The issue is that software is almost unique it the sense that you can use reinforcement learning to get the desired outcome - does it compile or not? A large amount of automated trial-and-error runs do that.
It's not true elsewhere. The definition of "is this a good outcome" is pretty nebulous in other areas.
"Is this what the law says?"
Yes, but that's not how the judge will interpret it, etc.
It's not true elsewhere. The definition of "is this a good outcome" is pretty nebulous in other areas.
"Is this what the law says?"
Yes, but that's not how the judge will interpret it, etc.