I miss the old days (which were just a few years ago?).
I was someone who got into tech because they enjoyed coding as a mental exercise, not because they wanted to get rich. I feel like the fun part of the job has been now been automated away. All that remains is the boring part that I previously tolerated as a necessary evil: aligning stakeholders, kissing management ass, obsessing over the performance review cycle.
I know that some people have the inverse relationship: they find the "boring" part to be energizing and the "fun" part to be tedious. And I've seen a couple coworkers experiment with using Claude to agentically caress their boss's derriere. That's good for them and I wish them all the best. As for me, hopefully I can last long enough to engineer an early retirement. Then maybe I'll contribute to open source, assuming that there still is a FOSS community left by then.
Doing coding professionally already stripped nearly all the fun out of coding for me years before AI got good at coding, so when it comes to "how fun is it to be a software engineer", AI didn't have any negative impact for me. Being a professional software engineer was already extremely boring. Cycle after cycle, ticket after ticket, feature after feature, meeting after meeting... constant blur of meaningless shit. Not totally meaningless - it delivers actual value for customers. But meaningless for me.
If anything, AI has made it more fun simply because it means not having to think as much about the drudgery of details: syntax, API schemas, boilerplate code... And by "more fun" I mean "somewhat less unfun". Cause it's still not fun. But AI makes my burnout somewhat more tolerable.
I guess there’s just two sides to this coin. I’m a little more on your side: it was never really all that fun to write the code that the organization wanted me to write for some kind of necessary reason.
Still, there were a lot of times where I was catching myself having fun.
If you work for the right organization and work remotely you can just use AI to be more lazy. This is especially true during this transitional period where many colleagues are bad at it or refuse to use it.
I guess it all depends where you’re at. If you’re still in the phase of life where improving yourself at work and producing work you are proud of is exciting, AI might make you feel like your work and value has been diluted. You’ve been turned into a product manager when you wanted to be an engineer.
If you’re in the phase of life where you’ve decided that work is a means to an end, and simultaneously your company doesn’t expect you to produce 2-3x output, maybe this new era is pretty awesome.
I fall into the latter category. I’ve learned a new musical instrument mainly during work hours. I’ve gotten a lot of quality time with the elder scrolls morrowind. I started going to the gym every day. In the short term, the fact that I can make the coding machine do all my tasks for me has been a net benefit. I’m just not sure how long it’s going to last.
Im not in a burn out phase but Ive been doing this long enough where the "work" really seems like an imaginary construct of my employer. I dont even relate the software development shop I work in to software anymore. I have a friend who describes it as "taking out the complicated trash". Thats close. But software development hasnt been a respectable career in a while.
Anyway, nothing has changed for me, except that I dont have to deal with google search results anymore for things. When the LLM gets stuck I do still go out there. But its rare.
Nothing has changed in the way I do my job, except having to put up with slop from people around me, or from business people pushing slop towards the dev team and wanting a pat on the back for it. That part has changed for the worse.
The complete willingness to traffic in nonsensical documentation that doesnt represent any real requirements as an excuse for doing a job is pretty sad, and I take note of who does it and I make sure to never trust anything they publish. And usually stop reading it.
I call it "crossing over", once I see someone do it and replace their work product with LLM slop and distribute it, they have created a barrier that will prevent me from working with them forever.
I am an OSS maintainer and I can tell you the job is terrible. Not because of the thankless work (I am sponsored by a corp for the work) but because we're still allowing public PRs. Plenty of people that "want to help" just push walls of text and code and 2 days later ask why hasn't their contribution been accepted yet. And when will it be? Because they need it ASAP!
"Closing as too broad / not focused enough in language and scope / not adhering to our PR submission guidelines (see [link])"
(where the top guideline on [link] clearly states: "we do not accept submissions that are either outright AI-generated or conceptually similar in terms of style, verbosity, generality, scope-creep or signal-to-noise ratio")
Meanwhile just a few minutes ago in an issue on a project not mine, I just wrote a simple 3 step set of directions answering another users problem. (how to get the Voice audiobook app to appear in Android Auto when installed from Fdroid instead of google play.)
He takes that, tells claude to add it to his docs, claude mangles my correct manually written directions and spits out a garbled and nonsensical version, and the project owner just throws that right into his docs.
Like where do you go from there? Even when you don't use any ai someone else just uses it upon you. It's like a bizarre new form of Bowdlerization.
Everyone sucks on all sides as far as I can tell. I really don't understand why people are accepting this paradigm of using ais for everything. Is everyone really that incapable of resisting the whatever it is you get from giving orders that they don't even care what results from the orders, just make me feel like I caused a lot of stuff to happen.
I don't know, sometimes I feel like I've been misled, I come back to the code that was generated, I find so many bad choices, claude never subtracts, it keeps piling on new code. Some minor things get declared as unbreakable rules, regardless of context. It memorizes the tiniest exception as some big decision and all new code follows those new rules it made up. Good luck cleaning up all the nonsense to all who just vibe coded their apps. It has to be steered constantly, if you care about understanding the code that is. The LLM will happily use gibberish variable names, long functions that mix responsibilities, inconsistent code style and so on. Give it a prompt to review the code over and over again, ten times in a row, it will always find something, that one race condition that wasn't handled, or it will warn about something that is clearly not important, but it won't realize that it showed rendering code into the core logic component. It can fix it once it's pointed out, it's good for refactoring and doing sweeping changes across the codebase, but it's better to be on top of things, because it can lose its way behind your back.
This has been exactly my experience so far, regardless of the model. A developer knows when to stop and ask questions (colleagues or stakeholders). Claude just keeps going and doubles down on its own interpretation.
For me, the biggest benefit is being able to produce an initial POC that I can learn from before writing my own code. And, as you pointed out, it’s also brilliant for refactoring work.
Claude absolutely subtracts if you ask it to. I just regularly ask it to look for simplifications and YAGNIs and give me a ranked list of them, some of which I get it to do.
I guess it was slow-boiling, but somewhere between 2010-2015 was the turning point.
It got dominated by brogrammers who judged you by whether you used vim or not, teams of 20 people producing the same output as a team of two, designers doing mockups for absolutely everything and trying to boss the programmers, product managers making mountains of useless documents for everything and trying to boss the programmers, some rando had Scrummaster certification (who doesn't) and wanted to boss the programmers.
