Are Students Suddenly Losing Interest in Computer Science as AI Coding Takes Off? (nerds.xyz) 149
On academic mentoring platform Nova Learning, Computer Science and AI, "once the dominant choice among students... is losing ground fast, while Engineering has nearly doubled its share." It's a small survey, but "The steepest proportional decline is in Software & Data Science, the most traditional 'learn to code' pathway, which lost 44% of its 2025 share. AI saw the largest absolute drop of any single subject in the survey, down 6.2 points.
"The category did not decline uniformly. Students moved away fastest from the pathway most associated with entry-level software work, while the more applied and human-facing subfields held their ground better."
Slashdot reader BrianFagioli writes: Nova Learning says the share of surveyed middle and high school students naming Computer Science and AI as their primary academic interest fell from 42.3 percent in 2025 to 27.9 percent in 2026, while Engineering rose from 11.9 percent to 23 percent. The biggest gains came from Mechanical and Aerospace Engineering...
[B]roader enrollment data points in the same direction. The National Student Clearinghouse Research Center reported declines in Computer and Information Science enrollment at four year institutions, even as Engineering grew. AI coding tools are not proven to be the cause, but as software development changes and AI handles more coding tasks, students may be starting to rethink what a future in technology should look like. [Undergraduate enrollment in CS programs at four-year institutions fell 8.1%, to approximately 606,000 students.]
"The category did not decline uniformly. Students moved away fastest from the pathway most associated with entry-level software work, while the more applied and human-facing subfields held their ground better."
Slashdot reader BrianFagioli writes: Nova Learning says the share of surveyed middle and high school students naming Computer Science and AI as their primary academic interest fell from 42.3 percent in 2025 to 27.9 percent in 2026, while Engineering rose from 11.9 percent to 23 percent. The biggest gains came from Mechanical and Aerospace Engineering...
[B]roader enrollment data points in the same direction. The National Student Clearinghouse Research Center reported declines in Computer and Information Science enrollment at four year institutions, even as Engineering grew. AI coding tools are not proven to be the cause, but as software development changes and AI handles more coding tasks, students may be starting to rethink what a future in technology should look like. [Undergraduate enrollment in CS programs at four-year institutions fell 8.1%, to approximately 606,000 students.]
Good riddance (Score:5, Insightful)
Re: Good riddance (Score:5, Funny)
Yeah, but we still have to figure out what to do with all those students we graduated without sufficient talent.
Re: (Score:2)
Yeah, but we still have to figure out what to do with all those students we graduated without sufficient talent.
Mommy and Daddy are supporting them. (I wish I was kidding)
Re: (Score:2)
Yeah, but we still have to figure out what to do with all those students we graduated without sufficient talent.
The arrogance of programmers really is astonishing. This is not an attack on the original poster, just in general. Take this with a grain of salt, I am venting here.
First of all, people's talents do not fit in a boolean. You are not either a great programmer or a loser. Try to use at least 2 to 3 bits. Then there are several dimenstions in programming you can be good at or suck at. One is people skills by the way, not the most important one, but oh boy it is needed. Keep it simple then 4 dimensions, 2 bit
Re: Good riddance (Score:5, Insightful)
So heres the thing. I've had a bunch of my work as the senior dev move from "Work on coding challenges with the younger devs and help them code high quality reliable software" to "code review this AI slop and tell me why it doesnt work" , and its fucking miserable work.
Theres a lot more of it, and its batshit code for the most part. Just illogical choices and no craftmanship (Though I'll give claude one thing;- It comments its code, something I've been yelling at junior coders to do for the better part of my career. Lets the greybeard know what the hell they are trying to achieve so I can look at their code and fix it so it actually achieves it.
Re: (Score:3, Insightful)
Re: Good riddance (Score:5, Interesting)
And even on solo projects it doesn't hurt. Sure, I can keep track of a few thousand lines of code (and to be clear I'm not paid by the line) in my head on a solo project. Above that - or throw one junior engineer in the mix as a "helper" - and all bets are off. Document, document, document, and then stick to the plan.
