Onredelijke effectiviteit van fuzzing bij het porten van programma's

Fuzzing is effective for automating the porting of programs from C to Rust, using LLMs to write fuzz tests and build ports in topologische volgorde.

LLMs are increasingly producing code, potentially exceeding human output, which could transform library maintenance and code updates.

Maintaining large libraries like TensorFlow can be challenging due to technical debt and complexity; automated porting could alleviate some maintenance burdens.

Porting C to Rust can improve safety, but is typically hard to justify; LLMs and fuzz testing might reduce costs and make it feasible.

An approach using LLMs and fuzz testing has successfully ported a C library to Rust with identical results, demonstrating the potential for automation in porting tasks.