IT Jobs Bound for Extinction

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
The IT industry sees rapid change and includes plenty of “hot today, gone tomorrow” positions: this article offers some predictions as to which jobs may become obsolete sooner rather than later. Included are C and C++ coders, who are expected to turn archaic like Cobol programmers, and PHP, WordPress, and LAMP programmers, as interest in such frameworks and languages are seeing decline. One of the major forecasts is that machine learning and AI will replace most coders after 2020.

Right now, the hottest jobs are going to developers, whether they’re front end, back end, mobile, or full stack, says Indeed Prime’s Chiu. But that could change relatively quickly thanks to improvements in artificial intelligence, warns PK Agarwal, regional dean of Northeastern University-Silicon Valley, which offers certificate and degree programs in business, management, and technology for busy professionals. “If I were to look at a crystal ball, I don’t think the world’s going to need as many coders after 2020,” notes Agarwal. “Ninety percent of coding is taking some business specs and translating them into computer logic. That’s really ripe for machine learning and low-end AI.”
 
“Ninety percent of coding is taking some business specs and translating them into computer logic. That’s really ripe for machine learning and low-end AI.”


Dude never dealt with customers who don't know what they want.

A real world example would be "take a spec, send it to offshore in Bangalore, see what you get back".
 
Obviously this person has no real world experience of how this whole game plays out!

Customer thinks shit up
Project Managers/Business Analyst over promises
Dev team tells PM/BA half of their requirements don't make sense or are impossible and/or stupid

Where does the machine "answer back"? What happens when poorly defined requirements are fed in? And ultimately who gets the shit when the customer says "this isn't what I asked for"?
 
I've been hearing C/C++ were going away for 20 years now. The demand has actually increased over that time.

It's always some language, process, or tool which will not only make those languages obsolete, but even the concept of engineers. A language so easy non-programmers can use it! Therefore we need no programmers, AMIRITE!?

Okay. Call us "obsolete" people up when you want a router that works, a PC that boots, or really anything besides a web page.
 
Star Trek computers in 3 years. This is what this guy is selling for whatever reason.

 
I'm sure it will eventually go the way of the dinosaur. Maybe not right away, but it will. As all technology does.

But, turning "We want it to do this, this, that, this, but not this, and maybe add this and that. Yea, and we have to have it secure but open to the whole world. And red but black and translucent. And multiplatform, but Windows exclusive." into a piece of software is an art. One that I could not do. I get requirements for "little" projects and I end up passing the buck. I can do the little projects, but they turn into much larger projects with the demands they have. Little projects are fun. They are out of my job scope, but I do them in my free time. Large projects? I'd have to give up my normal responsibilities to complete. So, I send them to our dev guys. :)
 
I could buy the PHP bit as other frameworks / methods (MEAN, Ruby) gained ground. But the latest releases of PHP (7) may have addressed some of the reasons people were leaving for other languages. The top5 active languages:

1 JavaScript
2 Java
3 Python
4 PHP
5 C#

Source: http://redmonk.com/sogrady/2017/03/17/language-rankings-1-17/

Usually with popular languages, unless their development is dead, they tend to adapt. So either someone comes along and develops an enhancement off of it or it gets a sizeable update.
 
there will always be a place. My company still staffs a few rpg/pascal/someotheroldlang because we still have green screen mainframe run applications.

talking stuff people wrote in the 70s and 80s and they are still writing new applications for.

when those people leave the company they could probably be contractors and make whatever they want.
 
20 years ago, COBOL was predicted to die just as soon as Y2K was in the rear view mirror. Nope, still here. Any tech that does its job well will stay around for a long while. C in its various versions will stay around for a long while, much for the same reason COBOL does. It is well understood, the compiler bugs are mostly worked out and developers have figured out many tricks to get things done. Both will still be in use long after the "Next New Thing" that was going to replace them is long forgotten.
 
This reminds me of what someone said, "IT people are the only ones actively working to put themselves out of business by their creations" or something like that.
 
I've seen the transition over time. I'm now working on the machine learning front and pushing python around. IT people learn and adapt. If you can't, you'd be out of the job years ago.
 
Alarmist bullshit for the most part. The assumption that specifications are so good you could just translate them to machine code are very wrong. I am not contesting that AI or whatever will make some coding obsolete, but it's a small, small fraction.
 