Re: (Score:2)
Exactly. That's all I've been doing lately. New Project, Start feeding Claude Code the notes. Build up the memory. Spec time comes, use him as a rubber duck, walking thru the whole process, we document everything first. I get a MD file, convert to a docx for the users, and submit. It's a good starting point, my ideas + boiler plate stuff from Claude. I could just give it to him to code out as well at this point, but I still need to go thru a review, and then there are contractors that need
Re: Good riddance (Score:4)
Re: (Score:2)
Don't bother reviewing it. Instead, work on tests so that you can be sure it does what you want (and doesn't do things you don't). Then, get it to spew out whatever crazed crap it wants, test it, if it works, ship it. If not, the agent will automatically go and fix bugs until it works.
This all falls to pieces on the "doesn't do things you don't [want]" step though. I don't want my app to leak data between users, but I also don't want it to leak data to anonymous users, or to users with an expired token, or
Re: Good riddance (Score:5, Interesting)
When I use AI to generate my code, I generally do it in small steps. I do all the design work myself and just break out the implementation for the AI, and I never give it too much work at once, for the very reasons you just complained about.
Its a bit more of an issue when I must code review work done by junior devs who literally did just try to tell the AI to do it all for them. That experience has been similar to what you are describing, for me at least. I am unsure if these devs are actually getting better at coding, or if we are moving towards a peculiar future in which the generation of senior devs who actually understand code well enough to make effective use of AI have all retired, leaving behind a generation of people who can only feed requirements to an AI and then flail when it doesn't work.
Re: Good riddance (Score:5, Insightful)
Yes. That Agile approach has really never worked. Competent people usually hate it and incompetent ones just produce trash a bit faster, which is not a win.
The fact of the matter is that when you produce stuff more complex than simple web interfaces or business logic, you cannot accommodate changing specifications during the project. It will mess up everything. And, funny thing, that has been known for a long, long time. It is just that we have far too many incompetents in the industry.
Re: (Score:2)
Yes. That Agile approach has really never worked. Competent people usually hate it ...
On a team of experts it can work. On one such team we also treated the "rules" as informal suggestions. So it was more of an iterative process with a list of things to deliver in two weeks. Skipping anything we considered performative silliness. Admittedly, this only worked due to the individual expertise and reliability.
The fact of the matter is that when you produce stuff more complex than simple web interfaces or business logic, you cannot accommodate changing specifications during the project. It will mess up everything.
I'm thinking of a project where the AI assistance is largely confined to certain modules. And that we are free to change the documentation for those modules and have the AI generate a new i
Re: (Score:3)
Yes. That Agile approach has really never worked. Competent people usually hate it ...
On a team of experts it can work. On one such team we also treated the "rules" as informal suggestions.
To be fair, a team of competent experts can work with any stupid management "methods" as long as they can basically ignore them. I have seen that with a pretty good team being forced to adopt Agile. They just made the "sprints" 6 months long and proceeded to use what they knew worked.
Re: (Score:3, Interesting)
Re: (Score:3)
On a team of experts it can work. [...] Admittedly, this only worked due to the individual expertise and reliability.
Exactly: there is no substitute for skill. You're absolutely right saying that's the only reason it worked.
Now with AI, the required skills change: we'd somehow have to teach students to learn what we learned though decades of grunt work and... choices that were good enough at the time, but now inadequate (requirements change). Learning that as a society won't be pretty...
I'm thinking of a project where the AI assistance is largely confined to certain modules. [...] And that we are free to [...] have the AI generate a new implementation from scratch if necessary.
Please go talk to a resident greybeard: this kind of software design already exists for years, and a kindred spirit will help you establi
Re: (Score:2)
Re: Good riddance (Score:5, Interesting)
Yes. That Agile approach has really never worked.
Kind of depends if you mean Agile(tm) or agile-small-a.
This isn't really the generic defense of "if your scrum isn't working, you're doing it wrong/it's not scrum or a true scotsman". After all, scrum is a bunch of large sweaty men (or women) all pushing against each other scrabbling after a moving target and then often just collapses in a mess. Kind of like Agile-scrum so maybe it's well named...
But really agile is an attempt to capture how small teams work on smaller projects with good people on large projects with mediocre people. Naturally it never works because (a) the teams are big and (b) the people are mediocre. It's meant to also offset some of the worst excesses of insane project management by doing things differently but unfortunately if you have an insane project manager who ran things in a crazy and unproductive way they are not going to become not crazy with a new framework of rules.