It was also when everyone started copying Google and Spotify and etc and a meeting was required for every decision, RFCs or other documents that would be produced but wouldn't survive reality, the software would have more trackers than features, and subscription was the default software model.
People keep saying "you only need 1 or 2 people to make a product now".
Well that was already the reality for me from 2000-2016.
> brogrammers who judged you by whether you used vim or not
When I hear that term I usually think of people who brag about their job and sling copy-pasted untyped Python and JS from StackOverflow while being willfully ignorant of infrastructure, architecture, security, algorithmic efficiency, and anything else that makes maintainable software. Micro-dosing K and listening to too many Huberman podcasts are probably solid tells too. When I hear vim, however, I think about greybeards.
Did I miss the memo? There has been a non-zero amount of performative TUI obsession happening in the technical space for the past five or six years but I didn't think that using a fifty-year-old modal text editor was a shibboleth for being a brogrammer...
>> It got dominated by brogrammers who judged you by whether you used vim or not, teams of 20 people producing the same output as a team of two, designers doing mockups for absolutely everything and trying to boss the programmers, product managers making mountains of useless documents for everything and trying to boss the programmers, some rando had Scrummaster certification (who doesn't) and wanted to boss the programmers.
That's because word got out that the field pays well but most people who had the sociopath desire to make a lot of money didn't also have the skill to get a programmer job. So they did what they always do: attach themselves to a living host and leech it's blood. In a pond of IT field water, for every fish (programmer) there were at least 20 leeches (project managers, product managers, scrum masters, designers, architects, you name it).
I thought several times why I chose to be a fish and not a leech and the answer was not that it's the moral thing to do or that I'm not that bothered as long as they let me with a little blood left. Rather pragmatically that it seemed safer being able to find and make my own food than rely on finding and sucking out the blood of someone who does it already.
For 20+ years they added more and more blood sucking parasites to the poor fish
and now, finally, they eliminate the fish altogether. Fine!
I'm just curious where will all the leeches go. What they'll suck on from now on. Can you suck the blood from an AI agent? The future looks exciting if only to see how the parasites adapt (or die).
As a maybe interesting contrast. I got into tech because I wanted to build things. Coding was just a necessary step. What I really liked was creating ideas for what we could build, thinking about UX, thinking about error conditions and what we do when two things disagree.
In the old days my team leader would review our PR and every time will remind us that he can “compile the code in his head”. This was 10 years ago, I still find it corny and disturbing.
I don’t want to go back to those corny bro programmers.
of course there'll be a FOSS community. with everything that's been happening I've yet to see anything even suggesting xkcd 2347 is being impacted here, except the thankless maintainer can bounce ideas off an agent now.
My favorite part about gushing about the new age of AI development is how oblivious the authors always seem to be about the next phase.
It's always some variation of: the LLMs are incredible and gaining logical and reasoning ability at an astounding rate, but fortunately they will always need me to sit here to hit the "Enter" key.
until 1) software is saturated or 2) llms or their creators will take accountability for any issues its code causes i think there will continuously be the next phase
Tbh i dont see software demand being saturated anytime soon (which is why im bullish SWEs, and we're seeing both demand and compensation increase as of late), we keep moving up abstractions and increasing ambition and scope for new projects.
If all the companies use {frontier model} then your competitive advantage is what you have that is not legible to the model
There are a bajillion apps on the app store that get little to no recognition. Most software startups fail to find a foothold and go under. An uncountable number of FOSS projects have been seen by at most a handful of people, and even fewer who have used/cloned the code.
Well I mean - I don’t think it’s so much that you’re needed to hit “enter” - anybody who’s used an agent on full auto has already experienced that lack of need.
What is needed is a human at the beginning, to decide what needs to be done - and a human at the end, to decide whether it’s finished.
For a while at least, specialized knowledge will be a huge asset, if not necessarily strictly necessary, for the beginning and end.
Possibly even that might pass away - but I m a little more skeptical there, because a fully automated proceeds of choosing what to work on is really a whole different world altogether. It’s not just one where humans aren’t required to do menial tasks, it’s one where humanity isn’t required period - at which case, what’s the point of any of this.
I don’t see us getting to that point within my lifetime.
While that sounds a little doomerish and probably for good reason, there is the obvious alternative.
Dont change. Software developers do not have to change anything they control their means of production and we dont HAVE to use any of this stuff. IF you are saying that there will be non-engineers willing to step in to hit enter (as in the business/product people) and LLMs will be good enough to do with it, Im ok with that.
I encourage business people to go down that path, especially since I have no organizational responsibility to help them. So if they run into even one little issue, I wont look at their code, I wont let it into a repo and I can just walk away.
There will be people who cross over and will want to help the people using LLMs and some will be technical. But it wont be in large numbers and it will be based on extraction and not pay very well at all. It will pay less every day as the LLM usage goes up to the CTOs desk.
More power to them! At the end of the day someone WILL have to hit enter, and it wont be me in some low paid slopshop.
What Im saying is the industrys future and mine are not intertwined. The tech industry is of no interest to me as a software engineer. They have jumped their own fake shark.
Ok but... You aren't planning to get paid to be a software engineer? Or you are? Surely this is the way in which the futures are entwined, for most of us?
> And yet, how also we have changed, and continue to adapt to the new ways of working.
Assuming this trajectory continues, the best case is the only adapting which we'll need is to sit in the assembly line chair: wages reduce to that of a skilled factory worker, little to no leverage in labor market, and economic precarity.
Enter key? Just talk to the LLM, no enter key needed ;-)
Now anyone who can talk can vibe code apps.
Personally I vibe code apps while driving. I still have to test them once no longer in motion, so im thinking maybe humans are still needed for testing but in a few years or more maybe not?
> fortunately they will always need me to sit here to hit the "Enter" key.
Nothing fortunate about it. That's the part I hate the most about today's LLMs. That you still have to do QA for them regarding the design. I should be able to arrive at cool result with 5% of current steering effort, if any. I still feel like a bottleneck between nothing and cool stuff existing. Unfortunately it's unavoidable until they read my mind and know what cool means for me.
Until either the enter key wears out and breaks, or we get pervasive sandboxes that act according to the users interests. It seems like Windows Vista UAC all over again for a new generation.
We'll keep doing it for as long as those making the blog posts keep making the same sort of blog posts we have seen on this cycle several times before. Slop begets slop, human or automaton made or a mix, and always has done.
