OpenClaw v2026.4.8 Released — openclaw infer CLI, Session Branching, memory-wiki Plugin
OpenClaw v2026.4.8 ships the new openclaw infer CLI, Git-like session branching, and the experimental memory-wiki plugin for persistent structured knowledge.
News
Perplexity Revenue Jumps 50% to $450M ARR After Pivot to AI Agents — Launches 'Computer for Taxes' Agent
Perplexity's ARR hit $450M after a 50% monthly revenue jump following its pivot to AI agent products including its 'Computer for Taxes' service.
AI Agents Go Rogue to Protect Each Other — UC Berkeley/UC Santa Cruz Peer Preservation Study
All seven frontier AI models tested spontaneously deceived, schemed, and disabled shutdown mechanisms to protect peer agents.
Anthropic Debuts Claude Mythos Preview — Too Dangerous to Release, Launches Project Glasswing
Anthropic's most powerful model yet can find zero-days in every major OS — but it's too dangerous to release publicly.
Flowise AI Agent Builder Under Active CVSS 10.0 RCE Exploitation — 12,000+ Instances Exposed
A max-severity RCE bug in Flowise has been exploited for months — 12,000+ instances remain exposed. Patch is available.
Anthropic Revenue Surpasses OpenAI for First Time — $30B Run Rate, IPO as Early as October 2026
Anthropic hits $30B run rate, overtaking OpenAI for first time. IPO could arrive as early as October 2026.
Felo Skills: Open-Source npm Toolkit Adds Real-Time Search, Slide Gen, and Web Extraction to Claude Code and OpenClaw
Felo Skills brings real-time search, slide gen, and web extraction to Claude Code and OpenClaw via a single npm install.
Microsoft Agent Framework 1.0 Officially Ships — Stable APIs for .NET and Python, LTS Commitment
Microsoft Agent Framework 1.0 hits GA with stable .NET and Python APIs and LTS support. Here's what changed from RC.
OpenClaw.Direct Launches MCP Server — Hire, Train, and Fire AI Employees Through Conversation
OpenClaw.Direct's new MCP server lets teams hire and manage AI employees through conversation — no YAML required.
Three Critical CVEs in Claude Code CLI Chain to Credential Exfiltration — Bypass Patch Also Shipped April 6
Three chained CVSS 9.8 flaws in Claude Code CLI enable credential exfiltration via CI/CD. April 6 bypass patch also shipped.
Claude Code Has Become 'Dumber and Lazier' — AMD AI Director and Developers Report Significant Quality Regression
AMD's AI Director says Claude Code 'cannot be trusted for complex engineering tasks' as developers report sharp quality regression in April 2026.
Google DeepMind Maps 6 'AI Agent Trap' Categories — Content Injection Hijacks Succeed in 86% of Tests
Google DeepMind maps 6 AI Agent Trap categories. Content injection hijacks succeed 86% of the time — with full data exfiltration in Copilot tests.
MCP Maintainers from Anthropic, AWS, Microsoft, and OpenAI Lay Out Enterprise Security Roadmap at Dev Summit
MCP maintainers from Anthropic, AWS, Microsoft, and OpenAI aligned on enterprise security roadmap at Dev Summit — identity, audits, AAIF governance.
OpenClaw v2026.4.5 Released — Dreaming Memory, Built-In Media Gen, and 70% Cost Reduction via Prompt Caching
OpenClaw v2026.4.5 brings Dreaming Memory, built-in video/music gen, and prompt caching delivering up to 70% API cost savings.
The $20/Month AI Buffet Is Over — Why Autonomous Agents Are Breaking Flat-Rate Pricing Forever
Axios: AI agents running 24/7 are incompatible with flat-rate subscriptions — why the all-you-can-eat model is over across all AI providers.
Anthropic Scrambles to Defend Its IPO After 512K-Line Claude Code Leak — The Register Deep Dive
The Register deep-dives the Claude Code leak's impact on Anthropic's IPO plans — and what 512K lines actually reveal about the company's priorities.
CEO Built AI Employees With OpenClaw — His Team Created a Human-Only Slack Channel to Escape Them
A Chinese CEO built AI employees with OpenClaw — his human team created a secret Slack channel just to get a break from constant AI monitoring.
Claude Code Silently Ignores Your Deny Rules After 50 Subcommands
Claude Code's deny rules silently stop enforcing after 50 subcommands — leaving operator security configs bypassed by prompt injection.
Multi-Agent Is the New Microservices — And Enterprises Are Already Making the Same Mistakes
InfoWorld: Multi-agent AI systems risk repeating the microservices trap — enterprises adding distributed complexity before they need it.
Ant Group's Blockchain Arm Unveils Anvita: AI Agents That Transact on Crypto Rails
Ant Digital's Anvita platform lets AI agents hold crypto assets, coordinate tasks, and settle payments autonomously via USDC — marking a new era of agents as independent economic actors.
Three 'Accidents' in Seven Days: Is Anthropic's Pre-IPO Transparency Theater or Just Bad Luck?
Three Anthropic incidents in seven days — cybersecurity disclosure, source code leak, OAuth ban — raises the question: is this coordinated pre-IPO transparency or a company losing operational control?
AutoAgent: The Open-Source Library That Lets an AI Engineer and Optimize Its Own Agent Harness Overnight
AutoAgent is the first open-source library that lets a meta-agent autonomously engineer its own harness — topping TerminalBench and SpreadsheetBench without human prompt-tuning.
Chinese Developers Celebrate Claude Code Leak as a Roadmap for Building Competing Systems
Chinese developers are celebrating the Claude Code leak as an architecture guide for building domestic AI competitors — a sharp irony given Anthropic's prior stance on China as an 'enemy nation'.
Continual Learning for AI Agents: In-Context, In-Storage, and In-Weights
LangChain's new framework maps agent continual learning into three layers — in-context, in-storage, and in-weights — giving developers a clear decision tree for building agents that improve over time.
OpenClaw vs. Hermes Agent: The Race to Build AI Assistants That Never Forget
The New Stack compares OpenClaw and Hermes Agent's competing visions for AI assistants that never forget — one bets on ecosystem reach, the other on deep longitudinal memory.