The Outlook for Programmers | Hacker News Reader

The Outlook for Programmers

cacm.acm.org

46 pointspseudolus86 comments

> “The notion going in was, you got a computer science degree, you were set. There was going to be a job waiting for you when you got out, and I think many people went into computer science for that reason,"

As much as the current market saturation in tech can be linked to interest rates, overhiring or AI, I think that this is actually the main reason. Too many people got into tech simply for the paycheque, social status and perceived job security who weren't necessarily very passionate about the field to begin with.

There was a great YC video about this around a year ago where they basically lamented that many of the people getting into tech nowadays are the same people who, 20 years prior, would've gone into law, medicine, finance or management consulting. This has made the field ridiculously competitive and has also made it really difficult for employers to tell the yuppies apart from the true hobbyists.

Some of the best programmers I know are competent, smart, and well adjusted people who chose this career and put in the work because of the demand. Some will tell you they don't even like programming. And some of the worst programmers I know are the type who centered their image and self esteem around being a true hobbyist.

There is absolutely nothing wrong with really liking, say, photography and looking around and saying “Nope, unless I basically win the lottery there’s not a lot of money in that and most of the jobs are actually pretty boring.”

Some people here probably disagree but nothing wrong with picking something you’re OK with, pays well, and you’re competent at.

...but you can't overwork them to the point of burnout and then replace them the next round of rubes.

YC and their startups were big perpetrators of the practice and spurred the trend further in software. I'm getting old so I was around for the late 2000's; YC used to con impressionable youths HARD. pg's cult of personality, his bad book, etc... it was all to get smart and passionate kids to overwork for him.

> Too many people got into tech simply for the paycheque,

A key factor in people overtraining for a particular field is a lack of alternatives.

People jam up a degree because it's the one reasonable option that has a real shot at paying for itself.

People who are not really in a position to make the decision tend to put themselves on at least a certain trajectory in college. Changes are possible—maybe less than in the past. But complete reinvention is hard and probably at least somewhat costly.

It has been painful working with yuppies in this way because most of them have no regard for correctness of the code. They push whatever runs. Code-consistency and maintainability are alien concepts to them, and most management isn't even aware of the existence of these concepts. Any product they develop turns to a wild, unmaintainable, unscalable, and incorrect mess within three years.

This has been an issue for decades.

Here are some numbers:

https://nces.ed.gov/programs/digest/d19/tables/dt19_325.35.a...

Note the peak in 1986 or so. Another peak right after the dotcom bubble burst (students started during the bubble). And the late 2010s surge in masters degrees as people are either trying to make themselves more competitive (than just having a BS) or coming in from another field like EE or Physics and trying to get a credential to get in the door.

These are some interesting numbers! Thanks for sharing!

No problem.

https://news.ycombinator.com/item?id=43778701 - You may like to read this, written by Bertrand Meyer in 1984 where a student of the time had lamented the apparent difficulty of earning the CS degree just so he could get a well-compensated job.

I appreciated this too! This was a pretty cathartic read. Definitely a bit ahead of its time.

> This has made the field ridiculously competitive and has also made it really difficult for employers to tell the yuppies apart from the true hobbyists.

This is an extremely funny way for a place like YC to put this problem. The same place that views employment as a purely contractual based relationship, with at-will firing and with a concept that people shouldn't stay in the same job for too long is lamenting the appearance of workers who view things the same way and couldn't care less about your vision, and definitely won't sacrifice their personal life for your business' success, since it's not their hobby.

There’s very much this bifurcation here, perhaps almost unique to jobs that also happen to generally pay well, at least in the US, of people for whom software development is a job you do like any other vs. it having to be some all-consuming passion.

YC being about startups and business and profit at all costs deriding people who go into it for money is a hell of a take. I’d say they’d love to see software become a passion industry so they can pay less like they do in most positions in every passion industry. The way software is developed for profit is a far cry from developing software from a hobbiest or passion perspective.

I realize I probably should've linked to the YC video: https://youtu.be/ia7IKW0yuG0?si=2a3xvKDZUZxZ_her&t=376

I don’t want to jeopardize my (public, state) university’s ABET accreditation but it was an open secret that the comp sci department just waved people through. My own experience supported this- any assignments that involved writing code (beyond introductory programming courses) were group projects, and I was one of very few who could write code- so after a couple of semesters everyone knew who to form groups around. All the groups had one (maybe two) people who could actually code. Most of my graduating class did not go on to be software engineers. They are mostly in the tech field, but as project managers, software sales, that sort of thing. One ended up at Boeing for a time as a software engineer- he was actually in my group for one of the classes and wrote, to this day (11 YoE) the worst code I’ve ever seen. That he was allowed to write code for Boeing proves to me that all the allegations about Boeing QA are true. That they decided he was competent enough totally explains why their MCAS software caused two of their planes to nose-dive into the ground.