My journey from hand-written code to agents is very similar. I think once a person exposed to agents goes through initial "I want to build anything I can think of" psychosis, it results in a much more productive workflow.
I've been spending less and less time glued to my screen, and more time thinking about concepts, as well as doing thing I like. The key for me was realizing that building with agents is very fun, but not every idea is worth building.
I am not too worried about the future of my career, as long as I stay in sync with the latest tools (agents). I think of it as elevator operators being misplaced by elevator controllers: the ones who refused to upskill into elevator technicians probably did not get a return offer. Not because nobody needed people to work with elevators, but because their skills were not economically valuable anymore.
every elevator needed an operator, how many elevators can a technician support? And what is the upskilling software developers can do? Not every software developer can be a good architect, PM or sales person (and even if they could be, there aren't going to be as many open positions anyways).
There are a LOT more elevators than when we had lift operators. Probably a lot more people employed globally in building, maintaining and operating them now too.
But yeah, disrupted careers is real. And the faster the economic or career discontinuity is, the harder it is for people to find new jobs or paths to new jobs.
I want the UI to be snappy. I used to dictate using
observables and other optimization preferences - and
sometimes still do. But I can also merely add
constraints - the image should process within 15ms
presses <enter>. Oh shoot - I meant to write 150ms
but… oh… I guess lets simply see i the model can do
it? Wow. I realize… I need to re-evaluate what I
think of as possible and not.
this resonates with me. Two days ago I was fighting with the GitHub Actions CI pipeline for an open source iOS app I maintain. It was stalling out after an hour, not completing. Sometimes, if I was lucky, it would finish after 45-50 minutes. I told Claude Code "make the CI pass in under 10 minutes. Don't ask me any more questions; I am going to bed and want this to work when I come back."
Claude Code wished me a restful night of sleep (seriously), and when I came back in the morning it had diligently chewed through the problem and was consistently getting the results I'd asked for.
The /goal command does a pretty dang good job if you tell it what will satisfy the goal and optionally what will not. It will still eventually stop if it finds the problem impossible (I too wanted faster CI, it could only get to a few minutes slower than my demand.)
Pretty minimal. For reference, I'm on a Claude Max 20x plan ($200/month). I expected to see about 15% usage of the plan overnight on the task but it was more like 5-6%? Well worth $2.50 of my monthly plan to get that speedup.
Just like people still use typewriters, make their own clothes, or cook themselves dinner etc. We can still hand craft code. There may always be a market for it too, like Rolls Royce.
However, I agree with OP, that the career in the main has changed. The cost of an engineer is probably going to fall precipitously once we hit saturation point. (I believe we're running on a backlog of software that was previously economically infeasible to make right now). IDK how long that will take (10 months? 10 years? 10 decades?)
But I do think LoC is going to fall in price to be like rice, a lot of the world just gives it away for free (to it's poor) because it's so plentiful. If AGI hits we may see many other goods follow.
We not only can, but we should. It will result in a much higher quality product, which works better, is easier to maintain, and has fewer bugs. Despite the hype, LLMs are not actually good at writing code.
i dont think we'll ever saturate the amount of software there is in the world unless we get some sort of scifi superintelligence scenario
Software allows us to build essentially infinite abstractions with infinite room to run. cheaper and better models released to the world means higher competition which drives up the cost for talent
however i think the value of 'hand crafted code' drops to zero unless we convince consumers to purchase software as a status symbol like rolls royces (seems unlikely, unless there is a concentrated marketing effort)
I feel it’s important to point out that the word “value” has different meanings to different people.
I think that the value of hand written code has gone up, and will continue to do so, as a direct consequence of anyone now being able to generate code with LLMs.
That’s how I feel, in any case. Win my heart with a euphoric 3am commit message, typos left in.
I think of it more like framework laptops, light phone etc which exist due to a fundamental decision to be a certain way and buyers value that fundamental feature.
I think it's heavily dependent on what type of dev role it is. There has always been a split between the startups and the corporate world.
In the case of working at startups, I'd agree that you don't (and never really needed to) hire extra devs.
Since this is HN, I know most hate corporate jobs, but that's where all the stability is and where AI is struggling to find footing. This rejection of AI puts a sorely needed spotlight on the cultural divide, and much of SV bet way too big on the wrong side of this misunderstanding.
It won't be a student on $10 replacing him. Who would hire an inexperienced student when there are piles of displaced experienced devs looking for work? One of them and a couple of AI licences could try to replace several people the way things are going.
At least until the investment capital starts to be called in and use of decent AI models becomes noticeable more expensive, assuming that happens long before open free models are good enough to replace them.
Of course that will lead to there being very few experienced people in a couple of decades, as most of the remaining ones age out, but I'll be long done by the time that becomes a big problem, I'm not on the right side of the mid-life divide!
> It won't be a student in $10 replacing him. Who would hire an inexperienced student when there are piles of displaced experienced devs looking for work?
Yeah it won't be a student making $10 an hour replacing him, it will be an experienced dev making $10 an hour because hey at least that beats starving to death.
The next step after "juniors are doomed, but seniors still get work" is pretty obviously that senior salaries are going to start rapidly dropping.
not sure where this meme of 'ai will become very expensive any time now!' comes from
consumer subscriptions are subsidized but the companies make very healthy margins via api pricing, which is what enterprises use, and cost per intelligence is dropping rapidly
Why does everyone keep framing this as if this is a decision this person is making?
This is what is happening to the industry. It's a fact of life. A new technology has disrupted the method through which software is created. There's no going back and there's no pretending we can continue not using it.
You either adopt the tools now and have trouble finding a job later or have trouble finding a job now.
> next time op comes up for air is when he finds out their employer replaced them with a college student making $10/hour.
I mean numerous firms seem to already be doing this, and they're making HN's front page regularly for week-long downtimes, blog posts tearing them apart for shockingly shit engineering practices, and their userbase in open rebellion from the sheer STATE of the product.
My employer has absolutely no interest in cutting any engineering staff in favor of AI. But we're also not a VC-funded human misery mill, and we have a sustainable business model involving selling actual products, so these facts may be related to that.
