AI News Roundup: Anthropic’s Code Leak Fallout, OpenAI Investors Flee to Anthropic, California Defies Federal AI Order

Anthropic’s Claude Code Source Code Leaks in npm Packaging Error

Anthropic is in damage-control mode after accidentally shipping the entire source code of Claude Code — 512,000 lines across 1,906 TypeScript files — to the public npm registry on March 31. A debug file mistakenly bundled into a routine update pointed to a zip archive on Anthropic’s cloud storage containing the full codebase.

The leak was discovered by Chaofan Shou, an intern at Solayer Labs, whose tweet with a download link drew 16 million views. Within two hours, a GitHub mirror of the code became the fastest repo in history to hit 50,000 stars. Anthropic pulled the npm package around 08:00 UTC and said “no sensitive customer data or credentials were involved.”

The fallout is significant: competitors can now reverse-engineer how Claude Code’s agentic harness works, and open-source clones are already appearing. Anthropic maintains this was “a release packaging issue caused by human error, not a security breach,” but it’s the company’s second major leak in days — coming right after the accidental exposure of its internal “Mythos” project.

OpenAI Shares Go Unsold as Investors Pivot to Anthropic

In a dramatic shift in AI market sentiment, OpenAI shares have become nearly impossible to unload on the secondary market. According to Bloomberg, about a half-dozen institutional investors — including hedge funds and venture capital firms — have approached Next Round Capital looking to sell roughly $600 million in OpenAI shares in recent weeks, with bids coming in at a 10% discount to OpenAI’s $852 billion valuation.

Meanwhile, buyers across Augment, Hiive, and other secondary platforms are sitting on $2 billion in cash ready to deploy into Anthropic. Hiive alone has registered more than $1.6 billion of demand for Anthropic shares at valuations around $600 billion — more than 50% above its last funding round of $380 billion. The investor exodus reflects growing confidence in Anthropic’s technical trajectory and concerns about OpenAI’s governance and valuation.

California Fires Back at Trump’s AI Preemption Order

Governor Gavin Newsom signed a “first-of-its-kind” executive order on March 30 requiring AI companies seeking state contracts to certify safeguards against illegal content, harmful bias, and civil rights violations. The move is a direct rebuke of President Trump’s December executive order that directed the DOJ to challenge state AI laws in federal court.

Newsom’s order also directs the California Department of Technology to create nationwide-first recommendations for watermarking AI-generated images and video. The state is additionally prepared to separate its AI procurement process from the federal government’s if needed. Colorado and other states have signaled they’ll join California in fighting federal preemption in court.

AI Crisis Contractor ThroughLine Expands to Combat Extremism

ThroughLine, the New Zealand startup that already redirects at-risk users of ChatGPT, Claude, and Gemini to crisis support services, announced today it’s developing a new tool to combat violent extremism. The tool would combine a chatbot trained to respond to signs of radicalization with referrals to real-world mental health services.

The startup, which operates a network of 1,600 helplines across 180 countries, is in talks with The Christchurch Call — the anti-extremism initiative launched after New Zealand’s 2019 terrorist attacks — for guidance on developing the intervention system. The initiative comes amid growing lawsuits accusing AI companies of failing to prevent violence enabled by their platforms.

Atlassian Cuts 1,600 Jobs to Fund AI Pivot

Atlassian is cutting 10% of its 16,000-person workforce to redirect resources toward AI development and enterprise sales. CEO Mike Cannon-Brookes framed the move as “self-funding” AI investment, noting that AI tools had reduced the need for manual QA testing by approximately 60%. The restructuring will cost $225–$236 million in severance and related charges.

The cuts hit QA roles and mid-level management hardest, with 40% of layoffs in North America, 30% in Australia, and 16% in India. Atlassian also replaced its CTO with two new AI-focused CTOs. The move follows similar AI-driven restructurings at Block, Salesforce, and Oracle earlier this year.