So in conclusion: agile is great if you can do it, but you can't do it for exactly the reasons why you need it in the first place. It is I believe therefore doomed by its very definition.
Re: Good riddance (Score:2)
It works fine when you actually have a skilled team. It essentially codifies how we used to work in the late 90s/early 2000s, before the web was hot and we were inundated by kids trying to make money.
Re: (Score:3)
It works fine when you actually have a skilled team. It essentially codifies how we used to work in the late 90s/early 2000s, before the web was hot and we were inundated by kids trying to make money.
It really depends. If you are working on a system that can get someone killed or financially ruin a large institution (or sometimes both), the precepts of the agile manifesto kinda fall apart. You don't want to prioritize speed, shipping, running code, and loose conventions over rigorous requirements, thoughtful design, and risk management when you are writing a clinical, financial, or on-mission military software system.
Re: Good riddance (Score:2)
That's true, but only describes a very, very tiny part of software development, even back in the 90s. Most software was being developed to support day-to-day business operations or to market shit online.
Re: (Score:2)
That "tiny" part has stopped bein tiny a while ago. The likelihood of that not working when it gets slapped together is growing all the time and you may want to look up "cost of cyber attacks" as a reference.
Re: (Score:2)
You don't want to prioritize speed, shipping, running code, and loose conventions over rigorous requirements, thoughtful design, and risk management when you are writing a clinical, financial, or on-mission military software system.
That is a silly chain of thoughts.
Shipping running code is obviously always a top priority. Non shipping is bad, not running is bad/worse.
and loose conventions No idea what that is supposed to mean.
rigorous requirements requirements in agile projects underly the same rigor than
Re: (Score:2)
Shipping running code is obviously always a top priority. Non shipping is bad, not running is bad/worse.
This is the kinda thinking that crashed two space shuttles. There are worse things than not shipping, even if it means the project gets scrubbed. Sometimes safety or durability has a higher priority than shipping. An older less efficient system that works is more valuable than a new one that is unreliable. A new capability isn't worth it if it gets you killed.
and loose conventions No idea what that is supposed to mean.
What I mean is that the following is from agile's founding document:
Individuals and interactions over processes and tools
Working software over com
Re: (Score:2)
This is the kinda thinking that crashed two space shuttles.
Nope it does not. It should be obvious that we are talking about working code that works perfectly fine - otherwise it would not be shipped.
Sometimes safety or durability has a higher priority than shipping.
As you only ship perfectly fine code: it is the same.
You can't freeze a design if your requirements are subject to customer change at the end of every 2 week sprint.
The point is: the fucking customer, IS CHANGING HIS REQUIREMENTS. So, your des
Re: (Score:3)
There is a real and large market for buggy software. Nobody likes the bugs, of course, but everybody loves for the software to be cheap. Many business ventures are deliriously happy to use cheap buggy software that sometimes creates problems for their team, so long as most of the time it makes a lot of money.
For such markets, Agile is great. Developers very quickly churn out buggy software which can be sold for cheap precisely because it was churned out quickly. The magic phrase is "good enough." It's
Re: (Score:2)
Yeah, but we still have to figure out what to do with all those students we graduated without sufficient talent.
AI coding assistants require a shift away from an incremental iterative agile approach, a move back towards having some documentation to feed to AI assistants to keep them on track. Too much of an agile approach will too often lead to something resembling "vibe coding slop." The less talented can work on the documentation that help reign in the AI assistants. The AI coding assistants can then be used where appropriate and with greater review by the more talented. Reducing the "slop."
You are on to something here, and I like it. Here's the interesting thing. You've proposed a solution. And it makes sense, and is doable. It offers a path for the less talented or newer staff. Most in here just complain without actually proposing anything.
Getting back to an organic level of CS majors (Score:5, Insightful)
Serious students of computer science only need apply.
In other words we are returning to "normal", "organic", CS student levels. There have always been two types of CS students.
(1) Those with a genuine interest and curiosity about computers. Those who studied computers in class and on their own for fun.
(2) Those who were told by someone that CS is a good career path. Those who did the minimum to get their "ticket" punched.
(2) was greatly inflated by various events, the internet going commercial, the learn to code movement, etc. If AI coding reduces their numbers that is not a great loss.
