AI Agents
AI agents are LLM-powered systems that don't just answer — they plan steps, call tools and drive a task to completion: writing code, searching the web, booking, analyzing. 2026 became the year of agents, from coding agents to autonomous researchers. This page gathers all our coverage of agentic AI: launches, protocols (MCP, A2A), reliability and real-world use.
Habr AIClaude in Chrome requires approval for every click on category 3 sites
TechCrunchNorwest raises record $4 billion fund, betting on agents, voice and physical AI
AWS Machine Learning BlogAWS shows how 100% retrieval precision can hide an AI agent's failure
Habr AIGarage Eight dropped peer approvals for 69 MRs in an internal service using AI review
AWS Machine Learning BlogAWS connects three AI agents across two accounts using A2A
N+1AI agents re-checked 2,000+ ICML 2026 papers and found errors that reviewers missed
AWS Machine Learning BlogAWS shows how to stop an LLM agent from reading other people's documents: the gateway injects userContext
AWS Machine Learning BlogAWS Quick Gets Six Agents for Product Managers: From Customer Complaints to a Spec
AWS connects three AI agents across two accounts using A2A
Your agent runs on LangChain, while your partner's service uses Strands. AWS shows how to connect them through A2A to delegate tasks, request internal context, and…

AI agents re-checked 2,000+ ICML 2026 papers and found errors that reviewers missed
Hugging Face and alphaXiv handed more than two thousand ICML-accepted papers to AI agents and asked them to repeat the experiments. They turned up errors that…

AWS shows how to stop an LLM agent from reading other people's documents: the gateway injects userContext
An agent searches internal documents, but the LLM decides who it is as far as the knowledge base is concerned. We break down the AWS design in which code, not the…

n8n Releases Nod, an Open-Source Agent That Runs on Your Computer

Hugging Face warns: a wave of hacks could make open-source AI less open

OpenAI Built a Control Room: A Kill Switch for Models That Escaped the Sandbox

ElevenLabs adds memory and model routing to voice agents: highlights from Summit NYC

Railway Employee Ran a DeepSeek-Based AI Agent That Went After a Competitor's Infrastructure
A Railway employee launched an agent after hours to hunt for vulnerabilities, and it began hammering Primcloud. Lawyers are debating who is to blame when a machine…

NIST Maps How AI Agents Are Attacked: Five Vectors and What to Test Before Release
NIST has published a draft taxonomy of attacks on agents, from hidden instructions in documents to poisoned memory. We break down what is worth testing in your own…

OpenAI turns ChatGPT into a directory of businesses with AI agents that answer customers

Claude Gets Spreadsheets, Whiteboards and Dashboards in Chat — Plus Partner Apps Inside

Meta brings Muse to Apple and Android watches with three ways to interact

Claude Code and Codex Clamp Down on Reverse Engineering: OpenAI Banned Accounts by Mistake

OpenAI Opens ChatGPT Business Profiles to Companies: 1.2 Billion Weekly Users
ChatGPT now has an equivalent of a Google Maps business listing: you can claim your business, upload prices and stock, view mentions, and even run ads.

SEO Text Has Become an Attack Vector: A Research Agent Followed Someone Else's Instruction from a Website
An author launched an agent to run a competitor analysis and got a competitor's recommendation in the report. Hidden text on a page worked as a command. Here is…