I would hate to think that I can’t find a job because all of these dead-weight engineers are occupying seats at companies who might otherwise hire me. I feel like it’s more likely that the dead weight let go from FAANG realized they have to actually do work now, and take their new roles “seriously enough.”

I thought it was an open secret that masters in CS was just a scheme to buy immigration status?

Most of my classmates were white, so I doubt they were there for immigration reasons. This was also undergrad. Is an M.S. an automatic ticket to permanent U.S. residency?

What I saw in school was a lot of foreign masters in CS students with zero experience with coding and often with very little interest in learning. Big cheating issues. The school liked it because they were paying full price, and IIRC the students liked it because they got residency while studying and they could turn that into work permits and longer term status by getting a job after graduating. Definitely not an automatic ticket, but a relatively low barrier way to enter, assuming you have money.

There were plenty of local masters students as well, but it just didn't make a whole lot of sense to do masters in CS at the time, not a lot of value there beyond undergrad unless you were on your way to PhD or maybe decided to study CS after completing some other undergrad discipline.

Sounds like one heck of a loophole for those individuals, propped up by the institutions themselves (as usual). It reminds me of all the loopholes with the H-1B visas- for example, employers can get around both the "you need to hire U.S. citizens ahead of H-1B visa-holders" and the "you need to pay H-1B employees the same as U.S. citizens" stipulations?

That said, I have personally only known Asian SWEs whom I worked with in person to be very talented and hardworking, perhaps to a fault. Definitely to the detriment of their health and time with their families.

I don't doubt there remains a significant number of "0.1x devs" that are somehow still employed despite everything, and it's a shame because that heavily implies there are a significant number of unemployed "10x" or "1x" devs who can't find a job- so, again, as always, the institutions are supporting the status quo (in this case, by not really giving a fuck if they have quality talent working for them).

Nevermind the MCAS, the doors flying off of airplanes shows that it's not just software, and that no quality disaster is really big enough by itself to create some renaissance of interest in engineering excellence. Clearly Google was absolutely never going to fix declining search quality until finally ChatGPT came along and pushed them towards doing it and despite that Anthropic may very well still win the AI race.

It's easy to look at this type of thing and moan that the MBAs finally won, or start talking about antitrust enforcement and too-big-to-fail. There's some truth to that but it fundamentally seems like a bigger shift where the social contract is broken almost everywhere and cooperation itself is just on the decline. Things are increasingly adversarial between corporations and labor, corporations and consumers, governments and citizens, governments and corporations, and the list goes on.

There's so much we can learn from history, but that takes time and we're a very apathetic and comfortable civilization now.

I think it's important and worthwhile, though. It will give you an edge over everyone else if you "know" what is going to happen- history repeats itself, in part because so few bother to learn it and learn from it.

It was only this month that I finally learned that the American Civil War was about BOTH states' rights and slavery. To some reading this (this is a particularly-learned forum after all), you may think "duh - it was about states' rights to own slaves"- but that wasn't how it was taught to me in my public, Florida high school. We weren't taught that the North was just as racist as the South, and that the ONLY reason they didn't have slaves in the North was that their economy was better-off without it. We weren't taught that Lincoln's main problem with the South seceding was simply that it was illegal- literally not allowed by the Constitution. It only became a fight over "slavery" when the Union Army was getting its ass kicked and Lincoln was forced to "free all slaves" so that some would join the Union Army to fight (this was also to gain support on the global stage and make it less likely for other countries to trade with and otherwise support the Confederacy).

The Confederacy decided to secede because they believed that slavery was going to become illegal at the federal level "at some point"- there was not a formal plan to do so at the federal level at the time.

This segue came about because I was seeing memes on Facebook about the Union beating the Confederates "a second time" if the current political trend continues. These people think they're on the right side of history by identifying with the Union but that's not really the case. I mean heck the Civil Rights Movement wasn't for another 100 years following Lincoln's Emancipation Proclamation.

All that to say- history implies there's another revolution on the horizon where, once again, we have to put the 1% in their place.

I also believe that the rise of code camps without having to put in the time, ZIRP, and FAANG level compensation led to a flood of software devs who were more about "making paper" than about the love of computer science.

Conversely I can say with a high degree of confidence that I've never met a M.Eng who was in it for the material perks.

Tangentially, it reminds me of something I heard about Scandinavian countries (though I can't vouch for the veracity of it). Since you were going to get the hell taxed out of you no matter what you did - it made more sense to focus on doing what you loved as a career.

Not a Scandinavian myself (Australian - halfway between the Nordic model and the US dollarocracy, and sadly sliding ever closer to the latter), so can't speak authoritatively. But I would have thought it's not so much about "getting the hell taxed out of you no matter what you do", but more about "getting decent pay and a massive social security safety net no matter what you do". Although either way, yes, makes sense to choose your career based on passion rather than paycheque. Which is why, for example, Finland (famous for its high standard of public education) AFAIK has a significantly less severe teacher shortage than most other OECD countries.