> blog posts tearing them apart for shockingly shit engineering practices, and their userbase in open rebellion from the sheer STATE of the product
So far I've seen relatively little anger of that level, except perhaps aimed at MS for the state of Windows. And unfortunately for the humans most companies high on AI juice are more likely to follow the path of finding better ways to use AI (or switching to what they think is a better model) rather than re-engaging with the meat workers, is such a user revolt seems imminent.
My DayJob is riding high on AI AMT, though quality is not currently falling (in fact by some metrics there is an improvement) because people are properly reviewing everything that comes out of it. I'm still planning on not touching the stuff though, if it comes to a choice of “AI or be redundant”, I'll be redundant, but even the Luddite that I am can see there is a benefit if it is used well.
I am transfixed watching gpt-sol-6-1 program all day non-stop. I am building a succesor to my web2.0 groupware platform and now on a low budget you can get the code built without having financial backing. Strategy is now what you need to get usage. How to grab customers who are unhappy with big players and want a more bespoke system.
This feels a bit like someone with one of the first cars thinking about how much faster they can distribute feed for all the horses.
That is, you haven't projected this forward just a bit further and are still thinking in terms of the old world where software is at all valuable and where you can have customers to sell it to.
I still feel myself falling into that trap, but haven't figured out what world replaces this.
Where a lot of this gets stuck is that AI software dev isn't a closed loop. They're still completely dependent on complex human written codebases to actually do anything. They're adept users of libraries and platforms but, at least for now, not so adept builders of the same. The ability cutoff being roughly where public documentation ends.
I don't think this will be a hill forever, after all folks have managed to get new
novel mathematics out of these models, but for the medium term
it seems like the good (software) business to be in is selling shovels to the people digging for AI gold.
I realized recently that I no longer get stuck. I don't sit and figure something out for a few hours, trying thing after thing. "Banging my head against the wall" isn't really part of my vocab anymore.
I learned a lot of what I know by banging my head against the wall.
Knowing these things might not really be needed going forward, but it was a good workout, and I'm grateful I got a chance to go through that even if it becomes largely unnecessary.
I learned a lot working with LLMs and successfully delivering ambitious projects I wouldn't even have approached before. However, whatever I learned with LLMs, I successfully forgot almost immediately.
Well said. I’ve often noticed how I have a huge tolerance for frustration and ability to persist compared to many people around me, and this is how I acquired these traits. They’ve turned out to be hugely useful, also beyond just coding. It’ll be a shame for our profession to lose them.
I guess the primary benefit of LLMs is that they give you more time to write navelgazy thinkpieces about LLMs. (And how long before the author starts reaching for the same tools to "assist" their writing?)
This person is not an engineer anymore, but it'll take a while for them to admit it to themselves.
Be glad you’re not a business copywriter. In my experience, that profession basically doesn’t exist anymore, and it has been subsumed into marketing/content marketing roles.
And I mean literally: the jobs that I had 5-10 years ago, that I built a copywriting career out of, simply don’t exist in 2026.
Sure: I can feel the inherent coolness radiating. Confessions of a mind witnessing its gradual metamorphosis into a cyborg. And sure, the productivity take-off sounds convincing. But each and every time I can't help but wonder: do these code-torrents water anything as frontier-worthy as the intelligence raining them? Where's the perfect, self-optimizing OS? Where's trading's Holy Grail? Where's the best-of-all-worlds Python-Rust-Zig hybrid?
In minds in danger of substituting counting, imitating, slop-spitting, relinquishing control for thinking, designing, daring in practice and not in prompting.
No hard feelings against the OP; they just happened to be the trigger. I fully understand I'm the shoemaker equivalent from 100 years ago. I just don't care--or rather care too much.
Remember when programmers viewed the guy who wrote Word documents and drew box-and-arrow diagrams all day with disdain because he was abstracted far away from the code base and didn't know how the real work got done?
Before he died, Tim Bryce had been trying to convey, since decades ago, that the document-and-diagram person is the real smart one in the room, the one contributing actual value. The programmers are just translators, turning those specifications into something a machine can execute. What's more, because they arrogate to themselves more value to the organization than they can actually justify, and because they resist discipline and organization, they're usually a liability to the company, and to society at large. Why does the TV and every device attached to it need a separate remote? Why don't they all just adopt a universal standard? It's programmers' fault. (No, really, that is a thing he said: https://patch.com/florida/palmharbor/amp/28237018/the-proble...)
Bryce was absolutely right. Now, with AI, there's barely any need for programmers anymore—just people who can describe the information systems the business needs. He'd be laughing his ass off had he lived to today, as the people who to whom every business had been beholden just watched their value drop to zero.
There's just no value anymore in being an obsessive technician anymore. There's no value in being a nerd. And here's a secret: the people you thought used to value you, secretly resented you and wanted you gone. Tim Bryce just said the quiet part out loud.
AI has freed you to focus on what's really important: being convincing in business meetings and charming at cocktail parties. Y'all need to go grind on your social skills, because AI won't be coming for those for a while yet.
I miss the old days (which were just a few years ago?).
I was someone who got into tech because they enjoyed coding as a mental exercise, not because they wanted to get rich. I feel like the fun part of the job has been now been automated away. All that remains is the boring part that I previously tolerated as a necessary evil: aligning stakeholders, kissing management ass, obsessing over the performance review cycle.
I know that some people have the inverse relationship: they find the "boring" part to be energizing and the "fun" part to be tedious. And I've seen a couple coworkers experiment with using Claude to agentically caress their boss's derriere. That's good for them and I wish them all the best. As for me, hopefully I can last long enough to engineer an early retirement. Then maybe I'll contribute to open source, assuming that there still is a FOSS community left by then.
Doing coding professionally already stripped nearly all the fun out of coding for me years before AI got good at coding, so when it comes to "how fun is it to be a software engineer", AI didn't have any negative impact for me. Being a professional software engineer was already extremely boring. Cycle after cycle, ticket after ticket, feature after feature, meeting after meeting... constant blur of meaningless shit. Not totally meaningless - it delivers actual value for customers. But meaningless for me.
If anything, AI has made it more fun simply because it means not having to think as much about the drudgery of details: syntax, API schemas, boilerplate code... And by "more fun" I mean "somewhat less unfun". Cause it's still not fun. But AI makes my burnout somewhat more tolerable.
I guess there’s just two sides to this coin. I’m a little more on your side: it was never really all that fun to write the code that the organization wanted me to write for some kind of necessary reason.
Still, there were a lot of times where I was catching myself having fun.
