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.
Microsoft attributes the ~24-hour M365 outage to a network capacity upgrade — not cyber or AI — as Anthropic's and OpenAI's M365 connectors recover. What connector-dependent agent teams should take from it.
Nov. 13, 2026: the Fed, FDIC, OCC and NCUA moved the third-party risk comment deadline to Dec. 16, the rule banks will use to vet every AI vendor they onboard.
Cloudflare's Workers-native agent browser is now Apache 2.0. What coding- and web-agent operators can do with a self-hostable browser engine, and the unauthenticated-deploy trap in its own README.
Claude Code 2.1.325 sends sandboxed Bash commands through the normal permission flow unless sandbox.autoAllowBashIfSandboxed is true. Scheduled skills, Always allow rules and read-only git also change. Here is what to set.
llama.cpp and ggml v1.0.0 promise major-version bumps for breaking C API changes, remove deprecated symbols and --defrag-thold, add FP8 KV caches, and make llama-server refuse unauthenticated tools on network addresses.
LangChain's CapitalG-led Series C at $3B doubles its valuation within a year, backed by 7,000 LangSmith customers. What the round and the Nov. 20 release event mean for agent platform buyers.
Mirendil raised a $1B Series A to scale Lor, a platform where hundreds of AI agents run the trial-and-error loop of model development. A single researcher already used it to pretrain a 100B+ parameter model in 3.5 months for $10M.
ClickHouse open-sourced Loom, a self-hosted MCP server storing agent memory and a live code index in one ClickHouse database for Claude Code, Codex and Cursor.
DeepSeek put the capabilities behind its consumer chat app on the public API: web search, deep thinking search, file parsing, link reading, safety checks, voice, plus standalone OCR/ASR/TTS and an MIT-licensed front end. Here is what shipped, the parameter trap in the docs, and what to do this week.
Anthropic's status page records two Nov 13, 2026 incidents: sign-in failed until 11:58 UTC, then elevated errors, with connectors still degraded at 14:36.
GitHub says hosted Actions runners stop ARM32 job execution on Windows and Linux by Dec. 8, 2026, after node20's end of support, because Node 24 dropped ARM32. Here is how to find affected workflows and what to migrate to.
Nvidia privately told a lawmaker chip tracking is feasible while Democrats open a second inquiry into China sales. What it means for GPU fleet operators.
Lightning Labs' pi402 pays for LLM inference in bitcoin per token from inside a coding agent, with built-in spend caps and no API keys, accounts or Lightning node needed.
Claude Code 2.1.323 lets Claude write JavaScript that calls its own tools, as an experimental opt-in that organizations can turn off. The release also changes the default context window for unrecognized model IDs to 1M, adds plugin version gating and closes several sandbox and permission gaps.
Expo opened cloud iOS and Android simulators to all paid customers on Nov 12, letting Claude Code, Cursor and Codex install, run and attach PR evidence.
Gemini 4 Hydrogen is rolling out today in the Gemini API, AI Studio, the Gemini app, Search and Antigravity. It costs $1.50/$7.50 per million tokens across its full 1M context. Here is who gets it and what the speed claims leave out.
OpenAI's alignment team reports that pretraining, RL, and model-to-model distillation each push two proxies for chain-of-thought monitor evasion upward, with direct implications for anyone auditing agents via visible reasoning.
Nvidia's board is preparing to oust founder-CEO Jensen Huang, Semafor and the WSJ report. Some directors favor CFO Colette Kress as interim chief if he is removed — no decision has been announced, and there is no succession plan.
LangChain details an artifact-generation workflow built around account context, leadership-owned templates and visual review, including a stopping rule that cut one one-pager run from 18 minutes to seven.