Re: MEng

Not so sure. A generally well-paying professional job? Probably. A deep passion for tinkering and building mechanical things? Not so obvious.

> Tangentially, it reminds me of something I heard about Scandinavian countries (though I can't vouch for the veracity of it). Since you were going to get the hell taxed out of you no matter what you did - it made more sense to focus on doing what you loved as a career.

I live in Scandinavia and it's not so much about being taxed the hell out of it, it's more that whatever your profession you will, most likely, have a decent income to live with dignity and some comforts, and getting "rich" by simply working for a FAANG-type/finance company with absurd amounts of bonuses through options/RSUs is not viable.

Also, there are safety nets in place in case you want to risk going self-employed/starting a company with your profession. If you are a carpenter and want to start a small business, failing is ok because you likely won't be homeless or starving.

My feeling is that it's more about income inequality being relatively low while still having enough range that it pays off in case you work with something that requires more extensive training (law, medical, etc.).

I agree and I think ultimately, all the layoffs are actually good for the industry (though not for the people obviously). While what I'm saying is callous, there are too many people who don't know anything and don't care to learn getting in the way of people who want to do a good job. It really hurts product quality and development practices, since people who don't care eschew best practices. I don't mind if somebody goes into the industry for money, as long as they actually care to do a good job, but those type of people are becoming more rare.

> I don't mind if somebody goes into the industry for money, as long as they actually care to do a good job, but those type of people are becoming more rare.

I think I've been very lucky to frequently have coworkers that want to do a good job, but what is remarkable is that I've never worked for a company that really wanted to do a good job. Leaders fail upwards, terrible companies often survive and thrive, and thrive even harder when they get more terrible. The product as such is often fake; the customer list is usually the really valuable thing, and the customers themselves cannot leave for whatever reason, because they have no alternatives or are somehow held hostage. Many B2B relationships are completely unnecessary, it's just a tactic for a person at the business to become or remain important. Most CEOs would much rather play power games with mergers and acquisitions than do the comparatively boring work of mastering fundamentals or being innovative in their industries. Some version of this kind of mentality is becoming common for investors and workers too. Real work just gets in the way of looting the corpse and arranging your exit plans.

As narratives become more important than substance at virtually every level of post-manufacturing economic activity in services, and incentive structures are screwed up enough for a long time.. it's sad but not surprising that workers eventually get the "no one else cares, why should I?" attitude.

"That has made it really difficult for employers to tell the yuppies apart from the true hobbyists."

In reading HN stories and comments for the past 17 years I have noticed in the last 5 years or so there seems to be a sizeable contingent of people who publlish and comment online who express skepticism or even disdain for anyone who writes programs as a hobby, so-called "hobbyist programmers".

Perhaps this is not what is meant by "true hobbyists" in the above quote as their employers would have no need to know about a programming hobby. These hobbyists are not employed as "developers". They are employed in fields where salaries are not routinely financed by debt.

I have never read such a thing. Care to link an example?

In 2013, IDC estimated 40% of software authors were "hobbyist programmers" not "developers".

https://www.computerworld.com/article/1513176/the-rise-of-ho...

https://www.pcworld.com/article/443076/idc-hobbyist-programm...

Fast forward 10 years. Stack Overflow conducted a survey and only 5.56% of respondents said they programmed as "hobby" and were not employed as "developers". But 70.4% said they were employed as "developers" but also programmed outside of work as a "hobby".

https://survey.stackoverflow.co/2023/

Perhaps there could be some ambiguity in the term "hobbyist" as applied to computer programming.

Those are some interesting findings. And while it probably doesn't need to be said: this is a Stack Overflow survey where the respondents voluntarily filled it out. I'd expect them to be a bit nerdier than the average! Indeed, 25% claiming to contribute to open source is pretty high haha.

Just to be precise, when I wrote "true hobbyist", I was referring to people who actually like programming, data science, IT, or UI/UX, etc and don't do it for purely career/financial reasons.

For example, my parents were originally musicians but were forced to change careers and became (music) teachers for financial reasons. However, they still practice and perform super often even though they don't make much (or anything) from it. It's kinda people like them that I'm referring to as "true hobbyists".

I’m freelancing for a lawyer right now and want out even though there’s nothing out there unless I want to commute 3-5 days/week and take a massive pay cut. The pay is good and consistent but… I would say the good lawyers are going to naturally be bad clients for software development services. “Can’t be bothered” is how I would sum it up. Can’t be bothered to flush out requirements, can’t be bothered to use any project management tools like JIRA or Basecamp, absolutely no interest in understanding the difference between Solution A and Solution B or the potential consequences of either.

I won’t commute. No point in being a software engineer if I have to drive to an open office floor plan just to utilize a bunch of tools to try to tune out my surroundings and focus (eg., noise cancelling headphones).