If you work for the right organization and work remotely you can just use AI to be more lazy. This is especially true during this transitional period where many colleagues are bad at it or refuse to use it.
I guess it all depends where you’re at. If you’re still in the phase of life where improving yourself at work and producing work you are proud of is exciting, AI might make you feel like your work and value has been diluted. You’ve been turned into a product manager when you wanted to be an engineer.
If you’re in the phase of life where you’ve decided that work is a means to an end, and simultaneously your company doesn’t expect you to produce 2-3x output, maybe this new era is pretty awesome.
I fall into the latter category. I’ve learned a new musical instrument mainly during work hours. I’ve gotten a lot of quality time with the elder scrolls morrowind. I started going to the gym every day. In the short term, the fact that I can make the coding machine do all my tasks for me has been a net benefit. I’m just not sure how long it’s going to last.
Im not in a burn out phase but Ive been doing this long enough where the "work" really seems like an imaginary construct of my employer. I dont even relate the software development shop I work in to software anymore. I have a friend who describes it as "taking out the complicated trash". Thats close. But software development hasnt been a respectable career in a while.
Anyway, nothing has changed for me, except that I dont have to deal with google search results anymore for things. When the LLM gets stuck I do still go out there. But its rare.
Nothing has changed in the way I do my job, except having to put up with slop from people around me, or from business people pushing slop towards the dev team and wanting a pat on the back for it. That part has changed for the worse.
The complete willingness to traffic in nonsensical documentation that doesnt represent any real requirements as an excuse for doing a job is pretty sad, and I take note of who does it and I make sure to never trust anything they publish. And usually stop reading it.
I call it "crossing over", once I see someone do it and replace their work product with LLM slop and distribute it, they have created a barrier that will prevent me from working with them forever.
[dead]
I am an OSS maintainer and I can tell you the job is terrible. Not because of the thankless work (I am sponsored by a corp for the work) but because we're still allowing public PRs. Plenty of people that "want to help" just push walls of text and code and 2 days later ask why hasn't their contribution been accepted yet. And when will it be? Because they need it ASAP!
"Closing as too broad / not focused enough in language and scope / not adhering to our PR submission guidelines (see [link])"
(where the top guideline on [link] clearly states: "we do not accept submissions that are either outright AI-generated or conceptually similar in terms of style, verbosity, generality, scope-creep or signal-to-noise ratio")
3 replies →
If you are not paid to work on it you can just ignore those.
3 replies →
Meanwhile just a few minutes ago in an issue on a project not mine, I just wrote a simple 3 step set of directions answering another users problem. (how to get the Voice audiobook app to appear in Android Auto when installed from Fdroid instead of google play.)
He takes that, tells claude to add it to his docs, claude mangles my correct manually written directions and spits out a garbled and nonsensical version, and the project owner just throws that right into his docs.
Like where do you go from there? Even when you don't use any ai someone else just uses it upon you. It's like a bizarre new form of Bowdlerization.
Everyone sucks on all sides as far as I can tell. I really don't understand why people are accepting this paradigm of using ais for everything. Is everyone really that incapable of resisting the whatever it is you get from giving orders that they don't even care what results from the orders, just make me feel like I caused a lot of stuff to happen.
This is all so self-inflicted too.
I don't know, sometimes I feel like I've been misled, I come back to the code that was generated, I find so many bad choices, claude never subtracts, it keeps piling on new code. Some minor things get declared as unbreakable rules, regardless of context. It memorizes the tiniest exception as some big decision and all new code follows those new rules it made up. Good luck cleaning up all the nonsense to all who just vibe coded their apps. It has to be steered constantly, if you care about understanding the code that is. The LLM will happily use gibberish variable names, long functions that mix responsibilities, inconsistent code style and so on. Give it a prompt to review the code over and over again, ten times in a row, it will always find something, that one race condition that wasn't handled, or it will warn about something that is clearly not important, but it won't realize that it showed rendering code into the core logic component. It can fix it once it's pointed out, it's good for refactoring and doing sweeping changes across the codebase, but it's better to be on top of things, because it can lose its way behind your back.
This has been exactly my experience so far, regardless of the model. A developer knows when to stop and ask questions (colleagues or stakeholders). Claude just keeps going and doubles down on its own interpretation.
For me, the biggest benefit is being able to produce an initial POC that I can learn from before writing my own code. And, as you pointed out, it’s also brilliant for refactoring work.
Claude absolutely subtracts if you ask it to. I just regularly ask it to look for simplifications and YAGNIs and give me a ranked list of them, some of which I get it to do.
For me that turning point was 10-15 years ago.
I guess it was slow-boiling, but somewhere between 2010-2015 was the turning point.
It got dominated by brogrammers who judged you by whether you used vim or not, teams of 20 people producing the same output as a team of two, designers doing mockups for absolutely everything and trying to boss the programmers, product managers making mountains of useless documents for everything and trying to boss the programmers, some rando had Scrummaster certification (who doesn't) and wanted to boss the programmers.
It was also when everyone started copying Google and Spotify and etc and a meeting was required for every decision, RFCs or other documents that would be produced but wouldn't survive reality, the software would have more trackers than features, and subscription was the default software model.
People keep saying "you only need 1 or 2 people to make a product now".
Well that was already the reality for me from 2000-2016.
> brogrammers who judged you by whether you used vim or not
When I hear that term I usually think of people who brag about their job and sling copy-pasted untyped Python and JS from StackOverflow while being willfully ignorant of infrastructure, architecture, security, algorithmic efficiency, and anything else that makes maintainable software. Micro-dosing K and listening to too many Huberman podcasts are probably solid tells too. When I hear vim, however, I think about greybeards.
Did I miss the memo? There has been a non-zero amount of performative TUI obsession happening in the technical space for the past five or six years but I didn't think that using a fifty-year-old modal text editor was a shibboleth for being a brogrammer...
5 replies →
>> It got dominated by brogrammers who judged you by whether you used vim or not, teams of 20 people producing the same output as a team of two, designers doing mockups for absolutely everything and trying to boss the programmers, product managers making mountains of useless documents for everything and trying to boss the programmers, some rando had Scrummaster certification (who doesn't) and wanted to boss the programmers.