(1) will see AI coding as just another tool. Sometimes appropriate, sometimes not, and will use it judiciously and properly review/rework the resulting code. Just as they would the code of the (2) intern.
Re: (Score:2)
Just like any other engineering field, really, only that the established ones have gotten pretty good at weeding out type (2). CS/IT will need to get there as well or software will remain crap.
Re: (Score:2)
Well, yes. A long time ago. When I studied CS, they had us take the Calculus and Linear Algebra for Mathematicians (same lecture, same exam). That did weed out about 50%. But it did nothing for CS/IT specific skills. Turns out mathematics is actually not that close to them. Friend of mine failed CS and then went on to get a Math Master's without any problems.
Re: (Score:2)
In Germany they also tries to use math to weed out students.
I passed a math test where ~96% failed. From roughly 500 students 23 or so passed the exam.
And to make it more funny, they sorted the people passing by grade, and not by their student ID. So I did not find myself so easy.
We do not know how many good programmers/computer scientists got lost that way.
Some of them however went to college (Fachhochschule) instead of university after they got kicked out, and made it there.
Re: (Score:2)
What if it's the opposite? (Score:2)
Yeah if you're one of the absolute top guys who can help Peter thiel spy on everybody sure there is still work for you.
There is a lot less work for somebody who can maybe score their masters but that's as far as they're going to go and they could be building a better tomorrow for you but they're not because they've got to put food on the table first.
How much are we losing out on because of ho
School Sucks (Score:4, Interesting)
Maybe... just maybe, it started before "AI" - maybe, just maybe, it is become what is taught as "computer science" and what they actual day job are absolutely nothing like each other. CS degrees force rote memorization of established algorithms. That's never the day job. 30 years of computer programming, and I've implemented more Mandelbrot renderers than I have inverted b-trees, or sorted link-lists. Schooling and industry are the furthest apart imaginable, and the younger generations have started to pick up on the fact that leetcode skills don't mean jack shit in the real world.
Re:School Sucks (Score:5, Insightful)
Not to defend computer science coursework, but the real world sucks too. Modern applications are an abomination, those "leetcode skills" are focused on gluing together a mile high stack of shit. AI can do that too.
But yeah, disillusionment with coding started before AI, for me it started with object-oriented programming. Code reuse makes sure everyone's work sucks with time.
Re: (Score:3, Insightful)
...for me it started with object-oriented programming. Code reuse makes sure everyone's work sucks with time.
You demonstrated in a sentence and a half that you don't have the slightest clue how object-oriented programming works.
Re:School Sucks (Score:4, Insightful)
He is right that OO never really delivered on its many promises. I regard its as mostly a failure these days and my perception is that doing OO often only serves to increase project cost. "OO for everything" is one of these bad ideas that refuse to die in low-insight environments. That said, OO has its place but it is not everywhere.
Re: (Score:3)
Good OO works, and works well, mostly because of the isolation of internal code.
But most OO code I see is procedural programming with classes.
Re: School Sucks (Score:4, Insightful)
Good procedural works, and works well, mostly because of the proper architecting of the problem space. OO isn't bringing anything special to the table here. Good devs write good code. Bad devs write bad code. And when management decides what language stack to use, it is primarily about who the cheapest developers are, and not about how effective the stack is.
Re: (Score:2)
Isolation is not actually an OO thing, because that isolation is fake. All you actually get is weak namespace separation and that you can get with other techniques as well.
Re: (Score:2)
uh dude (Score:2, Informative)
That probably means you aren't doing very sophisticated work. There are tons of jobs out there where implementing new algorithms is what gives your product an edge in the marketplace. Even something as boring as backup software is going to have a ton of proprietary algorithms which are based on ... you guessed it ... the stuff you learned in school.
I frankly don't see how it's possible that you started your job in the mid-1990s and didn't write algorithms. That was in the days before programming was slappin
Re: uh dude (Score:2)
Re: uh dude (Score:3)
AI makes the bolting so easy, many devs don't consider whether more bolts is going to help the product, or whether adding bolts to the bottom of a water tank is causing leaks.
Re: (Score:2)
CS degrees force rote memorization of established algorithms.
What kind of utterly crappy CS program did you attend?
Re: (Score:3)
You do not "rot memorize" algorithms.