I don't think it will go away, but as the article states, there will be far fewer jobs, and they're mostly maintenance. It goes along with the territory. You know you're the one putting yourself out of a job if you do your job well. And there's a reason why this industry has far fewer older people. They move out of the programming sector after a few years.
 
C/C++ is kind of in a special place all its own. Yeah its old, but its still the one and only language of choice if you need lightning fast software (like high speed trading systems). No other language comes close, and multiple attempts at replacing them (Rust being the most recent) have failed to gain significant traction. Being able to develop software using them requires a level of skill that also commands a high salary.
 
I've been hearing C/C++ were going away for 20 years now. The demand has actually increased over that time.

It's always some language, process, or tool which will not only make those languages obsolete, but even the concept of engineers. A language so easy non-programmers can use it! Therefore we need no programmers, AMIRITE!?

They already made this. In 1981.

http://stackoverflow.com/questions/1293278/what-became-of-the-last-one

A piece of software that takes requirements and generates computer code. You'll notice that it hasn't put the rest of the industry out of jobs, only the shit bog-simple enough to be handled by this.

AFAIK, the company didn't survive the dot-com crash, so that's a pretty good indicator of the future of software expert systems.
 
Dude never dealt with customers who don't know what they want.
A real world example would be "take a spec, send it to offshore in Bangalore, see what you get back".
This is a very sensationalist article for sure. This sounds more like an opinion piece based on lack of experience more than anything.
 
I've been hearing C/C++ were going away for 20 years now. The demand has actually increased over that time.

It's always some language, process, or tool which will not only make those languages obsolete, but even the concept of engineers. A language so easy non-programmers can use it! Therefore we need no programmers, AMIRITE!?

Okay. Call us "obsolete" people up when you want a router that works, a PC that boots, or really anything besides a web page.

Yep. The number of C/C++ programmers isn't likely to increase, but C/C++ won't become obsolete for a VERY, VERY long time.

Essentially, as long as the Linux kernel exists, C programmers ain't going anywhere.
And good luck coding something like uBoot or barebox in something other than C.
Even on Windows, the same can be said for device drivers that have any need of efficiency -- and for an ISR, even C++ isn't particularly useful from an efficiency standpoint if you use anything but the straight C parts of it (and, frankly, quite often I'll still code the most time critical parts of an ISR in inline assembly, especially if it's on something like an 8052 or Atmega part).
 
Last edited:
So many fear threads. Guess its the US mentality of looking at it.
 
I've been hearing C/C++ were going away for 20 years now. The demand has actually increased over that time.

It's always some language, process, or tool which will not only make those languages obsolete, but even the concept of engineers. A language so easy non-programmers can use it! Therefore we need no programmers, AMIRITE!?

Okay. Call us "obsolete" people up when you want a router that works, a PC that boots, or really anything besides a web page.
In my neck of the wood, I interviewed a few years ago for a company hiring C++ devs. Granted, most of the dev stuff in my area is Java or .net. Angular, React, etc are also pretty popular right now.
 
Exactly. The brunt of programming is translating gibberish into specs. Not translating specs into program code.

This... is the most accurate description of my job I've ever heard.

Thank you, sir or madam. My hat is off to you!
 
Right... AI will suddenly be able to write programs and maintain the billions of lines of existing C/C++. I'm not saying this won't happen in the next 100 years, but it certainly is not happening in the next three.

Sounds like the article from last week where we will all be driving EV cars in 6 years.
 
So many fear threads. Guess its the US mentality of looking at it.

Really, it's click bait articles that get us to look at ads.

Same reason why news across the globe is talking about the craziest shit that is happening all the time - and none of the amazing, wonderful things that happen everywhere each and every day. The human mind seems to be addicted to the horrific!
 
This is why it's very important to adopt and change with the way things are going. My company is moving to the cloud direction as well, and being a Server Admin, I see the future of my job not being needed as much. So I am adopting for the new way, and will have to change the way I work and what I do to keep myself employed.
 
the term "AI" is being used a bit lightly here, and AI and ML being full blown replacing devs by 2020, is EXTREMELY laughable.
 
Yep. The number of C/C++ programmers isn't likely to increase, but C/C++ won't become obsolete for a VERY, VERY long time.