That's because word got out that the field pays well but most people who had the sociopath desire to make a lot of money didn't also have the skill to get a programmer job. So they did what they always do: attach themselves to a living host and leech it's blood. In a pond of IT field water, for every fish (programmer) there were at least 20 leeches (project managers, product managers, scrum masters, designers, architects, you name it).
I thought several times why I chose to be a fish and not a leech and the answer was not that it's the moral thing to do or that I'm not that bothered as long as they let me with a little blood left. Rather pragmatically that it seemed safer being able to find and make my own food than rely on finding and sucking out the blood of someone who does it already.
For 20+ years they added more and more blood sucking parasites to the poor fish and now, finally, they eliminate the fish altogether. Fine!
I'm just curious where will all the leeches go. What they'll suck on from now on. Can you suck the blood from an AI agent? The future looks exciting if only to see how the parasites adapt (or die).
As a maybe interesting contrast. I got into tech because I wanted to build things. Coding was just a necessary step. What I really liked was creating ideas for what we could build, thinking about UX, thinking about error conditions and what we do when two things disagree.
I hear you. Sometimes my job feels like sitting and waiting on a bunch of different projects being run, with some pauses where I am able to evaluate
In the old days my team leader would review our PR and every time will remind us that he can “compile the code in his head”. This was 10 years ago, I still find it corny and disturbing.
I don’t want to go back to those corny bro programmers.
of course there'll be a FOSS community. with everything that's been happening I've yet to see anything even suggesting xkcd 2347 is being impacted here, except the thankless maintainer can bounce ideas off an agent now.
[dead]
My favorite part about gushing about the new age of AI development is how oblivious the authors always seem to be about the next phase.
It's always some variation of: the LLMs are incredible and gaining logical and reasoning ability at an astounding rate, but fortunately they will always need me to sit here to hit the "Enter" key.
I envy their optimism.
until 1) software is saturated or 2) llms or their creators will take accountability for any issues its code causes i think there will continuously be the next phase
Tbh i dont see software demand being saturated anytime soon (which is why im bullish SWEs, and we're seeing both demand and compensation increase as of late), we keep moving up abstractions and increasing ambition and scope for new projects.
If all the companies use {frontier model} then your competitive advantage is what you have that is not legible to the model
>until 1) software is saturated
Software has been saturated for a long time.
There are a bajillion apps on the app store that get little to no recognition. Most software startups fail to find a foothold and go under. An uncountable number of FOSS projects have been seen by at most a handful of people, and even fewer who have used/cloned the code.
3 replies →
Well I mean - I don’t think it’s so much that you’re needed to hit “enter” - anybody who’s used an agent on full auto has already experienced that lack of need.
What is needed is a human at the beginning, to decide what needs to be done - and a human at the end, to decide whether it’s finished.
For a while at least, specialized knowledge will be a huge asset, if not necessarily strictly necessary, for the beginning and end.
Possibly even that might pass away - but I m a little more skeptical there, because a fully automated proceeds of choosing what to work on is really a whole different world altogether. It’s not just one where humans aren’t required to do menial tasks, it’s one where humanity isn’t required period - at which case, what’s the point of any of this.
I don’t see us getting to that point within my lifetime.
> What is needed is a human at the beginning, to decide what needs to be done - and a human at the end, to decide whether it’s finished.
I've been thinking the same, this is very well put.
While that sounds a little doomerish and probably for good reason, there is the obvious alternative.
Dont change. Software developers do not have to change anything they control their means of production and we dont HAVE to use any of this stuff. IF you are saying that there will be non-engineers willing to step in to hit enter (as in the business/product people) and LLMs will be good enough to do with it, Im ok with that.
I encourage business people to go down that path, especially since I have no organizational responsibility to help them. So if they run into even one little issue, I wont look at their code, I wont let it into a repo and I can just walk away.
There will be people who cross over and will want to help the people using LLMs and some will be technical. But it wont be in large numbers and it will be based on extraction and not pay very well at all. It will pay less every day as the LLM usage goes up to the CTOs desk.
More power to them! At the end of the day someone WILL have to hit enter, and it wont be me in some low paid slopshop.
What Im saying is the industrys future and mine are not intertwined. The tech industry is of no interest to me as a software engineer. They have jumped their own fake shark.
Ok but... You aren't planning to get paid to be a software engineer? Or you are? Surely this is the way in which the futures are entwined, for most of us?
Exactly. Taken from this article:
> And yet, how also we have changed, and continue to adapt to the new ways of working.
Assuming this trajectory continues, the best case is the only adapting which we'll need is to sit in the assembly line chair: wages reduce to that of a skilled factory worker, little to no leverage in labor market, and economic precarity.
feels like the opposite to me. the 'factory worker' style code monkey job is vaporized
the new era of swe will demand more and more long term business + product direction and relationship building skills
right now the meme is we all became managers of agents. the trajectory leads us to all becoming mini-ceos directing our company of agents
7 replies →
Enter key? Just talk to the LLM, no enter key needed ;-)
Now anyone who can talk can vibe code apps.
Personally I vibe code apps while driving. I still have to test them once no longer in motion, so im thinking maybe humans are still needed for testing but in a few years or more maybe not?
>Personally I vibe code apps while driving
What a nightmare of a sentence.
7 replies →
> fortunately they will always need me to sit here to hit the "Enter" key.
Nothing fortunate about it. That's the part I hate the most about today's LLMs. That you still have to do QA for them regarding the design. I should be able to arrive at cool result with 5% of current steering effort, if any. I still feel like a bottleneck between nothing and cool stuff existing. Unfortunately it's unavoidable until they read my mind and know what cool means for me.
How many times are we going to cycle through these same arguments instead of actually discussing the content of the blog posts?
Circling back to these replies: Ah yes! The best slacktivism is having the bots do it for you!
As long as the daily propaganda articles to refresh the DRAM of people daily appear here.
Until either the enter key wears out and breaks, or we get pervasive sandboxes that act according to the users interests. It seems like Windows Vista UAC all over again for a new generation.
1 reply →
We'll keep doing it for as long as those making the blog posts keep making the same sort of blog posts we have seen on this cycle several times before. Slop begets slop, human or automaton made or a mix, and always has done.
1 reply →
My journey from hand-written code to agents is very similar. I think once a person exposed to agents goes through initial "I want to build anything I can think of" psychosis, it results in a much more productive workflow.
I've been spending less and less time glued to my screen, and more time thinking about concepts, as well as doing thing I like. The key for me was realizing that building with agents is very fun, but not every idea is worth building.