Either you comprehend them, then you can reiterate them, recite them.
Or you do not comprehend them, then you have nothing memorized.
Re: ?? Memorization of established algorithms ?? (Score:2)
Yeah, honestly I've never encountered an American institution of student that led me to believe CS was delivering rote memorization.
Maybe this guy just sucks at coder interviews and thinks doing them requires memorizing algos from college.
They're not losing interest. (Score:5, Insightful)
They're not losing interest, they're being TOLD not to go into coding because AI coding is taking off.
Re: (Score:3)
For good reason. I know people who, with zero coding knowledge .. like they can't print "hello world" let alone tell you what a for loop is, 100% vibe coded fully functional (has all the features they wanted it to have -- including a backend API on GCP/AWS) apps/games that are on the App Store (they aren't selling well, but that's due to marketing .. the features are fully implemented).
All they had is a little determination, a week or two to focus on it, and a few basics on prompting such as asking the AI t
Re: (Score:2)
This is what people said about outsourcing in the 2000s. Turns out you have to actually have organizational knowledge about what your code actually does.
And that's not even getting into the legal aspect of what is gonna happen when the AI does something like making a car's computer cheat on emissions tests. At the end of the day, enterprises will need a real person to blame for this stuff.
Re: (Score:2)
How does outsourcing reduce the need for a developer? The developer is still there, his paycheck is just being signed by someone else. How many people like that are they going to need, and will business analysts be qualified to do some of the work? Do they really need a CS grad? The number of coders, people who understand complex code and software, needed will reduce by a lot --- and it will accelerate as management starts to realize this. The effect is already happening. There are 20% less employed softwar
Re: (Score:2)
AI enthusiasts anthropomorphize coding agents, so they are just another developer that is significantly cheaper. They still have to pay something, its just by the token instead of by the hour.
As far as that graph goes, that's mainly because of high interest rates putting pressure on zombie companies.
Don't get me wrong, I don't think doubling the number of developers every 10 years like we have since Alan Turing is sustainable. And yea, once management discovers they don't need a real person to vomit 1000
Re:They're not losing interest. (Score:5, Interesting)
Outsourcing is actually an excellent comparison. Having had the misfortune of reviewing code from supposedly "quality" outsourcing several times about 15 years back, I now recognize this as an early form of "slop". Code produced with no insight, exactly matching the spec and the spec not being very good (which is normal).
The legal angle for LLM "code" is going to be even funnier. First, it does not have your copyright, as LLMs results do not qualify for copyright. But second, it may have somebody else's copyright, if the LLM produced training data too directly! One of the reasons the GCC project has prohibited LLM code completely and the Linux Kernel requires it to be clearly labelled (so they can rip it out later if needed).
When legal leverage finds the infinity stone. (Score:2)
..what is gonna happen when the AI does something like making a car's computer cheat on diesel emissions tests. At the end of the day, enterprises like VW will need a real scapegoat to blame for this stuff..
In case you needed a more realistic example of exactly how enterprises plan on not worrying too much about a "real person"..
Re: They're not losing interest. (Score:2)
Those apps aren't selling because they are schlock, not because of marketing. No one has ever developed a successful app using vibe coding, unless it's basically just a human management app (think Task Rabbit), where humans actually do all the work.
Re: (Score:2)
Unless you give sources, I'm going to assume you are stating your opinions rather than facts. ISTM that I've heard of a few "vibe coded" successes, though I'll admit I wasn't interested enough to either document or remember them. And lack of marketing is a sufficient explanation.
Re: They're not losing interest. (Score:3)
The null hypothesis for any software is that it's crap. The burden of proof lies with the guy claiming people are vibe coding good games that would be big hits if only they had a marketing team.
Re: (Score:2)
For any particular one, yes. For the claim for an entire category... that needs proof.
Re: (Score:3)
Re: (Score:2)
And in 20 years when those of us who are 40-50 are retired or dead and there is no pipeline of fresh talent that has spent 20 years becoming experts to guide the AI, then I guess society just collapses. Good job everyone!
Re: (Score:2)
Probably. On the other hand, any serious software project is like 10-20% initial coding and the rest is what makes it a success or failure. Well, with LLM use (and LLM supported attackers) the coding itself will probably also become a serious source of failures now. It is still hilarious how incapable LLMs are at producing code. Not that many supposedly professional coders are any better.
