It'd be neat if you could recall what IBM's programming language is called.
It does seem to be the case that a lot of things can be interpreted multiple different ways, which does require people to precisely state what they want.
I think a classic example if a house-cleaning robot told to keep the house as clean as possible, only for that robot to then try to murder the residents after correctly determining that the residents are a primary source of uncleanliness.
Within the system's own logic, I have to be super-careful to precisely specify exactly what I want something to do, much as someone working in any other system has to be careful to specify exactly what the code's supposed to do in its various use cases. But with user-facing applications, it'll tend to assume some default assumptions. For example, if you hire someone to clean your house and tell them to keep it as clean as possible, obviously they know not to murder people to accomplish that goal, as that intent has been communicated through other channels.
However, this also proves to be beneficial in a lot of cases. For example, if you say, "Go buy eggs from the store", but don't clarify which store, that can be interpreted multiple ways. Still, say that the first store happens to be out of eggs; then, it's good that a valid alternative interpretation exists, as the first attempt at interpretation wasn't viable.
It does seem to be the case that a lot of things can be interpreted multiple different ways, which does require people to precisely state what they want.
I think a classic example if a house-cleaning robot told to keep the house as clean as possible, only for that robot to then try to murder the residents after correctly determining that the residents are a primary source of uncleanliness.
Within the system's own logic, I have to be super-careful to precisely specify exactly what I want something to do, much as someone working in any other system has to be careful to specify exactly what the code's supposed to do in its various use cases. But with user-facing applications, it'll tend to assume some default assumptions. For example, if you hire someone to clean your house and tell them to keep it as clean as possible, obviously they know not to murder people to accomplish that goal, as that intent has been communicated through other channels.
However, this also proves to be beneficial in a lot of cases. For example, if you say, "Go buy eggs from the store", but don't clarify which store, that can be interpreted multiple ways. Still, say that the first store happens to be out of eggs; then, it's good that a valid alternative interpretation exists, as the first attempt at interpretation wasn't viable.