Raw LLM Responses

It's doubtful AI will ever replace programmers. I say this not because I think humans are special, but because programming requires specificity, which is driven by intentionality - we write code and design applications to do things we want to do, which are things that generally do not already exist. To do this, we use programming languages, which give us simplifications of operations we want to execute on a processor. This abstraction, alone, limits what we are able to do and our control over how it gets done; we let the compiler substitute tons of assembly for the few lines we wrote, which may or may not represent what we wanted to do (we don't have control over exactly how the program does what it does if we aren't writing the assembly, ourselves). If we expand on this abstraction, say to a "low" or "no-code" type of language, we surrender more control over what we are producing because we're using less "words" to describe how things should be done. If you ask AI to write you a program to do something, at best, the functionality of what it generates is limited by how well you describe what you want the code to do; else, what is the AI generating? You could spend hours describing exactly how the program should function and what specific details you need built in, but as you approach more specificity with your language, you approach the same complexity you would encounter if you had just wrote the code, yourself. Practically, you may think it doesn't matter, because AI can write you something that's maybe 80% of what you need or maybe you can't code and it's already helping you achieve something you couldn't do, but in the real and professional world, where an application has to do something complex and novel, with efficiency, accuracy, and reliability, there's no getting around the work required to describe that, be it through code or natural language.