AI should write rust and only rust ;) — by Patrick Gray — Seattle Rust User Group, January 2026

Patrick Gray's talk addresses the question of what code AI should write, advocating for Rust due to its safety features akin to seatbelts in self-driving cars. He shares his experience with FilePhoto, a raw file photo service, and his current project, FPExif, an open-source Rust library for EXIF data, developed with AI assistance. Gray details his AI-aided workflow, using tools like Cursor, Zed, and Claude Code, and emphasizes the importance of local development environments and efficient token usage. He also touches on side projects like TempDNS and Vibes.Bingo, highlighting the benefits of Rust's type system and compiler feedback in AI-driven development. The goal is to shift bug detection and validation earlier in the development process, leveraging AI to enhance code quality and efficiency.