Welcome to the main community in programming.dev! Feel free to post anything relating to programming here!
Cross posting is strongly encouraged in the instance. If you feel your post or another person's post makes sense in another community cross post into it.
Hope you enjoy the instance!
Rules
Follow the wormhole through a path of communities !webdev@programming.dev
Geoffrey Challen, a computer scientist at the University of Illinois at Urbana-Champaign, plans to offer a new course this fall in which he will teach students to develop software “without writing, reading, debugging, or viewing a single line of code,” he told me.
Is that meant to say reviewing? Either way, I can't see how this would lead to good results, even with a comprehensive test suite. Security? Scalability? Maintainability?
Writing code is not computer science. But instructing an LLM without knowing how to write code is it even less.
Computer science is maths, studying algorithms, understanding the underlying physics even. This does not require coding, though it helps IMO.
I use LLMs to automate boring tasks or generate starting points, but in my experience, i can't trust them to generate code that I'd be proud to share. If I use the code they spit out, I'm always adapting or rewriting it to meet my standards. I find they better at explaining code than generating it.. Anyway...
How will these students evaluate if the code they have generated is up to scratch?
You kind of have to have been a good coder to know what good code looks like.
I know, I know, another AI will be used to review the code...
Something feels a bit off here to me.
I'm sure I will be flamed.
Yeah, at most you can let them manage a 1k loc python script (the free tiers or Gemini Pro at least), but more than that and it starts to really eat your tokens without achieving what you asked or breaking functional behavior.
I extremely doubt that Coding Agents will see a future like promised. LLMs are still so expensive to run, and the useful larger models will probably never be affordable (if they charged for them what they cost). Apart from the fact that even their output can be utter garbage (and mediocre at best). You can already see it everywhere. Websites break in weird ways, ways in which it’s clear that either a complete beginner wrote that or an LLM did. Look at Shazam a few weeks ago. UI design? Horrific. Extremely inconsistent. Ugly. There are many other examples. It just shows that it doesn’t work. And no, the next model will not solve those issues. LLMs are flawed for this task from the ground up, the approach is outright wrong, we can make up so many bandaids and they will still suck, forever.
Inconsistency kills me. I've a colleague at work who insists on putting every doc through an llm to make it look nice then dumps it on the wiki. Now the wiki is a clusterfuck of assumptions and inconsistent styling.
Our documentation on some tedious process does not need to look like a magazine article nor does it help me find the salient code snippets that will help me understand why this awkward bit of code is the way it is.
Meanwhile he thinks he's achieved something.
Yeah, honestly if I notice something has AI in it, I don’t use it anymore. Open source projects with a CLAUDE.md or whatever in them? No, thanks.
The amount of assumptions they make are really one big issue what makes them suck so bad. In the end you just have more work. Instead of getting done 80% of the work in 20% of the time, now you get 30% of the work done in 1% of the time, but good luck getting the remaining stuff done at all.
Not even reading code? 🤣
A vibe coding class then?
So software development by magic - makes sense.
There was a post on here awhile back about a Japanese kids program teaching CS principles without a computer, using real-world examples. Maybe its something like that?
That's dangerously close to apologia, and at the very least hopefully naive, all due respect, fellow poor. 🙏🏼
'Computer Science is no more about computers than astronomy is about telescopes.'
Study computer science if you like it, it's never been about making good 'coders' or software engineers.
I don't think the number of software engineers will ever drop to zero, but the days of 'learn to code' to get a high paying job guaranteed are definitely over.
The jobs will just become (even more) horrible amagamations of fullstack development mixed with customer service. Meaning you'll be forced to sit on "urgent" live support lines and incident bridge calls, management will expect you to magically be able to answer any question about all facets of every integration and table the stupid thing touches, at the drop of a hat, including the vibecode garbage some other team just deployed yesterday, and you'll also have to do tier 1 type of tickets where you walk some illiterate fucktard through how to click the single sign-on button.
This is already my life.
LMAO. Are you a CEO? "Programming is unnecessary, AI will do everything".
I don't know how to break this to people but the vast majority of coding is boilerplate projects to solve trivial problems. Those jobs are disappearing (and have for years), what still exists is applying rigourous methods of computer science to solve specialised problems.
I've been writing code as a primary hobby and then as a profession for 26 years. The boilerplate has never been the bulk of any of my work, and we've had excellent tooling to eliminate the actual boilerplate for decades.
The work has always been the specialized parts, and the fun part of software dev work is that so much of it is bespoke and creative and unique beyond the grasp of the stochastic parrots.
I didn't say programming is unecessary, and I'm a proffesional software engineer with a degree in computer science. When I say 'learn to code' is over I mean the pressure for anyone and everyone to learn to code because there are so many well paying software engineer jobs.
This era is over undoubtedly, because all the people who never really cared about software engineering and are just there to collect a paycheck are going to be replaced - but the profession of software engineering will still be necessary, and the abstract maths of computer science isn't going anywhere as a field of research.
They're not over yet but they definitely have an uncertain future.
i’d say 2000-2002 pre and right after com bubble was a better time
Oh I was pissed when the bubble burst right before I graduated from high school
Imho only true if you can gain some sort of sovereignty by coding shit yourself, self-host, etc.
But that's is like saying it has never been a better time to learn to cook.
Sure it won't get you significant monies (except extremely lucky), but it's a life skill.
But yeah, companies in need of devs def think this is the best time to get more cheap labour to the field, how hard can it be to prompt what a mid/project manager (also AI) tells you to.
But we’re on the precipice of a new era when learning to develop software will be easier than ever, opening the door to students who might not otherwise have chosen to study computing. Perhaps a new golden age of CS education has only just begun.
It is good to have those students. They can fill roles in sales and customer relations and be the link to the actual software. These students won't be writing too much software besides some small, internal tools.
