📖 3 min read
Anthropic Pulls the Plug on Third-Party Claude Subscriptions
In a move that sent shockwaves through the AI developer community, Anthropic announced on Friday that Claude Pro and Max subscriptions will no longer cover usage on third-party tools — most notably OpenClaw, the popular open-source AI agent framework.
The change took effect Saturday at 12 PM PT. Boris Cherny, head of Claude Code, cited “outsized strain” on Anthropic’s systems from autonomous agent workloads. Users running Claude through OpenClaw and similar harnesses will now need to use API credits directly. The community reaction was swift and loud — developer and OpenClaw content creator Alex Finn called it a “massive mistake,” predicting open-source local models will close the gap with Claude Opus within months.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
Every Single xAI Cofounder Has Now Left Elon Musk
The last remaining cofounder of Elon Musk’s AI startup xAI — Ross Nordeen — has departed the company, completing an extraordinary exodus of all 11 original founding members in under three months.
The mass departure raises serious questions about xAI’s direction as it reportedly gears up for a combined IPO with SpaceX. Sources describe internal friction over strategy, though Musk has historically rebuilt leadership teams after similar shakeups at Tesla and SpaceX. Whether that playbook works for a company competing against OpenAI, Anthropic, and Google in the fiercest AI race in history remains to be seen.
Zuckerberg Is Writing Code Again — With Claude Code
According to The Pragmatic Engineer, Mark Zuckerberg submitted three diffs to Meta’s monorepo in March 2026 — his first code contributions in roughly two decades. The CEO is reportedly a heavy user of Claude Code CLI, Anthropic’s terminal-based coding agent.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
He joins a growing trend of tech leaders returning to the codebase thanks to AI coding tools. Y Combinator president Garry Tan has done the same. The takeaway: when an AI pair programmer can handle the boilerplate, even CEOs running trillion-dollar companies find reasons to ship code again. If you’re exploring AI-powered development workflows, check out our friends at BetOnAI.net for the latest analysis.
Microsoft Drops $10 Billion on Japan’s AI Future
Microsoft unveiled a four-year, $10 billion investment plan for Japan covering AI infrastructure, cybersecurity, and workforce development. The commitment runs from 2026 through 2029 and includes expanded cloud capacity, government cyber defense cooperation, and a goal to train 1 million engineers and developers by 2030.
Japan’s Sakura Internet jumped 20% on the news. The deal — part of Microsoft’s broader Asia AI push — signals that the infrastructure war is now just as important as the model war. For reviews of the tools being built on top of this infrastructure, AiToolCrush.com keeps a running list.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
Open-Source AI Agents Heat Up as Hermes Agent Challenges OpenClaw
The open-source AI agent space is getting crowded. Hermes Agent, a new multi-agent workflow system, launched this week as a direct alternative to OpenClaw, focusing on cross-platform automation and modular task pipelines.
Meanwhile, the OpenClaw ecosystem itself continues to grow rapidly — now powering over 770,000 active agents — though a recent security disclosure (CVE-2026-25253, CVSS 8.8) and the Moltbook database leak exposing 35,000 emails and 1.5M API tokens serve as a reminder that agent security can’t be an afterthought. If you’re building with AI agents, staying on top of security patches is non-negotiable.