Bun Rust 재작성: "코드베이스가 기본적인 miri 검사에 실패하고 safe Rust에서 UB를 허용"

Bun's Rust rewrite, using Anthropic's LLM-assisted code translation from Zig, encounters critical memory safety vulnerabilities that fail basic Miri checks—exposing unsafe code patterns with dangling references and undefined behavior. This raises fundamental questions about whether AI-generated code for systems-level software can meet safety requirements in production environments. For developers relying on Bun, this incident underscores the risks of aggressive AI-assisted rewrites without thorough safety auditing.