The Software Engineering 'Squeeze' (manager.dev) 64
Software developer Anton Zaides argues that software engineers have had it easy over the decades and the "best profession" on earth deserved the wake up call. He writes:It's not just one of the hardest times, it's also one of the most exciting.
I'm hugely optimistic about the software engineering career. All those companies started by vibe-coders all around you? Many will succeed, and will need great engineers to scale up.
Some engineers understand this, and use the chance to skill up. To succeed, you'll probably need all the skills of an engineer, some of a PM, and even a bit of design taste. It's not just about shipping code anymore.
But if you work as a code monkey, getting detailed tickets and just shipping them, you've done this to yourself. You won't be needed pretty soon.
I believe there are too many mediocre engineers, but also not enough great ones.
I'm hugely optimistic about the software engineering career. All those companies started by vibe-coders all around you? Many will succeed, and will need great engineers to scale up.
Some engineers understand this, and use the chance to skill up. To succeed, you'll probably need all the skills of an engineer, some of a PM, and even a bit of design taste. It's not just about shipping code anymore.
But if you work as a code monkey, getting detailed tickets and just shipping them, you've done this to yourself. You won't be needed pretty soon.
I believe there are too many mediocre engineers, but also not enough great ones.
I agree (Score:3, Insightful)
Yes AI may be generating a lot of code now. But you need someone to find where what was generated was weak, or inefficient.
Over time the quality of generated stuff will improve, but since so many companies are generating a lot of code today that is a LOT of technical debt that is building up rapidly.
I especially agree that now is the time to round out your skills - as stated, study design, study platforms you connect to but do not develop on. Study AI tools, find out when they work for things you work on and know well - and when they do not.
Good luck out there everyone!
Re: (Score:1)
Over time the quality of generated stuff will improve
that is a LOT of technical debt that is building up rapidly
The improvements will fix the technical debt. There is no silver lining here, no matter how much toxically optimistic people want to pretend there is one.
Re:I agree (Score:5, Insightful)
The improvements will fix the technical debt.
They will not. The only way to fix technical debt is with real, deep insight. AI cannot deliver that.
Re: I agree (Score:2)
Re: (Score:2)
Re: (Score:3)
Yep, that is probably what these people think up tech-debt is. No experience with _real_ problem and thinking they have seen it all.
Re: (Score:2)
I haven't seen it yet.
Re: (Score:2)
Re: (Score:2)
If you can't fix the architecture all you've done is kicked the can.
Re: (Score:3)
That is not what we are talking about. Seriously. That senior guidance is not a convenience. The model cannot do anything without it.
The other problem is that on many, many real world problems, that speed advantage will not even be there. Because here is a little known hard fact that has been established decades ago: Code review is harder and takes more time and skill than writing good code in the first place, above a pretty low complexity level.
Re: (Score:3)
Probably not. For very simple business code, code review by somebody really competent may be faster and cheaper than writing the code. At a pretty low threshold, code review gets more expensive and harder to do than assuring correctness by design. Generating code bt AI and then fix it is not an approach that can work beyond very simple things.
Re: (Score:3)
Yes AI may be generating a lot of code now.
I'd love to have a world graph of code reuse. I predict that the number of LOC under maintenance is increasing hugely as automated interns (AI) keeps making it much easier to make many slightly different and carefully incompatible variants of the same code. Whilst I'm sure you could also use the AIs to find duplication, that will be harder and more dangerous than just leaving it in most cases so won't get done.
Re:I agree (Score:4, Insightful)
Yes AI may be generating a lot of code now. But you need someone to find where what was generated was weak, or inefficient.
Ever work with a couple of clowns who can't code worth shit? I have. And most of the time it's easier to throw their stuff out and start over.
Re: I agree (Score:2)
There was one guy we had, every time I came across any of his code I just immediately deleted it. If it even remotely worked it was usually an accident.
Yea. (Score:5, Insightful)
But if you work as a code monkey, getting detailed tickets and just shipping them, you've done this to yourself.
I've seen this sort of victim blaming used to excuse automation sending people to unemployability in the past.
Fuck this jackass and every other one like him. Fuck them right in the head. I hope he vibe codes himself off a cliff. That's my sentiment and I'm not even a programmer/developer.
Re: (Score:3)
Here's the key:
Some engineers understand this, and use the chance to skill up.
Yeah, I've seen this before. They want you to get "skilled up" then don't give you any more pay for being a better worker. Thing to look out for are those very expensive training courses the company selected and then "pays for" which are utterly useless. But it puts you on the hook for a multi-year commitment and no raises because you got that very expensive training for free out of the goodness of their heart!
Re: (Score:3)
Being "skilled up" mostly just means that you keep your skills matched to the relevant skills for today's job market, not yesterday's job market. In a field that changes fairly quickly, that is a continuous process for anyone who wants to remain employed in the field over a full 45-50 year career.
