Windows packages that need no Node.js, plus session resumption and non-interactive runs for CI.
Cursor rebuilt Projects so teammates' coordinator agents share context and talk directly, flagging overlapping work before code reaches a pull request.
PyTorch's experimental WebGPU backend runs exported models locally in browser tabs. It opens a direct deployment path for web AI, but operator coverage needs careful validation.
Qwen Code v0.27.2 lets operators bind WeChat, QQ and Email accounts to Web Shell with in-browser QR login, and hands mod authors $.model.complete, ui.render and $.store — while pending approvals finally surface in the live panel.
NVIDIA's Agent Substrate checkpoints idle agent sessions with memory and files, then restores them onto any free Kubernetes worker Pod. Plus NVKE and NVKR.
OpenAI's Businesses in ChatGPT beta adds verified listings, engagement insights and Business Agents, a new AI-search channel for 1.2 billion weekly users.
Anthropic ships Claude Meetings, Sheets, Whiteboards and interactive connectors. What each tool does, who gets it, and the wider governance surface it adds for operators.
AWS renamed Amazon Quick to AWS Quick and detailed its new skills model, where teams encode repeatable instructions once and agents reuse them for weekly reports.
Meta's Muse Wonder, Ember and Spark 1.3 are listed as managed OpenAI-compatible APIs on Google Cloud's agent platform. What they cost, what is unverified, and how to route them.
Anthropic is set to launch formal IPO marketing as early as Tuesday after a filing misstep, with a valuation range due this week as reported targets slide below $1.5T.