> Can we ever trust a LLM the way we trust a regular computer program?

https://users.ece.cmu.edu/~ganger/712.fall02/papers/p761-tho...