Claude Becomes U.S. Enterprise Favorite as Anthropic Passes OpenAI in Business Adoption
For the first time, Anthropic leads OpenAI in U.S. business AI adoption — Ramp AI Index data shows Claude's agentic strength driving the shift.
News
Five OpenClaw Zero-Days Let Attackers Hijack Trusted AI Agent Access Across Platforms
Five newly disclosed OpenClaw zero-days let attackers impersonate trusted users and hijack AI agents across Slack, Discord, Teams, and more. Patch guidance inside.
Hermes Agent v0.16.0 'The Surface Release' — Nous Research Ships Cross-Platform Desktop and Web Admin Dashboard
Nous Research ships Hermes Agent v0.16.0 with native desktop apps for all platforms and a full browser-based admin dashboard — a major step toward OpenClaw feature parity.
OpenClaw Code Reuse Exposes AI Agent Accountability Problem — Developer Finds Own Code in Viral Framework
How a developer finding his own code inside OpenClaw sparked NanoClaw — the sandboxed, MIT-licensed alternative now backed by $12M in funding.
Apple Approves Poke as First Standalone AI Agent Inside Messages for Business — Third-Party AI in iMessage
Apple approves Poke as the first third-party AI agent inside iMessage via Messages for Business — a major platform gateway opening before WWDC 2026.
NSA Reportedly Using Claude Mythos for Offensive Cyber Operations — Anthropic Engineers Embedded Inside Agency
NSA is using Anthropic's unreleased Claude Mythos for offensive cyber ops, with ~6 embedded engineers — despite DoD flagging Anthropic as a supply chain risk.
Agentic AI Red Teaming Reveals Zero-Click Human-in-the-Loop Bypass Attack Chains — Microsoft Taxonomy v2.0 Ships 7 New Failure Categories
Microsoft's AI Red Team releases Taxonomy v2.0 with 7 new agentic failure modes. Zero-click HITL bypass chains now demonstrated end-to-end, with no human interaction needed.
Anthropic Warns Claude Is Building Itself Faster Than Expected — Calls for Coordinated Plan to Pause AI Development
Anthropic reveals Claude writes 80%+ of its own codebase and warns frontier labs need a coordinated pause plan as recursive self-improvement accelerates.
Cloudflare: AI Bots Now Outnumber Humans Online — 57.5% of Internet Traffic Is Automated, 18 Months Ahead of Forecast
Cloudflare data confirms bots now outnumber humans online at 57.5%, driven by agentic AI — 18 months ahead of when CEO Matthew Prince expected the crossover.
Microsoft Agent Framework at Build 2026: Agent Harness, Hosted Agents, and CodeAct Now Available
Microsoft Agent Framework ships at Build 2026 with Agent Harness, Hosted Agents (managed sandboxes), and CodeAct — the enterprise multi-agent runtime for Azure.
OpenClaw v2026.6.1: Self-Healing AI Agents and Multi-Channel Messaging Resilience
OpenClaw v2026.6.1 lands self-healing agents: automatic recovery from network flaps, failed tool calls, and multi-platform message collisions across Discord, Teams, Slack, and more.
Anthropic: 80% of New Production Code Now Authored by Claude — Enterprise Playbook Released
Anthropic reveals 80% of its production code is now Claude-authored, releasing a 3-step enterprise playbook for AI-native engineering.
Hermes Desktop Launches in Public Preview — Nous Research Ships Cross-Platform GUI for Hermes Agent v0.15.2
Nous Research launches Hermes Desktop in public preview—native macOS/Windows/Linux GUI for Hermes Agent v0.15.2, replacing CLI setup.
Microsoft Launches Scout — Always-On Autopilot Agent for M365 Built on OpenClaw, Secured by MXC
Microsoft Scout debuts at Build 2026 as an always-on M365 agent built on OpenClaw, with MXC OS-level sandbox security for enterprises.
OpenClaw v2026.6.1: Native Windows Node, Skill Workshop, and Multi-Agent Workboard Improvements
OpenClaw 2026.6.1 ships native Windows node support and Skill Workshop—a governed proposal pipeline for agent-created skills.
Google Releases Gemma 4 12B — Multimodal, Encoder-Free, Native Agentic Workflow Support
Google's Gemma 4 12B: encoder-free multimodal, Apache 2.0, runs on 16GB RAM with native agentic tool-use and Google AI Edge support.
h5i v0.1.5: Agent Radio Lets Claude Code and Codex Communicate in Real Time via Git
h5i v0.1.5 turns Git into an agent-to-agent message bus — Claude Code and Codex can now communicate in real time, no server needed.
NVIDIA Nemotron 3 Ultra 550B Released — Open-Weight MoE Agent Model Drops on June 4
NVIDIA Nemotron 3 Ultra 550B goes open-weight on June 4 — 55B active MoE, 1M context, top U.S. open model score.
OpenAI Codex Expands Beyond Coding to 'Every Role, Tool, and Workflow' with Role-Specific Plugins and Sites
OpenAI Codex evolves from coding tool to full agentic platform: 6 role plugins, Sites for hosted apps, and Annotations for doc refinement.
Researchers Demonstrate Self-Replicating LLM Worm That Runs Open-Weight Models on Hijacked Hosts
U of Toronto's CleverHans Lab demonstrates a self-replicating LLM worm that hijacks GPU hosts to sustain itself — no cloud API needed.
Anthropic Launches Services Track and Partner Hub for Claude Partner Network
Anthropic formalizes its Claude Partner Network with a three-tier Services Track and new Partner Hub portal, adding Deloitte, Accenture, and PwC as early Global Premier partners.
Claude Code v2.1.162 — Agents Now Report What They're Waiting For, Parallel Tool Resilience Improved
Claude Code v2.1.162 adds a `waitingFor` field to agent monitoring output and makes parallel tool calls resilient to individual failures.
OpenClaw Adds Auto Mode for Exec Approvals — \"Safer Than YOLO\"
OpenClaw's new 'auto' exec approval mode sits between YOLO and always-ask: allowlist first, AI reviewer for misses, humans for genuine risks.
OpenClaw Collaborates with NVIDIA for Stronger Agent Skill Security — ClawHub Skills Now Ship with Skill Cards and SkillSpector Scanning
OpenClaw and NVIDIA partner to add Skill Cards and SkillSpector scanning to all ClawHub skills, targeting AI-specific attack vectors like prompt injection.
OpenClaw v2026.6.2 Released — Skill Workshop Governs Agent-Created Skills
OpenClaw v2026.6.2 ships Skill Workshop: agents now propose skills for human review rather than writing directly, adding an auditable governance layer.