I am not too worried about the future of my career, as long as I stay in sync with the latest tools (agents). I think of it as elevator operators being misplaced by elevator controllers: the ones who refused to upskill into elevator technicians probably did not get a return offer. Not because nobody needed people to work with elevators, but because their skills were not economically valuable anymore.
every elevator needed an operator, how many elevators can a technician support? And what is the upskilling software developers can do? Not every software developer can be a good architect, PM or sales person (and even if they could be, there aren't going to be as many open positions anyways).
There are a LOT more elevators than when we had lift operators. Probably a lot more people employed globally in building, maintaining and operating them now too.
But yeah, disrupted careers is real. And the faster the economic or career discontinuity is, the harder it is for people to find new jobs or paths to new jobs.
this resonates with me. Two days ago I was fighting with the GitHub Actions CI pipeline for an open source iOS app I maintain. It was stalling out after an hour, not completing. Sometimes, if I was lucky, it would finish after 45-50 minutes. I told Claude Code "make the CI pass in under 10 minutes. Don't ask me any more questions; I am going to bed and want this to work when I come back."
Claude Code wished me a restful night of sleep (seriously), and when I came back in the morning it had diligently chewed through the problem and was consistently getting the results I'd asked for.
Did you look closely at what it had done? It's possible to save a lot of time on CI by omitting tests, for example...
https://github.com/OneBusAway/onebusaway-ios/pull/1531/chang...
We're not supposed to read code anymore!
The /goal command does a pretty dang good job if you tell it what will satisfy the goal and optionally what will not. It will still eventually stop if it finds the problem impossible (I too wanted faster CI, it could only get to a few minutes slower than my demand.)
What was the overnight cost on taking that route (tokens and dollars)?
Pretty minimal. For reference, I'm on a Claude Max 20x plan ($200/month). I expected to see about 15% usage of the plan overnight on the task but it was more like 5-6%? Well worth $2.50 of my monthly plan to get that speedup.
Just like people still use typewriters, make their own clothes, or cook themselves dinner etc. We can still hand craft code. There may always be a market for it too, like Rolls Royce.
However, I agree with OP, that the career in the main has changed. The cost of an engineer is probably going to fall precipitously once we hit saturation point. (I believe we're running on a backlog of software that was previously economically infeasible to make right now). IDK how long that will take (10 months? 10 years? 10 decades?)
But I do think LoC is going to fall in price to be like rice, a lot of the world just gives it away for free (to it's poor) because it's so plentiful. If AGI hits we may see many other goods follow.
> We can still hand craft code.
We not only can, but we should. It will result in a much higher quality product, which works better, is easier to maintain, and has fewer bugs. Despite the hype, LLMs are not actually good at writing code.
i dont think we'll ever saturate the amount of software there is in the world unless we get some sort of scifi superintelligence scenario
Software allows us to build essentially infinite abstractions with infinite room to run. cheaper and better models released to the world means higher competition which drives up the cost for talent
however i think the value of 'hand crafted code' drops to zero unless we convince consumers to purchase software as a status symbol like rolls royces (seems unlikely, unless there is a concentrated marketing effort)
I feel it’s important to point out that the word “value” has different meanings to different people.
I think that the value of hand written code has gone up, and will continue to do so, as a direct consequence of anyone now being able to generate code with LLMs.
That’s how I feel, in any case. Win my heart with a euphoric 3am commit message, typos left in.
hand crafted code:
I think of it more like framework laptops, light phone etc which exist due to a fundamental decision to be a certain way and buyers value that fundamental feature.
I think it's heavily dependent on what type of dev role it is. There has always been a split between the startups and the corporate world.
In the case of working at startups, I'd agree that you don't (and never really needed to) hire extra devs.
Since this is HN, I know most hate corporate jobs, but that's where all the stability is and where AI is struggling to find footing. This rejection of AI puts a sorely needed spotlight on the cultural divide, and much of SV bet way too big on the wrong side of this misunderstanding.
> where AI is struggling to find footing
As in struggling to find acceptance by the rule makers?
or As in struggling to be capable of delivering value?
(just asking to try and understand your reply)
1 reply →
> Just like people still use typewriters, make their own clothes, or cook themselves dinner etc.
One of these things is not like the others. Your silicon valley bubble privilege is showing.
Your privilege is showing by only saying one of those stand out.
At least with keyboards you can say they actually replaced typewriters but the other two are just paying someone else to do the same labor.
5 replies →
next time op comes up for air is when he finds out their employer replaced them with a college student making $10/hour.
I’m so happy I quit tech a year ago. this guy is driving straight off a cliff and doesn’t even seem to realize.
curious, is this the default culture at tech companies now?
It won't be a student on $10 replacing him. Who would hire an inexperienced student when there are piles of displaced experienced devs looking for work? One of them and a couple of AI licences could try to replace several people the way things are going.
At least until the investment capital starts to be called in and use of decent AI models becomes noticeable more expensive, assuming that happens long before open free models are good enough to replace them.
Of course that will lead to there being very few experienced people in a couple of decades, as most of the remaining ones age out, but I'll be long done by the time that becomes a big problem, I'm not on the right side of the mid-life divide!
> It won't be a student in $10 replacing him. Who would hire an inexperienced student when there are piles of displaced experienced devs looking for work?
Yeah it won't be a student making $10 an hour replacing him, it will be an experienced dev making $10 an hour because hey at least that beats starving to death.
The next step after "juniors are doomed, but seniors still get work" is pretty obviously that senior salaries are going to start rapidly dropping.
not sure where this meme of 'ai will become very expensive any time now!' comes from
consumer subscriptions are subsidized but the companies make very healthy margins via api pricing, which is what enterprises use, and cost per intelligence is dropping rapidly
5 replies →
Why does everyone keep framing this as if this is a decision this person is making?
This is what is happening to the industry. It's a fact of life. A new technology has disrupted the method through which software is created. There's no going back and there's no pretending we can continue not using it.
You either adopt the tools now and have trouble finding a job later or have trouble finding a job now.
> next time op comes up for air is when he finds out their employer replaced them with a college student making $10/hour.
I mean numerous firms seem to already be doing this, and they're making HN's front page regularly for week-long downtimes, blog posts tearing them apart for shockingly shit engineering practices, and their userbase in open rebellion from the sheer STATE of the product.