Re: (Score:3)
They're not losing interest, they're being TOLD not to go into coding because AI coding is taking off.
Employers are losing interest paying people a viable salary for the job. That's the reality of it.
If shoveling shit was suddenly paying $200K/year, you'd have PhD-accredited shovel pimps identified as college campus experts in Piled Higher and Deeper slinging undergrad degrees in it. The Master Bullshit Artist (MBA) would be a specialized field of study.
Nope, its not that (Score:5, Insightful)
The dabblers are wandering off. (Score:2)
Is this necessarily a bad thing? For quite some time, coding has been pushed as a way to make yourself marketable. For a lot of people who have no particular passion for it—nor for anything else, which is why they're dabbling in everything—but were only in it because they thought it was a good career move, it is smart to look for greener pastures now. The real question is, what adjacent fields are easy to slide into, but AREN'T equally vulnerable to the unblinking gaze of AI once there is the cr
Re: (Score:3, Informative)
Re: (Score:2)
Indeed. Sure long-term, some jobs will get reduced in numbers, but the only real area where that will happen is low-skill paper pushers. Hallucinations and slop will kill it for the other uses and these two cannot be fixed for LLMs.
Re: (Score:2)
Not at all. Getting rid of the semi-skilled is how an engineering discipline eventually grows up. The second way LLMs help with that is that incompetent attackers can not get to the level of a somewhat mediocre expert using LLMs. That is going to drastically increase the number of attackers and put a lot of (very desirable) additional pressure on makers of software. Of course, "slop" code has basically no chance of survival.
Hence while things are unpleasant at the moment, I think that longer-term this is a
Re: (Score:2)
Utter nonsense.
Many people use local LLMs which cost cents per hour, and will continue to do so.
Try to avoid talking about stuff you do not understand and if you cannot, at least look up things. You can, still, use an LLM AI to help with that.
Re: (Score:2)
Well, if you think running a local toy and doing your own training for that is good, sure, go for it. And good luck.
Re: (Score:2)
I have a local LLM working as a coding assistant. I don't do any training, I just use it.
You clearly have no idea what you are talking about. Time to stop talking.
Re: (Score:2)
Re: (Score:2)
You made a stupid remark and I pointed out how stupid it was. That is very much a sequitur.
"I'm questioning how much use that is, really, in the wider context of replacing programmers."
It was a stupid remark. You cannot try and justify it by saying you mean it in an entirely different context.
Up your game.
Re: You liked software better in the autistic 80s? (Score:2)
The 80s just had different computers and tools (Score:5, Interesting)
Counterpoint, Until the late 90s, programming was the domain of the autistic and severely mathematically inclined. In most areas, the pay sucked, and their output sucked even more. It was a hostile environment everywhere but in the bay area and Seattle...maybe a few others.
That's a long string of very bad guesswork. The top performers tended to be the more "math inclined" but plenty of people with natural talent and liberal arts degrees functioned as decent software developers. Pay and conditions pretty good for all.
My (much) older sister was a COBOL programmer...her life sucked.
Well she wasn't Computer Science, she was Computer Information Systems. Computer Engineering was also a thing. Now to be fair the names were used a little loosely, but in general a "Computer Science" degree was programming as envisioned by the School of Science at a University. "Computer Engineering" as envisioned by the School of Engineering. It differed by having more of a hardware focus, maybe trading off some math theory. "Computer Information Systems" programming as envisioned by the School of Business, focusing on corporate and government software development. COBOL confirming the latter.
Would you rather have the software of today or the software of 1995? Let's be real, most programmers before the internet couldn't find good work outside of working with programmers. They lacked the soft skills to work anywhere, but in front of a computer, FAR away from customers.
Un, no. Speaking as a programmer who started as a curious hobbyist in the early 1980s. Many of us could quite safely talk with customers, even customer not part of some other company's engineering department. We could even talk with members of the opposite sex who had no interest in computers. I agree that there are exceptions, but most I have run across in the decades are perfectly capable of joining a group of coworkers putting together a skiing/snowboarding trip, a beach camping trip, a water rafting trip, etc.
