Yes, although I can tell you that the idiocracy is already here for some large companies. I just caught up with a friend the other day who recently started working at a certain unnamed but well-known software company, and he was horrified to discover that the “engineering” department consists almost entirely of people using AI to generate all code and automated tests iteratively until the tests pass, then AI to review the code, then merge and release. No human in the loop, including for all practical intents and purposes the “engineer.” He said what might take a competent engineer 5 lines of code to accomplish and an incompetent one 10 lines is now 60 lines of barely functional garbage.
The result is frequent unreported downtime and an unmaintainable monstrosity that really doesn’t “need” software engineers in the same way you don’t “need” a structural engineer to inspect an addition to a house if it’s never reported to any authority.
I told him this is the future and to stop getting in the way, as this is obviously by design and no one cares. The industry isn’t hiring or training junior engineers, and the escalating demands of productivity from clueless execs are so outrageous that many engineers think this is the only way to survive. I advised him to make as much money as he can and escape in a few years after his kid graduates college, and while he can still look at himself in the mirror.
By the way, this isn’t a screed against generative AI in software. It has its place—if the best models are used by experts. But for companies that don’t use it wisely, which I suspect is most of them, it will just exacerbate the enshittification of everything.
It’s because they put the bean counters in charge. Boeing is a screaming example of why you don’t do that. It’s sad to see that software company move away from innovation and creative problem solving.
Yes, although I can tell you that the idiocracy is already here for some large companies. I just caught up with a friend the other day who recently started working at a certain unnamed but well-known software company, and he was horrified to discover that the “engineering” department consists almost entirely of people using AI to generate all code and automated tests iteratively until the tests pass, then AI to review the code, then merge and release. No human in the loop, including for all practical intents and purposes the “engineer.” He said what might take a competent engineer 5 lines of code to accomplish and an incompetent one 10 lines is now 60 lines of barely functional garbage.
The result is frequent unreported downtime and an unmaintainable monstrosity that really doesn’t “need” software engineers in the same way you don’t “need” a structural engineer to inspect an addition to a house if it’s never reported to any authority.
I told him this is the future and to stop getting in the way, as this is obviously by design and no one cares. The industry isn’t hiring or training junior engineers, and the escalating demands of productivity from clueless execs are so outrageous that many engineers think this is the only way to survive. I advised him to make as much money as he can and escape in a few years after his kid graduates college, and while he can still look at himself in the mirror.
By the way, this isn’t a screed against generative AI in software. It has its place—if the best models are used by experts. But for companies that don’t use it wisely, which I suspect is most of them, it will just exacerbate the enshittification of everything.
It’s because they put the bean counters in charge. Boeing is a screaming example of why you don’t do that. It’s sad to see that software company move away from innovation and creative problem solving.