Today in AI — April 1, 2026

📖 2 min read

Anthropic Accidentally Leaks Claude Code Source via npm

In what’s being called “the mothership of all code leaks,” Anthropic inadvertently exposed the entire source code of Claude Code — their flagship agentic CLI tool — through a source map file shipped in npm package version 2.1.88. Security researcher Chaofan Shou spotted the leak on March 31, and the internet had a field day.

The exposed TypeScript source reportedly spans nearly 2,000 files and reveals internal tooling, prompt engineering patterns, and system architecture. Worse, VentureBeat reports that a malicious axios dependency (versions 1.14.1 and 0.30.4) containing a Remote Access Trojan briefly made it into the supply chain between 00:21 and 03:29 UTC on March 31. Anthropic has acknowledged the issue and patched the package. If you run AI coding tools, this is your reminder to audit your dependencies.

Yupp AI Shuts Down Less Than a Year After $33M a16z Raise

AI feedback startup Yupp has closed its doors, barely a year after raising $33 million from a16z crypto’s Chris Dixon. The company, which aimed to build decentralized AI feedback mechanisms, couldn’t find product-market fit fast enough to justify its burn rate.

The shutdown is a sobering reminder that even well-funded AI ventures aren’t immune to the basics: you still need users who pay. Expect more consolidation in the AI startup space as investors tighten their criteria heading into Q2 2026. For ongoing coverage of which AI tools are worth your attention, check out reviews at AiToolCrush.com.

States Drop AI Regulation Bills After Trump Administration Threats

Louisiana has scrapped roughly a third of its nearly 20 proposed AI regulation bills after the Trump administration threatened to pull federal broadband funding from states that impose guardrails on AI development. Florida and Utah have similarly backed off their own proposals.

This follows the White House’s March 20 release of a National AI Legislative Framework calling for federal preemption — essentially arguing that states shouldn’t regulate AI development or impose developer liability for third-party conduct. The framework carves out exceptions for child protection, fraud prevention, and state procurement, but the message to state legislatures is clear: stand down or lose funding.

Fujitsu and Rapidus Team Up on Japan’s First 1.4nm AI Chip

Japan is making a serious play in the AI silicon race. Fujitsu announced plans to develop a dedicated AI inference chip using an advanced 1.4nm process, manufactured entirely domestically by Rapidus. Japan’s NEDO is expected to fund roughly two-thirds of the estimated ¥58 billion (~$363M) project.

The NPU (Neural Processing Unit) chip targets AI inference workloads in servers and is part of Japan’s broader strategy to reduce dependence on foreign semiconductor supply chains. If Rapidus can deliver on its ambitious 1.4nm manufacturing timeline, this could reshape the competitive landscape for AI hardware beyond NVIDIA and AMD. For AI market analysis and predictions, visit BetOnAI.net.