The programmers today certainly aren't as hardcore as those in the 80s, myself included, but we make MUCH MUCH MUCH better stuff.
That is not my appraisal. :-)
AI is really killing programming.
An AI assistant is just a tool. One needs to know when to use it and when not to. What requirements/documentation you need before turning it loose. Applying an agile-like iterative methodology is a potential path of disaster for an AI assistant. Basically AI works best with a specific well defined well documented task. Ideally something well discussed in professional level documentation, not limited to random stack overflow comments of questionable value. Every decision an AI is free to make is a hazard, best to have sufficient documentation to guide it rather than have it make its own design decisions. Try to keep it focused on implementation.
Otherwise your AI generated source code will on day one, look like that 10-year old project this is hack upon hack upon hack that has so incomprehensible it should just be abandoned and rewritten (as the Mythical Man Month describes). Aka "Vibe Coded Slop." No offense to those who developed the concept of "vibe coding", they described it as something for experimental code that will be thrown away, not something for production use.
The 80s just had different computers and tools, not so much different programmers. Well, at least once you get past HTML.
80s software was brittle and limited (Score:2)
The 80s just had different computers and tools, not so much different programmers. Well, at least once you get past HTML.
You don't remember the 80s well. Programmers did smaller scoped work that accomplished less. It was much harder to do back then and their tools were limited. it wasn't their fault. However, it wasn't fun. Languages that managed memory, like Java...as well as easier programming languages, like JavaScript and Python attracted a lot of people who enjoyed coding. It brought in a lot of scientists and creative individuals that wrote amazing stuff. The majority of code before 1995?...those SHITTY text-based
Re:You liked software better in the autistic 80s? (Score:5, Interesting)
Until the late 90s, programming was the domain of the autistic and severely mathematically inclined. In most areas, the pay sucked, and their output sucked even more.
I gotta call shenanigans on this. I got out of college in 1984. My starting salary in as a programmer was higher than my father's (a chemist working for the same company for 40+ years) had been when he retired 4 years earlier. And was quite well employed during that stretch from 1984 until your "late 90s" cut-off. And since.
I agree that it was mostly math folks going into CS (in fact my college didn't have an undergrad CS degree yet - you majored in math but chose a specific CS track), but the job market wasn't bad at all, let alone "sucked".
What city were you in? It sucked in the midwest (Score:2)
I gotta call shenanigans on this. I got out of college in 1984. My starting salary in as a programmer was higher than my father's (a chemist working for the same company for 40+ years) had been when he retired 4 years earlier. And was quite well employed during that stretch from 1984 until your "late 90s" cut-off. And since.
I agree that it was mostly math folks going into CS (in fact my college didn't have an undergrad CS degree yet - you majored in math but chose a specific CS track), but the job market wasn't bad at all, let alone "sucked".
If you were in Silicon Valley?...that tracks. Same with Seattle and probably a few other major cities with then current tech stalwarts. If you lived in middle America, the jobs sucked. They paid OK, but they were constantly laying people off. It was a high stress, low-security environment. Most programmers didn't work for tech companies, they worked in industry implementing software.
Most companies looked at software as an expense and few saw it as an investment and were eager to cut it whenever the
Re:You liked software better in the autistic 80s? (Score:4, Insightful)
The Age of Programming (Score:5, Interesting)
I wonder if the age of manual computer programming begins and ends within the lifespan of longtime Slashdotters.
Re: (Score:2)
First people employed as programmers date from, say, 1945, so born in ~1925 and now ~101 years old.
Re: (Score:2)
Didn't it arguably already end with learning autocomplete?
Re: (Score:2)
If it does then it signals that a lot of other jobs are going away as well. Personally I feel that AI isn't good enough to replace me yet, but I can't rule out that it will be sooner or later.
That in itself isn't a problem, the issue is the political and social ramifications in countries that are already rapidly ageing.
AI honestly makes this all less interesting. (Score:2)
Hopefully (Score:2)
It will make me more valuable as I grow older.
Sample size and sampling method? (Score:5, Interesting)
Nova Learning surveyed 463 students in 2025 and 677 students in 2026.
So at a standard 95% confidence level, a sample size of n = 463 carries a margin of error of roughly of + or - 4.55%. A sample size of n = 677 narrows that margin to about + or - 3.77%. While some of the changes are outside of this range, many are not.