One problem is that people who are attracted to a field may have unique skills and, more critically, interests that align with the field at the time the enter it. However as the field changes in what it requires, t
Re: (Score:1)
Yeah, I've seen this before. They want you to get "skilled up" then don't give you any more pay for being a better worker.
This is a very basic part of software development and always has been. There's always some framework, API, or concept that you have to learn to get some particular job done. They're paying you because you have the aptitude to pick up new skills quickly. I've only been doing this four'ish years and found that out before I even started (I am entirely self-taught -- no CS to speak of -- the only credential I can offer you is that I'm paid within the top 5% of software engineers at roughly $277k gross on my las
Re: (Score:3)
So much anger over something very few people even want to do in the first place and even fewer with the capability to do.
The reality is, writing the code has always been a small part of a software developer's job. I for one am quite happy that I will never have to fully type out "for (int i = 0; i size; i++) {" ever again, or have to Google how to filter the results from "element.querySelectorAll(...)" because some idiot decided that returning a list of items as an array was not exciting enough.
Re: (Score:2)
Re: (Score:2)
skill up to what? To being a middle management douchebag that sells and defends shit products?
Being a code monkey is a coping mechanism : give me work, I do work, I don't care about product. That's life in a bigtech company.
Re: (Score:3)
Life is competitive, and always has been. And nobody owes you a job.
Tech changes over time and everyone must adapt to it. That's going to put some people out of work. It's not a happy moment for them, but seriously, that's is how it has always been.
There is this nice theory that, by working together, we can all make sure everyone has enough without ever facing the horror of being put out of a job and needing to take unpleasant work to earn a living. Well, human nature doesn't really make this tenable, w
Sleazy McToolbag (Score:5, Informative)
Great software engineers are not afraid that AI will take their jobs - because they know it just can’t.
His logic in his writing invalidates his whole thesis. Rambling moron.
Also, I don't need to do 3 fucking jobs to succeed!
Not my pick as a submission (Score:2)
Re: (Score:1)
Oh just piss off you smug tosser (Score:2)
Not all of us can be the genius this guy clearly thinks he is. We do our job, work hard and earn money to pay the bill's. What we dont need is to be patronised by self regarding arrogant Gen Z foetuses like this who presumes the little people didnt do enough to help themselves when they had enough on their plate just making ends meet.
Well here's a heads up Anton - one day YOU will be surplus to requirements too. Enjoy.
My perspective (Score:5, Insightful)
I learned software engineering in the 70s. Back then there were not very many of us and even fewer who were good at it. Based on the laws of supply and demand, we were paid well. Then the word spread that software engineering paid well and a flood of people of varying talent jumped in. The word on the street was that everybody should learn to code. The truth is, talent is real. It takes a special kind of mind to be good at designing software. During the peak of boom times, lesser talented people got hired. Now, powerful tools are making the job easier. Change is coming and the best of the best will adapt and thrive. The others, not so much
Re: (Score:2)
Re: (Score:2)
20x productivity already today (Score:2, Funny)
yep. Vibe coding is no joke. 100% of my code is now generated, no more team needed, delivered in a fraction of time and money.
Re: (Score:2)
Re: (Score:2)
Prompt Engineer III?
Re: (Score:2)
Re: 20x productivity already today (Score:2)
And once your customers learn to do it - no more you needed either.
Re: 20x productivity already today (Score:3)
Exactly, in 2 year max 99.99% software engineers will be as obsolete as lamp lighters, switchboard operators or bowling pin setters. So what can you do: reskill, learn to vibe code, move up the food chain and build your own business, learn how to leverage AI tools for e.g marketing, customer support. Do it before the worth of your current skillset goes to zero and it will, sooner than you think.
Re: (Score:2)
LOL. Typical Slashdot bullshit.
Remember 15 years ago when everyone on Slashdot was paranoid that we were all going to lose our jobs to Indian outsourcing? Pepperidge Farm remembers -- and, oh look, we're still here!
History doesn't repeat itself, but it surely does rhyme :)
Re: (Score:2)
You write code to be 100% thrown away, do you? Because otherwise there is no chance in this universe that your statement is true.
Yeah, but does it work? (Score:2)
yep. Vibe coding is no joke. 100% of my code is now generated, no more team needed, delivered in a fraction of time and money.
100% of your code is generated...OK, did it actually work? Can you teach me how to make LLMs generate working code? Claude 4.0 only generates code that compiles about 50% of the time for me.
Re: Yeah, but does it work? (Score:2)
Looks like you already lost the race.
Re: (Score:2)
Smells like bullshit to me.
Re: (Score:2)
The above post was apparently written by AI, or by a marketing department of an AI company, or both!
manup and go union the big boss man gets big tax c (Score:2)
manup and go union the big boss man gets big tax cuts works need an union to get higher wages!
Re: (Score:2)
Software engineers are usually not engineers (Score:3)
Most are low-talent technicians that lack any real engineering education. This comes from lack of requirements and lack of liability.
