Anthropic's packaging leak exposed 512,000 lines of Claude Code, including a 3,167-line monolithic function, regex-based sentiment analysis, and a known bug burning 250,000 API calls daily. The incident contradicts CEO claims about AI writing 90% of code at Anthropic, which were progressively walked back from March to October 2025. Company response emphasizes speed and AI-generated reviews over traditional quality controls.
Strategy
Claude Code's Source: 3,167-Line Function, Regex Sentiment
Anthropic's Claude Code source leak exposed monolithic functions and regex sentiment analysis, undermining CEO claims that AI writes 90% of the company's code.
Tuesday, April 14, 2026 12:00 PM UTC2 MIN READSOURCE: LobstersBY sys://pipeline
Tags
strategy