My employer has absolutely no interest in cutting any engineering staff in favor of AI. But we're also not a VC-funded human misery mill, and we have a sustainable business model involving selling actual products, so these facts may be related to that.
> blog posts tearing them apart for shockingly shit engineering practices, and their userbase in open rebellion from the sheer STATE of the product
So far I've seen relatively little anger of that level, except perhaps aimed at MS for the state of Windows. And unfortunately for the humans most companies high on AI juice are more likely to follow the path of finding better ways to use AI (or switching to what they think is a better model) rather than re-engaging with the meat workers, is such a user revolt seems imminent.
My DayJob is riding high on AI AMT, though quality is not currently falling (in fact by some metrics there is an improvement) because people are properly reviewing everything that comes out of it. I'm still planning on not touching the stuff though, if it comes to a choice of “AI or be redundant”, I'll be redundant, but even the Luddite that I am can see there is a benefit if it is used well.
> I’m so happy I quit tech a year ago.
What do you do now? I'm also looking to get out of this industry, but it's tough to voluntarily give up the pay.
How much money do you need?! I don't find it a particularly hard sell to not sell off my time after having reached my retirement number.
4 replies →
Did you retire or do something else?
Might as well stick around until the option is no longer there, as long as they enjoy it.
Feels like AI is just exposing what was always glorified software-configuration
I am transfixed watching gpt-sol-6-1 program all day non-stop. I am building a succesor to my web2.0 groupware platform and now on a low budget you can get the code built without having financial backing. Strategy is now what you need to get usage. How to grab customers who are unhappy with big players and want a more bespoke system.
This feels a bit like someone with one of the first cars thinking about how much faster they can distribute feed for all the horses.
That is, you haven't projected this forward just a bit further and are still thinking in terms of the old world where software is at all valuable and where you can have customers to sell it to.
I still feel myself falling into that trap, but haven't figured out what world replaces this.
Where a lot of this gets stuck is that AI software dev isn't a closed loop. They're still completely dependent on complex human written codebases to actually do anything. They're adept users of libraries and platforms but, at least for now, not so adept builders of the same. The ability cutoff being roughly where public documentation ends.
I don't think this will be a hill forever, after all folks have managed to get new novel mathematics out of these models, but for the medium term it seems like the good (software) business to be in is selling shovels to the people digging for AI gold.
The world that replaces this has AI actually doing the work of gscott’s web2.0 groupware platform.
I realized recently that I no longer get stuck. I don't sit and figure something out for a few hours, trying thing after thing. "Banging my head against the wall" isn't really part of my vocab anymore.
I learned a lot of what I know by banging my head against the wall.
Knowing these things might not really be needed going forward, but it was a good workout, and I'm grateful I got a chance to go through that even if it becomes largely unnecessary.
I learned a lot working with LLMs and successfully delivering ambitious projects I wouldn't even have approached before. However, whatever I learned with LLMs, I successfully forgot almost immediately.
Well said. I’ve often noticed how I have a huge tolerance for frustration and ability to persist compared to many people around me, and this is how I acquired these traits. They’ve turned out to be hugely useful, also beyond just coding. It’ll be a shame for our profession to lose them.
2 replies →
I guess the primary benefit of LLMs is that they give you more time to write navelgazy thinkpieces about LLMs. (And how long before the author starts reaching for the same tools to "assist" their writing?)
This person is not an engineer anymore, but it'll take a while for them to admit it to themselves.
Keep your identity small[0].
[0] https://www.paulgraham.com/identity.html
This was a very illuminating read that put into words something I've noticed but never quite pinned down. Thank you.
Be glad you’re not a business copywriter. In my experience, that profession basically doesn’t exist anymore, and it has been subsumed into marketing/content marketing roles.
And I mean literally: the jobs that I had 5-10 years ago, that I built a copywriting career out of, simply don’t exist in 2026.
Sure: I can feel the inherent coolness radiating. Confessions of a mind witnessing its gradual metamorphosis into a cyborg. And sure, the productivity take-off sounds convincing. But each and every time I can't help but wonder: do these code-torrents water anything as frontier-worthy as the intelligence raining them? Where's the perfect, self-optimizing OS? Where's trading's Holy Grail? Where's the best-of-all-worlds Python-Rust-Zig hybrid?
In minds in danger of substituting counting, imitating, slop-spitting, relinquishing control for thinking, designing, daring in practice and not in prompting.
No hard feelings against the OP; they just happened to be the trigger. I fully understand I'm the shoemaker equivalent from 100 years ago. I just don't care--or rather care too much.
Can't relate at all. AI is only in your life if you let it.
I work 90% without AI, and don't plan to integrate it any further. It's an overhyped and overrated tech that will die out soon.
Another article of unbridled enthusiasm after finding a new religion.
[dead]
My job is extremely boring now.
Give claude the jira ticket, implement it, open the PR, review the PR, merge it. Rinse and repeat.
What is even the point of any of this anymore?
That's why you're about to get fired. That's the whole issue.
Remember when programmers viewed the guy who wrote Word documents and drew box-and-arrow diagrams all day with disdain because he was abstracted far away from the code base and didn't know how the real work got done?
Before he died, Tim Bryce had been trying to convey, since decades ago, that the document-and-diagram person is the real smart one in the room, the one contributing actual value. The programmers are just translators, turning those specifications into something a machine can execute. What's more, because they arrogate to themselves more value to the organization than they can actually justify, and because they resist discipline and organization, they're usually a liability to the company, and to society at large. Why does the TV and every device attached to it need a separate remote? Why don't they all just adopt a universal standard? It's programmers' fault. (No, really, that is a thing he said: https://patch.com/florida/palmharbor/amp/28237018/the-proble...)
Bryce was absolutely right. Now, with AI, there's barely any need for programmers anymore—just people who can describe the information systems the business needs. He'd be laughing his ass off had he lived to today, as the people who to whom every business had been beholden just watched their value drop to zero.
There's just no value anymore in being an obsessive technician anymore. There's no value in being a nerd. And here's a secret: the people you thought used to value you, secretly resented you and wanted you gone. Tim Bryce just said the quiet part out loud.
AI has freed you to focus on what's really important: being convincing in business meetings and charming at cocktail parties. Y'all need to go grind on your social skills, because AI won't be coming for those for a while yet.