I think I can get in somehwere via a referral but these friends/former coworkers don’t sell them at all. One is always complaining about a deadweight engineer on the team who hasn’t been fired because of nepotism. One always says how soul-sucking his job is, that the SRE basically decides how things are built rather than an architect, and it’s in the advertising space. The third one is a company I used to work for, and they’re full-remote now but apparently still rely heavily on cheap, garbage offshore “talent” which is why I left in the first place.

I love writing code and building apps but as time goes on it looks more and more like it’ll have to just be a hobby.

I thought about trying handyman type stuff. I have most of the tools needed, and I do things correctly. Like with code, I hate cutting corners and tend to be slower but I think deliver higher quality output. Some will say that is not appreciated anywhere- and perhaps least of all in the “handyman” space- I mean who wants to pay $80/hour for me when they can get another guy to do an “adequate” job for $60/hour?

I’ve also thought about building an app and trying to sell it/turn it into a revenue stream (I have a few ideas ) because it’s what I know how to do and if nobody is going to pay me I might as well write code “for me”.

Brass tacks, it’s a bad time to be a SWE.

It’s a bad time to be in any kind of business - except bankruptcy attorneys that is.

Bad time to be a SWE... Yet paragraphs of extremely nitpicky requirements for a job, turning up nose at opportunities for nonsense reasons.

SWEs truly do not recognize how extremely pampered and privileged they are. It might be a bad time to be a SWE but it's a worse time to be almost anything else.

What struck me was that the lawyer wasn’t interested in getting into all the nuts and bolts of software development. Of course they weren’t. That’s your job.

Of course I don't expect to wax poetic about the finer technical details of an implementation- what I mean is he tends to get frustrated at the technical limitations, but has no interest in understanding those technical limitations so that he might curb his expectations and align them more with reality.

Example: I wrote an ETL pipeline that processes "hundreds of millions" of records. It takes several hours to run (on my goddamn desktop computer). He doesn't understand why it takes "so long", and doesn't want to understand why. Also little interest/desire in paying for server compute that might get it done faster.

It’s empirically a bad time to be a SWE. We’ve seen more layoffs than any other field since 2022. Times is hard, but for us it feels unnecessary and artificial.

You can break my balls about having “high standards”, but the reality is our industry is absolutely plagued with MBAs and other non-technical folk who think they know better than we do about things like how we are most productive and how software should be written. It’s a plenty-toxic field made worse by, surprise, MBAs who don’t understand LLMs but have an unfounded belief that they can replace engineers outright (2024-)…. Or MBAs who thought hiring every SWE they possibly could would be a good idea because borrowing money was “basically free” and they could just throw things at the wall to see what stuck and ax the rest (2020-2022).

If being a software engineer means constantly swimming against the current with things like commuting to an office, allowing non-technical people to make architectural decisions, and getting yelled at and cursed at over the phone because an ML model is “only” ~90% accurate…. It’s not the career for me.

I guess by comparisons to extremely anomolous conditions, yes it is a bad time. But it is still a time in which it is better to be a SWE than so many other careers. I'm not making FAANG money and my life is so fucking easy and comfortable compared to basically anyone I interact with IRL (aside from other SWEs). It seems like dealing with a mild form of what every other office worker deals with is a very low price to pay.

Perhaps it is an especially bad moment to be a laid off tech worker due to the surge in supply, but that's a very small proportion of SWEs.

I'm sorry you're being shouted at and cursed at, that's totally unacceptable in any workplace and I hope you're able to find something better. FWIW there are still places that give their SWEs offices with doors that close.

I appreciate the empathy. I don't have much insight into "other industries", so I believe that even unemployed I would be better off than most (I have an emergency fund after all).

We should all be very worried when even ~5% of us can't find work. Sure, in the past it was just part of a cycle and the jobs "came back", but in the meantime it is bad for everyone- employed or not- when the employers have the leverage. Those who are employed are basically stuck- and employers know this. Employer abuse will run rampant. Stay late. Do the work of two people. Put up with your boss's attitude. Quit and they'll replace you within the week with someone a little more desperate than you.

There's also no guarantee those jobs are coming back, or coming back within the expected time-frame. How many years was the tech job market "bad" following 2000? 2008? By official metrics, this recession hasn't even formally started yet, even though we've arguably been in one since 2022 or 2023. Even if the timer were to start today and stocks fell 30%, it might be another 2-4 years before companies "recover" and feel like hiring SWEs again.

If all that's out there for SWEs are "scraps" that I have to vigorously fight over with other engineers via bullshit hoops like leetcode - I don't think it's worth it. I mean we're at the point where companies are just going to hire whoever they like more from a personality/appearance standpoint. That's just the natural outcome of having five, ten, or fifty otherwise-identical and equally-qualified candidates to choose from.

I really think the current best lane to be in in software is self-employed/startup. We’ve been in a recession for like a year, and will have another year before capital frees up. The big players and startups are dying from entropy and bad UX, and in 2 years there will be money on the table for the thing that replaces whatever shitty software you’re currently using.

