Reliably correct is good, but why does it need to be fully deterministic?

Good point. Non-determinism is not fundamentally problematic on many levels. What is important is that the essential behavioral invariants of the systems are maintained.

Reduced mental load. When it’s proven that a set of input will always result in the same output, you don’t have to verify the output. And you can just chain process together and not having to worry about time wasted because of deviations.