I also wonder who they asked " sample consists primarily of middle and high school students from the U.S. West Coast " So Bay Area, LA, Seattle? wealthy kids, middle class? private school, public?
I dunno, the numbers are interesting but I think a larger study with more distribution might be more helpful.
Mediocre students (Score:2)
...with no talent or passion who were lured by promises of riches are dropping out.
Talent is real. It takes a special kind of mind to be good at software and not everybody can be good at it.
In the new world of AI, the ability to learn and adapt quickly is even more valuable than it was in the past.
Here we go again (Score:4, Insightful)
No, they are losing interest in STUDYING CS (Score:2)
CS was the standard path .... (Score:2)
... for "I want to program for a living.". CS is way more than that, but that was the reason many people chose it. Since AI now can code as good as humans and explain to a human what it's doing, studying CS to program is less of a thing now.
CS and it's"applied math"thing won't go away, but that was less people will study it sounds quite plausible IMHO.
Where has "curiosity" gone? (Score:3)
Back in the day, many were so enamored with the how and why of things that their insatiable curiosity often led to pursuing tech-oriented hobbies and careers. Not just "doing IT", but embracing problem-solving, taking on new challenges, constantly deep-diving into new areas of study, etc.
People have instant access to all kinds of knowledge literally in the palm of their hands, spoon-fed to them with little need to face the mental challenge of working to get information, synthesize it, and understand it. It's the cycle of look it up, read it, move on, rinse, repeat. No depth. No retention. No real understanding.
Curiosity is seemingly no longer encouraged or desired. How do we get this back?
That's great! (Score:2)
That's actually great for serious practitioners, we only get more expensive and there will be less noise recruiting.
Bootcamps made hiring extremely annoying.
Coding is not computer science (Score:3)
Most want to go where there is money. (Score:2)
Real money is in actually doing things, the modern programmer/developer is not it. Notice interest is now moving (back?) towards engineering.
Sure. (Score:3)
Re: (Score:2)
It's been a long slide downward, but AI will accelerate the need for exohertz computers to render a Mandelbrot set in less than a week.
Re:And this too shall pass (Score:5, Insightful)
I do not think so, at least not with LLMs. The problem is that an LLM does not even begin to reach the level of reliability of a typical compiler. LLMs hallucinate, create bloated code, leave out functionality that was requested, have no clearly specified functionality, cannot deal with subtleties, etc. Hence LLM-generated code needs to be reviewed carefully and that is probably in the end more expensive than having somebody competent write the code from scratch. There are also indications that reviewing LLM code is mind-numbingly boring and people will burn out doing it, while even now there are not enough people that can do it in the first place.
Also note that the attempts of replacing programmers are _very_ old. When I studied CS almost 40 years ago, the 5GL project had just resoundingly failed. Its idea was to have program specifications in constraint logic and then have the machine create code for that. The problem was not that this did not work. The problem was that creating a specification on the level required was much harder and more time-consuming that writing the code directly with a typical informal or semi-formal specification. And, quite frankly, I see the same problem with LLMs as translation engine, as soon as you need actual production quality code. For producing low-quality slop, LLMs are fine. But that slop will not survive in the real world, especially as the one form of code LLMs are really good at creating is attack code, which does not need to be reliable, secure or maintainable. Hence low-skill hackers got a major upgrade here, while LLM use makes the defenders even weaker.
Re: Another generation gap (Score:2)
We have been blind to what is going on because nobody dreams of enslavement.
Re: (Score:2)
No. But I'm sure they expect there not to be many jobs in computer programming when an AI can do it cheaper than they can. And that's what's being advertised. And entry level jobs are already down.
Re: (Score:3)
dumb comment of the day right there. What children EVER thing of "creating a push-button government or massive unemployment"? Yeah, those toddlers are really taking notes.
First, not all children are toddlers. I shouldn't have to say that, but whatever.
Second, most kids today have all this information - and disinformation - at their fingertips on the web and in social media. They're also exposed to it via the mainstream "news" outlets that their parents and/or grandparents watch. Many of them also know of family members - and those of their friends - who are (not by choice) under-employed or unemployed.
So why would you think that they aren't thinking of massive unemployment?
Re: (Score:2)