What if the nondeterministic system gets it subtly wrong tho, like inputs a comma instead of a decimal somewhere, OCR reads an l instead of a 1; who takes the liability? And how can you ensure it doesn't have an issue like that without the same amount of effort as doing it yourself