Pentagon Workers Vibe-Coded 103,000 Gemini Agents in 5 Weeks — Largest Known Government Agent Deployment
DoD workers built 103,000 Gemini-powered AI agents in 5 weeks on GenAI.mil — the largest known government agent deployment to date.
News
Unit 42: Autonomous Multi-Agent System Can Independently Plan and Execute Cloud Infrastructure Attacks
Unit 42's 'Zealot' agent autonomously executed a full GCP cloud attack chain — and defenders need to respond now.
DeepSeek V4 Preview Released — Open-Source 1.6T MoE Model with 1M Token Context Tops Benchmarks, Optimized for OpenClaw
DeepSeek V4 Preview lands with 1M token context, MIT license, and explicit OpenClaw optimization — two models for local and cloud use.
IBM X-Force: Agentic AI Vulnerabilities Are Outpacing Security — Autonomous Capabilities Expand Attack Surface Faster Than CVE Tracking
IBM X-Force warns agentic AI attack surfaces grow faster than CVE tracking can handle — with a clear hardening roadmap for practitioners.
OpenAI Launches Workspace Agents for Enterprise — Codex-Powered, Always-On Bots Replace Custom GPTs in Slack, Salesforce, Google Drive
OpenAI's Workspace Agents (Codex-powered) replace Custom GPTs with always-on cloud agents for enterprise — free through May 6.
OpenClaw v2026.4.22 Released — xAI/Grok TTS, STT, Image Generation, Local TUI Mode, and Trajectory Bundles
OpenClaw v2026.4.22 ships Grok TTS/STT, image gen, trajectory bundles, local TUI mode, and Tencent Hy3 — biggest release since 4.21.
UAE Cabinet Approves World-First Plan to Run 50% of Government Services on Agentic AI Within Two Years
UAE becomes first government to mandate 50% agentic AI integration across federal operations by 2028 — a historic global benchmark.
Anthropic Publishes Claude Code Degradation Postmortem — Three Product-Layer Changes Identified and Fixed
Anthropic's April 23 postmortem reveals three product changes—not model degradation—caused weeks of Claude Code complaints. Full breakdown.
BAND Exits Stealth with $17M Seed — Universal Agent Orchestrator Targets Multi-Vendor AI Fragmentation
BAND exits stealth with $17M seed to bridge multi-vendor AI agent fragmentation. LangChain + CrewAI + custom Python — finally talking to each other.
Google Rebrands Vertex AI as Gemini Enterprise Agent Platform at Cloud Next 2026 — A2A Protocol Live in 150+ Organizations
Google rebrands Vertex AI to Gemini Enterprise Agent Platform at Cloud Next 2026. A2A protocol now in production at 150+ orgs.
Anthropic's Claude Models Are Generating Vulnerable Code — Cybersecurity Experts Raise Alarm
Cybersecurity experts warn Claude Opus 4.7 introduces vulnerabilities in 52% of coding tasks — root cause traced to reduced thinking effort.
Claude Code's Security Deny Rules Can Be Bypassed With 50+ Subcommands — Critical Flaw Exposed in Source Leak
Claude Code's deny rules silently fail past 50 subcommands — critical flaw found in March npm leak, patched by Anthropic after Adversa AI disclosure.
Google Launches Deep Research Max — Autonomous Research Agent on Gemini 3.1 Pro With Native MCP Support
Google's Deep Research Max runs 160 autonomous searches on Gemini 3.1 Pro with native MCP support — now in public preview via the Gemini API.
Google Gives Every AI Agent a Cryptographic Identity — Auditable Authorization Policies Now Baked Into Gemini Enterprise
Google Cloud Now gives every AI agent a unique cryptographic SPIFFE identity with auto-rotating X.509 certs for zero-trust agentic workflows.
Amazon Bedrock AgentCore Adds New Features to Help Developers Build Agents Faster
AWS AgentCore ships new developer features for CrewAI, LangGraph, and LlamaIndex — cutting prototype-to-production friction.
Anthropic's Subscription Plans Are Becoming Obsolete — Head of Growth Signals Structural Repricing
Anthropic's Head of Growth signals Pro and Max tiers are structurally outdated — heavy agent workloads are being pushed to API billing.
Google Reveals 75% of New Company Code Is Now AI-Generated
Sundar Pichai: 75% of Google's new code is now AI-generated — up from 25% in 2024, marking a coding tipping point.
Hugging Face ml-intern: Open-Source AI Agent Beats Claude Code on Scientific Reasoning
Hugging Face's open-source ml-intern agent outperforms Anthropic's Claude Code on GPQA scientific reasoning — built on smolagents.
Uber Burns Through Its Entire 2026 AI Budget in Four Months — Claude Code Is the Reason
Uber's CTO reveals the company burned its entire 2026 AI budget by April — 5,000 engineers adopting Claude Code drove runaway costs.
Google Cloud Next '26: $750M Agentic AI Partner Fund, Merck $1B Deployment, HCLTech Gemini Enterprise BU
Google Cloud commits $750M to agentic AI at Cloud Next '26 as Merck pledges $1B Gemini Enterprise deployment and HCLTech launches first Gemini BU.
OpenClaw v2026.4.21 Released — gpt-image-2 Default, Auth Hardening, and Slack Thread Fix
OpenClaw v2026.4.21 ships gpt-image-2 as default image model and patches a critical auth bypass in owner-enforced commands.
Alipay AI Pay Launches — Enables OpenClaw-Compatible Agents to Make Payments Autonomously
Alipay extends AI Pay to OpenClaw-type agents — autonomous payments with authorization controls now live in APAC.
Anthropic Removes Claude Code from $20 Pro Plan — A/B Test or Permanent Change?
Anthropic A/B tests removing Claude Code from $20 Pro plan for new users, pushing agentic coding to $100+ Max.
OpenClaw v2026.4.20 Released — Kimi K2.6 Default, Cron State Split, and OOM Prevention
OpenClaw v2026.4.20 ships Kimi K2.6 as default, splits cron state for git stability, and adds OOM prevention.
Datadog State of AI Engineering 2026: 5% of AI Requests Fail, Agent Adoption Doubled, Claude Up 23pts
Datadog's 2026 AI report: 5% failure rate in production, agent adoption doubled, Claude surged +23pts in market share.