This is a detailed technical critique of Anthropic/Bun's decision to rewrite a TypeScript runtime from Zig to Rust, framed as pushback against what the author sees as marketing theater disguising engineering shortcuts. The core claim: Bun's memory bugs stemmed from poor engineering practices (heavy reliance on AI agents for code generation/review, weak style guide enforcement, management dysfunction), not language limitations—and a Rust rewrite was chosen for optics rather than technical necessity. The author argues Bun never seriously tried alternatives like TigerBeetle's proven memory-safety approach in Zig, and that the project's embrace of AI-driven development actually undercuts the quality standards it now uses as justification. Actionable takeaway for practitioners: when evaluating major rewrites, demand explicit tradeoff analysis and consideration of alternatives before accepting the conclusion; watch for post-hoc narratives that conveniently serve business positioning; and be skeptical of claims that "AI is enough" when the same company simultaneously implements safety features (Rust's borrow checker, code review structures) that implicitly contradict that premise.