💻 Zig language creator Andrew Kelley has criticized the Bun project for its transition from Zig to Rust, which was carried out using Claude AI agents.
Rewriting over 1 million lines of code took only 11 days and cost approximately $165,000 via API. Despite passing 100% of the tests, Kelley called the result "unreviewed slop."
🌍 This case demonstrates the extreme development speed achievable by AI agents but raises questions regarding reliability and the necessity of human code review.
👤 While AI is capable of large-scale refactoring, automated tests do not guarantee the absence of architectural flaws.