And no, "vibe coding" is just one more crappy method that will make things worse.
Re: (Score:2)
Author ignores business fundimentals (Score:5, Interesting)
From a business perspective - All business care about is profit. period. If they can slash programmer jobs, they will. Using AI will allow them to easily reduce the number of developers by 1/5 or more. The fact the QA will suffer and in 10 years nobody will understand what's under the skin doesn't matter.- companies will be producing code at less $$ and time per application, which is all that matters and next quarters 10Q report will woo shareholders. The bottom line, there will be fewer developers, the entry market is on it's way to being decimated. And by the way, "vibe coding" is bullshit.
Re: (Score:1)
Re: (Score:2)
He used AI to generate his English, too.
Nothing is more expensive than a cheap programmer (Score:3)
Re: (Score:3)
Great, now you've got the POC done! Just 99% of the work left to be done!
I manage three dev teams at a company that builds and sells commercial time card entry systems. In all, we have a couple hundred developers working on the system. We have several million users spread among tens of thousands of companies. I guarantee your AI-built software isn't going to provide competition that our company needs to worry about.
Why are we listening to this guy? (Score:2)
He then meanders over to the theory that if you are a real actually-good software engineer your job is clearly safe, because
Hitting the high notes (Score:1)
This article reminds me of the many https:://joelonsoftware.com [https] blog articles from 20 years ago. (See the organized lists at the bottom of the front page.)
Especially relevant to Zaides' article:
"Hitting the High Notes" https://www.joelonsoftware.com... [joelonsoftware.com] is about the benefits of hiring the very top-notch developers, is especially relevant.
"The Law of Leaky Abstractions" https://www.joelonsoftware.com... [joelonsoftware.com] is about how developers still need to understand details about how lower-level abstractions actually
Are we ... (Score:2)
Many will succeed, and will need great engineers to scale up.
Don't deserve it (Score:3)
Very very few software developers deserve the title "engineer" at all, and not just because they haven't or couldn't pass the FE exam. Software quality has been in free fall for a couple of decades now for a variety of reasons. That is not all the software developers fault of course, but if you can't write code in some serious high performing programming language that could be put in ROM and perform according to spec without any serious bugs for at least a century you are basically incompetent compared to video game programmers who used to count cycles on 8 and 16/32 bit machines and produce reliable, nearly bug free software that was often burned into ROM and still works in emulation today or write code that could control safety critical embedded equipment where failure means death or injury, human spaceflight controls with similar consequences, or any kind of software where a major failure means a financial or other catastrophe that results in human suffering, major loss of life, disclosure of massive amounts of information like that you *definitely* do not deserve the title "software engineer" at all.
Ask a real engineer sometime, someone that deserves the title - they operate under such constraints as a matter of course, if they fail to do their jobs bad things like I described happen, and the professional ones bear personal and professional liability if a design for something like a bridge or a building catastrophically fails. There was a time when if an architect or engineer designed a structure that failed and killed someone, the consequence was the death penalty or at least permanent revocation of their professional license.
It would be nice to bring that level of seriousness and quality, reliability, and performance back instead of cutting and pasting random bits of (possibly low quality "AI' generated) code and tweaking it until it pretends to work like someone with a seventh grade education and then shipping the abysmally low quality result every couple of weeks and planning to fix the bugs sometime in the next decade if you or someone who works there or who calls the shots ever gets around to it at *all*. Some who claims to be a software developer (or worse a "software engineer") should act like they are smarter or at least more responsible than a fifth grader.
Re: (Score:2)
The quality level that you are talking about is not merely a matter of skill on the part of the developer, but also time. The formal methods used that produce that level of performance and perfection take a long time to apply, no matter how good one is.
The *primary* reason why most software doesn't perform at that level is this: time costs money. Software that is thrown together quickly, and as a result has bugs, will be available for purchase much sooner, and will cost far less (mainly because it cost fa
Re: (Score:2)
I generally agree with all that, but I am still not impressed. It is partly due to rapidly declining educational standards where newly graduating programmers need to have remedial classes so they know how and are not afraid to make a phone call or write an email in complete sentences. Writing a compiler for a serious programming language? Apparently that is just short of unheard of, and when I was in school you practically couldn't graduate in computer science without taking a class like that and almost
Janitors (Score:2)
"Vibe coding" makes POCs (Score:3)
It's valuable to be able to throw together a POC in short order. But the journey from POC to finished software, is a long and winding one. These new startups are going to quickly find out that scaling up is *hard* and vibe coding won't get them there.
scrum shops (Score:2)
A lot of people work in scrum shops that try to do agile development. I've been there myself, you betcha.
Everything is run using Jira. You attend a morning standup with your team and discuss the tasks and bugs related to the current sprint. You get assigned enough of it to keep you busy that day or the next. There is never a Jira ticket that says "design and build something cool that will take you several weeks, and use some new language or paradigm that the team isn't already familiar with but you would le