What makes you think that capital will free up in a years time? I only ask because this feels like the start of a winter to me. Not one that's approaching an end.

Donald Trump has been complaining about the Fed's interest rate being too high going all the way back to his first term (perhaps even earlier). It's no secret that he wants Powell to drop rates to zero or near-zero.

The "glass half-full" among us will tell you this tariff stuff is all part of Trump's 3D-chess plan to refinance U.S. debt to a lower rate. Personally I doubt it. Everything he does is for himself, and it's only coincidence that sabotaging the economy to get rates to zero could benefit the National Debt.

Anyway, the tech hiring spree we saw in 2020-2021 was fed by rates being at or near zero. If money is free or nearly-free to borrow, why not hire everyone you can and try to grow your company as quickly as possible?

So in theory, bringing rates back down will ease companies' recession fears - not necessarily because they would think a recession isn't coming - no, I'm sure we will all still be expecting a recession "any day" - but with money very cheap or free to borrow, there's no risk to hiring people, starting new projects - because when shit hits the fan, there's negligible additional financial burden from that hiring/starting of new projects. Maybe 2% interest on the debt taken out to pay those people and start those projects. That might as well be 0% to a company, especially one that is FAANG-sized.

I've been saying for years now that the Fed dropping rates to 0% would solve the tech job market woes, and I still believe that. BUT- I make no guarantees against terrible side effects of dropping the interest rate back to zero. I think it will revive the tech job market... and cause many other, bad side-effects (including heightened inflation).

Everyone is hot to trot for AI, but these companies (even FAANG) are working at partial-capacity for developing AI-driven products because they're afraid to hire right before a recession. If they could hire and otherwise expand "for free" I think we would see another 2020-2021 from a tech hiring perspective. Things would be stupid again. You'd have more of the "I work at Google and do absolutely nothing", more instances of senior engineers pulling $250k base salaries... It would be a great time for SWEs, but probably at the detriment to everyone else via things like inflation.

It’s hard to say how long we’ve been in a “recession”. My personal recession has been going since 2023. That’s when I last held a full-time job. I’ve been “1099 for table scraps for the last 2 years, and it doesn’t show signs of stopping.

So in that sense, absolutely self-employed has been the route for me that has yielded income.

> I would say the good lawyers are going to naturally be bad clients for software development services. “Can’t be bothered” is how I would sum it up. Can’t be bothered to flush out requirements, can’t be bothered to use any project management tools like JIRA or Basecamp, absolutely no interest in understanding the difference between Solution A and Solution B or the potential consequences of either.

Sounds pretty good to me TBH. I see this stuff as part of software development, though often there's a specialised role (e.g. project manager) that handles it, I think there's value in learning to enjoy it. If they don't want to use Jira, can you make it work for them via their preferred channel and write your own tickets? If they don't care about the difference between two solutions, does that mean they trust you? Can you communicate it in terms they care about (e.g. economics, security)?

I can see how you'd want to focus on the code, but I really do think the stuff between the stakeholder's brain and the code can be a satisfying part of the job as well.

Yeah fair enough. I do write my own tickets. I don't think that resolves the "can't be bothered" attitude- that they're very busy and value their time very highly.

I will add another interesting quirk- he will never admit fault and in written communication tends to "make assertions" and I feel like I have to defend them or the lack of defense means his assertions are correct "in a court of law" - I don't know if I'm describing that well. Basically he has it so ingrained in him never to say anything that may be used against him later, even though obviously I have no intention of suing him now or in the future.

It amounts to a sense of under-appreciation and that I am not respected.

We have a patent together through this project. It's very interesting work that I wouldn't be doing anywhere else. It definitely could be worse.

I don’t see a palpable effect from “AI”, I think software as an industry is just extremely sensitive to interest rates.

When money is cheap every company in every industry can soak up engineers to make speculative or vanity apps and websites. When money gets tight companies find ways to do without. So on top of software-qua-software companies decreasing hiring, other companies also hire fewer engineers, and the effect compounds.

New technologies always come with increased expectations but never to cut down the overall work. You got a car and you were expected to travel more distance in less time. You got mobile phone and internet, and you were expected to be always connected and informed.

Technologies were not meant to give you more free time or let people to be out of work and enjoy their idle time. Technologies are meant to increase productivity while utilizing ALL of the human resources at the disposal for a nation.

Heck, they don't even allow you to spend much time for family. They gave nursing homes for elderly and creches for the kids. The economies are built for workers and roads & cars are built to bring the workers to factories (offices).

Also they ensure that you don't dare to be out of job, by tying you down with the mortgages and keeping the house prices at a level that you need to work forever just to have a roof over your head.

You live in city because of work, and you work because you need to afford to live in city.

Blaming the bad market on AI is a marketing ploy pushed by AI startups to try to overstate the "AI revolution" and conveniently their companies.

