How is this an issue specifically with Rust and Haskell? Do you find that LLMs have an easier time proving global correctness with C, Python, or Typescript?
Yes, because those other languages all have much weaker type systems.
Yes, because those other languages all have much weaker type systems.