NEW: OpenAI slashes ChatGPT Pro to $100/mo — 5x more Codex usage, 10x promo through May 31, usage limits RESETCONFIRMED: Anthropic launches Project Glasswing with Claude Mythos Preview — everything ccleaks leaked on March 26 is now officialNEW: /ultraplan command is NOW LIVE in Claude Code — cloud-based interactive planning with Plan → Edit → Execute workflowBREAKING: Anthropic cuts third-party harnesses from Claude subscriptions — OpenClaw, OpenCode blocked starting April 4. Pay-as-you-go or leave.OpenClaw creator Peter Steinberger: 'They copy popular features into their closed harness, then they lock out open source'Anthropic offering $20-$200 one-time credits + up to 30% bundle discounts on extra usage — claim by April 17 or lose itNEW: Anthropic files 8,000+ copyright takedowns on GitHub after accidentally leaking Claude Code's proprietary source, per WSJFake news circulating that the Claude Code leak was an April Fools joke — the leak is realFake news about Anthropic making Claude Code open source and rebranding to OpenClaudeAnthropic team acknowledges hitting usage limits in Claude Code faster than expected — actively investigating1,884 TypeScript files analyzed from leaked cli.js.map sourcemapDevelopers are building modified versions of Claude Code using the leaked source — custom forks appearing on GitHubAnthropic DMCA retraction — initial sweep hit 8,100 repos, scaled back to 1 repo + 96 forks after backlashLeaked code mirror surpassed 84,000 stars on GitHub — claw-code clean-room rewrite crossed 100,000 starsSecurity researcher Chaofan Shou first flagged the leak — post hit 28M+ viewsAnthropic's Mythos model leaked 5 days before Claude Code — two leaks in one weekAxios npm package compromised same day as leak — users who installed Claude Code Mar 31 may have pulled trojanized dependencyOpenClaude community fork lets you run Claude Code's tools with any LLM — OpenAI, Gemini, DeepSeek, OllamaClaude Code's $2.5B annualized revenue exposed in leak — enterprise adoption accounts for 80% of revenue, per VentureBeatBoris Cherny confirms 'human error' — deploy process had manual steps, no one was firedLeaked source reveals 'undercover mode' — hides AI-authored commits from Anthropic employees in open-source projectsAnti-distillation system discovered — Claude injects fake tools into API calls to poison competitor training dataFrustration regex found in source — Claude monitors for expletives like 'wtf' and 'fuck' to adjust its behaviorPython fork hit 111K stars and 98K forks in one day — one of the fastest growing GitHub projects in history, per CybernewsSupply chain attack targets leak — fake npm packages color-diff-napi and modifiers-napi registered to trap developers compiling leaked code44 hidden feature flags discovered including always-on background agent KAIROS and a Tamagotchi-style AI pet called BUDDYBloomberg: Anthropic 'scrambling' to contain damage — leak hands competitors a literal blueprint for building AI agentsClaw-code creator Sigrid Jin profiled by WSJ — one of world's most active Claude Code users with 25B+ tokens processedAnthropic's total annualized revenue hit $19B as of March 2026 — Claude Code leak threatens competitive moatThe New Stack: 44 features Anthropic kept behind flags — swarms, daemons, and an always-on agent architecture