I’m sure it’s also a ploy by C suite leadership to explain/justify whatever they want (layoffs, price hikes, budget cuts, et al.)

My current conclusion about this topic:

- AI is great for any kind of prototyping. If you want to test something super quick, go for AI built programs. It’s a cheap and fast way to test ideas.

- AI is not great for actual products especially those that have evolved some well justified complexity over time.

This explains why YC sees something like +80% of their new batches using AI and why I haven’t seen any cases of larger/established companies using AI more than an advanced type ahead in confined areas like testing.

So will AI be an essential part of any develop and build process? Absolutely.

Will AI be able to conquer the second space over time? Maybe. But also maybe not.

And if many people avoid CS now because they (falsely) believe AI will take their job, we will have a bad shortage of devs in a couple of years.

How much of it is AI, and how much of it is that we have too many computer science graduates now because of the gold rush?

The same thing happened ten years ago with law school: everyone thought if they got a law degree they'd be set. And then there were too many degrees and not enough jobs

My ex was in that situation- the best she could find was a part-time gig at a tiny law firm that didn't pay much (and then she did a coding bootcamp and switched to software, and immediately tripled her salary)

Law was historically something that less employable undergrad majors did when they didn’t see a clear path to good employment. I knew a bunch who never really had a law career. For a long time the only really great path there has been partnership in big law though more successful local practices have probably done ok.

Rather poor writing for the ACM. It’s just a pile of anecdotes and overreaching conclusions. The downtrend in software dev hiring predates the AI boom.

Absolutely. Still, it’s possible that AI came at just the right time to “pick up” where FAANG’s “over-hiring purge” left off.

Personally, I think people will find a reason to justify their feelings, and I think the feelings since at least 2020 are that “a recession is imminent.” Nobody wants to be caught with their pants down, least of all FAANG, but when money was “virtually free” (Fed’s emergency interest rate drop to 0%) that was physically impossible. Now that interest rates are closer to where they probably should be and borrowing money isn’t really worthwhile if it can be avoided…. A recession is imminent.

The Fed’s soft landing may have done more harm than good in the long-run. Is delaying a recession (which by the way is a healthy part of any economy) really the preferred route when the economy has already had a record bull market? Now we’re in this weird limbo where the stock market and U-3 unemployment rate say “the economy is great!” and yet everything else seems to indicate the middle and lower classes are struggling at levels not seen since ~2009.

Part hype but also the technological singularity is a gradual, distributed, discontinuous trend moving in that direction partially. This sort of corporate hype is another excuse to suppress salaries and hollow-out another profession.

Can just echo the sentiment of others here: Tech people should stop it with their “Great Replacement” Conspiracy Theory.

Yes, the job market is terrible, but blaming it on AI is wishful thinking if not another round of flaming the AI craze.

Anyone looked at the job market charts during covid? There was a historical hiring spree going on and it was obviously bound for correction. Which it did and now we’re basically at pre-pandemic levels.

Pre-pandemic I was getting 2-4 unsolicited recruiter messages per week.

Now it’s roughly 1 per month.

There's still a large supply of relatively bad/inexperienced developers looking for a job from what I've seen (currently hiring for a mid level role). The last time I hired developers 2-3 years ago, there was a far better signal to noise ratio. So I suppose we still have some of the oversupply from that era left, not to mention that gen AI doesn't seem to be a net positive for hiring (less signal for both companies and good candidates). But at least here in Europe, I don't really know any base line competent devs looking for work.

It sounds like you're saying people like me just need to "wait out" the "mediocre" SWEs... that they, already with one foot out of the industry anyway (having never been particularly passionate), are more-likely to switch careers entirely in order to make ends meet.

If so, I would be playing right into that (to my detriment) by pivoting to another career myself. Therefore, I should stay put and trust that things will get better and eventually return to normal.

Sorry for the late response. No, that's not really what I tried to say. It's rather just a dilemma I was pointing out: If there's more noise and less signal, it's harder for the good people to find good employers _and_ vice versa. Not easier. Not to mention that a lot of "good enough" devs can manage to stay employed, even when better devs are available on the market. Any automatic correction to this will, I believe, take years.

The way I find clients for my company is exclusively by reputation, i.e. network. Quite often I get asked if I know good devs, and I'm starting to draw a blank. So if you feel like reaching out for a chat (see my profile), I could offer to reach out if something comes up. My circles are primarily in Europe though.

It's weird people attribute job cuts to AI when it clearly isn't replacing humans except in the most low values scenarios (e.g. ai slop where the work would just go undone otherwise). I don't think I've ever seen AI operating at "replace a human" level in development, despite the endless hype.

When you hear hoofprints, think horses, and the end of ZIRP is pretty clearly to blame.

I think everyone could see that the situation was unsustainable as far back as 10+ years ago - ad-supported apps being paid for by ads for other free ad-supporter apps. Below cost taxi rides and meals to your door. Companies paying $100,000 for engineers to do nothing just to make sure competitors are starved of talent. CRUD apps somehow getting million dollar valuations.