Essentially, as long as the Linux kernel exists, C programmers ain't going anywhere.
And good luck coding something like uBoot or barebox in something other than C.
Even on Windows, the same can be said for device drivers that have any need of efficiency -- and for an ISR, even C++ isn't particularly useful from an efficiency standpoint if you use anything but the straight C parts of it (and, frankly, quite often I'll still code the most time critical parts of an ISR in inline assembly, especially if it's on something like an 8052 or Atmega part).

Yep. I'd argue C/C++ (and some ASM, like you say) has seen a huge resurgence lately as efficiency matters again. The desktop market is certainly shifting a bit to more high level languages and web apps, but there are more embedded devices than ever these days. Those are running on very power constrained platforms, so you need to really tweak.

People want that low power router to work at gigabit speed. They want that new vid card to be faster than the competition. They want to see low DPC numbers (Hi Kyle!) in benchmarks. These are the tools we use to make that happen.
 
There are jobs I would predict the end to and programming would be the last of them. This person is an idiot.
 
So, REAL programmers will be retired, and development systems will instead require the ability to assemble software using Wood Blocks, Skittles and some sort of interlocking plastic shit. On screen, of course.

The code will be very small and efficient. Say, 10k for a complete word processor. But the necessary Runtime will be 9 Terabytes.

Got it :)

Anyone remember the Word Processor Write-Now for Macintosh? It was written in 100% Assembler and it was fucking BRILLIANT. It scrolled so fast they had to put timed delays in it LOL

This is what we need. Small, tight, efficient code. Sadly, the talent base for this is all but gone.
 
Let the machines code themselves, WHAT COULD POSSIBLY GO WRONG...
Kind've already happens.

Very few programmers actually do all the work today anyways. Compilers do much of the heavy lifting now and have increasingly done so for decades.

No that isn't the same as having a strong PI or weak AI do it all but it is an example of how automation has been creeping into the workforce in strangely unmentioned (at least publicly in most media sources) ways and how it might impact the blue/white collar desk jobs just as much if not more so than it will the physical labor stuff in the near future. And it illustrates quite clearly how a STEM degree, even in a "hot" field, will not save you from being negatively impacted by widespread automation in the work force.
 
Yep. The number of C/C++ programmers isn't likely to increase, but C/C++ won't become obsolete for a VERY, VERY long time.

Essentially, as long as the Linux kernel exists, C programmers ain't going anywhere.
And good luck coding something like uBoot or barebox in something other than C.
Even on Windows, the same can be said for device drivers that have any need of efficiency -- and for an ISR, even C++ isn't particularly useful from an efficiency standpoint if you use anything but the straight C parts of it (and, frankly, quite often I'll still code the most time critical parts of an ISR in inline assembly, especially if it's on something like an 8052 or Atmega part).
As long as you need a compiler for the next language made in something other than Machine language, C/C++ will be there first.
 
My job is safe. I'm lowly help desk support. People need another human to tell them to "reboot and call back if it doesn't fix the problem".
 
Kind've already happens.

Very few programmers actually do all the work today anyways. Compilers do much of the heavy lifting now and have increasingly done so for decades.

No that isn't the same as having a strong PI or weak AI do it all but it is an example of how automation has been creeping into the workforce in strangely unmentioned (at least publicly in most media sources) ways and how it might impact the blue/white collar desk jobs just as much if not more so than it will the physical labor stuff in the near future. And it illustrates quite clearly how a STEM degree, even in a "hot" field, will not save you from being negatively impacted by widespread automation in the work force.

It's not unmentioned. Tools becoming more user-friendly and interactive already had their impact many decades back.

From the complete removal of assistants and he mostly-removed secretary, that all happened thanks to Electronic Organization software, Office and Compilers/IDEs. This mostly happened in the 70s-80s.

But the offshoot was a huge demand for more experienced jobs, as people sought to create ever-more-complex software. You couldn't hope to have done that without all these new tools. And I'm not just talking about big desktop software - the web is a massive stack of software enabled by these tools, and embedded devices shot-off like a rocket because of the ease of creating them (standard BSPs with full IDE support, standard operating systems, both of which make embedded programming way easier).

I don't see AI removing this "demand for the next great piece of software" anytime soon. The small stuff will be automated, but if people keep reaching further, then there will never be a ceiling.

How can AI catch-up if the target is constantly moving? It conquered Chess and Go because those are fixed games. It conquered things like Legal Discovery because those haven't really changed for many years, and are really fucking repetitive. But I don't see the same pattern in Software, where we still have a big blue sky.
 
Last edited:
Back
Top