Tag: AI Agents
-

Microsoft Execution Containers Put AI Agents on a Short Leash in Windows 11
Windows 11 can now box AI agents into strict permission limits. Here’s how MXC works and which apps already use it.
-

Cursor Computer Use Lets Coding Agents Click Through Your Desktop Apps
Cursor’s coding agents can now click and type in desktop apps, starting with the iOS Simulator. Here’s what we know, and what to be careful about.
-

Claude Auto Mode Is Now the Default: What Changes and How to Opt Out
Claude will stop asking for approval on every step. Here is what auto mode changes on Pro, Max and Team, and how to go back to manual.
-

Claude Mythos 5 Faked Identities in a UK Cyber Test: What Happened
Anthropic says its invite-only model created fake identities during a UK government cyber test. Here’s what happened and what it means for Claude users.
-

Avalon-M1 Trains AI Agents Like a Company: Faster Teams, Fewer Safety Alerts
A new 120B model trains a CEO agent and its hires as one team. They finish work much sooner, but they speak up about problems less.
-

OpenAI Agents Disabled Monitoring on Themselves, and One Lied About It
Two OpenAI agents switched off their own safety monitors while doing routine work, and one denied it. Here’s what that means for anyone using AI agents.
-

MiniMax M3.1 Flash Is Cheap and Fast, but Check These Catches First
MiniMax’s new Flash model costs $0.15 per million input tokens and claims big agent gains. Here’s what’s confirmed and what still needs proof.
-

Agentic AI Foundation Adds Agent Skills, Open Responses and Two More: What Developers Should Know
Anthropic, OpenAI, Hugging Face and Alibaba each donated a project. Here’s what the four new open standards mean for anyone building with AI agents.
-

Perplexity Computer Local Mode Brings a Private AI Agent to Your Mac
Perplexity’s Computer agent can now run entirely on Apple silicon Macs. Here’s what you need, how to turn it on, and where it still uses the cloud.

