When I was exploring career options in high school in the late 90s, I regrettably didn't pursue computer science because I listened to these arguments:
"you have to like/be good at math" (not so much)
"all of those jobs will be outsourced to India soon" (nope)
"programming will disappear/be automated some day"
20 years later, argument #3 is still being made without noteworthy progress.
After I stumbled into this line of work at a bank, I encountered lots of low code/no code tools. Many of them even advertise "you don't need a programmer to make apps!" That sounds pretty good to execs & managers who love paying for software more than paying developer salaries; they'll often pick up a license for some project or get swayed on a sales call. In every instance I saw, a programmer was eventually assigned to the thing.
Why? My impressions:
Specifying what software should do is a fundamentally creative act - the same results won't be achieved by reducing it to some rote activity - and not everyone wants to do that kind of work, the debugging/refinement process in particular.
The PWBIs, to use your term ;), are never happy when some app-generator tool has limitations that conflict with their vision. In code, you can do just about anything.
Communicating a spec (whether to a person or machine) is a skill not everyone has (or wants to improve, seemingly). I've worked for people who requested a feature they couldn't have, or actually needed its opposite, and it was only my relationship with them as an individual that enabled me to counter the request.
As machine learning improves, I think our tools will get smarter. Maybe the process will change. What we consider an app will likely be different (Alexa skills!). I expect the job itself will stick around as long as the need for software customization exists.
In case we do get real AI, not just sophisticated pattern matching? There's always politics.
When I was exploring career options in high school in the late 90s, I regrettably didn't pursue computer science because I listened to these arguments:
20 years later, argument #3 is still being made without noteworthy progress.
After I stumbled into this line of work at a bank, I encountered lots of low code/no code tools. Many of them even advertise "you don't need a programmer to make apps!" That sounds pretty good to execs & managers who love paying for software more than paying developer salaries; they'll often pick up a license for some project or get swayed on a sales call. In every instance I saw, a programmer was eventually assigned to the thing.
Why? My impressions:
As machine learning improves, I think our tools will get smarter. Maybe the process will change. What we consider an app will likely be different (Alexa skills!). I expect the job itself will stick around as long as the need for software customization exists.
In case we do get real AI, not just sophisticated pattern matching? There's always politics.
This is an amazing response.