About a decade ago people were worried truck drivers would lose their jobs to AI; ignore the hype.

Obviously AI isn’t wholesale replacing anyone, but having a tool that can coach me through some library I’ve never used before can be the difference between a lot of trial and error and a 2 hour PR to finish a React feature that no one else is going to do. Whether that work (web dev, broadly) has real value or not is another question

Maybe, but that might fall into the bucket of "doing things that would have gone undone".

AI slop often doesn't displace designers because it is used in places you would never buy a design for. If the react feature wouldn't be done without AI... It just wouldn't be done and the business would probably survive.

Every software project has an infinite list of feature requests, bug fixes, etc. The cost of doing the work can shift a lot before it makes sense to start firing people. If there were actually a 100x productivity improvement it might be noticeable, the realistic estimate is closer to 20-30%, maybe 50 if you're writing very simple React exclusively. I think staffing levels is driven more by business/HR constraints than some ticket/worker calculation.

What I find curious in these AI discussions is how the tables have turned: Better tooling used to be pretty universally appreciated. I've never heard anyone say that IDEs, garbage collection, static code analysis, profilers, debuggers, frameworks or anything else that makes the life of a developer easier was "replacing" them.

This stuff replaces programmers just like LLMs do, they make the job easier/faster for experts, so you can do more with less.

They also make the job easier for laypeople, but there's specialised tooling for that already, mostly in the form of no code tools. Personally, I'd take an AI generated code base a non-developer whipped up over the no code nightmares I've seen _any_ day. But just like better tooling for experts, better tooling for laypeople isn't a particularly new thing.

How many if them can manage a clean git history ?

> 74% of 200 IT professionals surveyed worried their jobs will be replaced by AI tools

Seriously? I can't speak for that 74%, but personally, I'm not worried. AI is not that smart, and my job is not that dumb. And I'm just a garden variety (non-AI-specialising) Python back-end dev.

How about the rest of y'all? Are you worried or are you meh?

All the remote programmers do a worse job than AI, that’s for sure.

Not sure what you mean by this. Remote devs are usually better by necessity.

No that's complete nonsense. I work for 100% remote company and we're all better at the job than current gen AI ;-)

Yes I mistook remote for outsourced. Please understand it leniently, of course distance doesn’t make someone incompetent, but I’ve hired remote people and they clearly weren’t caring for my boat.

Wow, I haven't run across this degree of workplace toxicity since the last time I was in an office. Mind telling us where you work so we can steer clear?

Being a full remote dev for a decade now, I obviously disagree.

There is of course a difference in where the remoteness comes from though: is it because the talent is that much in demand that it can just require it?

Or is it because your employer is so cheap, he rather offshores the job to some emerging market, irrespective of the slop[1] he gets?

[1] why EM code might be slop is another topic in itself but let’s just say: Good ppl from EMs don’t need to slave through slopshops

Remote or offshore? I’ve worked with offshore teams (Infosys) that were vastly inferior to ChatGPT for writing code. We’re talking declares variables out of scope and checks that code in to source control. We’re talking forgets to remove comments in copy-pasted code that clearly refer to another F100 company. We’re talking takes 3 days to implement a one-liner code change.

I'm also a garden variety non AI-specializing back-end dev. I'm not worried.

I wouldn't be shocked if AI got so good that in 10 years my company of 200 or so can't compete with 6 AI-enabled tweens. But if that actually happens then it's a whole new world governed by radically different economics. There's no use taking anxieties built for this world and projecting them into a future they don't fit with.

I would imagine the wording is very important here.

Am I worried that my job will be replaced by AI tools? On one hand, I know that ChatGPT can’t replace me. On the other hand, companies seem seldom run by logic and empirical data these days.

I’m worried that companies will attempt to replace SWEs with AI tools (even if it’s just halving the SWE department and saying “use ChatGPT to pick up the slack”).

Offshoring comes in and out of fashion, but when it’s in fashion, it seems to take years before companies realize their mistake.

> On the other hand, companies seem seldom run by logic and empirical data these days.

Yeah, this is the real worry. The vast majority of the decision-making people in business are not at all trying to actually decide what is good for the business, they are trying to decide what is good for themselves.

As such it's not "can we effectively and efficiently fire people and replace them with robots and expect the business to continue to exist in a year". Instead it's more like "Can I get a bonus next quarter if I cut costs this quarter?".

People who are sitting around very rationally trying to discuss how much AI is improving or how fast are missing the point I'd say. Powerful people don't think in those kind of terms as far as the ground truth of things. Usually they are more interested in controlling the conversation and seizing initiative. They are happy to blow up the companies / countries / industries they operate in and depend on, because they can always find a way to profit from the chaos.

