A reasonable answer is that this is our future one way or another: the complexity of programs is exceeding the ability of humans to properly manage them, and cy

A reasonable answer is that this is our future one way or another: the complexity of programs is exceeding the ability of humans to properly manage them, and cybernetic augmentation of the process is the way forward. i.e. there would be a lot of value if an AI could maintain a detailed understanding of say, the Linux kernel code base, when someone is writing a driver and actively prompt about possible misuses, bugs or implementation misunderstandings.