Short-term thinking is a big problem. I'm not sure why the stock market continues to grow despite short-term thinking absolutely destroying stalwarts like Boeing, GE, Intel, et al. I suppose part of the stock market's perceived-success over the last 20 years is that short-term thinking- in the form of buybacks. I'll never forget that United Airlines ($UAL) needed that huge government bailout in 2020 when COVID hit because they had just spent ALL of their cash on stock buybacks. Actually, not to go on a tangent, the airline industry has decades and decades of proof that the U.S. government won't let them fail. United basically had the United States' blessing to blow all their cash reserves.

Another part of the equation are apps like Robinhood that have opened the floodgates and enabled the lower and middle classes to invest in the stock market (however recklessly). When you consider the stock market is just one big Ponzi scheme (the newcomers support the old-timers with fresh capital and enable them to retire) you see it as less of a "no-brainer, guaranteed road to wealth" and more of the hunting ground that it really is- where the 1% steal from the rest of us using "tools" like market volatility to scare people into selling at a loss.

Fun fact: Robinhood now offers IRAs, and allows you to TRADE OPTIONS with your IRA. This should 1000% be illegal.

When I look at the situation with AI I think that the insistence that it is in every product under the sun is also going to turn into one of the drivers slowing it's adoption as a method to replace people.

When something is sold as a magic tool that a decision maker is able to not be directly exposed to, they can go and trust the hype. Low-code/No-code were magic tools to make line workers into software Devs.

With LLM's and "AI all the things" being put in front of everyone's faces even decision makers are exposed to the wrinkles and can draw on their own experience of "would I hire Gemini or Co=pilot, the thing that struggles to follow seemingly simple instructions, to run my business" you start getting more speculation into the mystical powers of this human replacement machine.

With an opinion on AI being useful for low-risk activities but needing heavy supervision for stuff that matters, the calculous changes when considering handing that system the responsibility to deliver your paycheck.

Not worried about AI replacing my job in the foreseeable (10+ years) future, it's just not that impressive. I am worried about executives thinking it's the future and making even dumber hiring decisions based on that.

I'm worried about AI changing the shape of my job to the point of it being unstimulating and soul sucking. I enjoy the challenges, solving problems, achieving what others (peers) tried and failed to. Knowing that continuing to develop these skills also keeps my mind sharp is motivating, too. If I lose that, this is just a boring corporate role among dullards.

If one of the idiots in my reporting line decides AI should be mandated (e.g. Shopify dunce), that's the end for me. I have no interest in being spoonfed by a chatbot and being accountable for it.

> computer science majors are also competing for software engineering jobs with people who minored in the discipline

Uh if you think being a swe(i.e. programmer) requires having a CS background you're gonna have a bad time.

Most SW jobs are gluing shit together and dealing with the boogers, not designing an improvement to A* or implementing a red black tree. And I say that as an engineer on a team that actually does those sorts of things.

Hell, that's why AI has actually been slightly useful to save programmers work.

I think the opening sentence[1] is a big hint where the problem lies.

4 years (or more) in college is a ridiculously expensive investment both time and money wise to work in the software engineering field.

It shouldn't take more than 18 months of hands on training after/during high school to get started as an SWE. I don't know the origins of this 4 year degree but it shouldn't be the norm anymore. It robs the young adults precious years of real world experience and earning capability.

Also, it's a massive sunk cost which means they would be unwilling to explore other fields (e.g., masonry, plumbing, electrician..).

[1] ...working toward dual bachelor’s and master’s degrees in computer science.

Are you saying all the learning of the theory in the 3/4 year degree programme is pointless? Do you really think a 18 month coding bootcamp, without the theory, can make anyone a software engineer?

4 years degree programme is not a problem, the problem lies where the people holding the money think they can replace human with computer.

That's exactly what I'm saying. Let me explain.

The overemphasis on theory was probably justified when CSE program was created. Back then the field was quite new and engineers would often encounter problems that needed deep theoretical knowledge and mathematical modelling capabilities. In networking, compilers, distributed systems, OS, data storage systems etc., So a theoretical grounding was necessary to prepare them to solve those unknown problems.

However that doesn't seem to be the case anymore based on my experience. To build a typical business/consumer application there are well tested architecture patterns available along with technology and libraries.

Of course if one is operating at the bleeding edge (say chip design) then a deep grounding in theory (say physics of optics) is needed. But those are exception and not norm.

My frustration also stems from looking at the syllabus of typical CSE curriculum in India. In first year students learn a little bit of all other fields. Mechanical engineering, civil, electrical and so on. The CS courses start in real earnest from 4th semester. Why not start from CSE subjects right away? Why do they need to learn workings of an internal combustion engine!?

> the problem lies where the people holding the money think they can replace human with computer.

Those with money want to replace human workers with machines. Automation with robots and mechanisation has been happening in other fields (automobile, agriculture, construction...). That it'd happen to SWE was expected, no?

[1] https://vtu.ac.in/pdf/2022syll/csesch.pdf

Reply on news.ycombinator.com